2014 - Hamidi - Distinguished Ground Improvement Projects by Dynamic Compaction or Dynamic Replacement
2014 - Hamidi - Distinguished Ground Improvement Projects by Dynamic Compaction or Dynamic Replacement
Babak Hamidi
October 2014
Declaration
To the best of my knowledge and belief this thesis contains no material previously published
by any other person except where due acknowledgment has been made. This thesis contains
no material which has been accepted for the award of any other degree or diploma in any
university.
Babak Hamidi
Signature:
Date:
i
Abstract
Dynamic Compaction is a ground improvement technique that was invented by the late
French engineer, Louis Menard, in the mid‐1970s. In this technique deep layers of granular
soil are compacted by dropping a significantly heavy pounder from a considerable height.
The pounder impacts vibrate the ground, rearrange the soil grains in a denser configuration,
and densify a thick layer of soil that otherwise could have been subject to failure or
unacceptable deformations under loads.
In approximately the 40 years that has gone by since the invention of dynamic compaction
this technique has been become a well‐established and popular ground improvement
method, and has been the subject of research by various academic institutions. However,
consolidated and comprehensive information is outdated, and publications remain
scattered.
This thesis attempts to collect and compile a thorough and comprehensive documentation
of previous research that has been published on dynamic compaction and dynamic
replacement or related matters on these technologies to assist in the development and
progress of this branch of ground improvement technologies, with incorporation of views
that are closely related to the state‐of‐the‐art of the industry.
The author has had the opportunity to participate in numerous distinguished dynamic
compaction and dynamic replacement projects under the mentorship and guidance of the
world’s leading expert on the subject, and has had the unique opportunity to be exposed to
the views of the engineers who have developed and led the advancements of this ground
improvement technology. For documentary purpose and for enabling others to use the
knowledge and experience that has been acquired, a number of these projects have been
described, reviewed and analysed in this thesis. Discussions include construction methods,
quality control and verification processes, comparison of results with previously published
research, and development of new formulations.
ii
Acknowledgements
It is with the most sincere gratitude that I acknowledge the attention and guidance that my
supervisor, Professor Hamid Nikraz, has given to this work. His brilliance, patience, and good
nature have provided inspiration and motivation that shall guide me for many years. His
friendship and support will be always cherished.
I am also grateful to Menard for providing the data and information that has been used in
Chapter 3, without which this thesis would have been impossible.
Finally, I wish to acknowledge my parents, Amir and Narsin, for their encouragement
throughout my academic endeavors, my wife, Azin, for her support, and my mentor, Mr
Serge Varaksin, for reshaping my views on applied geotechnical engineering.
iii
List of Publications during the Course of this Thesis
Published Peer Reviewed Journal Papers
1. Hamidi, B., Nikraz, H. & Varaksin, S. (2009) A Review on Impact Oriented Ground Improvement
Techniques. Australian Geomechanics Journal, 44, 2, 17-24.
2. Hamidi, B., Nikraz, H. & Varaksin, S. (2009) Arching in Ground Improvement. Australian
Geomechanics Journal, 44, 4 (December), 99-108.
3. Hamidi, B., Nikraz, H. & Varaksin, S. (2011) The Treatment of a Loose Submerged Subgrade Fill
4. Hamidi, B., Varaksin, S. & Nikraz, H. (2013) Relative Density Concept is not a Reliable Criterion.
Ground Improvement, 166, GI2, 78-85.
5. Hamidi, B., Varaksin, S. & Nikraz, H. (2013) Relative Density Correlations are not Reliable Criteria.
Ground Improvement, 166, GI4, 96–208.
6. Hamidi, B., Debats, J. M., Nikraz, H. & Varaksin, S. (2013) Offshore Ground Improvement Records.
Australian Geomechanics Journal, 48, 4, 111-122.
Published Peer Reviewed Conference Papers
7. Hamidi, B., Nikraz, H. & Varaksin, S. (2010) Treatment of Thick Saturated Loose Subgrades Using
Dynamic Compaction. 3rd International Conference on Problematic Soils (PS10), Adelaide, 7-9
April, 121-128.
8. Hamidi, B., Nikraz, H. & Varaksin, S. (2010) Soil Improvement of a Very Thick and Large Fill by
Dynamic Compaction. 3rd International Conference on Problematic Soils (PS10), Adelaide, 7-9
April, 129-138.
9. Hamidi, B., Nikraz, H. & Varaksin, S. (2010) Application of Dynamic Compaction in Port of Ras
Laffan Expansion Project. 6th Australasian Congress on Applied Mechanics (ACAM6), Perth,
10. Hamidi, B., Varaksin, S. & Nikraz, H. (2010) Dynamic Replacement for Constructing
Embankments and Walls on Soft Soil. 3rd International Conference on Problematic Soils (PS10),
iv
11. Hamidi, B., Varaksin, S. & Nikraz, H. (2010) Treatment of a Hydraulically Reclaimed Port Project
by Dynamic Compaction. 3rd International Conference on Problematic Soils (PS10), Adelaide, 7-
9 April, 113-120.
12. Hamidi, B., Varaksin, S. & Nikraz, H. (2010) Correlations between CPT and PMT at a Dynamic
13. Hamidi, B., Varaksin, S. & Nikraz, H. (2010) Implementation of Optimized Ground Improvement
Techniques for a Giga Project. GeoShanghai 2010 Conference, ASCE Geotechnical Special
Publication No 207: Ground Improvement and Geosynthetics, Shanghai, 3-5 June, 87-92.
14. Hamidi, B., Varaksin, S. & Nikraz, H. (2010) Predicting Soil Parameters by Modelling Dynamic
Compaction Induced Subsidence. 6th Australasian Congress on Applied Mechanics (ACAM6),
15. Hamidi, B., Yee, K., Varaksin, S., Nikraz, H. & Wong, L. T. (2010) Ground Improvement in Deep
Waters Using Dynamic Replacement. 20th International Offshore and Polar Engineering
Conference, Beijing, 20-26 June, 848-853.
16. Hamidi, B., Nikraz, H. & Varaksin, S. (2011) A Case Study of Vibration Monitoring in a Dynamic
Compaction Project. 14th Asian Regional Conference on Soil Mechanics and Geotechnical
Engineering Hong Kong, 23-27 May, Paper No. 405.
17. Hamidi, B., Nikraz, H. & Varaksin, S. (2011) Ground Improvement Acceptance Criteria. 14th Asian
Regional Conference on Soil Mechanics and Geotechnical Engineering Hong Kong, 23-27 May,
Paper No. 404.
18. Hamidi, B., Nikraz, H. & Varaksin, S. (2011) Dynamic Compaction Vibration Monitoring in a
19. Hamidi, B., Varaksin, S. & Nikraz, H. (2011) The Application of Dynamic Compaction to HFO
Tanks. International Conference on Advances in Geotechnical Engineering (ICAGE), Perth, 7-9
November, 627-632.
20. Hamidi, B., Nikraz, H. & Varaksin, S. (2011) Application of Dynamic Replacement in a Steel Pipe
Factory. International Conference on Advances in Geotechnical Engineering (ICAGE), Perth, 7-9
November, 867-872.
21. Hamidi, B., Nikraz, H. & Varaksin, S. (2011) Advances in Dynamic Compaction. Indian
Geotechnical Conference IGC2011, Kochi, India, 15-17 December, Paper No. H-146.
v
22. Hamidi, B., Varaksin, S. & Nikraz, H. (2011) Predicting Menard Modulus using Dynamic
Compaction Induced Subsidence. International Conference on Advances in Geotechnical
Engineering (ICAGE), Perth, 7-11 November, 221-226.
23. Hamidi, B., Varaksin, S. & Nikraz, H. (2011) Dynamic Compaction for Treating Millions of Square
24. Hamidi, B., Varaksin, S. & Nikraz, H. (2011) A Case of Vibro Compaction Vibration Monitoring in
25. Hamidi, B., Varaksin, S. & Nikraz, H. (2011) Application of Dynamic Surcharging for Construction
of Tanks on Reclaimed Ground. International Conference on Advances in Geotechnical
26. Hamidi, B., Nikraz, H. & Varaksin, S. (2012) The Application of Dynamic Compaction on Marjan
Island. 11th Australia New Zealand Conference on Geomechanics - Ground Engineering in a
Changing World: ANZ 2012, Melbourne, 15-18 July, 1202-1207.
27. Hamidi, B., Nikraz, H. & Varaksin, S. (2012) Application of Dynamic Compaction in Reclaimed
Roads. 2012 AGS Symposium: Advances in Geotechnics of Roads and Railways, Sydney, 10
October, 115-124.
28. Hamidi, B., Varaksin, S. & Nikraz, H. (2012) The Effectiveness of Vibration Reduction Trenches in
29. Hamidi, B., Varaksin, S. & Nikraz, H. (2012) Construction of Raw Sugar Silos Using Dynamic
30. Hamidi, B., Varaksin, S. & Nikraz, H. (2012) Application of Dynamic Compaction in a Project with
31. Varaksin, S. & Hamidi, B. (2012) Ground Improvement Case Histories and Advances in Practice.
vi
32. Varaksin, S. & Hamidi, B. (2013) Pressuremeter for Design and Acceptance of Challenging
Ground Improvement Works. 18th International Conference on Soil Mechanics and Geotechnical
Engineering (18th ICSMGE), Parallel session: ISP6 - Pressio 2013, Paris, 2-6 September.
33. Hamidi, B., Nikraz, H. & Varaksin, S. (2015) Correlation between PMT & CPT after Dynamic
Published Non Peer Reviewed Conference Papers
34. Hamidi, B. & Varaksin, S. (2012) Lessons Learned from Millions of Square Metres of Ground
Improvement. International Symposium on Ground Improvement (IS-GI) Brussels 2012, 2,
Brussels, 29 May - 1 June, 29-39.
35. Varaksin, S. & Hamidi, B. (2014) Contributions of the Pressuremeter to Ground Improvement
and Perspectives for the Future. The International Scientific and Technical Conference Devoted to
the 80th Anniversary of the Geotechnical Department of St Petersburg State University, 1, St
Petersburg, Russia, 4-6 February, 550-555.
Accepted Conference Papers
36. Hamidi, B., Varaksin, S. & Nikraz, H. (2105) A Study on the Variation of PMT Parameters Ratio
after Dynamic Compaction of Saturated Sands. 16 African Region Conference on Soil Mechanics
and Geotechnical Engineering, Parallel session: ISP7 Pressio 2015, 60 Years of Pressuremeter,
Tunisia, 1-2 May.
Accepted Book Chapters
37. Hamidi, B. & Varaksin, S. (Accepted) Dynamic Compaction and Dynamic Surcharging at Dubai's
Palm Jumeira STP. In: Buddhima, I., eds. Ground Improvement: Case Histories and New Directions.
Oxford, Elsevier.
vii
Table of Contents
Declaration ................................................................................................................................ i
Abstract .....................................................................................................................................ii
Acknowledgements .................................................................................................................. iii
List of Publications during the Course of this Thesis ............................................................... iv
Table of Contents ................................................................................................................... viii
Notations................................................................................................................................ xix
List of Tables ....................................................................................................................... xxxii
List of Figures ...................................................................................................................... xxxv
Foreword .................................................................................................................................. lv
1 Introduction ..................................................................................................................... 1
1.1 Introduction ............................................................................................................. 2
1.2 Objectives ................................................................................................................. 4
1.3 Scope ........................................................................................................................ 4
1.4 Significance .............................................................................................................. 5
2 Literature Review ............................................................................................................. 7
2.1 Chapter Organisation ............................................................................................... 8
2.2 Tribute to Louis Ménard ........................................................................................ 11
2.3 Review of the Theory of Waves ............................................................................. 14
2.3.1 Waves in a Bounded Elastic Medium ............................................................. 14
2.3.2 Waves in an Infinite, Homogeneous, Isotropic, Elastic Medium ................... 15
2.3.3 Waves in a Homogeneous, Isotropic, Elastic, Half Space Medium ................ 16
2.3.4 The Wave Field Generated by a Vertical Oscillation ...................................... 19
2.3.5 Internal Damping and Attenuation ................................................................ 20
2.3.6 Elastic Waves in Layered Systems .................................................................. 22
2.3.7 Propagation of Waves in Mixed Media .......................................................... 28
2.4 Theory & Mechanism of Dynamic Compaction & Dynamic Replacement ............ 31
viii
2.4.1 Behaviour of Elastic Spheres Subject to Shear Loads .................................... 32
2.4.2 Behaviour of Elastic Spheres Subject to Normal Loads ................................. 33
2.4.3 Behaviour of Elastic Spheres Subject to Normal and Shear Loads ................ 35
2.4.4 Crushing of Soil Particles ................................................................................ 36
2.4.5 Liquefaction ................................................................................................... 36
2.4.6 The Mechanism of Dynamic Compaction in non‐Saturated Granular Soils ... 38
2.4.7 The Mechanism of Dynamic Compaction in Saturated Granular Soils .......... 45
2.4.8 The Mechanism of Dynamic Compaction in Saturated Cohesive Soils .......... 46
2.4.9 The Mechanism of Dynamic Replacement in Saturated Cohesive Soils ........ 50
2.4.9.1 Failure Mechanisms in Dynamic Replacement .......................................... 51
2.5 Design Guidelines for Dynamic Compaction .......................................................... 58
2.5.1 Suitability of Dynamic Compaction ................................................................ 58
2.5.2 Depth of Improvement .................................................................................. 60
2.5.2.1 Effect of Impact Energy .............................................................................. 60
2.5.2.2 Effect of Pounder size on Depth of Improvement ..................................... 70
2.5.2.3 Effect of Momentum .................................................................................. 72
2.5.2.4 Effect of Number of Blows and Energy Intensity ....................................... 75
2.5.2.5 Effect of Pounder Size and Shape .............................................................. 78
2.5.3 Crater Depth................................................................................................... 83
2.5.4 Grid Spacing, Number of Phases and Number of Drops ................................ 84
2.5.5 Improvement Profile ...................................................................................... 93
2.5.6 Ageing and the Effect of Time ........................................................................ 99
2.5.7 Extents of Improvement .............................................................................. 103
2.5.7.1 Conventional Applications ....................................................................... 103
2.5.7.2 Application for Liquefaction ..................................................................... 104
2.5.8 Design Methods ........................................................................................... 109
ix
2.5.8.1 Lukas (1995) ............................................................................................. 109
2.5.8.2 Poran and Rodriguez (1992) .................................................................... 112
2.5.8.3 Jahangiri et al. (2011) ............................................................................... 115
2.6 Design Guidelines for Dynamic Replacement ...................................................... 118
2.6.1 Soil Arching .................................................................................................. 118
2.6.1.1 British Standard ........................................................................................ 118
2.6.1.2 Hewlett and Randolph ............................................................................. 119
2.6.1.3 German Code EBGEO (2004) .................................................................... 124
2.6.1.4 Vertical Equilibrium Load Distribution ..................................................... 127
2.6.1.5 Numerical Methods ................................................................................. 131
2.6.1.5.1 Axisymmetrical Geometry Models .................................................... 132
2.6.1.5.2 Plane Strain Geometry ....................................................................... 134
2.6.2 Ultimate Load Capacity ................................................................................ 137
2.6.2.1 Ultimate Load Capacity due to Bulging .................................................... 139
2.6.2.2 Ultimate Load Capacity due to Bearing Capacity Failure ......................... 141
2.6.3 Liquefaction ................................................................................................. 145
2.7 Advances in Equipment........................................................................................ 150
2.7.1 Dynamic Compaction Rigs ............................................................................ 150
2.7.2 Pounders ...................................................................................................... 153
2.7.3 Alternative Impact Oriented Ground Improvement Techniques ................. 156
2.7.3.1 Rapid Impact Compaction ........................................................................ 156
2.7.3.2 Impact (Roller) Compaction ..................................................................... 157
2.8 Assessment of Dynamic Vibrations ...................................................................... 161
2.8.1 Development of Safe Vibration Limits ......................................................... 161
x
2.8.2 Air Vibrations and Damage .......................................................................... 168
2.8.3 Human Response to Vibration ..................................................................... 169
2.8.4 Standards ..................................................................................................... 172
2.8.4.1 British Standards ...................................................................................... 172
2.8.4.2 Australian Standards ................................................................................ 175
2.8.4.3 Beyond Code Recommendations ............................................................. 177
2.8.5 Evaluation of Dynamic Compaction Vibrations ........................................... 178
2.8.5.1 Vibration Frequency ................................................................................. 178
2.8.5.2 Measurement and Prediction of Peak Particle Velocity .......................... 180
2.8.6 Mitigation and Isolation of Vibrations ......................................................... 189
2.9 Quality Control ..................................................................................................... 196
2.9.1 Heave and Penetration Test ......................................................................... 196
2.9.2 Menard Pressuremeter Test ........................................................................ 200
2.9.2.1 Description of the Pressuremeter ............................................................ 200
2.9.2.2 Calibration ................................................................................................ 202
2.9.2.3 Testing Procedure .................................................................................... 203
2.9.2.4 Testing Frequency .................................................................................... 205
2.9.2.5 Calculation of PMT parameters ............................................................... 206
2.9.2.6 Estimation of Bearing Capacity ................................................................ 211
2.9.2.7 Estimation of Settlement under External Loading ................................... 222
2.9.2.8 Self‐Bearing .............................................................................................. 229
2.9.2.9 Relationships and Correlations ................................................................ 230
2.9.2.9.1 Moduli ................................................................................................ 230
xi
2.9.2.9.2 Estimation of Shear Strength ............................................................. 232
2.9.2.9.3 Estimation of Drained Friction Angle ................................................. 233
2.9.2.9.4 Correlation with CPT .......................................................................... 234
2.9.2.9.5 Correlation with SPT .......................................................................... 236
2.9.3 Dynamic Compaction Testing Methods of the Future ................................. 237
2.10 Special Topics ....................................................................................................... 244
2.10.1 Ground Improvement Specifications and Acceptance Criteria ................... 244
2.10.1.1 Acceptance Criteria Based on Work Quality ........................................ 244
2.10.1.2 Acceptance Criteria Based on Minimum Test Results ......................... 247
2.10.1.3 Acceptance Criteria Based on Design Criteria ...................................... 249
2.10.2 Unreliability and Unsuitability of Relative Density as a Ground Improvement
Criterion 250
2.10.2.1 History .................................................................................................. 250
2.10.2.2 Relative Density and Points of Concern ............................................... 251
2.10.2.2.1 The Definition of Relative Density and its Intended Range of
Application 251
2.10.2.2.2 Errors in relative density testing ...................................................... 254
2.10.2.2.3 4. A Second Glance at the Standards ............................................... 262
2.10.2.3 Relative Density Correlations ............................................................... 263
2.10.2.3.1 Relationship of Relative Density with Soil Characteristics ............... 263
2.10.2.3.2 Correlation of relative density with field tests ................................ 267
2.10.3 Reclamations ................................................................................................ 281
2.10.4 Offshore Dynamic Compaction and Dynamic Replacement ........................ 285
3 Distinguished Dynamic Compaction and Dynamic Replacement Projects .................. 290
xii
3.1 Choice of Projects ................................................................................................ 291
3.2 Abu Dhabi New Corniche Road ............................................................................ 297
3.2.1 Project Description ....................................................................................... 297
3.2.2 Ground Conditions ....................................................................................... 297
3.2.3 Application of Dynamic Compaction and the Challenges ............................ 298
3.2.4 Testing and Verification ............................................................................... 301
3.2.5 Lessons and Conclusion ............................................................................... 307
3.3 Al Quo’a New Township ....................................................................................... 308
3.3.1 Project Description ....................................................................................... 308
3.3.2 Ground Conditions ....................................................................................... 309
3.3.3 Development of an Alternative Solution ..................................................... 310
3.3.4 Application of Dynamic Compaction and the Challenges ............................ 312
3.3.5 Testing and Verification ............................................................................... 317
3.3.6 Lessons and Conclusion ............................................................................... 319
3.4 King Abdulla University of Science and Technology ............................................ 324
3.4.1 Project Description ....................................................................................... 324
3.4.2 Ground Conditions ....................................................................................... 324
3.4.3 Development of the Ground Improvement Solution ................................... 330
3.4.4 Application of Dynamic Compaction and Dynamic Replacement ............... 332
3.4.5 Testing and Verification ............................................................................... 336
3.4.6 Lessons and Conclusions .............................................................................. 339
3.5 Al Falah Community ............................................................................................. 341
3.5.1 Project Description ....................................................................................... 341
3.5.2 Ground Conditions ....................................................................................... 341
3.5.3 Application of Dynamic Compaction and the Challenges ............................ 343
3.5.4 Testing and Verification ............................................................................... 345
3.5.4.1 Calibration ................................................................................................ 345
xiii
3.5.4.2 Final Testing ............................................................................................. 359
3.5.5 Lessons and Conclusion ............................................................................... 360
3.6 Marjan Island Road Corridor ................................................................................ 362
3.6.1 Project Description ....................................................................................... 362
3.6.2 Ground Conditions ....................................................................................... 362
3.6.3 Development of Solution and Application of Dynamic Compaction ........... 367
3.6.4 Testing and Verification ............................................................................... 368
3.6.4.1 Calibration ................................................................................................ 368
3.6.4.2 Final Testing ............................................................................................. 379
3.6.5 Lessons and Conclusions .............................................................................. 381
3.7 Al Nakhilat Ship Repair Yard ................................................................................ 383
3.7.1 Project Description ....................................................................................... 383
3.7.2 Ground Conditions ....................................................................................... 383
3.7.3 Design and Acceptance Criteria ................................................................... 384
3.7.4 Development of Solution and Application of Dynamic Compaction ........... 385
3.7.5 Testing and Verification ............................................................................... 386
3.7.5.1 Heave and Penetration Test ..................................................................... 386
3.7.5.2 CPT ........................................................................................................... 392
3.7.5.3 CPT‐PMT Correlation for Carbonate Sand ............................................... 396
3.7.5.4 Zone Load Testing .................................................................................... 400
3.7.6 Lessons and Conclusion ............................................................................... 405
3.8 Abu Dhabi Ritz‐Carlton Hotel ............................................................................... 407
3.8.1 Project Description ....................................................................................... 407
3.8.2 Ground Conditions ....................................................................................... 407
3.8.3 Development of Solution and Application of Dynamic Replacement ......... 410
3.8.4 Testing and Verification ............................................................................... 415
xiv
3.8.5 Lessons and Conclusion ............................................................................... 418
3.9 Al Jazira Steel Pipe Factory ................................................................................... 420
3.9.1 Project Description ....................................................................................... 420
3.9.2 Ground Conditions ....................................................................................... 421
3.9.3 Development of Solution and Application of Dynamic Replacement ......... 422
3.9.4 Testing and Verification ............................................................................... 424
3.9.4.1 DR Calibration .......................................................................................... 424
3.9.4.2 Final Testing ............................................................................................. 428
3.9.5 Lessons and Conclusion ............................................................................... 430
3.10 Reem Island Causeway ......................................................................................... 432
3.10.1 Project Description ....................................................................................... 432
3.10.2 Ground Conditions and Fill Description ....................................................... 433
3.10.3 Development of Solution and Application of Dynamic Compaction ........... 435
3.10.3.1 Design and Acceptance Criteria ........................................................... 435
3.10.3.2 Application of Dynamic Compaction .................................................... 436
3.10.4 Testing and Verification ............................................................................... 437
3.10.4.1 DC Calibration ...................................................................................... 437
3.10.4.2 PMT and Acceptance ........................................................................... 439
3.10.5 Lessons and Conclusion ............................................................................... 443
3.11 Ras Laffan Heavy Fuel Oil Bunkering Facility ....................................................... 446
3.11.1 Introduction ................................................................................................. 446
3.11.2 Ground Conditions ....................................................................................... 447
3.11.3 Development of Solution and Application of Dynamic Compaction ........... 449
3.11.3.1 Preliminary Design ............................................................................... 450
3.11.3.2 Application of Dynamic Compaction .................................................... 451
3.11.4 Testing and Verification ............................................................................... 453
xv
3.11.4.1 DC Calibration ...................................................................................... 453
3.11.4.2 PMT and Acceptance ........................................................................... 456
3.11.5 Lessons and Conclusion ............................................................................... 459
3.12 Palm Jumeira Sewage Treatment Plant Tanks ..................................................... 461
3.12.1 Project Description ....................................................................................... 461
3.12.2 Preliminary Geotechnical Investigation ....................................................... 463
3.12.3 Development of Solution ............................................................................. 463
3.12.4 Supplementary Geotechnical Investigation in Lot A‐A ................................ 465
3.12.5 Ground Improvement in Lot A‐A & Verification of Results .......................... 465
3.12.5.1 Dynamic Surcharging ........................................................................... 465
3.12.5.2 Dynamic Compaction of the Tank Foundation .................................... 470
3.12.5.3 Post Improvement Verification in Lot A‐A ........................................... 473
3.12.6 Ground Improvement in Lot G‐G & Verification of Results ......................... 477
3.12.7 Lessons and Conclusion ............................................................................... 479
3.13 Al Khaleej Raw Sugar Silos ................................................................................... 482
3.13.1 Introduction ................................................................................................. 482
3.13.2 Ground conditions ....................................................................................... 482
3.13.3 The Foundation Solution: Dynamic Replacement ....................................... 483
3.13.4 Testing and Verification ............................................................................... 492
3.13.5 Lessons and Conclusion ............................................................................... 494
3.14 Trail for Quay Expansion in Southeast Asia ......................................................... 495
3.14.1 Introduction ................................................................................................. 495
3.14.2 Soil Softening ............................................................................................... 495
3.14.3 The Solution: Offshore Dynamic Replacement ............................................ 496
3.14.4 Testing and Verification ............................................................................... 498
3.14.5 Lessons and Conclusion ............................................................................... 501
3.15 Palm Jumeira Trial ................................................................................................ 503
xvi
3.15.1 Project Description and Ground Conditions ................................................ 503
3.15.2 Ground Conditions ....................................................................................... 504
3.15.3 Dynamic Compaction Trial ........................................................................... 504
3.15.4 Testing and Verification ............................................................................... 505
3.15.5 PMT‐CPT Correlations for Calcareous Sand ................................................. 508
3.15.6 Lessons and Conclusion ............................................................................... 513
3.16 Dynamic Compaction Vibration Monitoring ........................................................ 515
3.16.1 Fujairah Desalination Plant Phase 2 ............................................................. 515
3.16.1.1 Project Description and Ground Conditions ........................................ 515
3.16.1.2 Vibration Monitoring ........................................................................... 517
3.16.1.2.1 Improving PPV Estimation Accuracy ................................................ 519
3.16.2 Medina A’Zarqa (Blue City) .......................................................................... 520
3.16.2.1 Project Description and Ground Conditions ........................................ 520
3.16.2.2 Vibration Monitoring ........................................................................... 523
3.16.3 Um Quwain Marina ...................................................................................... 527
3.16.3.1 Project Description and Ground Conditions ........................................ 527
3.16.3.2 Vibration Monitoring ........................................................................... 528
3.16.4 Comparison of Vibrations Generated by Dynamic Compaction and Vibro
Compaction .................................................................................................................. 532
3.16.4.1 Previous Research on Vibro Compaction Generated Vibrations ......... 532
3.16.4.2 This Study: Vibration Monitoring of Vibro Compaction on Palm Jumeira
533
3.16.5 Lessons and Conclusion ............................................................................... 536
3.17 Predicting PLM and EM from Dynamic Compaction Induced Subsidence .............. 539
3.17.1 Introduction ................................................................................................. 539
3.17.2 The Relation between Induced Strain and Subsidence with PLM and EM ..... 540
xvii
3.17.3 Strain Distribution in Dynamic Compaction ................................................. 541
3.17.4 Developing the Procedure ........................................................................... 542
3.17.5 Verification ................................................................................................... 546
3.17.6 Conclusions .................................................................................................. 549
4 Conclusion .................................................................................................................... 550
References ........................................................................................................................... 560
Appendix: First Page of Publications .................................................................................... 584
xviii
Notations
Notation Definition
(PLM)i limit pressure before soil improvement
(PLM)j limit pressure after soil improvement
(Q"1)δ Ultimate footing inclined load
(Q'1)δ Ultimate footing inclined load
A Pounder base area
a Radius of contact
Horizontal radius of semi‐prolate spheriod
Acceleration
size (or diameter) of pile caps
Slope of the volume versus pressure calibration plot in PMT
percentage of strain induced for doubling PLM
A' Effective footing area
ac Area replacement ratio
Ac Pile cap or DR column area
AE Applied energy
AE Area of cell unit
amax Peak ground acceleration (PGA)
Ary Vibration reduction ratio
As Area of soil in cell unit
Aw Equivalent wall area
B Bulk modulus of elasticity of the mixture
Footing (loading) width
b Vertical radius of semi‐prolate spheroid
Pile cap width
Percentage of strain induced for doubling EM
B' Effective footing width
B" Effective footing width
B0 Reference footing width of 0.6 m
Ba Bulk modulus of elasticity of air
Bs Bulk modulus of solid particles
xix
Notation Definition
Bw Bulk modulus of elasticity of water
c Coefficient of depth of improvement
Cohesion of soil
Cd Error propogation factor
Cdmax Error propogation factor
Cdmin Error propogation factor
C0 Experimental coefficient for correlation of relative density to CPT cone
resistance
c1 Speed damping factor for depth of improvement
C1 Experimental coefficient for correlation of relative density to CPT cone
resistance
c2 Stratigraphic coefficient for depth of improvement
C2 Experimental coefficient for correlation of relative density to CPT cone
resistance
Ca Arching coefficient
cc Cohesionof DR column
Cc Compression index beyond the critical pressure
CD A parameter that reflects any densification by such disturbance mechanisms
as vibration and blasting, which are not related to the ideally static increase
in effective vertical stress.
cd Depth constant for depth of improvement
ce Energy constant for depth of improvement
ci Cohesion of segment i
Coc Over consolidation coefficient
CPT Cone penetration test
cR Proportion coefficient
CRRMw Cyclic resistance ratio for earthquake with magnitude Mw
cs Subsoil (soft soil) cohesion
CSR Cyclic shear ratio
CSRi Reduced cyclic shear ratio
ct Cohesion of trench
cu Undrained shear strength
xx
Notation Definition
Cα Secondary compression index
D Depth of improvement or influence
d Pounder or column diameter
Distance between pounder and monitoring point
D50 Mean particle diameter
DB Diameter of base of crater
dc Column diameter
Dc Crater depth
Constrained modulus of the DR column
DC Dynamic compaction
Dd Relative density
Df Depth of footing
di Inside diameter of the heavy duty steel casing or pipe in the pressuremeter
test
Di Interim depth of improvement
DR Dynamic replacement
Ds Constrained modulus of the soil
DT Diameter of top of crater
E Elastic (Young) modulus
Impact energy= WH
Arching efficacy or the portion of load that is supported by the piles
e Void ratio
Load eccentricity
E+ Spherical or compression modulus
Ec Elastic (Young) modulus of column
Weighted value of EM immediately below the footing
Ed Harmonic mean of EM in all layers down to the depth of 8B
Ei Harmonic mean of moduli of layer i
EM Menard modulus
EM+ Menard reload modulus
emax Maximum index void ratio or the reference void ratio of a soil at the
minimum index density/unit weight
xxi
Notation Definition
emin Minimum index void ratio or the reference void ratio of a soil at the
maximum index density/unit weight
Eoed Oedometric modulus
Es Elastic (Young) modulus of in‐situ soil
Ey Young modulus
F Safety factor for bearing capacity
f Frequency
Predominant frequency
f(z) Improvement factor at elevation z
f1 Maximum improvement factor observed at ground level
f2 Improvement factor obtained at the depth of influence of dynamic
compaction
fB Bottom factor
Fc Fines content
fC Cushion (crushed stone platform) factor
fM Material (soil) factor
fshell Factor for correlating relative density of calcareous sand to silica sand
g Gravity acceleration
G Shear modulus
Gc Shear modulus of the column
Gceq Equivalent shear modulus
Gr Ratio of column’s to soil’s shear moduli
Gs Specific gravity of the solid particles
Shear modulus of the soil
H Pounder drop height
Height of embankment
h test spacing in borehole
Hc Thickness of compressible layer
hc Critical depth of embedment
HDR High energy dynamic replacement
he Equivalent foundation depth
hg Arching height
xxii
Notation Definition
hk Pressuremeter test spacing in a borehole
Hp Depth of probe below the control unit
HT Isolation trench depth
Energy intensity= applied energy per unit area
I Improvement factor
ie Bearing capacity reduction factor for eccentric loading
iδ Bearing capacity reduction factor for inclined loading
iδe Bearing capacity reduction factor for inclined and eccentric loading
iδβ' Bearing capacity reduction factor for inclined loading near a slope
A bearing factor varying from 0.8 to 9 according to the embedment, the
k shape of the foundation level after construction
K Peak particle velocity intercept
Compression or volume change (bulk) modulus
K0 Earth pressure coefficient
Ratio of effective horizontal to vertical stresses when the soil is over
consolidated
Ka Active earth pressure coefficient
KG Shear reduction factor
KM Compression modulus
Ko Coefficient of horizontal earth pressure at rest
Ratio of effective horizontal to vertical stresses when the soil is normally
KONC consolidated
KP Passive earth pressure coefficient
Ks Ratio of vertical to horizontal stresses in the soil prior to loading
L Distance from edge of crater to where heave reduces to zero
Foundation length
l Length of PMT measuring cell
li Length of segment i
LR Wave length
LT Isolation trench length
m Pounder mass
Load distribution ratio
xxiii
Notation Definition
Number of measurements
Number of pressuremeter tests in the borehole within the improvement
zone
M Ratio of net limit pressure at depths of h and depth of h+B
MHHW Mean High High Water
MLLW Mean Low Low Water
MPM Menard pressuremeter
N Number of blows
Number of blows required to reach maximum depth of improvement
SPT blows per 0.3 m
Resistance of the foundation against normal loads
n Porosity
Vibration attenuation rate
Stress distribution ratio
Number of times PMT limit pressure has doubled
N60 SPT blow counts corrected to 60% efficiency
SPT blow count corresponding to an energy rod ratio of about 78% from the
N78 theoretical free‐fall energy
Nc Bearing capacity coefficient
NGL Natural ground level
Ni Interim number of blows
N q Bearing capacity coefficient
N γ Bearing capacity coefficient
OCR Over consolidation ratio
P Normal load
Number of passes
Pressure exerted by the PMT probe on the soil
Power
p Surface load stress on the soil
P*LM Net PMT limit pressure
p’ Effective normal stress
Pc Pressure loss correction in pressuremeter test
xxiv
Notation Definition
Pf Creep or end of pseudo elastic phase pressure in the pressuremeter test
PGA Peak ground acceleration (PGA)
Pi Pressure in pressuremeter test for reaching probe contacting borehole wall
PLM PMT limit pressure
PLM_k Pre‐improvement PMT limit pressure
PLM_min Minimum pre‐improvement PMT limit pressure
PMT Pressuremeter test
Po Total at rest horizontal earth pressure at the test level (at the time of the
test)
PPV Peak particle velocity
Pressure at which changes of volume change to pressure change; i.e.,
pr d(ΔV)/d(Δp), is a minimum in PMT
PR Pressure reading on the PMT control unit
Pu Pressure reading in PMT corresponding to a volume increase equal to the
initial volume of the borehole
PVS Pseudo vector sum
Pδ Hydrostatic pressure between the PMT control unit and probe
Q Footing load
q Shear Stress
Design normal pressure applied on the footing
Q' Footing load
Q" Footing load
Q"1 Ultimate footing vertical load
q* Net ultimate bearing capacity
q*a Net allowable bearing capacity
q*c Net CPT cone resistance
Q'1 Ultimate footing vertical load
qa Allowable bearing capacity
qc CPT cone resistance
qcNC CPT cone resistance of normally consolidated sand
qcOC CPT cone resistance of over consolidated sand
qcR CPT cone resistance at reference time
xxv
Notation Definition
qo Total overburden pressure at the periphery of the foundation level after
construction
Total vertical stress
qu Ultimate bearing capacity
qult Ultimate external pressure on footing width
r Distance of the wave front from the source
Distance of any point on the contact surface of two spheres
Pounder radius
(PLM)j/(PLM)i
R Sphere radius
Half width of foundation
R2 Coefficient of determination
RB Radius of base of crater
rd Depth factor in liquefaction evaluation
The reference dry density/unit weight of a soil in the densest state of
compactness that can be attained using a standard laboratory compaction
dmax procedure that minimizes particle segregation and breakdown
The reference dry density/unit weight of a soil in a standard state of
compactness at which it can be placed using a standard laboratory
dmin procedure which prevents bulking and minimizes particle segregation
RIC Rapid Impact Compactor
RL Reduced level
RT Distance from the source of excitation to the isolation trench
RT Radius of top of crater
ru excess pore water pressure to initial effective vertical stress ratio
S Degree of saturation
Shearing force
s Spacing between adjacent piles or columns
Displacement, settlement
Sd Standard deviation for d
Sdmax Standard deviation for dmax
Sdmin Standard deviation for dmin
xxvi
Notation Definition
sd Diagonal spacing between the piles or columns
SDd Standard deviation for Dd
Average heave (or settlement) for markers located at the same radius from
Si the centre of the print
Smax Assumed maximum heave at edge of crater
Creep or self‐weight settlement after n years
sn Settlement at measurement number n
SPT Standard Penetration Test
t Time
T wave period
ton Metric tonne
tR Reference time
TVS True vector sum
u Displacement of an element in the x direction
Pore pressure
ů Particle velocity
V Total volume
Corrected increase in volume of the measuring portion of the PMT probe
v Wave propagation velocity
Particle velocity
Pounder velocity
Va Air volume
vc Phase or longitudinal wave propagation velocity
VC Volume loss correction in pressuremeter test
Vcrater Volume of crater
Vheave Volume of heave
Vi Intercept of volume calibration line in PMT
Corrected volume reading at the pressure where the PMT probe makes
contact with the borehole
Vm Corrected volume reading in the centre portion of the ΔV volume increase
in the pressuremeter test
vmix Compression wave propagation velocity in the solid‐air‐water mixture
xxvii
Notation Definition
Vo Zero volume reading
vp Primary wave, P‐wave, or compression wave velocity
vR Rayleigh wave velocity
Vr Injected volume at the end of each pressure increment in PMT
Volume at which changes of volume change to pressure change; i.e.,
d(ΔV)/d(Δp), is a minimum in PMT
VR Volume of reading on PMT's readout device
vs Secondary wave, S‐wave or shear wave velocity
Vs Solid volume
Vv Void volume
VW Water volume
w Amplitude of the vertical component of the R‐wave at distance r from the
source
W Pounder weight
ws Uniformly distributed surcharge loading
WT Isolation trench width
y Stress distribution factor such as Boussinesq or Westergaard
z Level of point
Depth from ground surface
Zn Wave amplitude in the nth cycle
α Coefficient of attenuation (in terms of distance‐1)
Rheological factor for calculating settlement
αB Number of blows ratio
αz Depth of improvement ratio
β Angle between the footing and excavation level
Coefficient relating net limit pressure to undrained shear strength
β(F) A coefficient that is a function of the safety factor for bearing capacity
γ Unit weight
γa Unit weight of air
γc DR column density
Shear strain in the column
γd Dry density/unit weight of a soil deposit or fill at the given void ratio
xxviii
Notation Definition
γdmax The reference dry density/unit weight of a soil in a standard state of
compactness at which it can be placed using a standard laboratory
procedure which prevents bulking and minimizes particle segregation
γdmin The reference dry density/unit weight of a soil in the densest state of
compactness that can be attained using a standard laboratory compaction
procedure that minimizes particle segregation and breakdown
γeq Equivalent density
γs Subsoil (soft soil) density
Shear strain in the soil
γt Trench unit weight
γw Unit weight of water
δ Load inclination angle from the vertical
δ Logarithmic wave damping coefficient
ΔDd/Dd Relative deviation in relative density
Δe Change of void ratio
ΔNf Fines content correction
ΔP Corrected pressure increase in the centre part of the straight line portion of
the pressure‐volume curve of the pressuremeter test
Δqc Vertical stress increase at the compressible layer
δt Unit weight of test liquid in PMT
ΔV Corrected volume increase in the centre part of the straight line portion of
the pressure‐volume curve, corresponding to ΔP pressure increase
Δφ Change of internal friction angle
Δφb Change of internal friction angle beneath the pounder
Δφc Change of internal friction angle at the grid centre
Δφm Change of internal friction angleat the middle of the grid side
ε Strain
ε(z) Strain at depth z
εDC_k Dynamic compaction induced strain in layer (test spacing in borehole) k
εo_k The pre‐strain for each layer that demonstrates the strain difference of that
layer compared to the lowest PLM value
εR Rayleigh distribution strain
xxix
Notation Definition
ζ Reduction between the centres of the two spheres
ζ Lehr’s damping coefficient
η Shear displacement
κ Deformation modulus reduction factor
λ Lamé’s constant
Wave length
λc Shape factor for calculation of compressive component of settlement
λd Shape factor for calculation of deviatoric component of settlement
μ Coefficient of friction
ν Poisson ratio
νs Soil Poisson ratio
ϖ Average water content of soil
ρ Mass density
ρd Dry density of soil
ρd Dry density/unit weight of a soil deposit or fill at the given void ratio
ρtot Total mass density
σ Normal stress
Total normal stress on the failure plane
Scale parameter of the Rayleigh distribution
σc Stress in DR column
σ'c Vertical stress on the pile caps
σ'c‐H Horizontal stress in DR column
σ'c‐V Vertical stress in DR column
σh Lateral confining stress on the cylindrical surface of the DR column
σrL Limiting (lateral) stress
σro Total (initial) in‐situ lateral stress
σs Stress in soil
σu Upper yield value of the soil
σv Total vertical stress
σ'v Effective vertical stress
σv_max Maximum vertical stress below footing
σv_min Minimum vertical stress below footing
xxx
Notation Definition
σvo Total (initial) vertical stress at the level of the footing base
σ'vo Effective initial overburden pressure
σx Axial stress
Average normal stress
σ'zo Uniform stress at subsoil level
τ Shear strength
τave Input (equivalent average) shear stress
τc Shear stress in the column
τs Shear stress in the soil
φ Internal friction angle
φ' Effective internal friction angle
φc DR column internal friction angle
φs Subsoil (soft soil) internal friction angle
φt Trench friction angle
ω Undamped natural frequency
ωd Damped natural frequency
xxxi
List of Tables
Table 2‐1: Attenuation coefficient α for different soil descriptions and vibration frequencies
(Woods and Jedele, 1985)...................................................................................................... 22
Table 2‐2: Suitability of Deposits for Dynamic Compaction (Lukas, 1986) ........................... 59
Table 2‐3: Coefficient of depth of improvement for different soil types (Lukas, 1986) ........ 65
Table 2‐4: Predicted depth of improvement for different soil deposits (Luongo, 1992) ...... 66
Table 2‐5: Anticipated relative improvement for different types of soils (Lukas, 1986) ....... 94
Table 2‐6: Post dynamic compaction upper bound test values (Lukas, 1986) ..................... 94
Table 2‐7: Curve fitting coefficients for Equations (Poran and Rodriguez, 1992) ................. 98
Table 2‐8: Applied energy guidelines (Lukas, 1995) ............................................................ 110
Table 2‐9: Arching coefficient .............................................................................................. 119
Table 2‐10: Ranges of parameters ....................................................................................... 143
Table 2‐11: Safe levels of blasting vibrations for residential type structures, from USBM RI
8507 (Siskind et al., 1980) .................................................................................................... 166
Table 2‐12: Maximum allowable peak particle velocity for ground vibration (OSM, 1983) 168
Table 2‐13: Transient vibration guide for cosmetic damage (British Standards Institution,
1993) .................................................................................................................................... 174
Table 2‐14: Maximum BS 6472 Curve Numbers ‐ Human Comfort (8 to 80 Hz) ................. 175
Table 2‐15: Recommended ground vibration limits for control of damage (AS 2187.2) .... 176
Table 2‐16: Ground vibration limits for human comfort chosen by some regulatory authorities
(AS 2187.2) ........................................................................................................................... 177
Table 2‐17: typical probe and borehole diameters (ASTM, 2007, Centre D'Etudes Menard,
1975) .................................................................................................................................... 203
Table 2‐18: typical ranges of EM and PLM for some main types of soil and ground (Centre
D'Etudes Menard, 1975). ..................................................................................................... 211
Table 2‐19: Soil categories for determination of bearing factor (Centre D'Etudes Menard,
1975) .................................................................................................................................... 214
Table 2‐20: hc/R for different soil categories and foundation types (Centre D'Etudes Menard,
1975) .................................................................................................................................... 214
Table 2‐21: bearing factor for spread foundations, ENV 1997‐3 (European Standard, 2000)
............................................................................................................................................. 215
Table 2‐22: Rheological factor for different soils and ratios of EM/PLM (Centre D'Etudes
Menard, 1975) ..................................................................................................................... 224
xxxii
Table 2‐23: Rheological factor for rock (Centre D'Etudes Menard, 1975) ........................... 224
Table 2‐24: Shape factors λd and λc for different footing length to width ratios (Centre
D'Etudes Menard, 1975) ...................................................................................................... 225
Table 2‐25: P*LM (kPa) for self‐bearing condition (Centre D'Etudes Menard, 1975) ........... 230
Table 2‐26: q*c/P*LM for different soil types according to Baguelin et al. (1978) ................ 235
Table 2‐27: Correlation between PMT and CPT (Briaud et al., 1985) ................................. 236
Table 2‐28: ΔNf – fines content correlation (Tokimatsu and Yoshimi, 1983) ...................... 274
Table 2‐29: The ratio of qc measured in Villet and Mitchell (1981) to qc predicted by
Schmertmann (1978) (after Villet and Mitchell) .................................................................. 278
Table 3‐1: Case studies used for documentation, analysis and comparison with previous
publications .......................................................................................................................... 293
Table 3‐2: SPT blow counts acceptance criteria based on correlation to 80% relative density
............................................................................................................................................. 298
Table 3‐3: Table 1. Design criteria ....................................................................................... 312
Table 3‐4: Acceptance criteria for villa areas ....................................................................... 312
Table 3‐5: Acceptance criteria for non‐villa areas ............................................................... 312
Table 3‐6: Required treatment energy ................................................................................ 313
Table 3‐7: Ground conditions at KAUST ............................................................................... 326
Table 3‐8: Quick verification criteria .................................................................................... 345
Table 3‐9: Penetration test .................................................................................................. 358
Table 3‐10: Soil profiles in areas DDR4 and DDR6 ............................................................... 393
Table 3‐11: Settlements of the Dd design curve and the soil profiles subject to a load of 4,000
kN ......................................................................................................................................... 395
Table 3‐12: Summary of the ground profile ........................................................................ 409
Table 3‐13: Summary of design criteria ............................................................................... 422
Table 3‐14: Ground profile of reclamation area before ground improvement ................... 434
Table 3‐15: Dynamic compaction treatment area ............................................................... 450
Table 3‐16: parameters used in the numerical analysis model ........................................... 452
Table 3‐17: SPT blow counts and fines content in the preliminary and supplementary
boreholes ............................................................................................................................. 466
Table 3‐18: Ground settlement during static and dynamic surcharging ............................. 468
Table 3‐19: Description of prints ......................................................................................... 471
Table 3‐20: The principal characteristics and dimensions for the model ............................ 476
Table 3‐21: Ground layers and Ey values before ground improvement .............................. 483
Table 3‐22: Design parameters for numerical analyses ....................................................... 484
xxxiii
Table 3‐23: PMT Schedule ................................................................................................... 493
Table 3‐24: Pre‐treatment and post treatment PMT results ............................................... 500
Table 3‐25: Equivalent parameters for finite element model ............................................. 501
Table 3‐26: Results of vibration monitoring ........................................................................ 518
Table 3‐27: Summary of ground conditions ........................................................................ 522
Table 3‐28: Vibration monitoring summary for the case without the trench ..................... 529
Table 3‐29: Vibration monitoring summary for the case with the isolation trench ............ 531
Table 3‐30: Calculation for prediction of post dynamic compaction limit pressures .......... 547
xxxiv
List of Figures
Figure 2‐1: Summary of literature review topics presented in the thesis ............................. 10
Figure 2‐2: Louis Ménard (1931‐1978) .................................................................................. 11
Figure 2‐3: Dynamic compaction was used for the first time ever in 1969 for the ground
improvement works of a building construction site in Mandelieu‐la‐Napoule, France
(Communication Department of Menard, 2007) ................................................................... 12
Figure 2‐4: The Giga Machine holds the world record for lifting a 172 ton pounder at Nice
International Airport dynamic compaction project (Communication Department of Menard,
2007) ...................................................................................................................................... 13
Figure 2‐5: Relations between Poisson’s ratio and P, S and R‐wave velocities in an elastic half
space medium (Richart, 1962) ............................................................................................... 17
Figure 2‐6: Amplitude ratio versus dimensionless depth for R‐waves (Richart et al., 1970) 18
Figure 2‐7: Wave system from surface point source in ideal medium (Lamb, 1904) ............ 18
Figure 2‐8: Distribution of displacement waves from a circular footing on a homogeneous,
isotropic, elastic half space (Woods, 1968) ........................................................................... 19
Figure 2‐9: Attenuation of surface wave with distance from source of steady‐state excitation
(Richart et al., 1970) ............................................................................................................... 21
Figure 2‐10: Partition of elastic wave at interface between two elastic media (Richart et al.,
1970) ...................................................................................................................................... 23
Figure 2‐11: Example of amplitude ratio versus incident angle for P‐wave (McCamy et al.,
1962) ...................................................................................................................................... 27
Figure 2‐12: Example of amplitude ratio versus. incident angle for S‐wave (McCamy et al.,
1962) ...................................................................................................................................... 27
Figure 2‐13: Multiple wave reflections and refractions in a layered half‐space (Richart et al.,
1970) ...................................................................................................................................... 28
Figure 2‐14: Displacement and rearrangement of the soil grains in a denser configuration 31
Figure 2‐15: Modes of regular packing of equal spheres (Richart et al., 1970) ..................... 33
Figure 2‐16: Behaviour of equal spheres in contact. (a) Spheres just touching, (b) deformation
by normal force, (c) shearing forces between particles in cubic packing, (d) lateral
deformation by shearing forces (Richart et al., 1970). .......................................................... 34
Figure 2‐17: Hysteresis loop formed by shearing force‐displacement relations for two equal
spheres pressed together and subjected to shearing forces (Richart et al., 1970). .............. 35
xxxv
Figure 2‐18: (a) Distribution of relative density variations and (b) variation of relative density
versus depth under the pounder centre for different number of impacts (Hajialilue‐Bonab
and Rezaei, 2009) ................................................................................................................... 40
Figure 2‐19: Two‐dimensional axisymmetric mesh (Gu and Lee, 2002) ................................ 41
Figure 2‐20: Computed p’‐q (mean effective normal versus shear) stress path at point A
located 2 m below ground surface and 0∙1 m from centreline (Gu and Lee, 2002) .............. 42
Figure 2‐21: Computed p’‐q stress paths at point B located 6 m below ground surface and 0∙1
m from centreline (Gu and Lee, 2002) ................................................................................... 44
Figure 2‐22: Computed p’‐q stress paths at point C located 2 m below ground surface and 4
m from centreline (Gu and Lee, 2002) ................................................................................... 44
Figure 2‐23: Effects of waves in the densification of loose sands (Gambin, 1997) ............... 45
Figure 2‐24: Appearance of water fountains from the fissures due to dynamic compaction47
Figure 2‐25: Comparison of the classical and dynamic theories of consolidation (Menard and
Broise, 1975) .......................................................................................................................... 49
Figure 2‐26: (a) Changes in the soil conditions after one dynamic compaction pass, and (b)
changes in the soil conditions after a multiple number of DC pass (Menard and Broise, 1975)
............................................................................................................................................... 50
Figure 2‐27: The mechanism of load transfer to the improved DR – in‐situ soil system ...... 51
Figure 2‐28: Consolidation stresses on the clay before the test, and consolidation stresses on
the clay and the loading stresses on the column during the test (Hughes and Withers, 1974)
............................................................................................................................................... 52
Figure 2‐29: (a) Vertical displacement within the column versus depth (b) Radial displacement
at the edge of the column / initial column radius versus depth (Hughes and Withers, 1974)
............................................................................................................................................... 53
Figure 2‐30: Column failure mechanism in non‐homogeneous cohesive soil: (a) Soft layer at
surface – bulging or shear failure, (b) Thin very soft layer – contained local bulge, (c) Thick
very soft layer – bulging failure (Barksdale and Bachus, 1983) ............................................. 54
Figure 2‐31: Tracing of the superposition of markers due to after loading. A selection of the
markers are joined by arrows (Hughes and Withers, 1974) .................................................. 54
Figure 2‐32: Failure Mechanisms of a single (DR) column in a homogeneous soft layer: (a)
Long columns with firm or floating support – bulging failure, (b) Short column with rigid base
– shear failure, (c) Short floating column ‐ punching failure (Barksdale and Bachus, 1983) . 55
Figure 2‐33: Effect of compacted granular working platform on the behaviour of the column:
(a) Construction of column without compacted mat, (b) Construction of compacted granular
xxxvi
mat prior to the construction of the column, (c) Construction of the column prior to the
construction of the compacted granular mat (Barksdale and Bachus, 1983) ....................... 56
Figure 2‐34: Different loading types applied to columns: (a) Rigid footing loading, (b) Plate
load test, (c) Embankment loading (Barksdale and Bachus, 1983)........................................ 57
Figure 2‐35: Categorising of soil for dynamic compaction suitability (Lukas, 1986) ............. 58
Figure 2‐36: The process of dynamic replacement ................................................................ 60
Figure 2‐37: The effect of fines content on measured vertical stress(Chen and Lin, 2002) .. 63
Figure 2‐38: Depth of influence as a function of impact energy (Mitchell, 1981) ................. 64
Figure 2‐39: Trend between apparent maximum depth of influence and energy/blow (Mayne
et al., 1984) ............................................................................................................................ 64
Figure 2‐40: Correlation between depth of influence and crater depth (Hajialilue‐Bonab and
Rezaei, 2009) .......................................................................................................................... 69
Figure 2‐41: Variation of c with impact energy per drop (Ghassemi et al., 2009b) .............. 71
Figure 2‐42: Definition of depth and radius of compacted zone for any relative density
contour line (Oshima and Takada, 1997) ............................................................................... 74
Figure 2‐43: Contours of increase in relative density (%) for (a) W= 20 t, H= 20 m, (b) W=40 t,
H= 10 m, and (c) W= 80 t, H= 5 m (Gu and Lee, 2002) ........................................................... 75
Figure 2‐44: Depth of improvement with constant impact energy and varying momentum
(Hajialilue‐Bonab and Rezaei, 2009) ...................................................................................... 76
Figure 2‐45: Relationship between normalised depth of improvement and number of drops
for different energy intensities .............................................................................................. 77
Figure 2‐46: Interim depth of improvement ratio (Ghassemi et al., 2009a) ......................... 78
Figure 2‐47: Increase in relative density along centreline for various tamper radii (Gu and Lee,
2002) ...................................................................................................................................... 79
Figure 2‐48: Relation between tamper radius and depth of improvement (Ghassemi et al.,
2009a) .................................................................................................................................... 79
Figure 2‐49: Side view of cylindrical pounders (Feng et al., 2000) ........................................ 80
Figure 2‐50: Mini cone penetration resistance at below centre of pounder with fines content
(a) 3.4%, (b) 8% and (c) 14% (Feng and Ke, 2005) ................................................................. 82
Figure 2‐51: Mini cone penetration resistance for test series with 14% fines content.
Distances of impact points from the centre of pounder were (a) 1 pounder diameter, (b) 2
pounder diameters, and (c) 3 pounder diameters (Feng and Ke, 2005) ............................... 83
Figure 2‐52: Normalised crater measurements (Mayne et al., 1984) ................................... 84
Figure 2‐53: Application of Dynamic Compaction in a grid ................................................... 84
xxxvii
Figure 2‐54: Relationship between ratio of increase of internal friction angles and normalised
depths (Chow et al., 1994) ..................................................................................................... 86
Figure 2‐55: Increase in internal friction angles ratio for (a) grid centre point and (b) midpoint
of grid side (Chow et al., 1994) .............................................................................................. 87
Figure 2‐56: Displacement vectors after 12 blows when spacing= 5d (Hajialilue‐Bonab and
Zare, 2014) ............................................................................................................................. 89
Figure 2‐57: Displacement angle at different normalised depths and distances (Hajialilue‐
Bonab and Zare, 2014) ........................................................................................................... 89
Figure 2‐58: Displacement contours of 0.05d for different number of drops when pounder
spacing= 3d (Hajialilue‐Bonab and Zare, 2014) ..................................................................... 90
Figure 2‐59: Displacement contours for (a) pounder spacing= 6d and (b) pounder spacing=
3d (Hajialilue‐Bonab and Zare, 2014) .................................................................................... 91
Figure 2‐60: Strain diagrams after 12 blows with pounder spacing being equal to (a) 6d, (b)
5d, (c) 4d and (d) 3d (Hajialilue‐Bonab and Zare, 2014) ........................................................ 92
Figure 2‐61: Increase of relative density at the between prints centreline (Hajialilue‐Bonab
and Zare, 2014) ...................................................................................................................... 93
Figure 2‐62: (a) Initial stages of pounding, and (b) after densification, including the ironing
pass (Lukas, 1986) .................................................................................................................. 95
Figure 2‐63: Lateral movement (a) 3 m and (b) 6 m away from pounder drop point (Lukas,
1986) ...................................................................................................................................... 96
Figure 2‐64: Normalised (to pounder diameter) displacement vectors after 10th impact for
(a) vertical lines and (b) horizontal lines of soil (Hajialilue‐Bonab and Rezaei, 2009) ........... 97
Figure 2‐65: Experimental contour of plastic volumetric strains, modified from Poran and
Rodriguez (1992) .................................................................................................................... 97
Figure 2‐66: A semi‐prolate spheroid can be used to approximate the volume of density
contours (Poran and Rodriguez, 1992) .................................................................................. 97
Figure 2‐67: Comparison of relative density contours presented by Heh (1990) and Hajialilue‐
Bonab and Zare (2014) ........................................................................................................... 99
Figure 2‐68: Test model for studying seismic response of shallow foundations in liquefiable
sand (Liu and Dorby, 1997) .................................................................................................. 104
Figure 2‐69: Foundation settlement versus normalised compaction depth (Liu and Dorby,
1997) .................................................................................................................................... 105
Figure 2‐70: Footing settlement normalised by free field soil settlement versus normalised
compaction depth (Liu and Dorby, 1997) ............................................................................ 105
xxxviii
Figure 2‐71: Distribution of excess pore water pressure ratio at different transient times for
a ground model with width to depth ratio of compacted area being 1.6 and depth of
compacted area to total ground thickness equalling unity; modified from Akiyoshi et al.
(1993) ................................................................................................................................... 107
Figure 2‐72: Relationship between maximum excess pore water pressure ratio and width to
depth ratio of compacted area when treatment depth equals liquefaction thickness (Akiyoshi
et al., 1993) .......................................................................................................................... 107
Figure 2‐73: Ground improvement zone as specified by Japan Fire Defence (Port and Harbour
Research Institute, 1997), as cited by Hausler (2002). ........................................................ 108
Figure 2‐74: Stability of treated zone without contribution of a 30o wedge (Port and Harbour
Research Institute, 1997), as cited by Hausler (2002). ........................................................ 109
Figure 2‐75: Range of improvement in soil mass for (a) b/d and (b) a/d versus total normalised
specific energy (Poran and Rodriguez, 1992) ...................................................................... 113
Figure 2‐76: Proposed design chart for dynamic compaction in sandy soils (Poran and
Rodriguez, 1992) .................................................................................................................. 113
Figure 2‐77: Overlapping of semi‐prolate spheroids, modified from Poran and Rodriguez
(1992) ................................................................................................................................... 114
Figure 2‐78: Schematic scheme for evaluation of print spacing (Jahangiri et al., 2011) ..... 116
Figure 2‐79: Print spacing curves for Ir = 8%, 15% and 30% (Jahangiri et al., 2011) ........... 116
Figure 2‐80: Print spacing curves for Ir = 46% and 62% (Jahangiri et al., 2011) ................. 117
Figure 2‐81: Plane strain (2 dimensional) arching of soil (Hewlett and Randolph, 1988) ... 121
Figure 2‐82: Isometric view of a grid of pile caps (DR columns) and a series of domes forming
vaults spanning between them (Hewlett and Randolph, 1988) .......................................... 121
Figure 2‐83: Analysis of arching at the crown of a dome. The diagram on the right represents
a diagonal section through a pile cap and dome crown. (Hewlett and Randolph, 1988) ... 122
Figure 2‐84: Arching in the sand fill immediately over a pile cap (Hewlett and Randolph, 1988)
............................................................................................................................................. 123
Figure 2‐85: Point support definitions (rectangular‐triangular) (Kempfert et al., 2004) after
Zaeske (2001) ....................................................................................................................... 125
Figure 2‐86: Theoretical arching model (Kempfert et al., 2004) after Zaeske (2001) ......... 125
Figure 2‐87: Area of one pile cell (Satibi, 2009) after DGGT (2004) .................................... 127
Figure 2‐88: Stress distribution between soil and columnar inclusions (Aboshi et al., 1979,
1991) .................................................................................................................................... 128
Figure 2‐89: Influence of radial deformation and plastic strains on n (Castro and Sagaseta,
2009) .................................................................................................................................... 131
xxxix
Figure 2‐90: Axisymmetrical geometrical idealisation of one DR cell (Satibi, 2009) ........... 132
Figure 2‐91: Equivalent unit cell diameters (a) Square arrangement of columns, (b) Triangular
arrangement of columns, (c) Hexagonal arrangement of columns (Balaam and Booker, 1981)
............................................................................................................................................. 132
Figure 2‐92: Axisymmetrical model with concentric rings modified from Mitchell and Huber
(1985) ................................................................................................................................... 133
Figure 2‐93: Horizontal displacements in plane strain models (Tan et al., 2008) ............... 134
Figure 2‐94: Plane strain geometrical idealisation with equivalent stiffness (Satibi, 2009) 135
Figure 2‐95: Plane strain geometrical idealisation after Bergado (Satibi, 2009) ................. 135
Figure 2‐96: Plane strain geometrical idealisation with equivalent homogenized continuum
............................................................................................................................................. 137
Figure 2‐97: three dimensional geometrical idealisation (Satibi, 2009) .............................. 137
Figure 2‐98: Vertical stresses on a typical horizontal level in the column (Hughes and Withers,
1974) .................................................................................................................................... 138
Figure 2‐99: Stability condition of a DR column against bulging ......................................... 141
Figure 2‐100: Granular trench in weak clay (Madhav and Vitkar, 1978) ............................. 142
Figure 2‐101: Bearing capacity factor Nc (Madhav and Vitkar, 1978) ................................. 143
Figure 2‐102: Bearing capacity factor NΥ (Madhav and Vitkar, 1978) ................................. 144
Figure 2‐103: Bearing capacity factor Nq (Madhav and Vitkar, 1978) ................................. 144
Figure 2‐104: Effect of concentrated shear stresses on the CSR (Baez and Martin, 1993) . 148
Figure 2‐105: Dynamic compaction using the 700 tm rig .................................................... 151
Figure 2‐106: Dynamic compaction using 1,600 tm mega machine (Communication
Department of Menard, 2007)............................................................................................. 151
Figure 2‐107: Dynamic compaction using the Menard tripod ............................................. 152
Figure 2‐108: Dynamic compaction using the giga‐machine ............................................... 152
Figure 2‐109: Dynamic compaction using 25 ton pounder .................................................. 153
Figure 2‐110: Lampson crane modified by dynamic compaction ........................................ 154
Figure 2‐111: Dynamic compaction in China using cranes that have been modified into tripods
............................................................................................................................................. 154
Figure 2‐112: Reinforced concrete pounder encased in a steel box ................................... 155
Figure 2‐113: steel pounders composed of steel plates ...................................................... 155
Figure 2‐114: Marine dynamic compaction pounder used for seabed improvement at Kuwait
Naval Base (Communication Department of Menard, 2007) .............................................. 156
Figure 2‐115: Rapid Impact Compactor ............................................................................... 157
Figure 2‐116: Impact roller compaction .............................................................................. 158
xl
Figure 2‐117: Impact roller compaction depth of improvement (Kelly and Gil, 2012) ....... 159
Figure 2‐118: Schematic comparison of depth of improvement between impact roller
compaction, RIC and DC ...................................................................................................... 160
Figure 2‐119: Displacement versus frequency, combined date with recommended vibration
criterion (Nicholls et al., 1971) ............................................................................................. 163
Figure 2‐120: Particle velocity versus frequency with recommended safe limit criterion
(Nicholls et al., 1971) ........................................................................................................... 163
Figure 2‐121: Types of motion that can cause damage to a building (Siskind et al., 1980) 165
Figure 2‐122: USBM RI 8507 safe vibration limit criteria (Siskind et al., 1980) ................... 167
Figure 2‐123: OSM Safe vibration level criteria (Office of Surface Mining Reclamation
Enforcement, 1983) ............................................................................................................. 168
Figure 2‐124: Subjective response of the human body to steady state vibratory motion
(Nicholls et al., 1971) ........................................................................................................... 170
Figure 2‐125: Complaint history of Salmon underground nuclear detonation with superposed
subjective response (Power, 1966) ...................................................................................... 171
Figure 2‐126: Comparison of Human Response to steady state (Reiher and Meister, 1931) and
transient vibration (Wiss and Parmelee, 1974) ................................................................... 171
Figure 2‐127: Transient vibration guide for cosmetic damage(British Standards Institution,
1993) .................................................................................................................................... 174
Figure 2‐128: Summary guidance on vibration criteria given in British Standards. Damage
thresholds are those for domestic buildings (x‐axis is front‐to‐back, y‐axis is side‐to‐side, z‐
axis is head‐to‐toe) (Hiller and Hope, 1998) ........................................................................ 176
Figure 2‐129: Vertical, radial and tangential particle velocities’ histories at different distances
(Hwang and Tu, 2003, 2006) ................................................................................................ 179
Figure 2‐130: Fourier amplitude spectra of the vertical, radial and tangential velocities at
various distances from the pounder impact point (Hwang and Tu, 2003, 2006) ................ 180
Figure 2‐131: Attenuation relationship of PPV with distance in all three directions for nine
consecutive impacts (Hwang and Tu, 2003, 2006) .............................................................. 181
Figure 2‐132: PPV attenuation for different drop heights (Hwang and Tu, 2003, Hwang and
Tu, 2006) .............................................................................................................................. 182
Figure 2‐133: Vertical and radial Fourier spectra at 10 m distance for different falling heights
(Hwang and Tu, 2003, Hwang and Tu, 2006) ....................................................................... 183
Figure 2‐134: Vibration metre for measuring peak particle velocity ................................... 183
Figure 2‐135: Attenuation of ground vibrations measured on different dynamic compaction
projects (Mayne et al., 1984) ............................................................................................... 186
xli
Figure 2‐136: Observed attenuation of normalised PPV with normalised distance (Mayne,
1985) .................................................................................................................................... 187
Figure 2‐137: Peak Particle Velocity for impact energy in the range of 250‐300 tm (Chapot
and et al., 1981, Varaksin, 1981) ......................................................................................... 188
Figure 2‐138: Plot of peak particle velocity versus inverse scaled distance for collapsible soils
(Rollins and Kim, 1994) ........................................................................................................ 188
Figure 2‐139: Vibration isolation using an open trench (Shrivastava and Kameswara Rao,
2002) .................................................................................................................................... 189
Figure 2‐140: Schematic of test layout for active isolation in the field (Woods, 1968) ...... 191
Figure 2‐141: Amplitude of vertical displacement versus distance from source for five tests
(Woods, 1968) ...................................................................................................................... 191
Figure 2‐142: Effect of length of trench on vibration amplitude (Shrivastava and Kameswara
Rao, 2002) ............................................................................................................................ 192
Figure 2‐143: Effect of depth of trench on vibration amplitude (Shrivastava and Kameswara
Rao, 2002) ............................................................................................................................ 193
Figure 2‐144: The comparison of screening effectiveness between the open trench barriers:
(a) the amplitude of ground vibration behind the barriers and (b) the ratio of Ary was based
on the amplitude with and without the walls (Tsai and Chang, 2009). ............................... 193
Figure 2‐145: Effectiveness of shallow trenching for reduction of vibration (Varaksin, 1981)
............................................................................................................................................. 195
Figure 2‐146: Net effective volume change, modified from Lukas (1986) .......................... 196
Figure 2‐147: Inefficiency of pounder blows identifiable from measurement of net volume
change (Serridge, 2002) ....................................................................................................... 197
Figure 2‐148: heave and penetration test ........................................................................... 198
Figure 2‐149: Heaved ground section (Lukas, 1986) ........................................................... 199
Figure 2‐150: The main elements of the Menard pressuremeter (Baguelin et al., 1978) ... 201
Figure 2‐151: Pressure and volume calibration curves (ASTM, 2007) ................................. 202
Figure 2‐152: Typical pre‐drilled PMT curve (Amar et al., 1991, ASTM, 2007) ................... 206
Figure 2‐153: Pressuremeter pressure versus volume curve (Centre D'Etudes Menard, 1975)
............................................................................................................................................. 208
Figure 2‐154: Extrapolation of PLM (Amar et al., 1991) ........................................................ 209
Figure 2‐155: Determination PLM from inverse of volume versus pressure (ASTM, 2007) . 210
Figure 2‐156: Interpretation of EM and PLM from PMT from ENV 1997‐3 (European Standard,
2000) .................................................................................................................................... 210
xlii
Figure 2‐157: Bearing factor for isolated square or circular footings, piers and piles (Centre
D'Etudes Menard, 1975) ...................................................................................................... 213
Figure 2‐158: Bearing factor for strip footings and diaphragm walls (Centre D'Etudes Menard,
1975) .................................................................................................................................... 214
Figure 2‐159: Eccentric loading on footing, Baguelin (1978) after Meyerhof (1953) .......... 217
Figure 2‐160: Inclined load on footing (Baguelin et al., 1978) ............................................. 218
Figure 2‐161: Bearing capacity reduction factor for footings resting on sloping ground or near
excavations (Centre D'Etudes Menard, 1975) ..................................................................... 219
Figure 2‐162: Footing near slope or hillside (Baguelin et al., 1978) .................................... 220
Figure 2‐163: Footing subject to inclined and eccentric load (Baguelin et al., 1978) .......... 221
Figure 2‐164: Variation of vertical strain components with depth along the vertical axis below
a rigid circular footing (Baguelin et al., 1978) ...................................................................... 223
Figure 2‐165: Layers of ground under the footing that are taken into consideration for the
calculation of Ed (Centre D'Etudes Menard, 1975) .............................................................. 226
Figure 2‐166: Compressible layer embedded between two layers (Centre D'Etudes Menard,
1975) .................................................................................................................................... 228
Figure 2‐167: Settlement calculation for footings with overlapping stress bulbs (Centre
D'Etudes Menard, 1975) ...................................................................................................... 229
Figure 2‐168: Graph for estimating drained friction angle from net limit pressure (Etude
Pressiometrique Louis Menard, 1970) ................................................................................. 234
Figure 2‐169: Acceleration of pounder after impact (Adam et al., 2007) ........................... 239
Figure 2‐170: (Idealised) decaying wave (Adam et al., 2007) .............................................. 241
Figure 2‐171: Decaying acceleration wave measured after pounder impact (Adam et al., 2007)
............................................................................................................................................. 241
Figure 2‐172: Chart for estimation of Poisson’s ratio (Adam et al., 2007) .......................... 242
Figure 2‐173: Chart for estimation of Young modulus for the case of Dc= 0 (Adam et al., 2007)
............................................................................................................................................. 242
Figure 2‐174: Chart for estimation of κ (Adam et al., 2007) ................................................ 243
Figure 2‐175: Applicable range of relative density (non‐shaded areas) as defined by ASTM
............................................................................................................................................. 253
Figure 2‐176: Systematic and random errors (Selig and Ladd, 1973) .................................. 255
Figure 2‐177: Error propagation factors for relative density, redrawn from Yoshimi and Tohno
(1973) ................................................................................................................................... 256
Figure 2‐178: Error on the in‐situ relative density, redrawn in SI units from Tavenas et al.
(1973) ................................................................................................................................... 259
xliii
Figure 2‐179: Density limits as a function of grain shape for laboratory fractions with Cu=1.4
(Youd, 1973) ......................................................................................................................... 264
Figure 2‐180: Generalized curve for estimating emax and emin from gradational and particle
shape characteristics. Curves are only valid for clean sands with normal to moderately
skewed grain size distributions. (Youd, 1973) ..................................................................... 265
Figure 2‐181: Effect of particle shape on minimum and maximum void ratios (Holubec and
D'Appolonia, 1973) .............................................................................................................. 265
Figure 2‐182: Difference between effective peak and critical state friction angles in four
different uniformly graded silica soils with similar mineralogical compositions at different
relative densities; reconstructed from Liu and Lehane (2012) ............................................ 267
Figure 2‐183: Relative density versus penetration resistance curves for fine and coarse sands,
reconstructed from Gibbs and Holtz (1957) ........................................................................ 269
Figure 2‐184: Criterion for predicting relative density of sand from the penetration resistance
test (Bureau of Reclamation, 1998) ..................................................................................... 269
Figure 2‐185: Comparison of measured versus calculated relative density using Gibbs and
Holtz’s relationship (Haldar and Tang, 1979) ....................................................................... 271
Figure 2‐186: Figure 8: Comparison of measured versus calculated relative density using
Meyerhof’s equation (Hatanaka and Feng, 2006) ............................................................... 275
Figure 2‐187: Comparison of relative density – CPT cone resistance relationships for
Schmertmann (1975), Schmertmann (1978) and Villet and Mitchell (1981)...................... 277
Figure 2‐188: Dump truck tipping fill into the sea ............................................................... 282
Figure 2‐189: (a) subaqueous discharge by hopper or bottom dump barge (b) subaerial
rainbow discharge (c) pipeline discharge, redrawn from Lee et al. (1999) ......................... 283
Figure 2‐190: Application of dynamic compaction at Brest Naval Base using a specially
designed 11 ton marine pounder (Boulard, 1974) .............................................................. 285
Figure 2‐191: EM and PLM before and after offshore dynamic compaction at Brest Naval Base
(Boulard, 1974) .................................................................................................................... 286
Figure 2‐192: Cross section of rock fill at Udevella, Sweden (Menard, 1978) ..................... 287
Figure 2‐193: Application of offshore dynamic compaction at Kuwait Naval Base ............. 287
Figure 2‐194: Offshore DC at Kuwait Naval Base using a 32 t specially designed pounder 288
Figure 2‐195: Dynamic compaction at Lagos dry docks, Nigeria ......................................... 288
Figure 2‐196: Comparison of EM before and after dynamic compaction at Lagos Dry Docks
(Gambin, 1982) .................................................................................................................... 289
Figure 3‐1: Abu Dhabi Corniche: 6 km long project in a metropolitan area ........................ 293
xliv
Figure 3‐2: First application of MARS pounder release and reconnect mechanism in the world
............................................................................................................................................. 294
Figure 3‐3: Application of DC and DR at KAUST using 13 (9 shown in the photograph) DC‐DR
rigs working in two shifts per day. ....................................................................................... 294
Figure 3‐4: Treatment of 4.84 million m2 of loose sands at Al Falah Community. .............. 295
Figure 3‐5: First application of a double sided DC‐DR grater pounder in the world ........... 296
Figure 3‐6: Plan of Abu Dhabi New Corniche ....................................................................... 297
Figure 3‐7: Reclamation of New Corniche by pipeline discharged hydraulic fill ................. 297
Figure 3‐8: Typical crater size and depth in the deep treatment areas of Abu Dhabi New
Corniche ............................................................................................................................... 299
Figure 3‐9: Application of dynamic compaction at the shoreline ........................................ 300
Figure 3‐10: Application of dynamic compaction in front of Abu Dhabi skyline ................. 301
Figure 3‐11: PLM and EM before and after dynamic compaction at test point No. 1 ........... 303
Figure 3‐12: PLM and EM before and after dynamic compaction at test point No. 2 ........... 303
Figure 3‐13: PLM and EM before and after dynamic compaction at test point No. 3 ........... 304
Figure 3‐14: Average PLM and EM before and after dynamic compaction for three points . 304
Figure 3‐15: PLM and EM improvement ratios for Point No. 1 .............................................. 305
Figure 3‐16: PLM and EM improvement ratios for Point No. 2 .............................................. 305
Figure 3‐17: PLM and EM improvement ratios for Point No. 3 .............................................. 306
Figure 3‐18: Average PLM and EM improvement ratios ........................................................ 306
Figure 3‐19: Abu Dhabi Corniche ......................................................................................... 307
Figure 3‐20: Location of Al Quo’a ........................................................................................ 308
Figure 3‐21: Grand size cutting and dump filling of dune sands for the construction of the
second phase of Al Quo’a Township. ................................................................................... 309
Figure 3‐22: Sieve analysis of the dune sand ....................................................................... 309
Figure 3‐23: Free fall of 35 ton pounder using MARS .......................................................... 314
Figure 3‐24: Automatic grabbing of a 35 ton pounder by MARS ......................................... 315
Figure 3‐25: DC rig for lifting 15 ton pounders at Al Quo’a ................................................. 315
Figure 3‐26: DC rig for lifting 25 ton pounders at Al Quo’a ................................................. 316
Figure 3‐27: Menard 700 tm rig at Al Quo’a ........................................................................ 316
Figure 3‐28: 250 ton rig for lifting the 35 ton pounder using MARS ................................... 317
Figure 3‐29: Quick probing machine composed of an excavator base, a mast and a
penetrating tubular rod. ...................................................................................................... 318
Figure 3‐30: PLM and ratio of PLM before and after dynamic compaction at test point No. 1
............................................................................................................................................. 320
xlv
Figure 3‐31: PLM and ratio of PLM before and after dynamic compaction at test point No. 2
............................................................................................................................................. 320
Figure 3‐32: PLM and ratio of PLM before and after dynamic compaction at test point No. 3
............................................................................................................................................. 321
Figure 3‐33: PLM and ratio of PLM before and after dynamic compaction at test point No. 4
............................................................................................................................................. 321
Figure 3‐34: Average PLM and ratio of average PLM before and after dynamic compaction 322
Figure 3‐35: Plan of KAUST .................................................................................................. 324
Figure 3‐36: Variation of ground conditions for two test locations that were 30 m apart. 325
Figure 3‐37: Ground profile at KAUST .................................................................................. 326
Figure 3‐38: CPT‐19 .............................................................................................................. 326
Figure 3‐39: CPT‐178 ............................................................................................................ 327
Figure 3‐40: CPT‐27 .............................................................................................................. 327
Figure 3‐41: CPT‐12 .............................................................................................................. 328
Figure 3‐42: CPT‐138 ............................................................................................................ 328
Figure 3‐43: CPT‐206 ............................................................................................................ 329
Figure 3‐44: Finite element modelling for DR columns at (a) 5.5 m grid and (b) 3.8 m grid 331
Figure 3‐45: Vertical effective stresses for 5.5 m spaced DR grid elevations (a) 0.0 m RL, (b) ‐
1 m RL, (c) ‐2 m RL and (d) ‐3 m RL ...................................................................................... 331
Figure 3‐46: Vertical effective stresses for 3.8 m spaced DR grid elevations (a) 0.0 m RL, (b) ‐
1 m RL, (c) ‐2 m RL and (d) ‐3 m RL ...................................................................................... 332
Figure 3‐47: Foundation concept ......................................................................................... 332
Figure 3‐48: Application of 3 m surcharge above final grade level after dynamic replacement
in areas with sabkhah being deeper than 5 m. .................................................................... 333
Figure 3‐49: Dynamic Surcharging in KAUST ........................................................................ 334
Figure 3‐50: Flow chart for determining applicable ground improvement technique ........ 335
Figure 3‐51: Utilisation of 13 DC‐DR rigs in two shifts (photograph taken from the 13th rig)
............................................................................................................................................. 335
Figure 3‐52: PLM before and after dynamic compaction and PLM improvement ratio ......... 336
Figure 3‐53: PLM before and after dynamic replacement and PLM improvement ratio ........ 337
Figure 3‐54: The relationship between net limit pressure, fines content and improvement
energy .................................................................................................................................. 338
Figure 3‐55: Settlement induced by surcharging after implementation of DR in areas where
sabkhah depth exceeded 5 m. ............................................................................................. 338
Figure 3‐56: Settlement induced by static and dynamic surcharging .................................. 339
xlvi
Figure 3‐57: King Abudulla University of Science and Technology ...................................... 340
Figure 3‐58: Master plan of Al Falah Community ................................................................ 341
Figure 3‐59: Site plan and limit of loose silty sands (hatched areas) .................................. 342
Figure 3‐60: Two typical test results of initial ground conditions (a) CPT, (b) SPT ............ 342
Figure 3‐61: Implementation of 11 special cranes for DC (not all shown in this photograph)
............................................................................................................................................. 344
Figure 3‐62: Accumulative treated area .............................................................................. 345
Figure 3‐63: Calibration of phase 1 number of blows ......................................................... 346
Figure 3‐64: CPT cone resistance before and after DC (a) 6 blows, (b) 7 blows, (c) 8 blows
............................................................................................................................................. 347
Figure 3‐65: CPT cone resistance after compaction of prints .............................................. 347
Figure 3‐66: Dynamic compaction calibration layout .......................................................... 348
Figure 3‐67: CPT cone resistances before dynamic compaction ......................................... 349
Figure 3‐68: qc and average qc improvement ratio for pattern with 3 blows per print in phase
1 ........................................................................................................................................... 350
Figure 3‐69: qc and average qc improvement ratio for pattern with 4 blows per print in phase
1 ........................................................................................................................................... 350
Figure 3‐70: qc and average qc improvement ratio for pattern with 5 blows per print in phase
1 ........................................................................................................................................... 351
Figure 3‐71: qc and average qc improvement ratio for pattern with 6 blows per print in phase
1 ........................................................................................................................................... 351
Figure 3‐72: average qc and average qc improvement ratios of all four patterns ............... 352
Figure 3‐73: PLM and EM for pattern with 3 blows per print in phase 1 ............................... 353
Figure 3‐74: PLM and EM for pattern with 4 blows per print in phase 1 ............................... 353
Figure 3‐75: PLM and EM for pattern with 5 blows per print in phase 1 ............................... 354
Figure 3‐76: PLM and EM for pattern with 6 blows per print in phase 1 ............................... 354
Figure 3‐77: Comparison of PLM and EM before and after dynamic compaction ................. 355
Figure 3‐78: PLM and EM improvement ratios for pattern with 3 blows per print in phase 1
............................................................................................................................................. 355
Figure 3‐79: PLM and EM improvement ratios for pattern with 4 blows per print in phase 1
............................................................................................................................................. 356
Figure 3‐80: PLM and EM improvement ratios for pattern with 5 blows per print in phase 1
............................................................................................................................................. 356
Figure 3‐81: PLM and EM improvement ratios for pattern with 6 blows per print in phase 1
............................................................................................................................................. 357
xlvii
Figure 3‐82: Comparison of PLM and EM improvement ratios of all 4 patterns ................... 357
Figure 3‐83: Accumulative penetration and crater volume against number of blows ........ 359
Figure 3‐84: Percentages of accumulative penetration and crater volume against number of
blows .................................................................................................................................... 360
Figure 3‐85: Al Falah Community ......................................................................................... 361
Figure 3‐86: Master plan of Marjan Island (main road corridor shown in white) ............... 362
Figure 3‐87: Marjan Island satellite image (taken in 2009) with the main road identified in red
............................................................................................................................................. 363
Figure 3‐88: SPT borehole layout ......................................................................................... 364
Figure 3‐89: SPT blow counts of several boreholes at Marjan Island before dynamic
compaction .......................................................................................................................... 364
Figure 3‐90: PMT layout plan on the Peninsula ................................................................... 365
Figure 3‐91: Several PMT results at Marjan Island before dynamic compaction ................ 365
Figure 3‐92: Application of dynamic compaction at the vicinity of slope armour protection
............................................................................................................................................. 368
Figure 3‐93: Location of calibration area ............................................................................. 369
Figure 3‐94: Dynamic compaction calibration layout .......................................................... 369
Figure 3‐95: Measurement of crater depth in the heave and penetration test .................. 370
Figure 3‐96: Measurement of the crater’s upper diameter in the heave and penetration test
............................................................................................................................................. 370
Figure 3‐97: Heave and penetration test in phase 1 print of 7x7 m2 grid pattern .............. 371
Figure 3‐98: Heave and penetration test in phase 2 print of 7x7 m2 grid pattern .............. 371
Figure 3‐99: Heave and penetration test in phase 1 print of 5x5 m2 grid pattern .............. 372
Figure 3‐100: Heave and penetration test in phase 2 print of 5x5 m2 grid pattern ............ 372
Figure 3‐101: Ratio of Phase 2 to Phase 1 penetration and volume reduction in the two
patterns ................................................................................................................................ 373
Figure 3‐102: Ratio of compaction volume to accumulative penetration in HPTs .............. 373
Figure 3‐103: Ratio of square root of compaction volume to accumulative penetration in HPTs
............................................................................................................................................. 374
Figure 3‐104: PLM and EM values before and after dynamic compaction in pattern with 7x7 m2
grid ....................................................................................................................................... 375
Figure 3‐105: PLM and EM improvement ratios in pattern with 7x7 m2 grid ........................ 376
Figure 3‐106: PLM and EM values before and after dynamic compaction in pattern with 5x5 m2
grid ....................................................................................................................................... 377
Figure 3‐107: PLM and EM improvement ratios in pattern with 5x5 m2 grid ........................ 377
xlviii
Figure 3‐108: Ratio of Phase 1 to Phase 2 PLM values for Patterns 1 and 2 ......................... 378
Figure 3‐109: Ratio of Phase 1 to Phase 2 and in between prints PLM values for Pattern 2 378
Figure 3‐110: Two post dynamic compaction EM logs ......................................................... 379
Figure 3‐111: Average EM before and after dynamic compaction and average EM improvement
ratio ...................................................................................................................................... 380
Figure 3‐112: Marjan Island road after completion of dynamic compaction works ........... 382
Figure 3‐113: Plan of Nakilat Ship Repair Yard .................................................................... 383
Figure 3‐114: Comparison of CPT cone resistance before ground improvement with target
relative density..................................................................................................................... 384
Figure 3‐115: Free fall drop of a 35 ton pounder using MARS technology ......................... 386
Figure 3‐116: Heave and penetration test results ............................................................... 387
Figure 3‐117: Ratio of square root of compaction volumes to penetration........................ 388
Figure 3‐118: Percentage of compaction share of the crater and the peripheral concentric
rings ...................................................................................................................................... 389
Figure 3‐119: Variation of (DT‐DB)/Dc and crater side angle with number of blows ........... 389
Figure 3‐120: Ratio of square root of crater compaction volume to penetration............... 390
Figure 3‐121: Variation of (DT‐DB)/Dc and crater side angle with number of blows ........... 391
Figure 3‐122: Variation of (DT‐DB)/Dc with impact energy .................................................. 391
Figure 3‐123: Post ground improvement soil profiles in DDR4 ........................................... 394
Figure 3‐124: Post ground improvement profiles in DDR6 .................................................. 394
Figure 3‐125: CPT and PMT layout ....................................................................................... 396
Figure 3‐126: CPT cone qc used in the correlation ............................................................... 397
Figure 3‐127: PMT PLM values used in the correlation ......................................................... 397
Figure 3‐128: PMT EM values used in the correlation .......................................................... 398
Figure 3‐129: qc/PLM for Ras Laffan carbonate sand ............................................................ 399
Figure 3‐130: qc*/P*LM for Ras Laffan carbonate sand ...................................................... 399
Figure 3‐131: EM/qc for Ras Laffan carbonate sand ............................................................. 400
Figure 3‐132: Fines content of reclaimed soil with very high fines content ........................ 401
Figure 3‐133: results of CPT‐615 (before DC) and CPT‐828 (after DC) ................................ 402
Figure 3‐134: Results of PMT‐025 ........................................................................................ 402
Figure 3‐135: Location of zone load test in area with the most amount of fine soil ........... 403
Figure 3‐136: Location of zone load test and field tests shown in Figure ........................... 403
Figure 3‐137: Skid beam zone load test details (a) plan view, (b) side view, (c) front view 404
Figure 3‐138: Settlement monitoring of settlement monitoring points .............................. 405
Figure 3‐139: Layout of Abu Dhabi Ritz‐Carlton Hotel ......................................................... 407
xlix
Figure 3‐140: Layout of SPT tests ......................................................................................... 408
Figure 3‐141: The profiles of four SPT boreholes in a section through the site .................. 409
Figure 3‐142: Construction equipment sinking into the ground due to poor ground conditions
............................................................................................................................................. 410
Figure 3‐143: Pre‐excavation of DR prints ........................................................................... 412
Figure 3‐144: Release of excess pore water pressure in the form of sand boiling .............. 413
Figure 3‐145: Stability analysis of MSE Wall on pre‐excavated DR trench .......................... 414
Figure 3‐146: Backfilling of the lower part of the trenches beneath the wall with demolished
concrete debris .................................................................................................................... 414
Figure 3‐147: Comparison of PMT parameters before and after DR (in between and inside DR
columns) ............................................................................................................................... 416
Figure 3‐148: Comparison of PMT parameters ratios in between and inside DR columns . 416
Figure 3‐149: Construction of hills supported by MSE walls on one side ............................ 417
Figure 3‐150: The measurements of fill height and ground settlement under the fill during a
time interval ......................................................................................................................... 417
Figure 3‐151: Estimation of settlement and consolidation ratio using Asaoka’s method ... 418
Figure 3‐152: Completed chalets at Abu Dhabi Ritz‐Carlton Hotel ..................................... 419
Figure 3‐153: Layout of Al Jazira Steel Pipe Factory structures ........................................... 420
Figure 3‐154: Loading patterns in the main building ........................................................... 421
Figure 3‐155: Soil profile ...................................................................................................... 421
Figure 3‐156: Schematic illustration of the dynamic replacement process ........................ 423
Figure 3‐157: (a) Pre‐excavation, (b) backfilling with sand ................................................. 424
Figure 3‐158: Dynamic replacement HPT test layout .......................................................... 425
Figure 3‐159: Pounder penetration and volumetric changes during the DR HPT ............... 426
Figure 3‐160: Heave and penetration test at Dubai Airport (Serridge, 2002) ..................... 426
Figure 3‐161: DR calibration PMT results ............................................................................ 427
Figure 3‐162: DR calibration PMT improvement ratios ....................................................... 427
Figure 3‐163: Comparison of PMT parameters before and after DR ................................... 429
Figure 3‐164: Comparison of improvement ratios in between and inside DR columns ...... 429
Figure 3‐165: Al Jazira Steel Pipe Factory ............................................................................ 431
Figure 3‐166: Site plan of Reem Island Causeway ............................................................... 432
Figure 3‐167: Longitudinal profile of the approach road ..................................................... 433
Figure 3‐168: Project cross section at bridge level .............................................................. 433
Figure 3‐169: Reclamation by dumping sand and pushing it into the sea ........................... 435
Figure 3‐170: Application of dynamic compaction at Reem Island Causeway .................... 437
l
Figure 3‐171: Heave and Penetration Test results .............................................................. 438
Figure 3‐172: Penetration Test results ................................................................................. 438
Figure 3‐173: The relationship between the square root of volume to penetration and
number of blows .................................................................................................................. 439
Figure 3‐174: The relationship between (DT‐DB)/Dc and number of blows ........................ 440
Figure 3‐175: PLM and EM values before and after dynamic compaction ............................ 440
Figure 3‐176: PLM and EM improvement ratios .................................................................... 441
Figure 3‐177: Comparison of ground subsidence versus energy with previous research
(Mayne et al., 1984) ............................................................................................................. 443
Figure 3‐178: Abu Dhabi – Reem Island Causeway after completion .................................. 445
Figure 3‐179: HFO Bunkering Facility site in Ras Laffan ....................................................... 446
Figure 3‐180: Location of HFO Bunkering Facility in Ras Laffan .......................................... 447
Figure 3‐181: SPT blow counts in borehole BH2B................................................................ 448
Figure 3‐182: PLM and EM values in PMT T2‐01 .................................................................... 449
Figure 3‐183: Numerical analysis model geometry ............................................................. 452
Figure 3‐184: (a) Vertical displacement contours, (b) vertical displacement at the ground
surface .................................................................................................................................. 453
Figure 3‐185: Layout of Dynamic Compaction works and PMTs ......................................... 454
Figure 3‐186: Measured ground settlement during HPT‐01 ................................................ 454
Figure 3‐187: crater depth, crater, heave and net compaction volume changes due to
consecutive pounder blows during HPT‐01 ......................................................................... 455
Figure 3‐188: Variation of the square root of compaction volumes with consecutive pounding
............................................................................................................................................. 456
Figure 3‐189: PLM and EM values before and after dynamic compaction ............................ 457
Figure 3‐190: Ratio of PLM and EM before and after dynamic compaction .......................... 458
Figure 3‐191: HFO tanks under construction ....................................................................... 460
Figure 3‐192: Reclamation by rainbow discharge at Palm Jumeira ..................................... 461
Figure 3‐193: Location of the two sewerage treatment plant on Palm Jumeira ................. 462
Figure 3‐194: Layout of settlement monitoring plates ........................................................ 467
Figure 3‐195: Schematic illustration of the surcharge ......................................................... 467
Figure 3‐196: Ground settlement in Tank A‐A during static and dynamic surcharging ....... 468
Figure 3‐197: Pre‐excavation of dynamic surcharging prints .............................................. 469
Figure 3‐198: Dynamic surcharging ..................................................................................... 469
Figure 3‐199: Cross section of excavation for reaching working platform level ................. 470
li
Figure 3‐200: (a) Dynamic compaction of the tank’s foundation, (b) water boiling into the
craters .................................................................................................................................. 471
Figure 3‐201: Crater top diameter and depth in (a) the first and (b) the second phase of DC
............................................................................................................................................. 472
Figure 3‐202: Pre and post ground improvement PMT PLM and EM values ......................... 474
Figure 3‐203: Average PLM and EM improvement ratios ...................................................... 475
Figure 3‐204: (a) Finite element model, (b) zoom of model details .................................... 475
Figure 3‐205: Settlement contours ...................................................................................... 476
Figure 3‐206: Settlement variation under Tank A‐A ............................................................ 477
Figure 3‐207: Ground settlement in Tank G‐G during static and dynamic surcharging ...... 478
Figure 3‐208: Pre and post ground improvement PMT PLM and EM values ......................... 478
Figure 3‐209: Settlement profile under Tank G‐G ............................................................... 479
Figure 3‐210: Tank A‐A after construction ........................................................................... 481
Figure 3‐211: Satellite image of location of Al Khaleej Sugar Factory ................................. 482
Figure 3‐212: Schematic cross section of the sugar silos ..................................................... 484
Figure 3‐213: Model used for the numerical analysis of the sugar silo ............................... 485
Figure 3‐214: Total settlement after construction .............................................................. 486
Figure 3‐215: Total settlement after first loading................................................................ 487
Figure 3‐216: Total settlement after beginning of first unloading ...................................... 487
Figure 3‐217: Vertical displacement of the ground due to different load cases ................. 488
Figure 3‐218: Maximum horizontal stresses after construction .......................................... 488
Figure 3‐219: Maximum horizontal stresses after first loading ........................................... 488
Figure 3‐220: Maximum horizontal stresses after beginning of first unloading.................. 489
Figure 3‐221: Vertical displacement of the ground due to different reloading cases ......... 489
Figure 3‐222: Vertical displacement of the ground due to different loading and reloading
cases ..................................................................................................................................... 490
Figure 3‐223: Dynamic compaction of the shear ring .......................................................... 491
Figure 3‐224: Dynamic replacement next to an existing concrete silo ................................ 491
Figure 3‐225: Presence of very hard damping layer ............................................................ 492
Figure 3‐226: PMT modulus before and after ironing in the first silo ................................. 493
Figure 3‐227: Al Khaleej raw sugar silos after completion of construction ......................... 494
Figure 3‐228: Cross section of container terminal based on original foundation concept . 495
Figure 3‐229: Marine DR (bottom side) ‐ DC (top side) pounder ........................................ 497
Figure 3‐230: Marine pounder being utilised for dynamic replacement............................. 497
Figure 3‐231: Marine pounder used for dynamic compaction ............................................ 498
lii
Figure 3‐232: Pounder penetration at several DR print locations ....................................... 499
Figure 3‐233: Trial zone and PMT locations ......................................................................... 500
Figure 3‐234: Proposed method for estimation of rock fill friction angle from PMT .......... 502
Figure 3‐235: Location of dynamic compaction trial on Frond N (Al Naghal) ..................... 503
Figure 3‐236: Pre‐treatment PLM and EM values in the DC trial area ................................... 504
Figure 3‐237: Dynamic compaction trial area ...................................................................... 505
Figure 3‐238: After dynamic compaction PLM and EM values in the DC trial area ............... 506
Figure 3‐239: PLM and EM improvement ratios .................................................................... 506
Figure 3‐240: (a) EM/PLM before and after dynamic compaction, (b) Ratio of after to before
dynamic compaction EM/PLM ................................................................................................ 507
Figure 3‐241: After dynamic compaction CPT qc values ..................................................... 508
Figure 3‐242: qc/PLM and EM/qc correlations ....................................................................... 509
Figure 3‐243: qc versus PLM values of Palm Jumeira Trial and Al Nakhilat Ship Repair Yard 510
Figure 3‐244: qc versus EM values of Palm Jumeira Trial and Al Nakhilat Ship Repair Yard . 512
Figure 3‐245: Location of Fujairah Power and Desalination Plant ....................................... 515
Figure 3‐246: Before dynamic compaction CPT at Fujairah Desalination Plant .................. 516
Figure 3‐247: Dynamic Compaction at Fujairah Desalination Plant .................................... 517
Figure 3‐248: PPV and vibration frequencies ...................................................................... 519
Figure 3‐249: Comparison of PPV measurements with various methods for estimating PPV
............................................................................................................................................. 519
Figure 3‐250: Comparison of field measured PPV values with equations proposed by Mayne
(Equation 2‐152), Varaksin (Equation 2‐153) and the author (Equation 3‐21). .................. 520
Figure 3‐251: Location of Medina A’Zarqa .......................................................................... 521
Figure 3‐252: Plan of various phases of Medina A’Zarqa .................................................... 521
Figure 3‐253: Plan of Medina A'Zarqa Phase 1 .................................................................... 522
Figure 3‐254: Application of dynamic compaction at Medina A’Zarqa using a 23 t pounder
............................................................................................................................................. 523
Figure 3‐255: Dynamic compaction vibration monitoring at Medina A’Zarqa .................... 524
Figure 3‐256: PPV versus number of blow in DC phase 1.................................................... 524
Figure 3‐257: PPV versus number of blow in DC phase 3.................................................... 525
Figure 3‐258: PPV versus frequency in phase 1, phase 3 and ironing ................................. 526
Figure 3‐259: PPV versus distance and comparison with prediction equations ................. 526
Figure 3‐260: General location of UAQ Marina ................................................................... 527
Figure 3‐261: (a) Vibration monitoring without a trench, (b) vibration monitoring with a
trench ................................................................................................................................... 528
liii
Figure 3‐262: Comparison of measured and estimated PPV ............................................... 530
Figure 3‐263: PPV reduction factor when using a vibration reduction trench .................... 531
Figure 3‐264: : PPV generated by vibro compaction as per previous research .................. 533
Figure 3‐265: Location of vibro compaction vibration monitoring on Frond D of Palm Jumeira
............................................................................................................................................. 534
Figure 3‐266: PPV at various depths and distances from the vibroflot ............................... 534
Figure 3‐267: vibro compaction generated PPV versus distance ........................................ 535
Figure 3‐268: Rayleigh distribution strain as a function of pre‐treatment and DC induced
strain .................................................................................................................................... 543
Figure 3‐269: Ratios of actual and predicted limit pressure improvements for PMTs of
Figure 3‐30 to Figure 3‐33 .................................................................................................... 547
Figure 3‐270: Actual and predicted improvement ratios for (a) PLM, (b) EM ........................ 548
liv
Foreword
Mr Babak Hamidi has gone through an exceptional itinerary. Following the completion of his
civil engineering studies he commenced by working for a number of years in his country of
origin. He proceeded on to joining a ground improvement contractor in the Middle East
specialized in dynamic compaction techniques their design execution on turnkey basis; his
dream was to follow and advance the developments and ideas of Louis Menard, an objective
that was not only totally realized but even complemented.
During this period he was exposed to a number of the mega projects in the region and his
knowledge of soils under dynamic impacts became extensive. Even the questionable
theoretical concept of “dynamic surcharge” was first brought to reality and success on the
Dubai Palm project through his dedicated involvement. In summary, practice came before
the theoretical basis.
After his immigration to Australia, the idea to work on a document to assemble his and others
accumulated knowledge became more appealing. The aim of this work was to analyze limited
available theories on soil behavior under dynamic impacts, quality control and development
of technologies in terms of practical execution.
This “encyclopedia” can be considered as the first available complete document on the
“dynamic compaction or consolidation” process available for academics and practitioners for
the next decade. I would like to highlight within this document, the years of hard work and
persistence, as without these attributes this thesis would not have been written, this
combined with a full time activity.
Only experience is a lantern that is worn on the back and never illuminates the past path
(Confucius).
This work combined with further research and intuition will greatly contribute to the future
and to the state of the art.
Serge Varaksin
I.S.C. Chairman
TC211 – Ground Improvement Technical Committee
International Society of Soil Mechanics and Geotechnical Engineering
26 February 2014
lv
1 Introduction
1
1.1 Introduction
As construction and development progresses, available land in locations of interest becomes
scarcer and scarcer, and developers are forced to consider construction on sites with soils
whose in‐situ properties are not able to satisfy design requirements. Classically, structures
on such sites would have been built on piles, but construction of piles is expensive, and time
consuming. Hence, geotechnical engineers have been seeking other alternatives for
providing safe foundations, and have developed ground improvement.
Ground improvement can be categorised as (Chu et al., 2009):
Ground improvement without admixtures in non‐cohesive soils or fills
Ground improvement without admixtures in cohesive soils
Ground improvement with admixtures or inclusions
Ground improvement with grouting type admixtures
Earth reinforcement
Louis Menard, who had invented the pressuremeter test in 1954 (Communication
Department of Menard, 2007), developed dynamic compaction (DC) in 1969 (Menard
Soltraitement) and dynamic replacement (DR) in 1975 (Communication Department of
Menard, 2007).
In dynamic compaction the mechanical properties of soil is improved by applying high energy
blows to the ground by dropping a heavy pounder a multiple number of times from
significant heights. The impact creates waves that propagate in the soil, and re‐arrange the
particles in a denser configuration. The decrease of voids and increase in inner granular
contact will directly lead to improved soil properties. Dynamic compaction is classified as
ground improvement without admixtures in non‐cohesive soils or fills.
The efficiency of dynamic compaction decreases with the increase of fines content in
saturated soils due to the built‐up of pore water pressure and realisation of plastic
displacement in lieu of void reduction; hence, in such ground conditions the process of
dynamic replacement, which is classified as ground improvement with admixtures or
inclusions, can be adopted. In this latter technique the impact crater is backfilled with coarse
granular material and the column is further pushed into the ground and enlarged by applying
additional blows.
2
In the pressuremeter test a cylindrical probe is inflated in the ground. The measurements of
applied pressure and volume changes of the probe during the test realises the pressuremeter
limit pressure and Menard modulus. Foundation bearing capacity can be calculated using the
limit pressure (Centre D'Etudes Menard, 1975) without correlating the test results to the
parameters that are used in the classical Mohr‐Coulomb failure criterion based solutions and
bearing capacity factors (Bolton and Lau, 1993, Prandtl, 1920, Terzaghi, 1943). Foundation
settlements can also be calculated by using the Menard modulus and breaking down the
settlement to its volumetric and deviatoric components. The reliability of the pressuremeter
test and its calculation procedures (Centre D'Etudes Menard, 1975, Menard and Rousseau,
1962) has made this test the ideal verification method in ground improvement projects in
general and Menard invented techniques in particular.
Louis Menard died in 1978, but his legacy remained with his first generation of engineers
who turned the company that bore Menard’s name to one of the world’s leading ground
improvement specialist contracting firms. Serge Varaksin who joined Menard as a young
engineer was one of those first generation engineers who before his retirement became
overseas manager and deputy managing director of the company, and was appointed as the
chairman of ISSMGE (International Society of Soil Mechanics and Foundation Engineering)
Ground Improvement Technical Committee TC211 by ISSMGE’s President.
The author has had the honour of working for numerous years in industry under the
mentorship of Mr Varaksin. Combined with the more than 10,000,000 m2 of dynamic
compaction and dynamic replacement projects that the author was involved with has given
him the unique opportunity to participate in DC and DR state‐of‐the‐art projects with the
world’s leading expert.
Although information on dynamic compaction and dynamic replacement has been published,
the information appears to be scattered, and years has gone by since Lukas (1986, 1995) last
undertook the gathering of information on the subject. It is the author’s feeling that since
then significant advances have developed and evolved in the understanding of DC and DR, in
the equipment, and construction methods, which justifies a comprehensive review of these
technologies and their processes. Also, the author has been involved in a number of world
record projects, and first time projects that require documentation and preservation in order
not lost in history. Furthermore, the retirement of first generation of DC and DR engineers
carries the risk of knowledge loss, and it is imperative that their experience and skills be
maintained for future generations. Thus, this project is an attempt to gather and consolidate
3
information on DC and DR, and to document a number of distinguished projects based on
the author’s years of personal experience on the subject, and topics that, in his view, appear
to be frequently confused and misunderstood in practice.
1.2 Objectives
The objective of this thesis is to:
Collect the scattered information on dynamic compaction and dynamic replacement
Retrieve information on DR and DC that may have been lost or forgotten
Review and discuss state‐of‐the‐art research and publications on DC and DR
Collect, document and analyse distinguished state‐of‐the‐art DC and DR projects
Identify geotechnical problems and optimised treatment solutions
Identify optimised construction procedures in DC and DR projects
Document verification and testing procedures in DC and DR projects
Compare results of projects introduced and studied in this thesis with previous works
and data, with concentration on Menard Pressuremeter Tests
Develop new formulations and calculation procedures
Make recommendations and develop guidelines for future projects
1.3 Scope
This thesis will be presented in four chapters.
The current chapter, i.e., Chapter 1, is the introductory section of the thesis, and provides a
brief introduction on DC and DR for ground improvement and the pressuremeter test for
verification of results. The objectives of this endeavour, the thesis’ scope of works, and its
significance will also be presented in this chapter.
A comprehensive and thorough literature review and discussion will be presented in
Chapter 2. This review will commence with a brief introduction of Louis Menard’s
contributions and achievements before presenting the basics of the theory of waves and
vibrations in various cases. Next, the mechanisms of dynamic compaction and dynamic
replacement in different ground conditions will be reviewed. This will be followed by
reviewing the suitability of dynamic compaction in different ground conditions, and studying
4
DC parameters, which comprise of depth of improvement, crater depth, grid spacing, number
of compaction phases, number of pounder drops per phases, and extents of improvement.
Ground response to dynamic compaction, such as the improvement profile and ageing of
improved ground, will also be discussed. Similarly, design parameters of dynamic
replacement, which includes soil arching, column load capacity and liquefaction mitigation
analysis methods will be reviewed. As by this stage the reader will have sufficient
understanding of DC and DR mechanisms and parameters, the discussion will move forward
towards construction with the review of advances in equipment that is then followed by
assessing compaction vibrations and introducing methods for estimating DC induced
vibrations. A section in this chapter has also been allocated to quality control, verification
and testing, interpreting pressuremeter test results and calculating foundation bearing
capacity and settlement using the pressuremeter parameters. Chapter 2 in concluded with a
discussion of special topics that appear to be frequently confused and misunderstood in
practice.
A number of distinguished state‐of‐the‐art projects are documented, reviewed, and analysed
in Chapter 3, and the results will be compared with previous studies. This chapter will also
include new formulas for estimating DC induced vibration, estimating crushed rock internal
friction angle from the pressuremter test, and an innovative method to estimate
pressuremeter parameters based on DC induced ground settlements.
The conclusions of this thesis and recommendations will be presented in Chapter 4.
1.4 Significance
In general, ground improvement is an efficient and economical alternative to piling for
providing safe foundations, and in particular, dynamic compaction is one the most suitable
techniques for treating granular soils. This technology has been used for more than 40 years
to treat millions of square metres of soil. Although numerous researches have been carried
out to expand our understanding of this technique, most of the publications are scattered,
and some are no longer available to the public. This thesis will be an attempt to gather and
consolidate these publications as literature review.
As the author has had the unique chance to participate in a number of state‐of‐the‐art and
world record projects, this thesis is also an opportunity to document, review, analyse and
preserve these projects in an academic form that could be of practical use for future projects.
5
Review of previous work combined with presentation of the distinguished projects that will
be presented also provides an opportunity to compare results, update formulations, develop
new concepts, and make recommendations for future projects.
6
2 Literature Review
7
2.1 Chapter Organisation
This chapter has been allocated to the review of literature related to dynamic compaction,
dynamic replacement, and the pressuremeter test. The sections of this chapter include
information that is required to understand the processes of dynamic compaction and
dynamic replacement, performing and interpreting the pressuremeter test, and a number of
issues that the author has identified as subjects of general confusion among engineers who
do not have a thorough understanding of these ground improvement technologies.
Section 2.2 pays tribute to Louis Ménard, the inventor of the pressuremeter test, dynamic
compaction and dynamic replacement. Among other contributions, Ménard also formulated
the equations to interpret the results of the pressuremeter tests, and developed new
concepts for calculating bearing capacity and foundation settlements. This section also briefly
introduces dynamic compaction and dynamic replacement.
A basic knowledge of the theory of waves will be of great assistance to understand the
mechanism of dynamic compaction; hence, Section 2.3 will review waves in a number of
mediums, and in a homogeneous, isotropic, elastic half space medium. The review is then
further advanced by considering waves generated by a vertical oscillation, followed by the
introduction of damping. As in reality ground can be composed of layers of soil or rock with
various properties, elastic waves in layered system and mixed media will also be discussed.
The theory and mechanism of dynamic compaction and dynamic replacement will be
reviewed in Section 2.4. The discussion will be commenced in the micro level with the
simplified behaviour of elastic spheres subjected to shear, normal, a combination of the two
loads, and crushing of particles. Liquefaction is also discussed as it is a key process of dynamic
compaction in saturated soils. Further to the discussed basics, it will then be possible to
advance the discussion to the mechanism of dynamic compaction and dynamic replacement
in various ground conditions. At the end of Section 2.4 the reader will have an understanding
of the theory of wave propagation in the ground, and the mechanism of dynamic compaction.
Practical and design issues of dynamic compaction will commence at Section 2.5, and the first
discussion of the section will focus on ground conditions in which dynamic compaction can
be deemed as a suitable ground improvement technique. Then fundamental design aspects
including depth of improvement, crater depth, grid spacing, number of phases and number
of drops will be discussed. This section will also include a review of the improvement profile,
8
and the phenomenon of improvement continuation with time. The discussion in this section
will be concluded by determining the required extents of ground improvement and reviewing
several dynamic compaction design methods.
Similar to its preceding section, Section 2.6 will review dynamic replacement’s design issues,
and includes a comprehensive discussion on soil arching and calculation methods to
determine load distribution in the ground, load capacity of dynamic replacement columns,
and liquefaction assessment of composite ground.
Advances in dynamic compaction equipment will be reviewed in Section 2.7, and includes
discussions on compaction rigs, pounders and alternative impact oriented ground
improvement techniques that are sometimes erroneously referred to as dynamic
compaction or a derivative of dynamic compaction.
Noting that the basis of dynamic compaction and dynamic replacement is the propagation of
waves in the ground, the ability to assess the vibrations generated by dynamic compaction,
and their effects on nearby structures is of great importance. Hence, this subject will be
reviewed in depth in Section 2.8. Vibration assessment will commence with the review of the
vibration studies that eventually led to the current better known standards. Air vibrations
and human response to vibrations will also be reviewed. Estimation of vibrations based on
empirical formulas and mitigation of vibrations by application of vibration barrier or isolation
trenches will conclude this section.
Quality control and testing of dynamic compaction works will be reviewed in Section 2.9. This
important section of Chapter 2 includes the heave and penetration test and a complete
discussion on the pressuremeter test. The pressuremeter components and test procedure
will be explained, and the commonly used testing frequencies used in industry will be
defined. Also, a detailed review of calculation methods for determining bearing capacity and
foundation settlements will be presented. This section also includes correlations of
pressuremeter parameters with other testing methods.
Although Chapter 2 could have been concluded with Section 2.9, it is the author’s belief that
literature review of ground improvement methods would be incomplete if no discussion is
made on ground improvement acceptance criteria. Section 2.10 includes a brief discussion
of acceptance criteria specifications, and an in depth review of research that demonstrates
the unreliability of relative density concept and correlations as acceptance criteria. Other
9
special topics that will be discussed in this section are ground condition after sand
reclamations and the application of dynamic compaction in offshore projects.
A summary of the sections of Chapter 2 are shown in Figure 2‐1.
Figure 2‐1: Summary of literature review topics presented in the thesis
10
2.2 Tribute to Louis Ménard
Louis Ménard, shown Figure 2‐2, was born on 4 May 1931 in the Bay of Mont‐Saint‐Michel,
France (Communication Department of Menard, 2007). He attended the civil engineering
school of the prestigious École des Ponts et Chaussées in 1952. During the summer that he
was employed to carry out compaction tests on a new runway in Paris, Ménard realised that
while strength‐deformation properties were important, field tests and measurements were
not able to measure them. Consequently, he developed the first prototype of the Ménard
pressuremter as his dissertation and filed for a patent for it in 1954 at the age of 23.
Figure 2‐2: Louis Ménard (1931‐1978)
Ménard later improved his invention and carried out the first tests with the new probe while
studying for a Master’s degree (Menard, 1957) under the supervision of Professor Peck at
the University of Illinois and filed for a second pressuremeter patent in 1959.
The year 1962 was a big year for Ménard. By then, the Pressuremeter Test, commonly
abbreviated to PMT had become an internationally recognised in‐situ soil testing method,
and Ménard was revolving the accepted theories of soil mechanics. In the same year he set
up a bilingual technical journal, Sols‐Soils and split his company in two, Techniques Louis
11
Ménard (Louis Ménard Technologies) and Études Pressiométriques Louis Ménard (Louis
Ménard Pressuremeter Measuring).
There was still more to come for Ménard in 1969. Although others may have tried to compact
the soil by dropping weights onto the ground surface, it was Ménard who made a major
breakthrough by lifting a substantially heavy weight, or pounder, and dropping from a
considerable height, the technique he called compactage dynamique, translating to dynamic
compaction and commonly abbreviated to DC. Ménard first implemented dynamic
compaction at Bormes‐Les‐Mimosas in the French Riviera by renting a crane. This project
never became famous because there are no photographs available of that endeavour;
however, there is a photograph of Ménard’s third project (Figure 2‐3) at the marina in
Mandelieu‐la‐Napoule, also in the Riviera, that has made that project better known. This
latter project was 110,000 m2 of reclamation, and was to be become a development with five
storey buildings. Ménard used an 8 ton pounder with a drop height of 10 m. The energy
applied in that project was 10 times what is typically used in more recent times
(Communication Department of Menard, 2007).
Figure 2‐3: Dynamic compaction was used for the first time ever in 1969 for the ground improvement
works of a building construction site in Mandelieu‐la‐Napoule, France (Communication Department
of Menard, 2007)
Louis Ménard invented and promoted Dynamic Compaction (DC) as early as 1969, but it was
not until 29 May 1970 that he officially patented his invention in France. Ménard obtained
two patents, firstly compactage dynamique or dynamic compaction for damped fills, then
12
consolidation dynamique or dynamic consolidation for liquefaction mitigation (Varaksin,
2014). These techniques were later also patented in many other countries, including
Argentina, Australia, Austria, Bahrain, Bangladesh, Belgium, Canada, Egypt, France, Germany,
Great Britain, Greece, Hong Kong, Italy, Japan, Lebanon, Malaysia, Mexico, the Netherlands,
New Zealand, Norway, the Philippines, Saudi Arabia, Singapore, Spain, Sweden and the
United States.
It was in 1973 and during the Compagnie Française de l’Azote project in Le Havre, France that
Ménard became a true contractor. He soon signed overseas contracts for carrying out
projects in Sweden and Switzerland.
Ménard invented dynamic replacement (DR) in 1975 (Menard Soltraitement), but still had
more new ideas and innovations to offer until his death on 15 January 1978. Ménard also
had a mind set for developing equipment that were capable of lifting very heavy and
extremely heavy pounders, including the mega machine that was capable of lifting 40 ton
pounders 40 m (Mayne et al., 1984) and the Giga machine, shown in Figure 2‐4, that still
holds the world record for lifting a 172 ton pounder during the ground improvement works
of Nice International Airport in France (Gambin, 1983).
Figure 2‐4: The Giga Machine holds the world record for lifting a 172 ton pounder at Nice
International Airport dynamic compaction project (Communication Department of Menard, 2007)
Ménard’s contribution to ground improvement has not gone unnoticed and based on the
proposal of TC211, the Technical Committee of Ground Improvement, the International
Society of Soil Mechanics and Geotechnical Engineering (ISSMGE) board voted unanimously
(12 in favour, none abstaining, none opposing) to create a ISSMGE TC211 lecture called “Louis
Ménard Lecture” (Briaud, 2011).
13
2.3 Review of the Theory of Waves
The concepts of dynamic compaction and dynamic replacement are based on the
propagation of energy waves in the soil medium, and it is not possible to properly
comprehend these techniques without an introduction and understanding of the theory of
waves.
The equation of motion in a bounded elastic medium, such as a rod, can be expressed by the
partial differential equation:
2‐1
u = the displacement of an element in the direction
t = time
v = the wave propagation velocity
The phase velocity or longitudinal wave propagation velocity for a longitudinal wave, vc, can
be determined to be:
2‐2
E = Young’s modulus
ρ = mass density; i.e., ρ γ/g
γ = unit weight
g = gravity acceleration
If σx is the axial (compressive or tensile) stress induced by the longitudinal wave, then particle
velocity, ů , will be:
2‐3
14
Although many people confuse longitudinal velocity, vc, with the particle velocity, ů,
comparison of Equations 2‐2 and 2‐3 clearly demonstrates that these parameters are two
different entities that are only proportional to one another. It can also be understood from
the same equations that vc and ů are in the same direction when the stress is compressive,
but are in opposite directions when the stress is tensile. Another important observation that
can be made from Equations 2‐2 and 2‐3 is that while vc is only a function of material
properties, ů is additionally a function of the applied stress.
It can also be proven that for a torsional wave the shear wave velocity is:
2‐4
G = shear modulus; i.e.:
2‐5
2 1
ν = Poisson’s ratio
The equations of motion and wave propagation have also been solved (Richart et al., 1970)
for an infinite, homogeneous, isotropic, elastic medium. It can be determined that the
velocity of the dilatational wave vp, also known as the primary wave, P‐wave, compression
wave or irrotational wave, is equal to:
2
2‐6
λ = Lamé’s constant.
Similarly, the velocity of the distortional wave, vs, also known as the secondary wave, S‐wave,
shear wave or equivoluminal wave, can be determined to be equal to:
15
2‐7
It can be noted that while the particle motion associated with the compression wave in the
bounded elastic medium and the dilatational wave in the infinite medium are the same, the
wave propagation velocities are different. The comparison of Equation 2‐2 and Equation 2‐6
indicates that compression waves travel faster in an infinite medium than in a rod because
while there can be lateral displacements in a rod, the same is not true in the infinite medium.
The comparison of Equation 2‐4 and Equation 2‐7 demonstrates that the shear waves
propagate in the rod and the infinite medium with the same velocity.
Particle motion associated with the compression wave is a push‐pull motion parallel to the
direction of the wave front; however, the particle motion associated with the shear wave is
a transverse displacement normal to the direction of the wave front.
In addition to the compression and shear waves that propagate in an infinite medium, there
is also a third wave in a homogenous, isotropic, elastic half space medium. This wave, which
is called the Rayleigh or R‐wave in recognition of the first person who studied it, (Rayleigh,
1885) is confined to a zone near the boundary of the half space, and has later been described
in detail (Lamb, 1904).
8 24 16 16 1 0 2‐8
where
1 2
2‐9
2 2 2
From Equation 2‐8, it is evident that K2 and consequently vR is independent of the frequency
of the wave.
16
Figure 2‐5: Relations between Poisson’s ratio and P, S and R‐wave velocities in an elastic half space
medium (Richart, 1962)
(Richart, 1962). These ratios are graphically shown in Figure 2‐5.
As shown in Figure 2‐6, the horizontal and vertical displacement components of the R‐waves,
respectively denoted by U(z) and W(z), can be determined for different Poisson ratios and
depths. It can also be observed in this figure that the largest displacements and consequently
most of the vibration energy are within a depth that is about one third to one half of the
wave length.
Under the conditions considered by Lamb, the waves spread out from the point source in the
form of a symmetrical annular wave system. The characteristic wave system shown in
Figure 2‐7 will develop if input is of short duration. This wave system has three main features
corresponding to the arrivals of the P, S and R‐waves. A particle at the surface first undergoes
a displacement in the form of an oscillation due to the arrival of the P‐wave. After a relatively
quiet period the particle experiences another oscillation due to the arrival of the S‐wave.
Lamb refers to these events as the minor tremor. The major tremor, which has oscillations
with much larger magnitudes occur when the R‐waves arrive. Obviously, the time interval
between the wave arrivals becomes greater with increasing distances from the source.
17
Figure 2‐6: Amplitude ratio versus dimensionless depth for R‐waves (Richart et al., 1970)
Figure 2‐7: Wave system from surface point source in ideal medium (Lamb, 1904)
18
2.3.4 The Wave Field Generated by a Vertical Oscillation
(Richart et al., 1970)also discuss the wave field generated by a circular footing undergoing
vertical oscillations at the surface of the half space. The basic features of this wave field at a
relatively large distance from the source, i.e., a distance of 2.5 times the wave length (Lysmer,
1966), is shown in Figure 2‐8. In this figure, the distances from the source of waves to the
wave front are drawn in proportion to the velocity of each wave for a medium with a Poisson
ratio of 0.25 (also refer to Figure 2‐7 for the comparison of the wave velocities).
The body waves, i.e., the P and S‐waves, propagate radially outward from the source along a
hemispherical wave front that is shown by the heavy black lines of Figure 2‐8. The R‐waves
propagate radially outward along a cylindrical wave front.
The volume of material that is encompassed by each of the waves increases as the wave
travels away from the source. Hence, the energy density, i.e., the energy per unit volume, in
each wave front decreases with distance from the source. This decrease in energy density
and consequently the decrease in displacement amplitude is called geometrical damping,
which should not be confused with material damping that is a result of energy loss due to
hysteresis damping and internal sliding of soil particles (Thevanayagam et al., 2006).
Figure 2‐8: Distribution of displacement waves from a circular footing on a homogeneous, isotropic,
elastic half space (Woods, 1968)
Material damping is the decrease in amplitude of vibration with distance from a source due
to energy losses in the soil (designated as attenuation). Attenuation should be distinguished
from geometrical damping which occurs in elastic systems because of the spreading out of
the wave energy from a source.
19
If r= the distance of the wave front from the source, then it can be shown (Ewing et al.,
1957)that the amplitudes decrease proportionally to:
1/r 2 for P and S‐waves along the surface of the half space
1/r for P and S‐waves at other places
1/r ½ for R‐waves
This means that the body waves’ geometrical damping is highest along the surface of the
medium.
The shaded zones along the wave fronts for the body waves indicate the relative amplitude
of particle displacement as a function of the dip angle or the angle measured downwards
from the surface at the centre of the source (Hirona, 1948, Miller and Pursey, 1954). The
region of shear wave front in which the larger amplitudes occur is referred to as the shear
window.
It has been determined (Miller and Pursey, 1955) that for a vertically oscillating, uniformly
distributed, circular energy source on the surface of a homogeneous, isotropic, elastic half
space the distribution of the total energy among the three elastic waves is:
P‐wave: 7%
S‐wave: 26%
R‐wave: 67%
In the previous section of this chapter it was stated that the amplitude of the R‐wave
decreases proportionally with 1/r ½; i.e.:
2‐10
r1 = distance from source to point of known amplitude
r = distance from source to point in question
20
w1 = amplitude of the vertical component of the R‐wave at distance r1
w =amplitude of the vertical component of the R‐wave at distance r from the source
However, the wave amplitude decreases faster in soils because soil is not an ideal elastic
medium and there is an internal or material damping. Both geometrical damping and
material damping can be taken into account for R‐wave attenuation (Bornitz, 1931):
2‐11
α= coefficient of attenuation (in terms of 1/distance)
Figure 2‐9: Attenuation of surface wave with distance from source of steady‐state excitation (Richart
et al., 1970)
Values for α have been suggested to be from 0.03 to 0.13 m‐1 (Barkan, 1962). Figure 2‐9
suggests a wider range from 0.02 to 0.26 m‐1 (Richart et al., 1970). Based on results
measurements of ground vibration due to dynamic loads such as dynamic compaction,
21
vibrocompaction and ball dropping, Woods and Jedele (1985) proposed α values for different
soil descriptions, vibration frequencies and SPT N values aas shown in (Table 2‐1).
α (m‐1)
Soil Description and SPT blow counts
5 Hz 50 Hz
0.01 – 0.03 0.1 – 0.3 Weak or soft soils: lossy soils, dry or partially saturated peat and
muck, mud, loose beach sand, recently plowed ground, etc.
N<5
0.003 – 0.01 0.03 0.1 Competent soils: most sands, sandy clays, silty clays, gravels, silts
and weathered rock
5<N<15
0.0003 – 0.003 0.003 – 0.03 Hard soils: dense compacted sand, dry consolidated clay,
consolidated glacial till and some exposed rock
15<N<50
<0.0003 <0.003 Hard, competent rock: bedrock and freshly exposed hard rock
N>50
Table 2‐1: Attenuation coefficient α for different soil descriptions and vibration frequencies (Woods
and Jedele, 1985)
The coefficient α increases with dominant frequency, as a higher frequency wave will pass
through more motion cycles than will low frequency waves when travelling the same
distance. For material damping, decay is a function of energy loss per cycle of deformation,
not distance per second. This frequency dependency explains why in a general sense
dominant frequency declines with distance for the same wave type. The lower frequency
components have travelled fewer deformational cycles and have lost proportionally less
energy (Dowding, 2000).
Although material damping occurs in soil, it can be understood from Equation 2‐11 and
Equation 2‐12 that it is geometrical damping that produces most of the attenuation of the R‐
wave. Prange (1977) has also concluded that material damping is generally small for granular
media. Energy attenuation is related to the square of the amplitude of vibration (Wolf, 1994).
When a body wave that is propagating in an elastic medium reaches and interface with
another elastic medium, some of the incident wave energy will reflect into the first medium
and some of the energy wave will be transmitted into the second medium.
22
The nature of the reflected and refracted waves have been determined (Zoeppritz and
Kosten, 1919) using the theory of elasticity. In the study, shear wave was separated into two
components before consideration of energy partition at the interface. The two components
considered were the SV and SH‐waves, respectively designating the motion in a plane
perpendicular to the plane of the interface and a plane parallel to the interface. Upon
encountering an interface, an incident P‐wave, shown in Figure 2‐10a will have four
resultants:
Figure 2‐10: Partition of elastic wave at interface between two elastic media (Richart et al., 1970)
1. A reflected P‐wave (P‐P1)
2. A reflected SV‐wave (P‐SV1)
3. A refracted P‐wave (P‐P2)
4. A refracted SV‐wave (P‐SV2)
Similarly, the resultant waves from an incident SV‐wave will be (see Figure 2‐10b):
1. A reflected SV‐wave (SV‐SV1)
2. A reflected P‐wave (SV‐P1)
3. A refracted SV‐wave (SV‐SV2)
4. A refracted P‐wave (SV‐P2)
However, an incident SH‐wave only produces SH‐waves. P‐waves are not produced because
the SH‐wave does not have a component normal to the plane of the interface. The resultant
waves from an incident SH‐wave, shown in Figure 2‐10, are:
23
1. A reflected SH‐wave (SH‐SH1)
2. A refracted SH‐wave (SH‐SH2)
The angle at which a resultant wave leaves the interface depends on the angle at which the
incident wave approaches the interface and the ratio of the wave velocities of the two media.
From Snell’s law:
Where the angles a, b, e and f are measured from the normal to the interface (see
Figure 2‐10) and:
vp1 = velocity of the P‐wave in medium 1
vs1 =velocity of the S‐wave in medium 1
vp2 =velocity of the P‐wave in medium 2
vs2 =velocity of the S‐wave in medium 2
Noting that the energy transmitted by an elastic wave is proportional to the square root of
the displacement amplitude, the equations for the distribution of energy among the resultant
waves can be written with consideration of the below parameters:
A= amplitude of incident P‐wave
B= amplitude of incident S‐wave
C= amplitude of reflected P‐wave
D= amplitude of reflected S‐wave
E= amplitude of refracted P‐wave
F= amplitude of refracted S‐wave
ρ1= density of medium 1
ρ2= density of medium 2
24
For incident P‐wave:
For incident SV‐wave:
For incident SH‐wave:
0 2‐21
25
cos
0 2‐22
cos
The above equations demonstrate that the amplitude of each resultant wave is a function of:
1. The angle of incidence of the incident wave
2. The ratio of the wave velocities in the two media
3. The ratio of densities of the two media.
In the event that the media and consequently the media’s densities and wave velocities are
known, the amplitude of each resultant wave can be calculated. As an example and for a
situation where wave velocity and density in medium 1 are greater than medium 2, e.g. when
waves travelling in a very dense soil reach a very loose soil, Figure 2‐11 and Figure 2‐12 show
the amplitude ratios of reflected and refracted waves over a range of incident angles.
If there is more than one interface and as shown in Figure 2‐13, waves may be reflected back
to the surface from each interface. When reflected waves return to the surface of the half
space they encounter the interface between the solid medium and void where they will be
totally reflected. Multiple total reflections within the upper layer can generate another type
of surface wave called Love or L‐wave in recognition of the first person who described this
wave type. (Ewing et al., 1957) describe the L‐wave as a horizontally polarized shear wave
trapped in a superficial layer and propagated by multiple total reflections. L‐waves propagate
with a velocity that is between the S‐wave velocity of the superficial layer and the S‐wave
velocity of the next lower layer.
It is necessary for the phase velocity of the L‐wave to be less than the shear wave velocity in
the next lower layer for the L‐waves to be confined to the superficial layer. L‐wave will not
occur if the superficial layer is the higher velocity layer.
26
Figure 2‐11: Example of amplitude ratio versus incident angle for P‐wave (McCamy et al., 1962)
Figure 2‐12: Example of amplitude ratio versus. incident angle for S‐wave (McCamy et al., 1962)
27
Figure 2‐13: Multiple wave reflections and refractions in a layered half‐space (Richart et al., 1970)
Up to this point, this section has been about ideal solids, but soil is not such an ideal medium,
and is composed of solids and voids, which may be filled with air or water and described
either by porosity, n, or void ratio, e. By definition porosity is the ratio of void volume, Vv, to
total volume, V. Likewise void ratio is the ratio of void volume to the solid volume, Vs.
2‐23
2‐24
1
In a completely saturated soil, the voids are totally filled with water, and the degree of
saturation, S, is 100 per cent. The theory of propagation of waves in fluids is relatively simpler
than in elastic solids because ideal fluids do not have shear stiffness, cannot develop shearing
resistance, shear waves do not occur, and only compression waves need to be considered.
Depending on the water temperature and salinity, the bulk modulus of elasticity of water,
Bw, it can be calculated to be in the order of 2,200 MPa which suggests that, in comparison
to soil, water is quite incompressible. Solids suspended in liquids may produce an influence
on the propagation of the compression wave because their presence influences both the
mass density and the compressibility of the mixture. The effect of small amount of air in the
28
water portion of the mixture is to reduce the compression wave velocity significantly. For any
given saturation, S, the volumes of air, Va, water, Vw, and solid , Vs, per unit volume V of the
soil are:
1
2‐25
1
2‐26
1
1
2‐27
1
The total mass density of the solid ‐air‐water mixture, ρtot, is:
1
2‐28
1 1 1
GS= specific gravity of the solid particles
γw = unit weight of water
γa =unit weight of air
The bulk modulus of elasticity of the mixture, B, is:
1 1 1 1
2‐29
1 1
where:
2‐30
1 1
Bs=bulk modulus of solid particles
29
Ba=bulk modulus of elasticity of air
The compression wave propagation velocity in the solid‐air‐water mixture, vmix , is:
2‐31
It has been calculated (Richart et al., 1970) that only a small percentage (0.1%) of air in a
saturated mixture can reduce the bulk modulus of elasticity of water by 16 times.
Theory and research on wave propagation in porous saturated solids (Biot, 1956) shows the
strong influence of the structural coupling involved in the compression waves and the lack of
structural coupling for the shear waves. Hence, it can be confidently stated that field
measurements of shear waves in saturated soils determine the shear wave velocity in the
soil structure.
30
2.4 Theory & Mechanism of Dynamic Compaction & Dynamic
Replacement
The concept of this dynamic compaction is improving the mechanical properties of the soil
by transmitting high energy impacts to loose soils that initially have low bearing capacity and
high compressibility potentials. The impact creates body and surface waves that propagate
in the soil, and as shown in Figure 2‐14, re‐arrange the particles in a denser configuration.
The decrease of voids and increase in inner granular contact will directly lead to improved
soil properties.
Figure 2‐14: Displacement and rearrangement of the soil grains in a denser configuration
Impact energy is delivered by dropping a heavy weight or pounder from a significant height.
The pounder weight is most often in the range of 8 to 25 tons although lighter or heavier
weights are occasionally used. Drop heights are usually in the range of 10 to 20 m although
40 ton pounders have been dropped from 40 m by Menard’s Mega machine.
The behaviour of soil under the impact of the dynamic compaction or dynamic replacement
pounder is very complex and a function of a long list of parameters such as soil grading and
fines content, water content, soil layers, loading history, liquefaction and dissipation of pore
pressures, pounder drop height, pounder weight, pounder shape, pounder size, etc.
The deformation of granular materials upon loading involves particle rearrangement and
particle deformation (Cascante and Santamarina, 1996). Particle rearrangement and
consequent fabric changes occur at middle and large shear strains (in general greater than
10‐4). Fabric changes are affected by surface roughness, degree of frustration of particle
rotation, particle size distribution and inter‐particle forces. On the other hand, particle
deformation is the prevailing deformation mechanism at low strains and depends mainly on
31
the inter‐particle contact response and material properties including elastic, visco‐plastic and
brittle behaviours.
As a first approximation to a mass of cohesionless soil, an array of identical spheres with
radius, R, shall be assumed (Richart et al., 1970). The spheres are packed in a simple cubic
packing pattern in which each sphere is in contact with six neighbouring spheres, and there
are no large voids due to the absence of any missing spheres (Figure 2‐15a). This is the loosest
possible arrangement for equal sized spheres (Figure 2‐15c), and produces a void ratio equal
to 0.91.
4
8
3 0.91
4
3
If the layer on top of the first layer of spheres is translated by a shear load along a path
indicated by A‐A’ in Figure 2‐15a, then each sphere of the second layer will fit in a pocket
formed by the depression between four spheres of the first layer. If such a translation is
applied to all layers, then the pyramidal or face‐centred cubic packing of Figure 2‐15b will be
realised. The void ratio of the new packing can be calculated to be 0.35.
32 4
4
√2 3
0.35
4
4
3
The void ratio for the tetrahedral or close‐packed hexagonal packing (Figure 2‐15e) is also
0.35. This reduction in void ratio will decrease the volume or thickness of the packing system
by 29.3%.
∆
∆ 0.293
1
Even further reduction of void ratio is possible if all the spheres are not the same size, and
smaller spheres (or in soils, the grains) can be re‐arranged to fill in the remaining voids.
32
Figure 2‐15: Modes of regular packing of equal spheres (Richart et al., 1970)
The behaviour of elastic spheres subject to a normal load has also been studied (Hertz, 1881,
Timoshenko and Goodier, 1951). Figure 2‐16a and Figure 2‐16b show respectively two
spheres of equal radius R just touching and the deformation that occurs when a normal force
presses them together.
Upon application of the normal load, P, the radius of contact, a, will be:
3 1
2‐32
8
33
Figure 2‐16: Behaviour of equal spheres in contact. (a) Spheres just touching, (b) deformation by
normal force, (c) shearing forces between particles in cubic packing, (d) lateral deformation by
shearing forces (Richart et al., 1970).
The reduction between the centres of the two spheres, ζ, will be:
2
2‐33
The normal stress, σ, at any point on the contact surface that is at a distance r from the
contact surface can be calculated to be:
3
1 2‐34
2
The maximum stress (in the centre of the contact surface) is 1.5 times the average stress
value, and can be calculated to be:
3 4
2‐35
2 1
The average stress equals to the force transmitted through a row of spheres divided by the
effective square cross section of the row (see Figure 2‐15c).
34
2‐36
4
A more complicated situation may arise when, as shown in Figure 2‐16c, the force is applied
along a diagonal direction, creating normal and shear forces. This problem has been studied
theoretically (Cattaneo, 1938, Deresiewicz, 1974, Mindlin, 1949, Mindlin and Deresiewicz,
1953).
The equations for normal loading were explained in Section 2.4.2. If, as shown in
Figure 2‐16d, the two spheres of Figure 2‐16b are subjected to a normal load and a shearing
force, S, then:
2‐37
4
Figure 2‐17: Hysteresis loop formed by shearing force‐displacement relations for two equal spheres
pressed together and subjected to shearing forces (Richart et al., 1970).
The shear force produces a shear displacement, η, between the centres of the spheres. If η
is plotted against S for loading and unloading cycles (without exceeding the limit of μP, where
μ= the coefficient of friction between the spheres), the hysteresis loop of Figure 2‐17 will be
produced.
These studied show that a concave stress‐strain will be recognised when lateral expansion of
the mass of spheres is permitted, and that the stress‐strain behaviour depends on the
amount of lateral strain developed during axial loading. This discussion also demonstrates
that energy is dissipated through friction before sliding.
35
2.4.4 Crushing of Soil Particles
In addition to the strain‐strain behaviour that was modelled and described by elastic spheres
that are subjected to normal and shear loads, soil particles can be crushed or broken during
loading. The latter will increase volume reduction. For the type of deformation where soil
grains move past or around each other, crushing at sliding contacts or breakage of particles
decreases the rate of dilation corresponding to a given principal stress ratio (Hardin, 1985).
Hardin (1985) assesses that the amount of particle crushing that occurs in an element of soil
under stress depends on the particle size distribution, particle shape, state of effective stress,
effective stress path, void ratio, particle hardness (hardness of cementing material or
weakest constituent of a particle and weakest particles of an element) and the presence or
absence of water. Based on the equations that Hardin has developed, he demonstrates that
the crushing hardness of angular shaped particles is more than round shaped grains.
The potential for breakage of a soil particle increases with its size. The normal contact forces
in a soil element increase with particle size. Also, the probability of a defect in a given particle
increases with size. The breakage of soil particles in a rock fill under moderate stresses may
be quite evident, whereas very high stresses are required to crush silt size particles.
In principle, it would be possible to crush all soil particles under extremely high stresses to
an extent that no particles remain larger than 0.074 mm. Thus, it may be expected that
crushing and breakage of soil particles will increase with the magnitude of the state of
effective stress.
The amount of crushing produced by a hypothetical stress may be expected to decrease with
increasing particle concentration because the particle contact forces will be reduced by
increased concentration. Hence, it can be visualised that loose soils with higher void ratios
are subject to more breakage than dense soils.
2.4.5 Liquefaction
Liquefaction is the phenomenon in which the soil loses its strength due to increase in pore
water pressure. In cohesionless soils the shear strength, τ, of the soil is due to internal friction
only. In a saturated state it may be expressed as:
36
tan ′ 2‐38
σ= total normal stress on the failure plane
u = pore pressure
φ'= effective angle of internal friction
It can be well observed that an increase in u will result in a decrease in τ to the point where
the soil’s strength becomes zero and the soil liquefies.
Saturated soils can liquefy due to repetitive pounder impacts. The dynamic forces disturb the
(granular) soil skeleton, and tend to force the particles to compact into a denser
arrangement, which consequently will mitigate liquefaction potential.
In granular DR columns, conceptually, even if high excess pore water pressure build‐up occurs
within the remediated soil, the imposed shear stresses during seismic loading will be shared
between the dense granular columns and the in‐situ soil in proportion to the relative stiffness
of the two materials; thus increasing overall stability (Adalier and Elgamal, 2004, Baez and
Martin, 1993).
As is the case for dynamic replacement columns, according to Murali Krishna and Madhav
(2009) granular inclusions mitigate liquefaction because:
Granular columns function as drains and permit rapid dissipation of induced pore
pressures by virtue of their high permeability with the additional advantage that they
tend to dilate as they get sheared during an earthquake event.
Excess pore water pressures generated by cyclic loading are dissipated almost as fast
as they are generated due to significant reduction in the drainage path.
Granular columns compact, reinforce, and improve the deformation properties of
the in‐situ soil.
Granular columns are installed in to a very dense state, and are thus not prone to
liquefaction. The columns also replace a percentage of in‐situ liquefiable soil.
Granular columns modify the nature of earthquake experienced by the in situ soil.
37
Furthermore, Adalier and Elgamal (2004) report a number of advantages for maintaining low
excess pore water pressure to initial effective vertical stress ratio (ru):
1. A large portion of the overall soil strength and stiffness is preserved. This will enable
the stratum to continue providing the necessary vertical and lateral support to the
overlying structure. Maintaining strength may substantially decrease the extent of
lateral ground deformation due to dynamic excitation.
2. Preventing large total and differential settlements, which are often associated with
ru values above 0.5 to 0.6 (Ishihara and Yamazaki, 1980, Lee and Albeisa, 1974,
Nagase and Ishihara, 1988, Seed et al., 1976, Tokimatsu, 1979). Large settlements at
high ru values are caused by the high volume compressibility of liquefied soil at low
confining stresses.
3. Reducing high hydraulic gradients that may cause migration of fine soil particles into
the gravel drain; thus reducing the drainage capability.
Seed and Booker (1977) have proposed that the permeability of gravel columns should be at
least two orders of magnitude larger than the surrounding soil to avoid significant generation
of excess pore water pressure within the columns.
In non‐saturated soils, wave energy produces a stress that overcomes the inter‐granular
friction resistance of the soil grains, and displaces the soil. The grains are consequently
arranged in a denser configuration, and the soil undergoes a settlement (see Figure 2‐14).
The re‐arranged soil matrix will have more inter‐granular connections; i.e., more friction and
thus better mechanical properties.
Similarly, Lukas (1986) describes the process of dynamic compaction densification in moist
(partially saturated) soils above the water table primarily due to compaction. Without
entering any theories, he states that the energy imparted into the ground causes the particles
to move closer together, and consequently results in an increase in the unit weight and
strength, and a reduction in the compressibility of the soil.
38
Nashed (2005) notes that impact energy loss is responsible for the improvements
experienced by the deposits due to densification techniques.
Airey et al. (2012) have performed small scale laboratory dynamic compaction tests using
high speed photography and digital image correlation techniques to investigate the
deformation patterns, calculate soil strains and observe strain localisation. In the experiment
a 2 to 1 dry mixture of Sydney sand and non‐plastic feldspar silt was used. The test results for
drop heights that were 150 to 300 mm suggest that densification occurs due to a series of
shock or compaction bands that propagate through the soil, starting from just below the
tamper base. Likewise, the sequence of the photographs suggest the compaction bands
move through the soil, gradually slow down and ultimately come to a stop at some distance
away from the pounder. The conclusion is that as expected, in this process the displacements
do not diminish steadily with depth below the pounder. Also the bands were still detectable
in the cumulative displacement vectors at the end of the series of drops, and most of the
volumetric compression beneath the pounder was concentrated in these bands. Airey et al.
reason that shear and volume strain are negligible in between the bands because the positive
strains in front of the band are counteracted by negative strains behind it, so that the net
effect on the soil after the compaction band has passed is negligible. With further impacts
additional compacted bands that lead to the numerous shear strain bands are formed. This
experiment was carried out to six blows, and the expectation of Airey et al. was that with
more flows, there will be more shear bands filling in much of the space between the pounder
and the furthest shear band.
The findings of Airey et al. are very interesting, but do not agree with the personal experience
of the author and conclusions of other dynamic compaction physical modellings. The author
has performed dynamic compaction with limited number of blows in loose dry sands, but the
verification tests, including Cone Penetration Tests (CPT) and Pressuremeter Tests, have not
shown any signs of banded improvement patterns.
Hajialilue‐Bonab and Rezaei (2009) and Hajialilue‐Bonab and Zare (2014) have also carried
out particle image velocimetry of dynamic compaction to detect soils displacements at high
speed. In their experiment Hajialilue‐Bonab and Rezaei used fine dry sand from Ana Khatoun,
east of Tabriz, and applied up to 15 blows per test. In order to model an axisymmetrical half
space, in this experiment pounders in the shape of half discs and weighing 6.5 to 20.04 N
were dropped tangentially to the transparent Perspex facing from heights of 567.5 mm to
912.5 mm. Hajialilue‐Bonab and Rezaei have not reported strain bands as reported by Airey
39
et al. (2012). The initial relative density of the test shown in Figure 2‐18 was 25%. Depth and
radius in Figure 2‐18 were normalised by dividing them by the pounder diameter.
Figure 2‐18(a) shows the variation in relative density after 15 blows. Figure 2‐18(b) shows the
variation of normalised depth of improvement for different number of pounder drops. The
results of this experiment is compatible with what has been reported by Lukas (1986), and it
appears that the depth of improvement changes little after the first several pounder blows.
Hajialilue‐Bonab and Zare (2014) who have advanced the research of Hajialilue and Rezaei
(2009) by performing similar particle image velocimetry of dynamic compaction physical
modelling have also arrived at similar results to Hajialilue and Rezaei for single and double
point poundings.
(a) (b)
Figure 2‐18: (a) Distribution of relative density variations and (b) variation of relative density versus
depth under the pounder centre for different number of impacts (Hajialilue‐Bonab and Rezaei, 2009)
The author has understood from his personal experience that many people are under the
incorrect impression that dynamic compaction is caused by compression. This may be due to
a simple misunderstanding that compaction is a result of compression. However, this is not
the case; it has already been explained in the previous section that (shear) displacement can
cause substantial amounts of settlement simply by rearranging the soil matrix. The contact
stress and deformation of soil grains under normal loading was also explained and
formulated. The amount of plastic compression that a soil grain will undergo is negligible.
Some grains may crush due to the high impact stresses, but that also results in additional
settlements originating from shear displacement of the grains and smaller grains fitting into
the smaller voids of the soil matrix.
40
Gu and Lee (2002) have studied the mechanics of dynamic compaction in dry sand using two‐
dimensional finite element analyses with a large‐strain dynamic formulation and a cap model
for soil behaviour. Included in the study was the mechanism of ground improvement for a
case in which the pounder weight and drop height were respectively 20 tons and 20 m, and
the changes at four points, marked as A to D in Figure 2‐19, were examined. Points A and B
were 0.1 m away from the centreline, and points C and D were 4 m away from the centreline.
Point A and C were at the depth of 2 m, and point B and D were 6 m deep.
Figure 2‐19: Two‐dimensional axisymmetric mesh (Gu and Lee, 2002)
The stress path of point A during the first three blows is shown in Figure 2‐20 . On the first
impact, the stress path initially follows the elastic K0 compression line closely until the first
cap is reached. From this point onwards, plastic strains become more and more dominant.
As plastic strains are governed by the associated flow rule rather than Hooke’s law for elastic
strains, the earth pressure coefficient and stress path are different in the two conditions.
Therefore, as plastic strains begin to dominate soil deformation, the stress path stirs towards
the plastic K0 compression line, and follows the latter line closely for much of the loading
phase, and as the cap expands. Towards the end of the loading phase, the stress path moves
towards the shear yield surface, and this path continues for the initial unloading phase.
41
Figure 2‐20: Computed p’‐q (mean effective normal versus shear) stress path at point A located 2 m
below ground surface and 0∙1 m from centreline (Gu and Lee, 2002)
Gu and Lee explain the stress path by the effects of lateral inertia as explained by Goh et al.
(1998) who have noted that lateral inertia causes lateral strains to lag behind vertical strains
in a sand column. Prior to the onset of lateral strains, the sand column compresses in nearly
K0 condition. For the soil beneath the tamper, the onset and rate of increase of lateral strains
are also much slower than those of vertical strains, so that the nearly K0 condition will persist
until a sufficiently large lateral strain has occurred. In the first impact a significant proportion
of the void ratio reduction occurs at less than 0∙2% lateral strain magnitude. As lateral strain
accumulates, the stress path diverges from the K0 line towards the shear failure line under
triaxial compression. This effect continues into early unloading, even when the mean
effective normal stress, p’, is decreasing, being sustained by the lateral outward momentum
of the soil beneath the pounder. As the soil dilates, the cap shrinks to the point where lateral
outward momentum is dissipated. Towards the end of unloading, the high K0 induces a
passive (or extension) stress state. Since q, the shear stress is always positive, as shown in
the inset of Figure 2‐20, the transition from a compression to an extension state is
demonstrated by kinks in the stress paths at the hydrostatic axis.
42
Further blows produce larger peak compressive stresses at the same location. Also, the
computed zone of improvement deepens with successive blows because, by then, the
overlying soil responds elastically more during the loading phase, which results in less energy
dissipation. Therefore, additional blows create a deepening compacted soil plug, which
enhances propagation of the impulsive stress further downwards. However, the incremental
improvement at a given point reduces with successive blows. The net decrease in void ratio
attenuates significantly by the third impact, but the crater depth continues to increase. The
rapid attenuation of void ratio decrease occurs because progressively higher stresses are
needed to enforce further plastic volumetric compression. Furthermore, with successive
blows, progressively larger portions of the stress path remain in the elastic zone during
loading. By the time plastic volumetric strain commences, significant lateral strains have
already occurred, and the stress path starts veering towards the shear yield surface. Thus,
with successive blows, the state of the soil along or near the centreline moves from K0
compression to active shear upon loading. Hence, for regions directly below and near to the
crater surface, limiting improvement is reached when plastic volumetric compression is
entirely offset by shear‐induced dilatancy.
As shown in Figure 2‐21, the compaction process is similar at point B; however, more blows
will be required to achieve a yield surface of equal size, which indicates that the rate of
improvement will be slower. Gu and Lee reason that this is due to the lower impulsive stress
and the smaller rate of increase in peak dynamic stress, which arises from the dissipative and
dispersive effects of the overlying soil. Thus the number of blows needed is determined
largely by the soil near the outer fringe of the zone of improvement. Furthermore, there will
be little or no dilatancy after the peak stress is reached because the dynamic stress level is
lower, and the geostatic stress level at greater depths is higher.
The stress path at point C is shown in Figure 2‐22. Here, the early loading stress path of each
impact is much steeper than those directly beneath the tamper. Furthermore, the peak
lateral stress is higher than the peak vertical stress, indicating that the soil is in a passive
state. Therefore, the soil outside the tamper footprint is compacted largely by passive
compression arising from the lateral outward momentum of the soil beneath the tamper.
Also, there is a delay in the onset of improvement in point C compared with point A, which,
according to Gu and Lee, can be explained by the time needed to overcome the lateral inertia
and build up the lateral compressive wave. On the other hand, there is also a smaller delay
43
between points B and D, which is consistent with the change in the shape of wavefront from
planar to spherical.
Figure 2‐21: Computed p’‐q stress paths at point B located 6 m below ground surface and 0∙1 m from
centreline (Gu and Lee, 2002)
Figure 2‐22: Computed p’‐q stress paths at point C located 2 m below ground surface and 4 m from
centreline (Gu and Lee, 2002)
44
2.4.7 The Mechanism of Dynamic Compaction in Saturated Granular Soils
In saturated granular soils the pounder impact creates a wave that increases the pore
pressure until the soil liquefies, and the soil grains rearrange in a denser configuration.
According to Lukas (1986), in saturated soils the energy applied to the soil causes an
immediate increase in pore pressure and a reduction in effective stress. As the pore pressure
dissipates the soil particles move into a denser state of packing under the confining pressures
of the overburden.
Figure 2‐23: Effects of waves in the densification of loose sands (Gambin, 1997)
Liquefaction of saturated granular soils is a process involving energy dissipation due to
frictional loss along grain contacts during cyclic loading, leading to contact slips and instability
of the soil structure and an increase in excess pore pressures (Thevanayagam et al., 2006).
Studies have shown that the magnitude of induced excess pore pressure due to undrained
cyclic undrained loading in a saturated energy soil is related to cumulative energy dissipated
per unit volume of soil. If the energy dissipated in the saturated loose soil due to dynamic
compaction exceeds the energy required to cause liquefaction on a per volume soil basis,
then the soil liquefies. Thervanayagem et al. (2006) add that the objective of a vibratory
densification scheme is to impart sufficient energy to the soil to be improved by repeated
applications of vibratory energy to repeatedly liquefy and densify the soil until its density
increases sufficiently.
Without reference, Slocombe (2004) states that laboratory and in‐situ tests have consistently
shown that in order to achieve maximum density, the lowest number of stress impulses to
attain the required energy input will provide the optimum result and further adds that
saturated granular soils normally require higher treatment overall energy, in a larger number
of tamping passes, than if the soils were essentially dry.
45
2.4.8 The Mechanism of Dynamic Compaction in Saturated Cohesive Soils
Irrespective of the published case studies (Menard and Broise, 1975, Perucho and Olalla,
2006) that have demonstrated the effectiveness of dynamic compaction for the treatment
of saturated cohesive soils, the application of this technique to such soils has always been
controversial, and it is the personal experience of the author that the improvement obtained
by dynamic compaction in soils classified as Zone 2 and Zone 3 in Lukas’ diagram
(Figure 2‐35), if any, is only a fraction of what could have been achieved in a granular soil of
Zone 1.
It is a fact that the classical consolidation theory (Terzaghi and Frohlich, 1936) that was
anticipated for explaining the behaviour of saturated cohesive soils does not describe how
dynamic compaction could treat saturated soils. However, the reason may be understood
once the theory’s assumptions are reviewed and compared to the process that takes place
during dynamic compaction.
Louis Menard, himself, explains how dynamic compaction is able to treat saturated cohesive
soils in the famous paper that he wrote with Broise (1975). This historic paper was the first
publication about dynamic compaction in English and many people mistakenly believe that
it is the first paper to have ever been published about dynamic compaction (in fact papers
were previously published in French e.g. Menard (Menard, 1972). Menard explains his
consolidation theory based on compressibility, liquefaction, permeability and thixotropic
recovery.
Terzaghi assumes that fine saturated soils are incompressible when subject to rapid loadings
because their low permeability does not allow the rapid drainage of pore water and the
settlement due to volume changes, however Menard’s early observations indicated that
performing dynamic compaction did cause the saturated impermeable soils to settle
immediately. According to Menard and Broise, subsequent research showed that most
quaternary soils contained between 1 to 4% of gas in the form of micro‐bubbles. As a first
approximation, Menard and Broise assumed that the variations in volume of the micro‐
bubbles may be essentially governed by the laws of Marriotte and Henry, but went on to
state that in fact other less known phenomena also play a fundamental role.
The gas volume in the soil gradually reduces as the soil undergoes repeated impacts. As the
percentage of gas approaches zero the soil begins to react as an incompressible material and
46
liquefies. Menard and Broise define the energy level to reach this stage as the saturation
energy.
Menard and Broise note that liquefaction in natural soils will often occur gradually. Most
natural deposits are layered and structured, and the silty or sandy components will liquefy
before the clayey material. It is important that liquefaction of these layers occur while
liquefaction of the main clay body must be avoided in order to prevent remoulding of the soil
mass. It is therefore imperative to know the exact level of energy corresponding to this
threshold condition which is essential to develop high pore‐water pressures as to reach high
permeability. Once the saturation energy has been reached, further introduction of energy
would be totally wasted.
A particular feature that Menard and Broise observed on dynamic compaction sites was the
initial very rapid dissipation of pore water pressure, which could not be explained by the
coefficient of permeability measured before dynamic compaction. Menard and Broise
consider this phenomenon as general and apparent in all soils when conditions tend towards
liquefaction, regardless of the grain size. They state that a very slight local increase of pore
water pressure is sufficient to commence tearing of the solid tissues by splitting, and the flow
of liquid quite naturally concentrates in these newly created fissures.
Figure 2‐24: Appearance of water fountains from the fissures due to dynamic compaction
By concentrating the pounding energy at regular grid locations, vertical fissures that are
regularly distributed around the impact points are created. These preferential drainage areas
are generally perpendicular to the direction of the lowest stress. As shown in Figure 2‐24,
fountains of water that appear near the craters sometime after pounding under certain
geological conditions are initiated and fed by this flow network.
47
It is the opinion of Varaksin (2014) that rapid dissipation of pore water pressure could be
expected and is not unusual when loading is removed. Also, regardless of the air bubbles
compression, the experience of the author suggests that improvement in clay is small and in
the order of a few per cent.
Menard and Broise have noted that irregular and disorderly pounding can disrupt the
continuity of the fissures and render reconstitution more difficult for later and better planned
dynamic compaction passes. They also hypothesise that the impact waves transform the
absorbed (solid) water into free water. This would encourage an increase in the sectional
area of the fissures. The reverse of this phenomenon occurs when the soil resets due to
thixotropy.
A considerable reduction in shear strength can be initially observed during dynamic
compaction. The minimum value will occur when the soil is liquefied or approaching
liquefaction. At that time the soil is completely torn, and the absorbed water that plays an
important part in stiffening of the soil structure is partially transformed to free water. As pore
water pressure dissipates, a substantial increase in shear strength and deformation modulus
can be noted. This phenomenon that may continue for several months, and in the opinion of
Menard and Broise is due to the closer contact between the particles as well as the gradual
fixation of new layers of absorbed water is called thixotropy.
The mentioned above discussion can be presented as shown in Figure 2‐25 by using a
modified presentation of Terzaghi’s hydraulic system of a cylinder filled with incompressible
fluid and supported by a spring. Unlike the classical consolidation model, in the dynamic
consolidation model the pore water filling the cylinder is considered to be partially
compressible due to the presence of micro‐bubbles. In the classical model there is no friction
in between the piston and the cylinder. In Menard and Broise’s model there is friction, which
results in hysteresis in the interaction between the hydraulic pressure increase and the
intensity of piston surcharge. Hence, a reduction in the pressure of the liquid does not
automatically result in a piston movement or a change in the spring. This illustrates a fact
that is often observed in foundation soils; i.e., the reduction of pore water pressure without
a corresponding settlement of the construction being investigated. In the classical model the
spring stiffness (the soil’s modulus of deformation) is assumed to be constant. However,
considerable modifications of the modulus can be observed under the influence of
alternating loading. The absorbed water plays an important part in the process as it becomes
partially free. This results in a weakening of the mechanical bond between the solid particles.
48
Consequently, there will be a reduction in the overall strength of the material. In the classical
model, permeability is assumed to be constant, but in Menard and Broise’s model,
permeability is represented by a variable nozzle section.
Figure 2‐25: Comparison of the classical and dynamic theories of consolidation (Menard and Broise,
1975)
Figure 2‐26(a) illustrates the changes in the soil after one pass. The topmost diagram shows
the energy applied to the soil by a series of impact at the same print location. The diagram
below it presents the soil’s volumetric changes, which is then followed by the diagram
showing the changes in the pore water pressure. The bottommost diagram refers to the
changes of the bearing capacity as a function of time.
Figure 2‐26(b) presents diagrams similar to what is shown in Figure 2‐26(a) but for multiple
numbers of passes. Menard and Broise note that although energy follows an arithmetic
progression, the volume changes and bearing capacity do not follow the same law.
While in highly permeable soils, such as sands and gravels, the excess pore water pressure
usually dissipates within minutes after the impact, in semi‐pervious soils such as silty soils,
the time required for excess pore pressure dissipation can range from a few days to a few
weeks. In saturated fine soils, it may take weeks for the pore water pressure to dissipate.
49
Consequently, this has discouraged some researchers such as Lukas (1986) to use dynamic
compaction in (pseudo) impervious soils.
Figure 2‐26: (a) Changes in the soil conditions after one dynamic compaction pass, and (b) changes
in the soil conditions after a multiple number of DC pass (Menard and Broise, 1975)
There is almost no publication describing the process of dynamic replacement in saturated
cohesive soils. What is known by the author is that if sufficient time is not given for the pore
water pressure to dissipate, the pounder impacts will cause mostly plastic displacement
rather than volumetric change in the soil. In other words, the soil will deform plastically with
minimal, if any, volumetric changes, which will result mostly in soil displacement rather than
reduction of voids and consolidation. If the overburden pressure of the soil is high, then the
displacement will be more laterally oriented. Otherwise the soil will also be pushed towards
the ground, and the working platform will heave and perhaps even crack.
50
If the soil does not displace or only somewhat displaces due to the impact, then the pounder
impact must have compacted it. This means that the in‐situ soil is behaving more granular
(or pervious) than anticipated, and the result of pounding is a mixture of dynamic compaction
and dynamic replacement.
When dynamic replacement is implemented, the foundation system usually includes a well
compacted granular blanket on top of the DR columns. As illustrated in Figure 2‐27, in such
a case the loads will be transferred to the soft soil – DR column system by arching in the
transitional granular layer that is called the load transfer plataform. The arching effect and
the load distribution by arching will be further discussed in Section 2.6.1.
Figure 2‐27: The mechanism of load transfer to the improved DR – in‐situ soil system
The dynamic replacement crater is backfilled with crushed stone that is dynamically
compacted, and pushed into the ground. Additional blows will further drive the material
down, with penetration decreasing per blow, to a point where further blows will negligibly
compact or drive the crushed stone down any further.
Hughes and Withers (1974) have performed experiments on model sand columns that were
150 mm long. Column diameters were from 12.5 to 38 mm and as shown in Figure 2‐28, loads
were only applied to the top of the columns. It can also be noted that the columns were
terminated within the soft clay layer with the same cohesion throughout the layer. As shown
in Figure 2‐29, Hughes and Withers noticed that irrespective of the column length, the
pattern of vertical and radial deformation of the columns were basically the same, and
51
considerable vertical and lateral distortion that occurred at the top of the column rapidly
diminished with depth. At failure a length of about four column diameters were significantly
strained.
Figure 2‐28: Consolidation stresses on the clay before the test, and consolidation stresses on the
clay and the loading stresses on the column during the test (Hughes and Withers, 1974)
52
Figure 2‐29: (a) Vertical displacement within the column versus depth (b) Radial displacement at the
edge of the column / initial column radius versus depth (Hughes and Withers, 1974)
The results showed that the ultimate strength of an isolated column loaded at its top only
was governed primarily by the maximum lateral reaction of the soil around the bulging zone,
and that the extent of vertical movement within the column was limited. In this particular
experiment the vertical movement did not go below four column diameters, and Hughes and
Withers concluded that it appeared that if the column length to diameter ratio was less than
4, then the column would fail by end bearing before bulging.
It is interesting to note that although the clay properties at the top and at depth were both
the same, column bulging happened only at the top. By equating the forces on a horizontal
element of the column and with the assumption that the undrained cohesion was constant
over the depth of the column, Hughes and Withers calculated that the vertical stress at depth
reduces to zero at a critical depth (4.1 column diameters when the undrained cohesion is
constant throughout the treatment depth). As the column length is reduced some of the
applied vertical stress will be taken by the soil at the base, and the columns will utilise end
bearing resistance. If the columns are made so short that the stress at the base exceeds the
bearing capacity of the soil (about 9 times the undrained cohesion), then end bearing failure
mode will occur before the bulging. When the length of the column is equal to the critical
depth end bearing and bulging failure occur simultaneously, and beyond the critical depth,
the stresses in the column reduce to below the ultimate bearing of the cohesive soil and the
column will fail by bulging.
Another conclusion of this experiment was that when the undrained cohesion was constant
in the soft soil, once the column exceeded about four times its diameter, then any additional
length did not increase the column’s load capacity because the column failed by bulging, and
not in end bearing. However, the author notes that although the ultimate load capacity of
the column will not increase beyond about four times its diameter, it may still be necessary
to construct longer columns to solve other problems such as settlements and liquefaction.
The failure mechanism as discussed by Hughes and Withers is an idealised condition with the
assumption that soil properties are constant throughout the treatment depth. However, this
hypothetical ground condition may not be case in many instances. As shown in Figure 2‐30,
isolated zones of very soft cohesive soils can result in significant bulging at both shallow and
deep depths.
53
Figure 2‐30: Column failure mechanism in non‐homogeneous cohesive soil: (a) Soft layer at surface
– bulging or shear failure, (b) Thin very soft layer – contained local bulge, (c) Thick very soft layer –
bulging failure (Barksdale and Bachus, 1983)
Hughes and Withers also noticed that, as shown in Figure 2‐31, only the clay within a cylinder
with a diameter of about 2.5 times the column diameter is significantly affected by the
loading. This suggests that the columns could act independently if placed more than 2.5
diameters apart.
Figure 2‐31: Tracing of the superposition of markers due to after loading. A selection of the markers
are joined by arrows (Hughes and Withers, 1974)
54
Figure 2‐32: Failure Mechanisms of a single (DR) column in a homogeneous soft layer: (a) Long
columns with firm or floating support – bulging failure, (b) Short column with rigid base – shear
failure, (c) Short floating column ‐ punching failure (Barksdale and Bachus, 1983)
Based on the works of Hughes and Withers (1974), and as shown in Figure 2‐32, Barksdale
and Bachus (1983) have considered the case where a column is loaded just over the area of
the column. Without explaining the reason for the difference between what they understand
is the critical length where either bulging or end bearing will govern, they report that when
the column length is greater than three (but show a variable length of two to three in their
illustrations) diameters, the column will fail by bulging (Figure 2‐32a). This may be because
they have envisaged a more general case where the undrained cohesion increases and not
constant with depth.
The column length to diameter ratio in dynamic replacement is usually from about 1 to 3.5,
which is less than 4, and columns are generally pushed down onto a supportive layer. Hence,
it could be expected that DR columns in soft soils will usually fail in end bearing mode.
Barksdale and Bachus suggest that the granular working platform serves a dual purpose by
not only spreading the load in arching to the column, but also by improving the performance
of the columns. The granular blanket forces the bulge to a lower depth where the overburden
pressure is greater, and hence will probably result in a larger ultimate load capacity of the
column. This is shown in Figure 2‐33.
55
Figure 2‐33: Effect of compacted granular working platform on the behaviour of the column: (a)
Construction of column without compacted mat, (b) Construction of compacted granular mat prior
to the construction of the column, (c) Construction of the column prior to the construction of the
compacted granular mat (Barksdale and Bachus, 1983)
Apparently referring to the experiments of Hughes and Withers (1974), Barksdale and Bachus
comment that small scale models studies show that the bearing capacity and settlement
behaviour of a single column is significantly influenced by the method of applying the load.
As shown in Figure 2‐34, applying the load through a rigid foundation over an area greater
than the column increases the vertical and lateral stress in the surrounding soft soil. The
larger bearing area and the additional support of the column result in less bulging and greater
ultimate load capacity.
Other possible failure modes of dynamic compaction columns are shown in Figure 2‐32(a)
and Figure 2‐32(b). Shear failure can occur in short columns or columns that are founded on
dense soil. Madhav and Vitkar (1978) assess that if the width of a trench (that is a plane strain
equivalent of a axisymmetrical DR column) is more than three times the width of the loaded
area, then a general shear failure mechanism can be considered, but for larger loaded areas
they propose an alternative analysis. Alternatively, short columns on soft ground can fail in
punching shear.
Hughes and Withers assess that the behaviour of a typical column within the group is the
same as that of an isolated column, and justify that this approach produces an upper bound
to an estimate of the settlements and hence can be considered to be conservative.
56
Figure 2‐34: Different loading types applied to columns: (a) Rigid footing loading, (b) Plate load test,
(c) Embankment loading (Barksdale and Bachus, 1983)
57
2.5 Design Guidelines for Dynamic Compaction
Varaksin (1981) identifies the main limits of dynamic compaction to be low soil permeability,
depth of low permeability soils, thickness of compressible soils, organic content and very soft
initial state of soil. He further clarifies that dynamic compaction becomes less and less
effective as the soil permeability reduces to less than 10‐7 m/s when the impermeable layer
is too thick. If the soft layer is not located well within the depth of influence, then the degree
of densification may become negligible. The depth of influence is limited by the impact
energy (and thus the lifting capacity of the equipment). When soil to be compacted includes
layers of high organic content, then dynamic compaction can only be applied if the allowable
settlement of the structure is less than the long term settlements. Also, if natural soils have
a very low resistance and high water content, it is often not possible to obtain sufficient
practical improvement.
As shown in Figure 2‐35, Lukas (1986) has categorised the applicability of dynamic
compaction to soils into three zones of improvement based on soil’s grading; i.e., pervious
soils, semi pervious soils and impervious soils. Figure 2‐35 supplements Table 2‐2.
Figure 2‐35: Categorising of soil for dynamic compaction suitability (Lukas, 1986)
58
General Soil Type Most likely Most Degree of Suitability for DC
fill class likely saturation
AASHTO
soil type
Pervious deposits in the Building A‐1‐a High or Excellent
grain size range of boulders rubble low
to sand with 0% passing the Boulders A‐1‐b
0.075 mm sieve Broken A‐3
Coarse portion of Zone 1* concrete
Among factors influencing the depth of improvement, Luongo (1992) notes that groundwater
level and fines content generally influence the effectiveness of dynamic compaction. He
further adds that experience indicates that dynamic compaction is generally less effective
when the percentage of clay content is greater than 15%, which is a much larger percentage
than the personal experience of the auhtor.
59
Inefficiency of dynamic compaction in soils with high fines content has led to the
development of dynamic replacement. As shown schematically in Figure 2‐36, in DR the
pounder is dropped several times until a crater, which is more plastic displacement than
compaction, is formed in the ground. Then the crater is backfilled with crushed stone, and
the stone is compacted by dropping the pounder onto it several times. Backfilling of the
crater may be done once or several times depending on the ground conditions.
Figure 2‐36: The process of dynamic replacement
The depth of influence or depth of improvement is a loosely defined term that signifies the
depth where there is no more significant improvement in the soil. Menard and Broise (1975)
developed an empirical relationship in which as a first approximation the energy per blow,
being the product of the pounder weight, W (in tons) and drop height, H (in m), was greater
than the square of the depth of influence, D :
2‐39
Varaksin (1981) later proposed that depth of improvement can be predicted by using
Equation 2‐40:
√ 2‐40
c1= speed damping factor
c2= stratigraphic coefficient, which in the studied case was 0.7
60
Others such as Leonards et al. (1980), Lukas (1980), Mayne et al. (1984), and Luongo (1992)
also reported lower depth of improvement values than the square root of impact energy and
Equations 2‐39 and 2‐40 gradually became better known as Equation 2‐41 with a single depth
of improvement coefficient, c:
√ 2‐41
Although the effect of pounder weight and drop height are of the same order in Equation 2‐
41, and thus it would appear that the depth of improvement would be equally sensitive to
either of them, Menard and Broise (1975) noted that a marked increase in energy efficiency
was observed when the impact velocity became greater than the velocity of the wave
transmission in the liquefying soil, and thus the tendency to increase the drop height to
optimise energy efficiency. Menard and Broise also stated that the maximum depths that
correspond to a good efficiency for dynamic compaction are distinctly greater for partially
immersed soils than for soils completely out of the water, and added that efficiency is a
function of the shape and dimensions of the pounder, the drop height, and the periods of
delay observed between phases.
Varaksin (1981) proposed that in Equation 2‐40, c1 should be taken as 1 if the pounder is
dropped in free fall, and in the special case of his paper c2 was taken as 0.7. Later Varaksin
and Racinais (2009) further clarified the coefficients by proposing that c1 should be taken
respectively as 0.9 or 1 when the pounder is dropped with a cable or dropped in free fall, and
that c2 should be taken respectively as 0.5 or 0.7 when the soil is heterogeneous or granular.
Lukas (1986) assesses that as the impact energy is proportional to the square of the impact
velocity, since the measured velocity of a cable dropped pounder was 90% of the theoretical
impact velocity, then the energy delivered to the ground would be approximately 80% of the
theoretical potential energy. However, the velocity of free fall pounders appeared to be 97%
of the theoretical velocity, and thus free fall impact would deliver 94% of the theoretical
impact energy. Therefore, the impact energy of a free falling pounder would be 18% more
than a cable dropped pounder. Lukas notes that this has been taken into account in
Equation 2‐41. Similarly, Yee (1999) suggests that c1 should be taken as 0.9 for cable drops or
1.2 for free falls, and c2 should be assumed to be from 0.3 to 0.7 depending on the soil type.
61
Lukas (1980) observed from post dynamic compaction pressuremeter or Standard
Penetration Tests (SPT) that the measured depth of improvement was about 65 to 80% of
the maximum value predicted in Equation 2‐39. Lukas also reports that the number of passes
applied to the treated areas did not appear to affect the depth of improvement.
Leonards et al. (1980) proposed that an increase of more than 3 to 5 SPT blows could be a
suitable criterion for determination of depth of influence, and reported that Equation 2‐39
expressed as an equality appeared to overestimate the depth of improvement (this is of no
surprise because Equation 2‐39, was an inequality in the first place that gave an upper bound
value, but had turned into the form of an equality in American publications), and that
applying a coefficient of 0.5 in Equation 2‐41 was able to relate to field measurements more
accurately.
The presence of fine soil reduces the efficiency of dynamic compaction. Leonards et al. (1980)
observed that clay layers or seams greatly attenuated the effective depth of compaction.
They also noted that a suitable criterion for the depth of improvement would depend on the
soil type and its initial state of compaction.
Chen and Lin (2002) carried out laboratory scale dynamic compaction on Mai‐Liao sand with
various fines content to understand the effect of fine soil layers. In their experiment they
inserted a manometer under the centre of the pounder’s impact point at the bottom of the
testing tank. As part of the experiment they varied the fines content of the soil. Without
being clear on the soil’s water content, Chen and Lin present Figure 2‐37 that shows that the
relationship of the dynamic compaction induced stress measured by the manometer and
fines content. It can be observed that the measured vertical stress reduces as fines content
increases. Beyond a certain point, increase of impact energy or momentum does not appear
to have any effect. Chen and Lin also experimented by placing thin dry fine soil layers with
various thicknesses on the ground surface and above dry sand. The result of this series of
trials also showed that increasing the thickness of the fine soil layer reduced the measured
vertical stress. Similarly, when the water content of the fine soil layer on ground surface was
increased, it was observed that the measured vertical stress reduced. The reduction of
vertical stress was more when the fine soil layer was clay rather than silt. Chen and Lin then
varied the location of the thin fine layer of soil and experimented with different water
contents in the fine soil. This series of experiments indicated that the measured vertical stress
reduces as the fine soil layer becomes more superficial.
62
Figure 2‐37: The effect of fines content on measured vertical stress(Chen and Lin, 2002)
Mitchell (1981) states that the definition of depth of influence is itself subjective and depends
both on the method of measurement and the engineer’s definition of what constitutes a
measurable ground improvement. He also notes that depth of influence should depend on
factors other than just the impact energy; soil type being the most important. Other factors
that have been mentioned by him include type of drop (frictionless free drop or dropped with
a cable, which is frictional due to the drum and sheaves), and the presence of soft layers that
have a damping influence on the dynamic forces. Figure 2‐38 shows the depth of influence
as a function of impact energy.
Mayne et al. (1984) compiled data from several sites, and as shown Figure 2‐39, concluded
that the coefficient for depth of improvement is in between 0.3 and 0.8.
Lukas (1986) assessed that other than the pounder weight and drop height other parameters
such as soil type, energy applied, base pressure, cable drag, and layer hardness also influence
the depth of improvement. Soil type indeed has an important influence on the depth of
improvement, and Lukas categorises soils based on their grading into three zones of pervious,
semi pervious and impervious (see Figure 2‐35). Based on data review, he proposes that the
coefficient of depth of improvement should be in the order of 0.35 to 0.4 for partially
saturated cohesive fills situated above groundwater level. Lukas notes that optimum water
content of fill for dynamic compaction is not known, but speculates that it is probably near
the optimum water content as determined by the modified Proctor compaction, which can
be estimated as being equal to the soil’s plastic limit minus 7 percent (NAVDOCS, 1962).
63
Figure 2‐38: Depth of influence as a function of impact energy (Mitchell, 1981)
Figure 2‐39: Trend between apparent maximum depth of influence and energy/blow (Mayne et al.,
1984)
64
below the crust, and from the other end base pressures higher than those mentioned above
could result in the pounder punching into the ground upon impact, which reduces energy
efficiency. It is the author's experience that nowadays the commonly used pounder static
pressure is more than what was envisaged by Lukas as an upper bound, and pressures in the
magnitude of 90 kPa are used in most projects. However, the author agrees with Lukas that
heavy pounders with smaller base areas punch through the soil, and are thus used in dynamic
replacement to create displacement cavities.
When hard or cemented layers occur within the upper portions of the deposits being
compacted, these layers will distribute the impact stresses; therefore, the impact energy will
not propagate as deep as in a loose deposit. On the other hand, if the hard layer is located
near the bottom of the improvement zone, then the presence of the hard layer can improve
the densification as a portion of the compression and shear waves reflect and cause
additional compaction (Lukas, 1986).
Lukas (1986) proposes that Table 2‐3 can be used for conditions when pounders are dropped
using single cable lines, with total impact energy ranging from 100 to 300 tm/m2, and static
base pressure varying from 39 to 82 kN/m2.
As compaction mechanism is similar to Proctor compaction test, depth of influence is greater
in partially saturated cohesionless soils. The depth of improvement in saturated soils is less
because the sudden large energy release gives rise to immediate build‐up of pore pressure,
which greatly attenuates dynamic loading with depth (Luongo, 1992).
65
Based on data obtained from 30 sites, Luongo also proposed Equation 2‐42 as a modification
for prediction of depth of improvement. It can be noted that Luongo's equation is of the first
order of energy and not a function of the square root of the impact energy.
Deposit
Category Pervious Semi‐pervious Partially saturated
impervious fill
LB1 D= 5.6 + 0.009WH D= 1.4 + 0.016WH D= 2.0 + 0.012WH
General Avg1 D= 7.9 + 0.009WH D= 3.4 + 0.016WH D= 2.6 + 0.012WH
UB1 D= 10.1 + 0.009WH D= 5.3 + 0.016WH D= 3.3 + 0.012WH
sample range 125<WH<400 tm 150<WH<400 tm 250<WH<400 tm
High LB1 D= 4.6 + 0.012HW insufficient data dynamic compaction
Groundwater Avg1 D= 7.4 + 0.012HW not recommended
Level UB1 D= 10.1 + 0.012HW
sample range 125<WH<400 tm
Low LB1 insufficient data D= 1.7 + 0.016WH D= 2.0 + 0.012WH
Groundwater Avg1 D= 3.3 + 0.016WH D= 2.6 + 0.012WH
Level UB1 D= 4.9 + 0.016WH D= 3.3 + 0.012WH
sample range 200<WH<400 tm
LB1 2
D= ‐0.2 + 0.024WH D=‐0.4 + 0.024WH 2
D= 4.9 + 0.004WH
Pounder base
Avg1 2
D= 1.5 + 0.024WH D= 1.3+ 0.024WH 2
D= 5.1 + 0.004WH
area 4‐5 m2
UB1 2
D= 3.2 + 0.024WH D= 3.0+ 0.024WH 2
D= 5.4 + 0.004WH
sample range 200<WH<325 tm 150<WH<400 tm 250<WH<350 tm
1
LB= lower bound values, AVG= average values, UB= upper bound values
2
Because of the small number of samples these equations should be interpreted with caution.
Table 2‐4: Predicted depth of improvement for different soil deposits (Luongo, 1992)
2‐42
cd = depth constant
ce = energy constant
66
The mean depth of improvements which have been derived from Luongo's (1992) analysis is
shown in Table 2‐4. Luongo warns that the equations of Table 2‐4 should be used with great
caution, and only as a guideline since depth of improvement observed will not only depend
on the variables discussed, but also on specific site characteristics, the definition of depth of
improvement, and the method of measurements used to quantify depth of improvement. In
Table 2‐4 rock fill deposits are not considered, impact energy is from 100 tm to 400 tm, the
reduction of impact energy due to cable drag and absorbing layers is considered, high
groundwater level is equal to or less than half of the depth of improvement, low groundwater
level is greater than half of the depth of improvement, and data used for depth of
improvement for pounder area category includes both high and low groundwater levels.
Rollins and Kim (1994) studied 6 mostly semi impervious, primarily silt, collapsible deposits
and found the coefficient of depth of improvement, c, to be 0.4; however, the coefficient of
determination, R2 was a low 0.4. They also tried a second order polynomial, which turned out
to give a much better prediction with R2 of 0.9. This relationship is shown in Equation 2‐43.
In comparison with the best fit line, the best fit curve predicted somewhat greater depths of
improvement at lower energies but somewhat lower depths at the very highest energies. 16
years later Rollins and Kim (2010) published an updated study which included an additional
9 case studies. While this study remained in line with their earlier publication, R2 for the
polynomial regression reduced to 0.75.
Van Impe and Bouazza (1996) reviewed data of dynamic compaction in solid waste materials,
and concluded that the coefficient in Equation 2‐41 was from 0.35 to 0.65. The lower limit
corresponded to old landfills; i.e., fills older than 15 to 20 years, and the upper limit was for
young landfills; i.e., fills that were less than 15 years old.
Vuola and Hartikainen (1999) somewhat modified Equation 2‐41, and proposed the depth of
improvement equation for layered ground to be as given in Equation 2‐44.
√ 2‐44
fC = cushion (crushed stone platform) factor
fM = material (soil) factor
67
fB = bottom factor
According to Vuola and Hartikainen fc can be taken as 1 if the cushion is proper. An unproper
cushion reduces maximum depth of improvement. In their specific research they report fc to
be approximately 0.8 to 0.9 in one of the investigated sites. Vuola and Hartikainen propose
from literature sources that fM is in the range of 0.65 to 0.75 for saturated sand, and derive
a value of 1.5 to 2 for fB when they encountered rock on one of the sites. They state that rock
at a finite depth increases the maximum compaction depth, and contribute the reason to
impact waves reflecting from the bottom. Wave interference increases dynamic stresses and
strains in soil, and increases compaction depth.
Hajialilue‐Bonab and Rezaei (2009) have proposed one of the most rational criterion for
depth of influence by relating this depth to volumetric strain rather than in‐situ field tests or
relative density values. In their study, they assume that depth of influence is the depth at
which volumetric strain reduces to 1.5%. In order to predict this depth, Hajialilue‐Bonab and
Rezaei developed a laboratory scale physical model of dynamic compaction, and used
particle image velocimetry to plot soil (fine poorly graded dry sand) deformations, and based
on energy intensities (pounder impact energy divided by pounder base area) of 0.837 to
5.566 kN.m/m2 they developed a relationship between the depth of improvement, D (in m),
pounder weight, W (in tons), pounder drop height, H (in m), diameter, d (in m), number of
drops per print, N, and pounder base area, A (in m2) which is presented in Equation 2‐45;
however, the author notes that there is an inconsistency in their paper as D appears in two
forms in the publication, sometimes in the numerator and sometimes in the denominator of
the right hand side of the proposed equation.
With D appearing in the denominator, the form of parameters presented in Equation 2‐45 is
similar to Equation 2‐70 that will be presented later, but the author has inserted values for
the parameters, and the equation yields unrealistic results.
Inserting D in the numerator, as shown in Equation 2‐46, yields more acceptable results. The
author has corresponded with Hajialilue‐Bonab and Rezaei and awaits their feedback at the
time of compilation of this thesis.
68
0.7669Ln 3.004 2‐46
Hajialilue‐Bonab and Rezaei noticed that, as shown in Figure 2‐40, normalised depth of
influence and normalised crater depth appear to have a linear relationship, and thus also
proposed alternative formulas, to predict the depth of influence based on crater depth, Dc.
These formulas are presented in Equations 2‐47 and 2‐48.
Figure 2‐40: Correlation between depth of influence and crater depth (Hajialilue‐Bonab and Rezaei,
2009)
Or:
d = pounder diameter
It is the author’s experience that the estimated depth of influence from Equations 2‐47 or 2‐
48 is an underestimation of what should be expected in the field, and what is predicted by
other methods. For example, with a conservative value of c= 0.63 for granular soil,
Equation 2‐41 predicts that the depth of improvement for a 15 ton pounder, with a diameter
of 1.6 m, that is dropped from 20 m would be 10.9 m. Equations 2‐47 or 2‐48 predict the
69
crater depth to reach such a depth of improvement would have to be about 2.6 m, which is
much more than what is measured on site.
Ghasessemi et al. (2009b) performed a numerical stimulation of dry and moist sands using a
finite element programme to determine depth of improvement. In the studied conditions,
pore flow did not exist, and a cap model was assumed to define the soil behaviour. Ghassemi
et al. conducted a series of numerical models using different combinations of pounder
weights ranging from 15 to 25 tons, drop heights varying from 10 to 20 m and pounder radii
of 0.8 to 1.8 m. In all analyses 15 drops were modelled, and the depth of improvement was
determined based on the increase in relative density.
The obtained results suggested that the influence of sand type and initial relative density was
small when the results were presented in terms of change in relative density; hence, they
focused their study on cases in which initial relative density was 40%. The study also showed
that the soil did not undergo any significant compaction after 15 blows. Assuming that SPT
blow counts (with 60% efficiency) increase by 3 at depth of improvement, and by considering
a relationship between SPT blow counts and relative density (Skempton, 1986), they
calculated that after 15 blows, relative density would increase by 4% at depth of
improvement.
As shown in Figure 2‐41, Ghassemi et al. observed that the coefficient of depth of
improvement in Equation 2‐41 reduces as impact energy increases. Ghassemi et al. have not
proposed an equation for the relationship between c and WH (impact energy per drop), but
it can be approximated from Figure 2‐41 that:
0.65 2‐49
2500
By replacing the Equation 2‐49 into Equation 2‐41, it can be derived that:
0.65 √ 2‐50
2500
70
Figure 2‐41: Variation of c with impact energy per drop (Ghassemi et al., 2009b)
By reviewing the results of the analyses, Ghassemi et al. proposed a relationship in which, as
formulated in Equation 2‐51, the depth of improvement is a function of pounder weight,
pounder drop height and pounder radius, r (in m).
2√ 2‐51
Ghassemi et al. have not attempted to relate Equation 2‐51 to Equation 2‐41, but the author
has done so. If the 4th root of WH is multiplied in the numerator and denominator of the right
hand side of Equation 2‐51, the same equation can be presented in the form of Equation 2‐
522‐54.
2 √ 2‐52
In other words, the coefficient of depth of improvement is as per Equation 2‐53
2 2‐53
Although it could be expected that depth of improvement would increase with the pounder
radius and stress bulb size, it is rather unexpected that the efficiency of penetration has an
71
inverse relationship with the impact energy itself. The same trend could be observed from
Equation 2‐49.
Zou et al. (2005) has cited that Fei et al. (2002) has performed a series of model tests of
dynamic compaction on loess, and have developed Equation 2‐54 for depth of influence.
2‐54
1
In addition to the parameters that have previously been defined
ρd = dry density of soil (in kg/m3)
= average water content of soil.
Input of real figures into Equation 2‐54 yields unrealistically low values of depth of
improvement. For example if W= 16.5 m, H= 20 m, d= 1.54 m, N= 15 blows, ρd = 1650 kg/m3
and = 10%, then depth of improvement will be a mere 0.92 m. At the same time, depth of
improvement can be predicted to be approximately 8 m using the same values for the
parameters that are used in Equation 2‐51. The author has tried to communicate
unsuccessfully with Zou et al. for a clarification.
Even though most research has been concentrated on impact energy (WH), and Menard and
Broise had observed a marked increase in energy efficiency when the impact velocity became
greater than the velocity of the wave transmission in the liquefying soil, and thus the
tendency to increase the drop height to optimise energy efficiency, it is of interest to study
the influence of impact momentum on the volume and depth of improvement.
Oshima and Takada (1994, 1997, 1998) have studied the influence of momentum in dynamic
compaction. They reason that if the ground is plastic and the pounder does not rebound after
impact; i.e., the impact is perfectly inelastic, then pounder penetration into the ground could
be considered to be governed by the law of conservation of momentum rather than by the
law of conservation of energy. They further add that this is because of the large loss of energy
due to plastic deformation of the ground and the emission of heat and sound. Based on
results of measurements of actual field work and centrifuge model tests and laboratory
72
compaction tests, Oshima and Takada conclude that the period of pounder impact is
proportional to the quotient of the pounder weight divided by the pounder base area, and is
not influenced by the drop height. They also conclude that maximum acceleration of the
pounder during penetration into the ground is proportional to the square root of the drop
height.
Oshima and Takada performed a number of centrifuge dynamic compaction tests using
various pounder weights under constant pounder drop heights, different pounder drop
heights under constant pounder weights, and combined pounder weight sand drop heights
with constant kinetic energy per blow. Oshima and Takada did not observe any substantial
differences among the tests when pounder drop height, pounder base area, total kinetic
energy (mgHN , where m is pounder weight) and total pounder momentum (mvN, where v
is given in Equation 2‐57) were kept constant, but the pounder weight was made variable.
However, they did notice deeper and enlarged improvement volume when the pounder
weight, pounder base area and total kinetic energy were kept constant, but total momentum
was increased by varying the drop height and number of blows. When the pounder weight,
drop height, and base area were combined in such a way that the kinetic energy per blow
and pounder weight per unit area of pounder base remained constant, it was also observed
that the depth and volume of improvement increased with the increase of momentum. From
these results Oshima and Takada concluded that it was clear that pounder weight had a
predominant effect on compaction of sandy soils, followed by the ram drop height and
number of blows, and that the compacted volume size was more influenced by pounder
momentum rather than pounder kinetic energy. They also concluded that the depth and
radius of improvement was proportional to the logarithm of total pounder momentum, and
that the volume of compaction could be estimated by the total pounder momentum. Oshima
and Takada proposed that it is possible to estimate the maximum depth and radius of the
bulb shown in Figure 2‐42 for any increase of relative density by using Equations 2‐55 and 2‐
56.
log v 2‐55
log v 2‐56
73
v 2 2‐57
Figure 2‐42: Definition of depth and radius of compacted zone for any relative density contour line
(Oshima and Takada, 1997)
Gu and Lee (2002) studied the mechanics of dynamic compaction using two‐dimensional
finite element analyses, and similar to Oshima & Takada (1998), assumed that depth of
improvement was the depth at which the increase in relative density reduced to 5%. The
computed results indicated that the soil under study did not undergo any significant further
compaction after about 15 blows. The coefficient of depth of improvement in the numerical
study was in the range of 0.55 to 0.625, and comparison of results with the findings of Oshima
& Takada (1998) showed that the error between the numerical and experimental depth of
improvement was less than 10%. The analyses also showed that in the initial blows stress
wave propagation induces transient elasto‐plastic Ko compression due to lateral inertia,
which preserves the plane wavefront, and reduces the attenuation rate of the dynamic
stresses with depth. With multiple blows, the effect changes to one of triaxial compression,
which sets a limit on the degree of improvement that can be achieved in the near field. At
greater depths, the wavefront adopts a bullet shape, and the attenuation rate rises, which
sets a limit on the depth of improvement. Both of these phenomena are consistent with the
existence of a threshold.
Similar to the conclusions of Oshima and Takada (1994, 1997), the numerical analyses of Gu
and Lee also showed that, in addition to energy per blow, depth of improvement is also
74
dependant on momentum per blow. As can be seen in Figure 2‐43, even if energy per blow
is kept constant, an increase in momentum will enlarge the zone of improvement, and
deepen the depth of influence.
Figure 2‐43: Contours of increase in relative density (%) for (a) W= 20 t, H= 20 m, (b) W=40 t, H= 10
m, and (c) W= 80 t, H= 5 m (Gu and Lee, 2002)
Hajialilue‐Bonab and Rezaei (2009) who have performed low energy laboratory scale
dynamic compaction physical modelling have also assessed the importance of impact energy
and impact momentum by performing three tests on dry sand. In these tests, impact energy
per blow (WH) and pounder base area were kept constant; however, pounder momentum
was made variable. Figure 2‐44 shows the variation of depth of influence with the number of
blows. It can be observed that with constant energy, a greater improvement and influence
depth can be achieved with greater impact pressure (WH/A) in comparison with low
pressure impact, and it was concluded that the influence of pounder weight was more
significant than the drop height. In other words, the influence of momentum was greater
than the influence of applied energy.
Research funded by the Federal Highway Administration (Lukas, 1986) has shown that depth
of improvement does not significantly increase in sandy and silty sand formations after 2 to
7 blows. Therefore, dropping the weight up to 30 or 40 times will improve the degree of
treatment within the depth of influence, but not the improvement depth itself. The greater
75
the total energy used, the greater the level of improvement will be, but only within the zone
of maximum influence.
Figure 2‐44: Depth of improvement with constant impact energy and varying momentum (Hajialilue‐
Bonab and Rezaei, 2009)
Hajialilue‐Bonab and Rezaei (2009) who carried out low energy laboratory scale dynamic
compaction physical modelling on dry sand, have also studied the relationship of normalised
depth (ratio of depth of improvement to pounder diameter) with the number of blows. In
the series of tests the impact energy and momentum of the print were kept constant, but
the impact energy intensity or energy per unit area of impact (not the average energy
intensity over the treatment area) was made variable. Figure 2‐45 shows the relationship
between the normalised depth of influence depth and the number of drops for one of these
series of tests. What can be observed is for all curves, depth of improvement initially
increases rapidly with the number of blows, but the rate of improvement then reduces until
the increase in depth of influences appears to be levelling off after the 12th to 15th blow. The
same can also be concluded from the test results presented in Figure 2‐44.
It can also be seen that with the same impact energy and momentum, depth of improvement
is greater when the impact energy intensity is more (in other words, when the pounder
diameter is less). Hajialilue‐Bonab and Rezaei observed that for larger impact energy
intensities, larger displacements occur at shallow depth beneath the pounder compared to
when impact energy intensity is lower, but displacements decrease significantly at greater
depths. Thus, they concluded that most of the applied energy is consumed to create large
deformations in the soil mass beneath the pounder. As the pounder base area decreases and
the stress bulb become smaller, the influence radius consequently decreases.
76
Figure 2‐45: Relationship between normalised depth of improvement and number of drops for
different energy intensities
As part of their numerical study of dynamic compaction application to dry sands, Lee and Gu
(2004) note that maximum depth of improvement will be achieved if sufficient number of
blows are applied. If the actual (interim) number of blows is insufficient to reach the
maximum achievable depth of improvement, then the interim depth of improvement will be
less than the maximum achievable depth of improvement. The interim depth of
improvement (strictly for dry sand) can be calculated as
2‐58
Di = interim depth of improvement
αZ = depth of improvement ratio (dimensionless)
2 2‐59
2‐60
αB = number of blows ratio (dimensionless)
N= number of blows required to reach maximum depth of improvement
Ni= interim number of blows
Ghassemi et al. (2009a) suggest that Equation 2‐61 is able to fit numerical and experimental
results better than Equation 2‐59, and compare the two equations in Figure 2‐46.
77
2 2‐61
1.2
Figure 2‐46: Interim depth of improvement ratio (Ghassemi et al., 2009a)
Gu and Lee (2002) have carried out a numerical study on dynamic compaction of dry sands
have noted that, as shown in Figure 2‐47, as the radius increases, the depth of improvement
initially increases, but then decreases. In other words, the maximum depth of improvement
is achieved with neither the smallest nor the largest radius, but rather with an intermediate
radius. Gu and Lee reason that if the pounder radius is too small, then the lateral confinement
on the soil directly beneath the footing will be maintained only for a relatively short duration,
thereby limiting the depth at which one dimensional wave propagation occurs. On the other
hand, if the pounder radius is too large, then the impact force will be distributed over too
large an area, thereby reducing the impact stress, and thus limiting the depth of
improvement.
Ghassemi et al. (2009b) who have carried out a numerical study on dynamic compaction have
proposed Equation 2‐51, also repeated hereunder, which relates the depth of improvement
to pounder radius.
78
Figure 2‐47: Increase in relative density along centreline for various tamper radii (Gu and Lee, 2002)
2√
Figure 2‐48: Relation between tamper radius and depth of improvement (Ghassemi et al., 2009a)
Ghassemi et al. (2009a) speculate that there should be an optimum pounder radius whose
value depends on the pounder weight because most pounder’s static pressure are in a certain
range (the author believes that impact energy could have been a better basis than pounder
weight). Figure 2‐48 shows the relationship between depth of improvement and pounder for
a series of numerical cases in which one drop has been applied from 20 m height. It can be
observed that the variation of depth of improvement depth with pounder radius for each
pounder weight appears to have a peak at which the maximum depth of improvement has
been achieved.
79
Laboratory scale dynamic compaction physical modelling using circular flat bottom and
conical bottom pounders have been carried out to determine the effect of pounder bottom
shape on the efficiency of dynamic compaction (Arslan et al., 2007, Feng et al., 2000, Feng
and Ke, 2005, Feng and Yuan, 2009). Feng et al. (2000) performed their study on fine poorly
graded Mai‐Liao sand with 8% fines content and commercial grade Ottawa sand, with no
fines, poorly graded and coarser than Mai‐Liao sand that had relative densities of 30 to 40%
in dry states and 60% in wet states. Feng and Ke (2005) report that Mai‐Liao sand particles
are platy shaped rather than the more common bulky shaped grains. In addition to the flat
bottom pounder, conical pounders with apex angles of 60, 90 and 120o (Feng and Yuan, 2009)
were also used in the study; however, Feng et al. (2000) have only published the results for
the flat bottom and conical bottom pounder with an apex angle of 90o (see Figure 2‐49). Feng
and Yuan (2009) explain in their discussion about Arslan et al.’s (2007) paper that that the
90o conical bottom pounder worked the best, regardless of the type of sands used in the
model test programme. The 60o conical bottom pounder penetrated excessively into the
specimen, and the 120o conical bottom pounder created craters smaller than that created by
the 90o conical bottom pounder. The flat bottom and conical pounders reported by Feng et
al. both weighed the same.
Figure 2‐49: Side view of cylindrical pounders (Feng et al., 2000)
While the craters in the dry Mai‐Liao sand specimens all had bowl shapes, the craters in the
wet Mai‐Liao sand specimens were nearly cylindrical in shape, and heave was absent around
the craters in these tests. The craters in the dry Ottawa sand specimens also had bowl shapes;
however, unlike the Mai‐Liao sand specimens, heave was clearly appearing around those
craters. Feng et al. observe that the ratio of net crater volume, being the difference between
crater and peripheral heave volumes, was 1.29 to 1.95 times larger when using a conical
bottom pounder compared to when a flat bottom pounder was used in dry Mai‐Liao sand.
The same ratio varied from 1.33 to 1.73 when the sand was wet. However, when dry Ottawa
80
sand was used, it was observed that net crater volume was 3 to 4% larger for the flat bottom
pounder rather than the conical bottom pounder.
Feng and Ke (2005) also performed laboratory scale dynamic compaction on dry Mai‐Liao
sand with relative density of 50% and fines content of 3.4%, 8% and 14% using flat bottom
and conical bottom pounders shown in Figure 2‐49. Although the net crater volume in this
series of tests was also consistently larger in conical bottom pounders compared to flat
bottom pounders; however, the ratio was not as pronounced as what Feng et al. (2000) have
reported. Here the net crater volume ratio of conical to flat bottom pounders was 1.08 to
1.21 for 3.4% fines content, 1.09 to 1.18 for 8% fines content and 1.05 to 1.12% for 14% fines.
Figure 2‐50 shows the mini cone resistance profiles under the pounder centre for tests with
fines content of 3.4%, 8%, and 14%. It can be seen that the cone penetration resistance
increases sharply at depths about 10 to 20 cm, which is right below the bottom of the crater.
Feng and Ke note that the depth of the crater was at about 10 cm, and that the sharp increase
was due to a direct compaction from the impact of the flat bottom pounder. The lesser
amount of cone resistance increase at depths greater than 20 cm was attributed to the
transmission of body waves downwards through the test specimen. Figure 2‐50 indicates that
the conical bottom pounder induces somewhat lower maximum cone resistances than the
flat bottom pounder in all tests. Feng and Ke suggest that this may be due to the conical
bottom pounder inducing lesser amounts of downwards compaction. The differences in the
post dynamic compaction cone resistances are even less at greater depths between the
conical and flat bottom pounders, and the depth of improvement between the two cases
appears to be the same.
Figure 2‐51 shows mini cone resistances measured at distances of 1, 2 and 3 pounder
diameters from the pounder centres for tests carried out on Mia‐Laio sand with 14% fines
content. It can be observed that at the distance of one pounder diameter from the pounder
centre the maximum cone resistance induced by the flat bottom tamper is about the same
as that of the conical bottom tamper; however, as the distance from the pounder centre
increases the maximum cone resistance induced by the conical bottom pounder becomes
greater than the flat bottom pounder. Feng and Ke suggest that this is because the conical
bottom pounder moves more sand particles laterally as it penetrates into the sand than what
is achieved by the flat bottom pounder. Feng and Ke note that they have also observed the
same behaviour for Mia‐Laio sand with 3.4% and 8% fines content.
81
(a) (b) (c)
Figure 2‐50: Mini cone penetration resistance at below centre of pounder with fines content (a)
3.4%, (b) 8% and (c) 14% (Feng and Ke, 2005)
Arslan et al. (2007) have also compared the efficiency of conical bottom pounders with flat
bottom pounders by performing laboratory scale dynamic compaction on loose, medium
dense and dense Sakarya River sandy gravel. There is a discrepancy in the reported soil
grading as one of the figures in Arslan et al.’s paper shows coarse gravel size material on the
testing surface, but the maximum grain size of the soil is reported to be less than 2 mm, and
material has been classified by the paper’s authors as well graded sand. The internal friction
angles of the loose, medium dense and dense soil used in the tests were respectively 30o, 35o
and 40o.
The pounders used in Arslan et al.’s experiment all weighed about the same. The conical
bottom pounders used for each density of soil had the same side angle with the horizon as
the internal friction angles; i.e., 30o, 35o and 40o, which would be equivalent to apex angles
of 120o, 110o and 100o. It is the author’s opinion that this will make efficiency comparison of
any one apex angle in looser or denser soils somewhat difficult.
82
(a) (b) (c)
Figure 2‐51: Mini cone penetration resistance for test series with 14% fines content. Distances of
impact points from the centre of pounder were (a) 1 pounder diameter, (b) 2 pounder diameters,
and (c) 3 pounder diameters (Feng and Ke, 2005)
Arslan et al. measured the crater depths and crater top diameters in two orthogonal
directions, and assumed that the compaction volume is equal to the volume of a cylinder
with the crater diameter at its base and the crater height. The author believes this is an
oversimplification because even in the figures provided in Arslan et al.’s paper the crater
sides are clearly not vertical, and the is no reference to potential surface subsidence or heave
around the crater. The results of this study suggest that crater depth and ground area are
from ‐5% up to about 38% larger for conical bottom pounders compared to flat bottom
pounders.
Crater depth may also be an indication of the dynamic compaction performance. Mayne et
al. (1984) observed that, as shown in Figure 2‐52, when crater depth measurements are
normalised with respect to the square root of energy per blow, the results fall within a narrow
band.
Hajialilue‐Bonab and Rezaei also noticed that, as shown in Figure 2‐40, in laboratory scale
dynamic compaction tests, normalised depth of influence and normalised crater depth
appeared to have a linear relationship, which led to the formulation of Equations 2‐47 and 2‐
48.
83
Figure 2‐52: Normalised crater measurements (Mayne et al., 1984)
Applying dynamic compaction in spaced out grids is a more efficient treatment method than
dropping the pounder contiguously because in the latter pattern a very dense intermediate
layer will be formed, which will attenuate energy propagation into the soil mass. Menard &
Broise (1975) note that efficiency depends quite closely on progressive consolidation of the
layers, commencing with the deepest and finishing with the surface, by means of an
adequate distribution of impacts.
Figure 2‐53: Application of Dynamic Compaction in a grid
As shown in Figure 2‐53, the initial phase of treatment is carried out on a wide grid with the
maximum amount of impact energy or drops per impact point (more commonly called print).
84
The objective of this phase is to treat the deepest soil layers. The second phase, which is also
a deep treatment, and is intended to treat the intermediate soil layer, may be carried out
with less energy or drops. If necessary, the final phase, called ironing will comprise of closely
spaced grid points with one or two low energy blows per print for improving the superficial
loose soil layer above the crater depth.
Mayne (1984) comments that grid spacing is usually at least equal to the thickness of the
compressible layer (the author interprets that it was probably meant to be as first phase
spacing). Lukas (1986) states that in semi‐pervious soils with high groundwater level, spacing
between prints should be wide enough and in the order of 10 to 15 m to allow for pore water
pressure dissipation. He further adds that, as a rule of the thumb, print spacing is taken equal
to the treatment thickness in some projects. Mayne suggests that in granular deposits where
groundwater level is not close to the ground surface, print spacing can be reduced, and
energy can be applied in one or two passes. For shallow depth treatment, print spacing as
close as 2 m have been used effectively. Later, Lukas (1995) noted that a spacing of 1.5 to 2.5
times the diameter or width of the pounder is common (the author interprets that is was
probably meant as inclusive of all deep phases of compaction), and adds that in fine grain
soils where there is a concern with pore water pressures developing in the soil, compaction
should be done in two or more phases. The first phase would involve dropping the pounder
at every second or third drop point location. After a period of time to allow for dissipation of
pore pressures, the intermediate drop point locations could be compacted as part of the
second or third phase.
Although 7 to 15 drops of high level energy are applied normally at each print (Lukas, 1995),
the author notes that the number of blows is a parameter that should be determined or
defined as a part of the design process, and has personally been involved in projects with the
number of blows per print ranging from 3 to 30 blows.
Chow et al. (1994) have proposed a method for predicting the change in the soil’s internal
friction angle based on a trial case study (Harada and Suzuki, 1984) in which a 12 ton pounder
was dropped from 20 m onto a single print to treat 7 to 8 m of mainly loose sand with pockets
of silty clay. CPTs were carried out before and after dynamic compaction at 0, 3, 6, and 9 m
away from the print location as part of the trial DC.
Chow et al. used a correlation (Meyerhof, 1976) to estimate the soil’s internal friction angle
from the CPT values, and assessed the increase of internal friction angles (Δφ) of the soil at
85
various distances. Denoting Δφb as the increase of internal friction angle beneath the
pounder, Δφ/Δφb versus normalised distance is plotted in Figure 2‐54. The distance has been
normalised using the pounder diameter or its equivalent diameter. Δφ/Δφb is shown for
depths of 0d, 1d, 2d and 3d beneath the pounder. It can be observed that the ratio is
insignificant for depths below 3d. Chow et al. also used data from tips of driven piles and
caissons (Meyerhof, 1959) to show the similarity of trends.
Figure 2‐54: Relationship between ratio of increase of internal friction angles and normalised depths
(Chow et al., 1994)
Δφ/Δφb obtained at equal lateral distance but at different depths showed some variations,
but they did not appear to be significant; hence, Chow et al. assumed that Δφ/Δφb was a
function of distance and independent of depth. Consequently they proposed that:
∆
1.0 for 0.5 2‐62
∆
∆
0.642 1.180 log for 0.5 3.5 2‐63
∆
∆
0 for 3.5 2‐64
∆
86
Chow et al. note that although Equations 2‐62 to 2‐64 are basically meant for the prediction
of the lateral extent of soil improvement around the pounder at a single drop point,
modification of Δφb by considering the densification of soil within the grid brought about by
the pounder impact on the surrounding prints can allow the extension of the equations for
evaluating the effect of multiple points.
For design purposes, the two most critical locations to achieve a uniform level of
improvement are at the centre of the grid and at the middle of the side of the grid. Chow et
al. have provided the relationship between Δφ/Δφb and the ratio of distance to grid spacing
for these two points in Figure 2‐55. As before, in the analysis Δφ/Δφb is assumed to be
independent of depth. When spacing is close, the influence of impacts at neighbouring prints
will be significant, and should be included in the analysis following the sequence of pounding.
Chow et al. note that they have carried out the analyses for s/d > 1, since when s/d ≤ 1 the
centre of the grid and the middle of the side of the grid are located directly beneath the
pounder and Δφ/Δφb is assumed to have values of 1.
Figure 2‐55: Increase in internal friction angles ratio for (a) grid centre point and (b) midpoint of grid
side (Chow et al., 1994)
According to Chow et al., for design purposes the effect of print spacing on the increase of
internal friction angle at the grid centre, Δφc , can be approximated by two linear equations:
∆
1.0 for 2.1 2‐65
∆
87
∆
1.60 0.29 for 2.1 5.5 2‐66
∆
Similarly, for design purposes the effect of print spacing on the increase of internal friction
angle at the midpoint of the grid side, Δφm, can be approximated by two other linear
equations:
∆
1.0 for 2.1 2‐67
∆
∆
1.49 0.23 for 2.1 6.5 2‐68
∆
Hajialilue‐Bonab and Zare (2014) have also studied the effect of grid spacing by advancing
the physical modelling scheme and particle image velocimetry method (Hajialilue‐Bonab and
Rezaei, 2009) to detect soils displacements. In their experiment Hajialilue‐Bonab and Zare
used poorly graded fine dry Sofian sand, and dropped pounders in the shape of half discs,
with 100 mm diameter and weighing 12.5 N, tangentially to a transparent Perspex facing
from 750 mm height. As a base, displacements were recorded for a single print that was
subjected to 12 consecutive blows, and for comparative purposes two prints with varying
spacing of 6d, 5d, 4d and 3d were subjected to 12 blows.
Figure 2‐56 shows the displacement vectors after 12 blows when pounder spacing was 5d,
and Figure 2‐57 shows the angle of displacement vectors at various normalised depth for the
same case. It can be observed that ground beneath the pounder centrelines and the spacing
midpoint line displaced vertically. The displacement angle decreases as distance from the
pounding centreline increases, and then begins to increase. It can also be seen that deeper
points appear to displace more vertically than horizontally.
88
Figure 2‐56: Displacement vectors after 12 blows when spacing= 5d (Hajialilue‐Bonab and Zare,
2014)
Figure 2‐57: Displacement angle at different normalised depths and distances (Hajialilue‐Bonab and
Zare, 2014)
Figure 2‐58 shows the displacement contour of 0.05d after applying 12 blows to the left print
(n1 series) followed by 12 blows to the right print (n2 series) when pounder spacing was 3d.
It can be observed that, as expected, progression of pounding at the first print increases the
displacement bulb volume under the first print and at its periphery, but displacement at the
location of the second print is much less and appears more parallel than bulb shaped.
89
Figure 2‐58: Displacement contours of 0.05d for different number of drops when pounder spacing=
3d (Hajialilue‐Bonab and Zare, 2014)
Pounding of the second (right) print does not induce any observable or significant
displacement under the first print, but increases the displacements between the two prints.
Progression of pounding generates further displacement under the second print and in
between the prints, and at higher blows the displacement contour appears to move towards
a uniform horizontal distribution.
What is also noticeable is that, generally, displacement contours under the second print were
deeper than the contours beneath the first print. Hajialilue and Zare reason that this is
because the soil in the zone of the second print has been compacted slightly by the first
print's pounding, and compaction of the second print would commence on the compacted
soil. Thus, it could be understood that in practical dynamic compaction works pounding of
the second phase prints will not improve the soil under the first phase prints, but pounding
of the first phase prints will (assuming the grid spacing is close enough) effect treatment of
the second phase, and lesser blows could be required for reaching the same depth of
improvement. This conclusion of Hajialilue and Zare is in compliance with the author's field
experience.
Figure 2‐59 shows the displacement contours for tests in which pounder spacing were 6d and
3d. It can be observed that in both tests displacement of superficial layers under the pounder
is greater than the displacement in between the prints; however, displacement tends to
become uniform as depth increases. Reducing pounder spacing increases displacement
uniformity of the upper layers, and suggests that it will be possible to realise reasonably
uniform ground layers by selecting an appropriate dynamic compaction grid size. Hajialilue
90
and Zare's experiment does not include an ironing phase; however, it can be expected that
after filling the craters, levelling the site, and performing an ironing phase with a close grid
spacing the superficial layers will also become reasonably uniform.
(b)
(a)
Figure 2‐59: Displacement contours for (a) pounder spacing= 6d and (b) pounder spacing= 3d
(Hajialilue‐Bonab and Zare, 2014)
Figure 2‐60 shows the strain diagrams for tests in which pounder spacing were 6d, 5d, 4d,
and 3d. Once again it can be observed that reduction of pounder spacing increases the trend
of uniformity, and points at the same depth will tend to have the same compaction.
Hajialilue and Zare also studied change in relative density at the spacing midpoint line for
various pounder spacing. The result of this part of the study is shown in Figure 2‐61. As
expected, increase of relative density reduces as (normalised) depth increases; however, the
rate of reduction decreases as print spacing is reduced. Figure 2‐61 also suggests that
increase of relative density curves for print spacing of 4d, and 3d are almost the same
91
Considering that the magnitude of energy intensity in this physical model is much less than
field dynamic compaction, noting that this experiment does not extend to the midpoint
between 4 prints on a grid (that is further away from the prints that the midpoint between
two prints), and with recognition that print spacing (inclusive of phases 1 and 2 of deep
dynamic compaction) of field work is usually less than 4d, it is the author's opinion that this
experiment should not necessarily lead to a conclusion that additional improvement will
always be negligible when grid spacing is reduced to less than. However, the experiment
results are an indication that additional improvement could be negligible when spacing is
reduced to less than a certain figure.
(a)
(b)
(c)
(d)
Figure 2‐60: Strain diagrams after 12 blows with pounder spacing being equal to (a) 6d, (b) 5d, (c) 4d
and (d) 3d (Hajialilue‐Bonab and Zare, 2014)
92
Figure 2‐61: Increase of relative density at the between prints centreline (Hajialilue‐Bonab and Zare,
2014)
To understand the significance of pounding sequence, Hajialilue and Zare extended their
experiment by carrying out pounding with spacing of 3d in two ways; i.e., in the first test the
first print was initially subjected to 12 consecutive blows followed by 12 consecutive blows
to the second print, and in the second test prints were alternatively subjected to blows until
they both received 12 non‐consecutive blows. The result of this phase of the experiment
showed that alternating the blows between the first and second prints was able to create
larger displacements than compacting each print by consecutive pounding. The author notes
that common practice in industry is consecutive pounding of each print and then moving on
to the next print. Although common practice may result in lesser treatment, and alternating
pounding is certainly worth further investigation, it may however, not be practical to
alternate the blows between the prints as this will greatly reduce production due to increase
of setting up time.
The magnitude of soil improvement by dynamic compaction is dependant, among a number
of parameters, on the amount of applied impact energy and initial density of the soil. For
commonly used energy ranges; i.e., 100 to 300 tm/m2, improvement in the properties of soils
is in the order of 100 to 400% increase in strength as measured by tests such as SPT, CPT or
PMT (Lukas, 1986). Table 2‐5 presents the magnitude of soil improvement for representative
soil types. The higher values are for soils that are initially loose and the lower values are for
soils that are initially denser.
93
Soil Type Anticipated amount of
improvement*
Pervious coarse grained soils – sands and gravels 300 to 400%
Semi pervious soils
Silty sands 100 to 400%
Silts and partially saturated clayey silts 100 to 250%
Partially saturated impervious soils – clay fills and mine spoils 200 to 400%
Landfills 200 to 400%
Building rubble 200 to 300%
* For applied energies of 100 to 300 tm/m2
Table 2‐5: Anticipated relative improvement for different types of soils (Lukas, 1986)
Based on the data that he had studied, Leonards (1980) speculates that it appears there may
be an upper bound to the densification that can be achieved, which corresponds
approximately to CPT cone resistance, qc, of 15 MPa. Lukas (1986) proposes that typical upper
bound values for SPT, CPT and PMT would be as summarised in Table 2‐6. The highest values
occur in coarse pervious soils and the lowest values can be envisaged for landfills.
Figure 2‐62 schematically shows a soil profile for any test before and after dynamic
compaction (Lukas, 1986). Figure 2‐62(a) shows improvement after the deep phases of
compaction. It can be seen that while the soil parameters have increased beyond a certain
94
depth, the superficial layer of soil above the crater depth has loosened. Maximum
improvement as suggested in Table 2‐6 usually occurs at depths of about D/2 to D/3.
Implementation of light pounding or ironing after the deep compaction phases is able to
increase the superficial layer’s compaction, and as shown in Figure 2‐62(b), the soil
parameters increase consequently. The depth of influence and depth of maximum
improvement does not change at this phase, and remains at where they were.
Varaksin and Racinais (2009) propose that the ground improvement profile after dynamic
compaction can be formulated in the form of a parabola, which can be written as shown in
Equation 2‐69.
NGL 2‐69
Figure 2‐62: (a) Initial stages of pounding, and (b) after densification, including the ironing pass
(Lukas, 1986)
z= point level
NGL= natural ground level
f1= fmax= maximum improvement factor observed at ground level, which is a unit dependant
parameter that was equal 0.008 times total compaction energy (in tm/m2) in the case studied
by Varaksin and Racinais.
f2= 1= improvement factor obtained at the depth of influence.
f(z)= improvement factor at elevation z
95
As illustrated in Figure 2‐63, Lukas (1986) has studied the lateral ground movement in
dynamic compaction, and has measured lateral movement in three types of soils at distances
of 3 m and 6 m from the pounder’s drop point. From his records, it appears that the ground
movement initially increases to a point, but then drops off until at some point soil particle
movements become negligible.
(a) (b)
Figure 2‐63: Lateral movement (a) 3 m and (b) 6 m away from pounder drop point (Lukas, 1986)
Similar results have also been achieved in laboratory scale dynamic compaction by Bonab
and Rezaei (2009). As shown in Figure 2‐64, a sickle shape curve characterises the soil
displacement profile along vertical lines of soil. The curves become flatter as they go farther
away from the impact centre. It was observed that the displacement of soil in any given
horizontal plane has a bell shaped curve.
Poran and Rodriguez (1992) cite Heh (1990) and Poran et al. (1991) for a laboratory model
dynamic compaciton test programme on dry sand. As shown in Figure 2‐65, Heh computed,
and plotted the density contours for the final state of the sand for each test. Based on these
results it appeared that density contours could be approximated as upside‐down
semi‐prolate spheriods with dimensions of a and b respectively for the horizontal and vertical
radii, as shown in Figure 2‐66. The parameters a and b were measured from the centre of the
original ground surface, above the centre of the base of the crater.
96
(a) (b)
Figure 2‐64: Normalised (to pounder diameter) displacement vectors after 10th impact for (a)
vertical lines and (b) horizontal lines of soil (Hajialilue‐Bonab and Rezaei, 2009)
Figure 2‐65: Experimental contour of plastic volumetric strains, modified from Poran and Rodriguez
(1992)
Figure 2‐66: A semi‐prolate spheroid can be used to approximate the volume of density contours
(Poran and Rodriguez, 1992)
97
Total specific energy and semi‐prolate spheroid dimensions were respectively normalised to
NWH/ A.b , a/d and b/d to minimise the effects of model size, and a study was conducted
to determine a relationship between normalised total specific energy and the normalised
dimensions for different relative density (volumetric strain) contours.
d = pounder diameter
N= number of drops per print
A= pounder base area
The best linear fits for these semi‐log relationships are:
log 2‐70
log 2‐71
j, k, l and m are regression curve fitting parameters that are presented in Table 2‐7 along with
their coefficient of determination, R 2. NWH/ Ab is in kPa. For non‐circular pounders, d is
the width of the pounder.
State of improvement
Unaffected Slight Moderate Strong
Volumetric strain (%) 0 1.5 5.2 7.7
Relative density (%) 25 35 65 85
j ‐12.59 ‐13.22 ‐15.27 ‐15.27
k 8.08 7.91 7.79 6.25
2
R 8.84 8.84 0.90 0.53
l ‐2.49 ‐2.39 ‐4.25 ‐4.43
m 1.97 1.90 2.32 1.99
R2 0.79 0.71 0.81 0.65
Table 2‐7: Curve fitting coefficients for Equations (Poran and Rodriguez, 1992)
For simplicity, Berry et al. (2004) have proposed to model the void reduction of the soil profile
using the Rayleigh distribution in the form of a Rayleigh distribution, which can be
mathematically written in the form of Equation 2‐72:
98
2‐72
z= depth from surface
= depth of maximum strain
ε z = strain at depth z
it is noted that Hajialilue‐Bonab and Zare (2014) who have also carried out physical modelling
of dynamic compaction have not observed the same improvement pattern as Heh (1990)
that has been used as the basis of Poran and Rodriguez's method (see Section 2.5.8.2). A
comparison of results obtained by Heh and Hajialilue‐Bonab and Zare is shown in Figure 2‐67.
Hajialilue‐Bonab and Zare explain that the results of the two studies were not similar because
the characteristics such as applied energy and momentum of the two experiments were very
different. Noting that Poran and Rodriguez (1992) propose their research as a design method
(see Section 2.5.8.2), the author is in the opinion that further research should be carried out.
Figure 2‐67: Comparison of relative density contours presented by Heh (1990) and Hajialilue‐Bonab
and Zare (2014)
Varaksin (1981) has reported a dynamic compaction case history at an industrial site in
Ashuganj, Bangladesh. In that project a Menard Tripod was used to lift a 40 ton pounder 23
metres. During the post ground improvement verification testing excess pore water pressure
with low dissipation rates was observed at depth in the lower silt and sand (8 to 18 m deep).
99
The phenomenon was monitored in the upper silt layer where the drainage conditions were
more favourable. Testing was performed 25, 40 and 60 days after dynamic compaction. It
was observed that the PMT limit pressure, Menard modulus of deformation and ratio of
Menard modulus to limit pressure had increased as time had gone by. Varaksin notes that
this phenomenon is often observed without noticeable volume change after dissipation of
total pore water pressure, and notes that it is analogous to thixotropic recovery.
Varaksin (2014) recalls that Mitchell reports a vibro compaction project in Nigeria in which
CPT results improved after 6 months by approximately 20%, but there could be doubts to
whether the re‐tests were carried out at the exact locations of the original tests or by an
offset that was closer to the vibroflot. In Varaksin’s opinion clean sand will probably not
improve at all just 1 week after ground improvement, but silty soils may experience a
continuance of improvement.
Mitchell and Solymar (1984) reiterate that the specific mechanisms leading to an
improvement in the soil parameters was not well understood, but speculate that the
observed behaviour is, in some respects, similar to thixotropic strength gain in fine‐grained
soils. They note that in the case of sands, the extent of strength loss on disturbance followed
by strength gain at rest was not known, nor was it evident that the process occurred without
chemical changes, and add that the most probable cause of the observed phenomena
seemed to involve the formation of silica acid gel films on particle surfaces and the
precipitation of silica or other material from solution or suspension as a cement at particle
contacts.
Mitchell and Solymar cite Denisov and Reltov (1961) who had suggested that hydrolysis can
cause disruption of the silicate particle surface accompanied by formation of soluble
compounds and silica gel. The gel adheres to the surface in a thin layer and has cementing
properties. Denisov and Reltov had performed experiments to study the force required to
100
dislodge a quartz sand particle from a quartz plate on which the particle rested. The
investigation showed that the adhesion force increased with the time of contact between
particle and plate. For a test in air the force approximately doubled during the first 20 hours
after the particle was placed on the plate. If after this period the particle and plate were
placed in water for 14 days, then the force increased by a factor of more than three.
Mitchell and Solymar note that pressure and sand particle surface characteristics may also
be important in this phenomenon. They cite Pettijohn et al. (1972) who had noted that
pressure solution may be important due to high stress at grain contacts, and that due to the
high pressure, higher solubility leads to a preferential solution. They also cite Bely et al.
(1975) who had noted that sands with polished grains have fewer opportunities for structural
connections, chemical and biological processes are not so active, and the content of
physically linked moisture is less than for particles with scaly surfaces.
As a discussion to Mitchell and Solymar’s (1984) paper Schmertmann (1987) appears to
doubt that dissolution and precipitation of silica, leading to cementation at inter particle
contacts, is the likely major factor in explaining the strengthening behaviour of sand, but
briefly proposes two alternative explanations. Analogous to thixotropic behaviour in fine
soils, Schmertmann considers time dependant particle re‐orientations in sands important
enough to be included as a secondary ageing contribution to the computation of settlements.
Schmetmann notes that a dispersive particle re‐orientation takes place in all soils with time,
and that this results in a seemingly purely frictional gain in modulus and strength with time.
He also gives increases in horizontal stresses due to time dependant recovery as a second
explanation.
Similar to Schmertmann (1987), in Mesri et al.’s opinion (1990) a possible mechanism for the
time dependant increase of clean sand strength is the continued rearrangement of sand
particles during secondary compression. They note that improved frictional resistance to
deformation develops with time with a gradual increase in interlocking of particle surface
roughness and increased geometrical particle interference through more efficient packing.
Mesri et al. propose that in the absence of continued cone penetration testing subsequent
to compaction, the empirical equation that is presented in Equation 2‐73 can be used for
predicting and extrapolating post soil densification CPT cone resistance from measurements
that are carried out immediately after ground modification.
101
2‐73
qc = cone resistance at any time greater than reference time, tR
qcR = cone resistance at reference time
tR = reference time, which must be greater than end of primary consolidation time
t = any time greater than the reference time
CD = a parameter that reflects any densification by such disturbance mechanisms as vibration
and blasting, which are not related to the ideally static increase in effective vertical stress.
Cα = secondary compression index
CD = compression index beyond the critical pressure.
These two indices can be calculated from Equations 2‐74 and 2‐75:
∆
2‐74
∆ log
∆
2‐75
∆ log ′
∆ = change in void ratio
′ = effective vertical stress
Michalowski and Nadukuru (2011a, 2011b, 2012) comment that rearrangement of particles
and restructuring are the results of an underlying process, and cannot be considered to be
themselves the cause of time dependant strength gain. To explain particle rearrangement
and strength gain they propose the hypothesis of static fatigue also known as stress corrosion
cracking (as the process is dependent on the presence of moisture).
When two soil grains come into contact, the micro‐morphologic features undergo
compression, shear and bending, which will cause fracturing in the microscopic asperities
and crystalline debris fragments. This fracturing, which is referred to as static fatigue by
Michalowski and Nadukuru is a time delayed process, and does not occur immediately after
the contact is loaded. Micro cracking causes the grains to move microns closer together by a
distance comparable with a fraction of the size of the asperities on the grain surface. Grain
convergence leads to firmer contacts, and consequently an increase in the contact stiffness,
102
which propagates through the spatial scales to be manifested at the macroscopic scale as an
increase in macroscopic shear modulus; i.e., the elastic property in macro‐scale. This process
is time‐delayed, as stress corrosion cracking does not occur at all contacts simultaneously.
As a consequence of the change in elastic moduli, the horizontal stress in a sand bed (under
one‐dimensional deformation conditions) will increase; however, the vertical stress will
remain unchanged because it is governed by gravitational forces. Consequently, the
deviatoric stress will reduce, and the stress state will move further away from the yield line,
which will lead to an increase in the cone penetration resistance.
USACE (US Army Corps of Engineers, 1999) defines conventional ground improvement as
applications of soil treatment for intents such as improving bearing capacity, slope
stabilisation, increasing the rate of consolidation settlement and improving seepage barriers.
When a site is compacted uniformly over the entire area, a peripheral band appears with
characteristics that are intermediate between the non‐compacted surrounding zone and the
compacted interior zone. This band follows the line of the perimeter and has a width equal
to approximately two times the thickness of the layer that is being improved (Menard and
Broise, 1975). Thus, the uniformly improved zone will have dimensions that are reduced by
two times the treatment thickness, and an offset width from net treatment area has to be
included in the treatment program to obtain effective treatment throughout the target area.
Menard and Broise comment that the offset width may be reduced by implementation of
specific measures such as increasing the energy applied at the periphery, which involves
complementary inspections to avoid the appearance of frontiers in the densification of the
soil with consequent differential settlements beneath the construction.
Lukas (1995) lightly touches the issue of extent of ground improvement, and suffices to note
that the offset width should be equal to the depth of the loose deposit.
USACE (1999) states that for conventional ground improvement, increasing the rate of
consolidation settlement and improving seepage barriers, the depth of treatment should
extend either to the depth of influence of the structure or to the bottom of the layer requiring
improvement. Based on an approximate load spread of 2 to 1, USACE also specifies that the
103
areal extent of treatment for conventional applications should extend outside the perimeter
by at least a distance equal to half the thickness of the treated layer.
Mitchell et al. (1995) note that loss of strength in improved ground can result during
earthquakes due to inward migration of pore pressure plumes from adjacent liquefied soil,
and add that common design is to extend the treatment a distance equal to the depth of the
layer being compacted.
Liu and Dobry (1997) performed a series of centrifuge tests to study the effect of improved
depth on foundation settlement induced by liquefaction. For their experiment they
constructed model sand deposits with relative densities of approximately 52%. They then
used a vibrating tube to vibrate the soil in 19 points per test to compact areas equal to 1.6
times the diameter of test footing to a relative density of approximately 90%. The testing
model is shown in Figure 2‐68.
Figure 2‐68: Test model for studying seismic response of shallow foundations in liquefiable sand (Liu
and Dorby, 1997)
The improvement depth was made variable, and its ratio to the footing diameter was made
variable from as low as 0 to as high as 2.76. As shown in Figure 2‐69, it was observed that
footing settlement decreased as ratio of improvement depth to footing diameter increased.
The rate of settlement reduction was initially high to about when ratio of improvement depth
to footing diameter reached a value of approximately 1.5, beyond which the footing
settlement increased only marginally. As shown in Figure 2‐70, the experiment also
demonstrated that the ratio of footing settlement to free surface (non‐footing) soil
decreased at a higher rate up to about when ratio of improvement depth to footing diameter
reached 2 (the thickness of treated ground at that ratio was about 70% of the liquefiable
layer) beyond which the settlement ratio remained at unity. Liu and Dobry noted that they
were not able to confirm that what they have observed in their specific experiment would
also be applicable when conditions are different, but their experiment has been able to
104
provide enough evidence to demonstrate that liquefiable soil foundations will settle less
when the ratio of treatment depth to footing width increases.
Figure 2‐69: Foundation settlement versus normalised compaction depth (Liu and Dorby, 1997)
Figure 2‐70: Footing settlement normalised by free field soil settlement versus normalised
compaction depth (Liu and Dorby, 1997)
Liu and Dobry plotted the ratio of average foundation settlement to thickness of thickness of
liquefaction, compared it with the ratio of building width to thickness of liquefaction for the
1964 Niigata and Luzon earthquakes. Their conclusion was that as the centrifuge data points
without soil compaction fell within the earthquake field band, then using building width to
normalise field data and footing width to normalise centrifuge data could be linked, and that
it seemed reasonable to assume that the entire width of the building rather than the sizes of
the individual footings determined the overall pore pressure response under the buildings.
Akiyoshi et al. (1993) have also conducted a series of tests using a shaking table and by
performing numerical analyses to assess the effect of ground improvement with sand
compaction piles for mitigation of liquefaction. In their experiment the saturated sand
models initially had relative densities of about 22%. Then each model was compacted in 16
locations with mini sand piles that were 70 mm in diameter and 500 mm long. Depending on
the compaction energy that was applied, the average relative density of the compacted areas
were 40 to 50%.
105
It was observed that depending on the amount and extent of compaction the ground still
liquefied gradually. However, the numerical analyses showed slow accumulation of excess
pore water pressure in the compacted area and rapid accumulation in the uncompacted
peripheral areas. After the excess pore water pressures reached a peak in the uncompacted
areas, they began to dissipate. Akiyoshi et al. note that the results indicated that the
compacted area gradually liquefied due to seepage flow from the adjacent liquefied area.
Figure 2‐71 shows the distribution of excess pore water pressure ratios at four transient
times for a ground model with width to depth ratio of compacted area being 1.6 and depth
of compacted area to total ground thickness equalling unity. It can be understood from
Figure 2‐71(d) that the angle between the boundary of compacted and non‐compacted areas
and the line corresponding to the excess pore water pressure ratio curve of 0.5 is
approximately 30 to 35%, in the case when most of the compacted area does not liquefy.
Akiyoshi et al. comment that the compacted triangular zone between these two lines can be
considered to liquefy readily as a result of the influence of the adjacent liquefied
uncompacted area. The results of this study appeared to be consistent with the experimental
results that were performed by Japan's Port and Harbour Research Institute, PHRI (Iai et al.,
1987).
Figure 2‐72 shows the relationship between the maximum excess pore water pressure ratio
at two depths of the centre point of the treated area and the width to depth ratio of the
compacted area. It can be observed that the excess pore water pressure ratio decreases as
the width to depth ratio of the compacted area increases. It appears that in Akiyoshi et al.'s
experiment the maximum excess pore water pressure ratio of the centre point of improved
soil was less than 0.5 when the width to depth ratio of the compacted area was greater than
1.5. Akiyoshi et al. note that although the amount of data was limited, there may be an
optimum compaction width for design purposes.
106
Figure 2‐71: Distribution of excess pore water pressure ratio at different transient times for a ground
model with width to depth ratio of compacted area being 1.6 and depth of compacted area to total
ground thickness equalling unity; modified from Akiyoshi et al. (1993)
Figure 2‐72: Relationship between maximum excess pore water pressure ratio and width to depth
ratio of compacted area when treatment depth equals liquefaction thickness (Akiyoshi et al., 1993)
USACE (1999) states that for liquefaction mitigation, the depth of treatment should generally
extend to the bottom of the layer that requires improvement, particularly for large or heavily
loaded structures. USACE seems less stringent for lightly loaded structures, and notes that it
may not be necessary to treat the entire liquefiable layer for such structures. Yet, USACE
further adds that design procedures for an improved crust over liquefiable soils are not well
established, and refers to Ishihara (1985) who presents correlations between minimum
thickness of a non‐liquefiable surface layer, the maximum thickness of an underlying
liquefiable soil layer and surface manifestations of liquefaction for free field conditions or
lightly loaded structures. USACE also specifies that areal extent of treatment for liquefaction
protection should extend generally outside the perimeter of the structure at least by a
distance equal to the thickness of the treated layer.
107
Hausler (2002) cites Cooke and Mitchell (1999) for also commenting that standard practice
for improvement of liquefiable soil beneath most structures is to extend ground
improvement to the bottom of the liquefiable material. Hausler (2002) also cites PHRI, (1997)
for recommending ground improvement to be carried out through the full thickness of the
potentially liquefiable soil, but at the same time cites Japanese Geotechnical Society, JGS,
(1998) that ground improvement contractors in Japan are not required to extend treatment
to below the depth of 20 m for most buildings, bridges and railway structures, and to below
the depth of 15 m for oil storage tanks and LNG storage tanks. According to Hausler, JGS
(1998) recommends that lateral extent of ground improvement be equal to ½ of the
improved depth. Hausler also cites PHRI for reporting that as shown in Figure 2‐73, the Japan
Fire Defence Agency, which regulates hazardous facilities such as oil tanks, has regulations in
place since 1974 that specifies that the lateral extent of the ground improvement area should
be equal to or greater than 2/3 of the compaction depth.
Figure 2‐73: Ground improvement zone as specified by Japan Fire Defence (Port and Harbour
Research Institute, 1997), as cited by Hausler (2002).
Hausler (2002) notes that shaking table tests and seepage analyses performed by PHRI shows
that, as shown in Figure 2‐74, an area of softening or instability exists above a failure plane
at 30o with the vertical plane between the improved and unimproved zones, and adds that
according to the simplified design procedure set forth by PHRI, the strength of this area
should not be expected to contribute to bearing capacity, and the treated area should be
designed wide enough to obtain sufficient bearing capacity from shear resistance along the
Terzaghi failure surface.
108
Figure 2‐74: Stability of treated zone without contribution of a 30o wedge (Port and Harbour
Research Institute, 1997), as cited by Hausler (2002).
Although the oldest and most relied on rules of thumb, the design guidelines proposed by
Lukas (1995) are the closest to industry practices. He proposes six steps:
1. Selection of pounder weight and drop height based on the required depth of
improvement (refer to Section 2.5.2).
2. Determination of applied energy to achieve the required depth of improvement:
109
For this purpose he proposes the utilisation of Table 2‐8 for the selection of the unit
energy for the proper deposit classification. Volumetric compaction energy can be
converted to planar energy intensity by multiplying the volumetric energy by the
treatment thickness.
3. Determination of the treatment area:
3.1. For level sites, use a grid spacing throughout the area in need of improvement and
an offset area beyond the project boundaries equal to the depth of improvement
(refer to Section 2.5.7 for a more comprehensive discussion and alternative offset
widths).
3.2. If slope stability is a concern, improvement over a wider plan area may be required.
3.3. At load concentration areas, apply additional energy as needed.
4. Determination of grid spacing and number of drops
For this purpose Lukas proposes the implementation of Equation 2‐76:
2‐76
AE= Applied energy
I= energy intensity= applied energy per unit area
N= number of blows
P= number of passes
W= weight of pounder
110
H= drop height
s= grid spacing
In Equation 2‐76, grid spacing is the final distance inclusive of phase 1 and phase 2 prints.
The number of passes is not the number of phases; it refers to applying (equal amount
of) blows to the same prints of the grid.
4.1. Select a grid spacing ranging from 1.5 to 2.5 times the diameter of the pounder
4.2. Enter W and H from Step 1 and applied energy intensity from Step 2 into Equation 2‐
76.
4.3. Use Equation 2‐76 to calculate the product of N and P. Lukas suggests that generally
7 to 15 drops are made at each print, and adds that if the calculations indicate
significantly more than 15 or less than 7 blows, then grid spacing should be adjusted.
5. Multiple passes
Lukas speculates that prediction of crater depths or ground heave in advance of dynamic
compaction is difficult, and assesses that multiple passes should be applied when very
loose deposits like landfills are present or where silty deposits are nearly saturated.
He suggests that:
5.1. Crater depths should be limited to the height of the pounder plus 0.3 m to allow for
non‐complicated extraction of the pounder from the ground.
5.2. Energy application should stop if ground heave occurs. Although heaving is a sign of
non‐volumetric plastic displacement, the author does not believe that any amount
of heaving necessitates the termination of compaction efforts. Indeed, it is
preferable to observe net volumetric change, being the difference between volume
of compaction and volume of heave, and to terminate pounding when impact
energy does not lead to any significant net compaction.
5.3. If 5.1 and 5.2 occur before the required number of blows has been applied, then
multiple passes should be used to allow for dissipation of pore water pressures.
6. Surface stabilising layer:
111
Not needed for Zone 1 soils (refer to Section 2.5.1), but may be possibly required for
nearly saturated Zone 2 soils. The surface stabilising layer is usually required for landfills.
6.1. When a surface stabilising layer is used, the thickness ranges generally from 0.3 to
0.9 m.
6.2. Lukas (1995) does not mention the ironing phase in his guideline, but does mention
it elsewhere in his publication.
Poran and Rodriguez (1992) have proposed a design method for dry sands based on test
models reported by Heh (1990) and Poran et al. (1991), which has already been explained in
Section 2.5.4. In this method, Poran and Rodriguez estimate the soil mass that has been
compacted to a certain strain or relative density using semi‐prolate spheriods, with side, top
and approximate shape shown in Figure 2‐66. Based on Equations 2‐70 and 2‐71, and
Table 2‐7, Poran and Rodriguez developed the charts shown in Figure 2‐75 to demonstrate
the relationship between the normalised semi‐prolate spheriod’s horizontal and vertical radii
(respectively a and b, as shown in Figure 2‐66) and normalised total specific energy,
NWH/ Ab (in kPa). The definitions for the variables in the chart are given in Section 2.5.4.
The least and most improvement lines refer to DC induced plastic volumetric strains of
respectively 1.5% and 7.7%. Poran and Rodriguez contemplate that the parameter range
obtained from the model test results in the laboratory is also adequate for actual dynamic
compaction works, and as more field data is investigated, it will be possible to further
calibrate these relationships for different soil conditions.
Poran and Rodriguez (1992) recommend that the numerical average improvement lines
shown in Figure 2‐75 and redrawn in Figure 2‐76 be used for the purpose of dyanmic
compaction design. In this process, the vertical radius of the semi‐prolate spheriod should be
taken as equal to the depth of improvement. Poran and Rodriguez do not explain why they
make such a recommendation, and the significance and consequences of this decision. These
lines that correspond to strain values are indeed a fundemental component of the design,
and all other steps of the design will be impacted greatly and directly by this choice. In the
end, it is not clear what criteria are to be satisfied by these choices of lines.
112
(a)
(b)
Figure 2‐75: Range of improvement in soil mass for (a) b/d and (b) a/d versus total normalised
specific energy (Poran and Rodriguez, 1992)
Figure 2‐76: Proposed design chart for dynamic compaction in sandy soils (Poran and Rodriguez,
1992)
113
Poran and Rodriguez propose that the following steps should be used for the preliminary
design of dynamic compaction in dry sandy soils:
1. Examination of soil data and defining the required depth of improvement.
2. Determination of W, H and d, and calculation of WH and A. The equivalent diameter for
non‐circular pounders is to be taken as the pounder's width.
5. With the value of a, the grid pattern for the first and subsequent passes may be designed
(based on the semi‐prolate spheroid shape) to obtain the most complete overlap of
improved soil mass to the desired depth.
Figure 2‐77: Overlapping of semi‐prolate spheroids, modified from Poran and Rodriguez (1992)
Poran and Rodriguez do not explain why they are targeting the most complete overlap; as
there is no discussion about the amount of improvement in the overlapping zone, the author
speculates that it is their intention to minimise the volume of soil outside of the semi‐prolate
spheriods in between the prints, shown in Figure 2‐77, which does not receive the same
amount of treatment as in the semi‐prolate spheriods themselves. They are also not clear on
114
the procedure for determining the most complete overlap of improved soil mass to the
desired depth, and suffice by presenting an example in which they deem that considering a
grid spacing equal to 0.9a is appropriate.
Based on two dimensional finite element analyses using a cap model, Jahangiri et al. (2011)
have developed a design method for dynamic compaction. In their approach, it is assumed
that the soil beneath the pounder centre reaches its maximum achievable compaction, and
subsequent pounder blows at adjacent points does not induce further improvement at that
location. Therefore, it is also assumed that the degree and depth of improvement along the
vertical axis of prints is unlikely to be significantly affected by the compaction of adjacent
prints. However, there is significant overlapping of lesser improved zones that are further
away from the print location.
∙ 100 % 2‐77
100
Dd = relative density
Ddo = initial relative density
As can be seen in Figure 2‐78, for a single DC print (left point), relative degree of
improvement, Ir, is achieved at depth z. The second print (right point) is assumed to be at a
distance away from the first print in such a way that two improved bulbs with ½ of the relative
degree of improvement, Ir/2, intersect one another also at depth Z. Jahangiri et al. assume
that by using the relative degree of improvement in the overlapping area will also become Ir
using the law of superposition. Noting that the left and right prints have equal bulbs, it is
concluded that grid spacing should be double the distance between the centre of each print
and the intersection point.
115
Figure 2‐78: Schematic scheme for evaluation of print spacing (Jahangiri et al., 2011)
Jahangiri et al. have performed a number of analyses with various drop heights, pounder
weights, pounder radii and number of drops to develop charts for grid spacing when Ir is 8,
15, 30, 46 or 62%. Review of the analyses suggested that trends of variation of grid spacing
were tightly bounded when Ir was 8, 15 and 30%. These were considerably different from
when Ir was 46 or 62%, but the average difference for cases with Ir values of 46 and 62% was
minor and only 5.7%. Thus, two charts were developed for determination of grid spacing
based on NWH/A, pounder radii, r, and Ir. Figure 2‐79 and Figure 2‐80 show the
relationships when Ir is respectively equal to 8, 15 and 30%, and 46% and 62%. Jahangiri et
al. note that for pounder radii other than what has been indicated in the figures; i.e., 0.5, 1
and 1.5 m, interpolation may be used, but do not clarify how to proceed for other Ir values.
Figure 2‐79: Print spacing curves for Ir = 8%, 15% and 30% (Jahangiri et al., 2011)
116
Figure 2‐80: Print spacing curves for Ir = 46% and 62% (Jahangiri et al., 2011)
In the opinion of the author the method proposed by Janhangiri et al. is very promising, but
application of the law of super position to soil that has undergone potentially nonlinear very
large plastic strains is very questionable, and without further study, it is doubtful that a
simple summation of Ir halves will yield true and accurate results in the overlapping zones.
117
2.6 Design Guidelines for Dynamic Replacement
It was mentioned in Section 2.4.9 that loads are distributed between the in‐situ soil and DR
columns by arching through the load transfer platform. A number of different methods are
available for estimating the amount of load that is transferred to the DR columns. The most
relevant methods are discussed in this section.
Punching of the column into the upper granular fill may be improved by using geotextiles or
higher grade compacted fill in the lower layers of the embankment that will act as the load
transfer platform.
One of the first researchers who studied soil arching were Marston and Anderson (1913) who
evaluated soil loadings on buried pipelines. Their formula was later adopted and modified in
BS 8006:1995 (British Standards Institution, 1995) for the two dimensional calculation of
average pressure acting on a pile cap.
Although BS 8006 assumes plane strain behaviour and does not consider the actual three
dimensional arching that occurs in reality when there are no beams, this method is
nevertheless used by engineers to determine the ratio of load that the piles (or columns)
support. Van Eekelen & Bezuijen (2008) have reviewed BS 8006, and have proposed an
adaptation of the equations so that the method becomes three dimensional and vertical
equilibrium is satisfied. Van Eelelen and Bezuijen’s modifications are especially of interest
for calculating the tensile stresses in geotextiles that could be used for basal reinforcement
in conjunction with the rigid inclusions.
To ensure localised differential deformations cannot occur at the surface of embankments
(which can be a problem with shallow embankments) BS 8006 recommends that the
relationship between embankment height and pile cap spacing be maintained to
0.7 2‐78
Where
118
a= size (or diameter) of pile caps (assuming full support can be generated at the edges of the
caps),
s = spacing between adjacent piles, and
H = height of the embankment.
According to BS 8006 the ratio of vertical stress exerted on top of the pile caps to the average
vertical stress at the base of the embankment can be estimated by
′
2‐79
′
Where
σ’c = vertical stress on the pile caps,
σ’v = equal to γH ws and is the average vertical stress at the base of the embankment,
γ = unit weight of the embankment fill,
ws = uniformly distributed surcharge loading, and
Ca= the arching coefficient and is as shown in Table 2‐9
Pile Arrangement Arching coefficient
End‐bearing piles (unyielding) 1.95
0.18
Friction and other piles (normal) 1.5
0.07
Table 2‐9: Arching coefficient
Hewlett and Randolph (1988) studied two and three dimensional soil arching of piled
embankments in granular fills. Analogous to dynamic replacement, Hewlett and Randolph
assume in their analysis that no slab is used and the piles (columns in the case of DR) are
placed at a relatively wide spacing.
According to Hewlett and Randolph field evidence suggests that pile caps (columns in the
case of DR) covering as little as 10% of the area beneath the embankment may carry more
than 60% of the weight of an embankment due to arching action in the fill.
119
Hewlett and Randolph also note that in the model test that they used, with a constant ratio
of column spacing to column width, the settlement of the surface of the sand fill was less for
the smaller width columns. The observed deformities of the sand indicated that arching
occurred across adjacent columns. Between the columns, sand close to the subsoil (foam
rubber in the test model) underwent significant settlement (as compared to the settlement
of the fill’s surface). Shear distortion was concentrated in fans originating from the corners
of the columns. Sand was observed to settle uniformly well above the pile caps; however,
the bottom layer of sand remained straight between the piles, which showed uniform
settlement, and suggested that the pressure on the subsoil was uniform.
Based on pile cap (column in the case of DR) size and spacing, height of embankment fill and
friction angle of the granular fill that formed the embankment, Hewlett and Randolph
developed two dimensional (plane strain) and three dimensional expressions for determining
the proportion of weight of the embankment that is carried directly by the pile cap. Having
estimated the proportion of load carried by the columns, the overall performance of the
embankment may be deducted from a separate analysis of the subsoil and membrane.
In the plane strain case, where long arches are supported by continuous ledges, as shown in
Figure 2‐81, it is assumed that the horizontal band of soil that contains the arch is weightless,
and the sand in the infilling regions beneath the arches and the cusps between the arches
does not mobilise any strength (isotropic stress state). Within each arch, the tangential
direction is the direction of major principal stress, and the radial direction is the direction of
minor principal stress. Static equilibrium requires that the arches are semi‐circular, of
uniform thickness, and span adjacent ledges with no overlap of arches.
For an embankment of height, H, spanning pile caps of width, b, at centreline spacing, s, the
efficacy, E, or the portion of load that is supported by the piles can be calculated to be
expressed by
1 1 1 2‐80
2
Kp is the Rankine passive earth pressure coefficient, and is equal to
120
Figure 2‐81: Plane strain (2 dimensional) arching of soil (Hewlett and Randolph, 1988)
1
45° 2‐81
1 2
φ is the internal friction angle of the sand fill. The remainder of the fill weight can be assumed
to be uniformly distributed on the subsoil.
As shown in Figure 2‐82, the case of most relevance for embankment piling (or DR
installation) is for three dimensional spatial arching above a grid of piles where sand vaults
form. The vaults are comprised of a series of domes. The crown of each dome is
approximately hemispherical, and its radius is equal to half of the diagonal spacing ( /√2) of
the piles.
Figure 2‐82: Isometric view of a grid of pile caps (DR columns) and a series of domes forming vaults
spanning between them (Hewlett and Randolph, 1988)
121
In this case, the crown of the dome is not necessarily the weakest region of the system of
vaulting. The limited area of support at the pile caps may lead to a bearing failure at that
point. The approach adopted by Hewlett and Randolph follows the analysis used for
consideration of equilibrium at the crown of the arches. Integration of the tangential stress
in the arch at pile cap level allows an estimate to be made of the overall force that may be
taken by the pile cap.
Analysis of the two regions; i.e., the crown and base of the arches (bearing capacity of the
pile cap punching into the granular fill) leads to two separate estimates of the efficacy of the
pile support of which, the lower estimate should be used for design.
Figure 2‐83: Analysis of arching at the crown of a dome. The diagram on the right represents a
diagonal section through a pile cap and dome crown. (Hewlett and Randolph, 1988)
At the crown of the arch (Figure 2‐83), the form of analysis is similar to that for the plane
strain case, except for the spherical geometry and the inclusion of self‐weight. It can be
demonstrated that based on the analysis of the crown, the efficacy of the piles will be
1 1 2‐82
1 2‐83
2 2
2‐84
√2 2 3
122
2 2
2‐85
√2 2 3
As shown in Figure 2‐84, the vault comprises of four plane strain arches at the pile cap, with
each occupying a quadrant of the cap. It can be analytically demonstrated that efficacy will
be:
2‐86
1
Figure 2‐84: Arching in the sand fill immediately over a pile cap (Hewlett and Randolph, 1988)
Where:
2 1
1 1 2‐87
1 1
123
Efficacy will be the minimum of the values calculated from Equations 2‐82 and 2‐86. As in the
case of plane strain, the remainder of the fill weight can be assumed to be uniformly
distributed on the subsoil.
As might have been expected, at low embankment heights relative to the spacing of columns
(b/s), the performance of the columns is governed by the condition at the crown of each
arch. However, as the height of the embankment increases, the critical region transfers to
the columns.
EBGEO 2004 Section 6.9 is a recommendation for piled embankments design procedure
issued by Deutsche Gesellschaft für Geotechnik or the German Geotechnical Society (DGGT,
2004) in July 2004. The method adopts the so‐called multi‐shell arching theory based on
Zaeske’s (2001) research. Satibi (2009), Kempfert et al. (2004), and Rainthel et al. (2008) have
reviewed the German Code recommendations. Their papers have been used as references
for this section.
The German method EBGEO 2004 is recommended for the design of embankment on rigid
end‐bearing piles. For the design of embankment on floating piles, further elaboration of the
design procedure is required (Satibi, 2009).
As shown in Figure 2‐85, in EBGEO 2004 it is assumed that the piles are spaced at diagonal
distances. Denoting the pile (or DR column) diameter and diagonal spacing between the piles
respectively with d and sd, in rectangular grids:
2‐88
In triangular grids:
124
max , 2‐89
Figure 2‐85: Point support definitions (rectangular‐triangular) (Kempfert et al., 2004) after Zaeske
(2001)
Similar to Hewlett and Randolph (1988), EBGEO 2004 assumes that the arches have the shape
of hemispherical domes spanning between the pile caps. However, as shown in Figure 2‐86,
EBGEO 2004 considers that the arches consist of multi‐shell domes. The topmost arching
shells take the shape of hemispherical domes with the radius being equal to 0.5sd. Inside the
topmost shells, there are multi‐spherical shaped arching shells with radii larger than 0.5sd to
infinity for the lowest arching shell. The lowest arching shell is tangential to the surface of
the soft subsoil.
Figure 2‐86: Theoretical arching model (Kempfert et al., 2004) after Zaeske (2001)
By evaluation of forces equilibrium of an element, and solving the differential equation it can
be derived that the vertical stress at the lowest arching shell (at the soft subsoil) will be:
125
′
2‐90
4
1
2‐91
8
2
2‐92
2
1
2‐93
σ'zo= uniform stress at subsoil level,
γ = unit weight of the embankment fill,
ws = uniformly distributed surcharge loading, and
Kp = passive earth pressure coefficient, and
hg = arching height and can be calculated from
; 2‐94
2 2 2
When the pile cap is not circular, the diameter may be taken as the equivalent diameter
based on pile cap area, Ac:
4
2‐95
The effective pressure acting on top of pile cap can be calculated to be:
′ ′ ′ 2‐96
126
AE = area of one unit cell, see Figure 2‐87 (Balaam and Booker, 1981, Goughnour and Bayuk,
1979).
Figure 2‐87: Area of one pile cell (Satibi, 2009) after DGGT (2004)
Similar to Hewlett and Randolph (1988), this analytical model is also based on the lower
bound theorem of plasticity (Satibi, 2009).
ABGEO 2004 requires that the embankment height be at least 0.7sd to ensure that soil
arching fully develops.
Aboshi et al. (1979, 1991) cite Murayama (1962) for establishing equations, which
demonstrate that vertical stresses on a (DR) column and the surrounding soil can be
calculated through the equilibrium of vertical stresses acting on the composite ground.
Defining the area replacement ratio, ac, as the area of one DR column (sand column in the
original research) to the total area of the DR unit cell, and stress distribution ratio (also
referred to as the stress concentration ratio), n, as the ratio of the stress in the soil to the
stress in the DR column:
2‐97
2‐98
σ = average stress on unit cell
σc = stress in DR column
σs = stress in soil
Ac = area of DR column
127
As = area of soil in unit cell
Then, if:
σh = lateral confining stress on the cylindrical surface of the DR column, and
σu = upper yield value of the soil
c = soil cohesion
The parameters defined above are shown schematically in Figure 2‐88. It can be calculated
that:
Figure 2‐88: Stress distribution between soil and columnar inclusions (Aboshi et al., 1979, 1991)
2‐99
1 sin
2‐100
1 sin
2‐101
2 2‐102
From Equations 2‐100 to 2‐102:
1 sin
1 2‐103
1 sin
128
2‐104
1 1
2‐105
1 1
At the beginning of loading the composite ground under ko consolidation process, n equals
unity; i.e., the stresses in the soil and the DR column are the same; however, in the course
of consolidation reaches a yielding value (Aboshi et al., 1979).
While the area replacement ratio and average stress on the unit cell are known, it is also
necessary to determine the value of n to calculate σs and σc. However, n itself is the ratio
between σs and σc, and an additional assumption is required before the stresses in the soil
and DR column can be determined. Varaksin (1981) assumes that the settlement of the
in‐situ soil and column are the same, which is a reasonable assumption for large dimension
foundations, but does not explain his calculation method. If the DR column and the in‐situ
soil deform one dimensionally (i.e., the DR column is fully confined with no lateral
deformation), then at the end of consolidation n should be equal to the ratio of the column
modulus to the soil modulus; i.e.:
2‐106
Ec = elastic (Young) modulus of the DR column
Es = elastic (Young) modulus of the soil
This assumption is still used, and can be found in commercial software; however, even the
software producers caution that while n can theoretically reach high values, experience
shows that it is smaller. In reality, n is not the ratio of the elastic modulus although it has
been suggested that the elastic modulus could be used as an upper bound (Barksdale and
Bachus, 1983). Barksdale and Bachus note that using elastic theory it can be demonstrated
that:
129
2‐107
Dc = constrained modulus of the DR column
Ds = constrained modulus of the soil, which is the elastic modulus in one dimensional
deformation, and for three dimensional deformation will be (Lambe and Whitman, 1979):
1
2‐108
1 1 2
νs = soil Poisson ratio
Although no value for n has been proposed in literature for DR columns, Barksdale and
Bachus note that while application of Equation 2‐108 will yield n values in the range of 25 to
500 (Castro and Sagaseta (2009) put the range of constrained modului ratio for stone
columns at 10 to 50), field measurements suggest that n is between 2 to 5 for stone columns.
Enoki et al. (1991) proposes n to be from 1 to 3 for sand columns. Hence, Barksdale and
Bachus do not recommend the use of the compatibility method and Equation 2‐108 in soft
clays.
By comparing equations that are based on total column confinement and when columns are
able to deform laterally, Castro and Sagaseta (2009) demonstrate that lateral deformations
always reduce the stress concentration ratio by subtracting a term from the oedometric
modulus of the column and adding another term to that of the soil.
Based on numerical studies, the influence of the horizontal deformation and plastic
behaviour of a (stone) column on the distribution of stresses between the soil and the column
is shown in Figure 2‐89. With lateral confinement, n starts from zero and reaches a final value
equal to the ratio of constrained moduli (40); however, as already discussed, this does not
agree with reality. The consideration of elastic radial deformations reduces the final value n
of to 25, with a non‐zero initial value due to the presence of immediate settlement. Plastic
strains in the column further reduce the final value of n to realistic values (around 5), with a
small influence of the dilation angle of the column material (Castro and Sagaseta, 2009).
130
Figure 2‐89: Influence of radial deformation and plastic strains on n (Castro and Sagaseta, 2009)
In addition to the analytically derived methods that can predict the distribution of loads
between the columnar inclusions and the soft soil, it is also possible to make similar
evaluations based on finite element software. This type of calculation is becoming more and
more preferable as commercially available software are becoming more and more capable
of calculating settlements, stresses, pore pressures and weak planes in various complicated
scenarios. Needless to say, the accuracy of such analyses is based on the constitutive laws,
geometrical model and assigned properties of material.
According to Satibi (2009), Zaeske (2001) shows that three dimensional behaviour can be
well approached using plane strain analysis with the geometrical idealisation as suggested by
Bergado and Long (1994). Axisymmetrical or plane strain analysis of soil arching requires less
computer capacity, is faster to compute and is much easier to perform compared to three
dimensional analysis. Hence, a practical but at the same time well modelled axisymmetrical
or plane strain analysis may have significant advantages.
Several methods can be used for modelling the geometry for the purpose of soil arching
analysis in DR columns. The geometrical idealisation models include axisymmetrical, plane
strain and three dimensional.
131
2.6.1.5.1 Axisymmetrical Geometry Models
In axisymmetrical geometry modelling one three dimensional DR cell (see Figure 2‐87) is
transformed into a circular cell with the area of DR column and soil remaining the same. The
transformation of a squared cell into a circular cell is shown in Figure 2‐90. Figure 2‐91 shows
the equivalent unit cell diameters when the column arrangements are square, triangular or
hexagonal (Balaam and Booker, 1981). The axisymmetrical finite element analysis uses one
radian of the circular DR column cell in its calculation.
Figure 2‐90: Axisymmetrical geometrical idealisation of one DR cell (Satibi, 2009)
Figure 2‐91: Equivalent unit cell diameters (a) Square arrangement of columns, (b) Triangular
arrangement of columns, (c) Hexagonal arrangement of columns (Balaam and Booker, 1981)
In this model, it is assumed that each unit cell works independently and with only vertical
strain. Thus, this assumption is valid when a very large area is loaded, such as tanks, storage
132
areas, etc. This model may also is used for modelling a single footing on one DR as the
problem can be envisaged to be a large area with a central load and zero load elsewhere.
Mitchell and Huber (1985) have tried to include the effects of the surrounding inclusions on
the central unit cell in an axisymmetrical model in an innovative way. In their method, it can
is assumed that in addition to the central DR column, there are also concentric DR rings with
radii that increase according to the DR column spacing. The thicknesses of the DR rings are
calculated based on the area replacement ratio.
As shown in Figure 2‐92, the limit of study for the first ring is assumed to be a concentric
circle in the middle of the first and second DR rings. Once the area replacement ratio is
known, the total DR column area can be determined. Consequently, the area of the first DR
ring will be the difference between the total DR area in the limit of ring study and the central
DR column. It can be assumed that the area of the first DR ring is equal to the product of the
thickness multiplied by the ring circumference. Michel and Huber have performed an
example calculation, but the thickness of the first ring can be calculated to be as per
Equation 2‐109. The same concept can be applied for determining the thickness of the other
rings.
1.5
2‐109
2
Figure 2‐92: Axisymmetrical model with concentric rings modified from Mitchell and Huber (1985)
133
2.6.1.5.2 Plane Strain Geometry
The assumption that the unit cells work independently with only vertical strain and without
any lateral displacement is not always correct. In some problems, such as construction of
embankments on soft ground, as shown in Figure 2‐93, lateral deformation and spreading
can occur, and its value can be an indication of the ground stability condition. In such cases,
plane strain models are more meaningful, and will represent the actual conditions more
realistically.
Figure 2‐93: Horizontal displacements in plane strain models (Tan et al., 2008)
One of the methods that can be found in literature for transforming three dimensional grids
of columns into a continuous wall in plane strain condition is by assuming an equivalent wall
stiffness (Kempfert and Gebreselassie, 2006, Satibi, 2009). As shown in Figure 2‐94, in this
method the thickness of the wall is the same as the width (equivalent width for circular DR
columns) of the original inclusion. However, the equivalent wall stiffness (modulus) Eeq is
taken as the weighted average of column and soil stiffness based on an elastic approach.
2‐110
Ec = DR column Young modulus
Es = soft subsoil Young modulus
Aw = equivalent wall area
134
Satibi (2009) notes that when this method is used, the improved area ratio (in other words
the area of the DR column) becomes larger. As a consequence, the volume of fill below the
arch becomes less as compared to the actual three dimensional conditions.
Figure 2‐94: Plane strain geometrical idealisation with equivalent stiffness (Satibi, 2009)
According to Bergado and Long (1994), and as shown in Figure 2‐95, the three dimensional
grid of DR columns can be transformed into rows of continuous walls with equivalent
thicknesses, teq, in plane strain geometry. The thickness of the continuous wall is calculated
based on the assumption that the ratio of improved area; i.e., the ratio of DR column area to
one unit cell area, Ac/AE , is constant. In a rectangular grid:
2‐111
or more simply:
2‐112
Figure 2‐95: Plane strain geometrical idealisation after Bergado (Satibi, 2009)
135
2.6.1.5.2.3 Plane Strain with Equivalent Homogenised Continuum
As shown in Figure 2‐96, in this model the DR columns and in‐situ soil are assumed to have
equivalent properties:
1 2‐113
1 2‐114
1 2‐115
ceq = equivalent cohesion
φeq = equivalent internal friction angle
γeq = equivalent density
cc = DR column cohesion
cs = subsoil (soft soil) cohesion
φc = DR column internal friction angle
φs = subsoil (soft soil) internal friction angle
γc = DR column density
γs = subsoil (soft soil) density
ac = area replacement ratio
m = load distribution ratio, and is equal to
2‐116
1 1
Dimaggio (1987) has expressed Equation 2‐114 slightly differently and in the form of:
1 2‐117
Equation 2‐117 is based on the original understandings of the composite column‐soil
behaviour, and with the thought that that for shear loading undrained conditions were
appropriate. However, as additional projects and knowledge progressed the understanding
of importance of load distribution ratio greatly increased, and a move toward a drained
136
strength approach was developed; i.e., Equation 2‐114 (DiMaggio, 2009). Thus it is generally
Equation 2‐114 that is used in calculations.
Figure 2‐96: Plane strain geometrical idealisation with equivalent homogenized continuum
As shown in Figure 2‐97, in three dimensional analyses, a 3D geometry is used. As in previous
cases, a material constitutive model is still an important constituent part of the analysis, and
will govern the material stress‐strain behaviour in the analysis.
Figure 2‐97: three dimensional geometrical idealisation (Satibi, 2009)
After determining the DR column load and stress, it is also possible to determine its stability
by comparing the load (stress) with the ultimate load (stress) capacity.
Hughes and Withers (1974) assume that in a floating column the vertical stress (load) that is
transferred to the column (through arching) reduces as the depth increases. With the
assumption that the cohesion in the soft layer is constant, and by equating forces on a
horizontal level within the column (Figure 2‐98), they calculated the vertical stress at any
level of the column to be:
137
4 2‐118
σvz = vertical stress in the column at level
σv = vertical stress at the column
z = distance of horizontal level from the top of the column
d = column diameter
ρ = density of the column
c = soft soil cohesion
Figure 2‐98: Vertical stresses on a typical horizontal level in the column (Hughes and Withers, 1974)
Equation 2‐118 can be readily modified for the condition in which the cohesion is not
constant throughout the column length, but is rather equal to ci throughout a length segment
li, and consequently:
4∑
2‐119
Equations 2‐118 and 2‐119 suggest that if the column is long enough, then the stress in the
column would become zero at a point.
As reported by Greenwood (1991) although the form of the results of Hughes and Withers is
valid, full scale tests carried out on single stone column going down to firm ground at the
depth of 11 m in Uskmouth indicated that while in the early stages of loading little stress was
transferred deep into the column (because skin friction against the strong soil crust sustained
stress distributed from the plate load), at higher loadings the stone column sustained higher
138
stress levels at the depth of 1.8 m (where the stress gauge was installed) than those applied
at the surface. A possible explanation was given as stress being redistributed due to
deformation of the soft clay below the crust and hence causing the crust to transfer its weight
to the column by skin friction. Greenwood (1991) also reports a loading test on a group of
columns constructed under an embankment load in East Brent (Somerset) where the effect
of containment on soil between columns enhanced their bulging resistance. Consequently,
the columns transmitted stress to full depth.
If the column is long or resting on a dense layer, then failure can be expected to be by bulging
in the top of the column (or a lower softer layer). On the other hand, if the column is short
and not resting on a dense layer, then it may fail in end bearing (punching).
According to Barksdale and Bachus (1983) the lateral confining stress that constrains a
column is usually taken as the ultimate passive resistance that the soft surrounding soil can
mobilise as the column bulges outward against the soil. Since the column is assumed to be in
a state of failure, the ultimate vertical stress that the column can support will be the product
of the reciprocal of the active pressure (which is the passive pressure) of the column and the
lateral confining stress.
The passive resistance developed by the surrounding soil can be better modelled as an
infinitely long cylinder that expands about the axis of symmetry. The expanding cylindrical
cavity simulates approximately the lateral bulging of the column into the surrounding soil.
Even if the column bulges outward within a length of 2 to 4 column diameters (Barksdale and
Bachus, 1983, Hughes and Withers, 1974), the model of an infinitely long expanding cylinder
appears to give reasonably good results.
Hughes and Withers (1974) idealised the bulge as a cylindrical expansion into the clay, and
concluded that it resembles a pressuremeter test in which a cylinder is expanded against the
side of a borehole. They then assumed the soil to be an elasto‐plastic material, and used the
limiting stress theory of Gibson and Anderson (1961) to calculate the limiting stress:
1 2‐120
2 1
σrL = limiting (lateral) stress
139
σro = total (initial) in‐situ lateral stress
E = elastic modulus
ν = Poission’s ratio
σro can be taken as (Greenwood, 1991):
2‐121
Ks = the ratio of vertical to horizontal stresses in the soil prior to loading and is from 0.6 to 1
in soft clays.
γ = effective unit weight at depth
p = surface load stress on the soil
y = stress distribution factor such as Boussinesq or Westergaard.
Hughes and Withers assess that based on detailed examination of many field records of quick
expansion pressuremeter tests it is possible to approximate Equation 2‐120 to:
4 2‐122
u is the pore pressure.
At the same time if the granular material in the bulged region near the top of the column is
in a critical state of stress then the relationship between the vertical stress, σ’c‐V, and
horizontal stress, σ’c‐H, in the column could be expressed by:
′ ′ 2‐123
Ka is the active earth pressure, and equal to:
1
45° 2‐124
1 2
Hence, Hughes and Withers conclude that:
140
1
′ 4 2‐125
1
As shown in Figure 2‐99, stability can be assessed directly by comparing the horizontal
stresses in the DR column with the restraining strength of the soft soil using the Menard
pressuremeter. The horizontal stress in the DR column must be less than the soil’s strength
for the column not to fail (Varaksin, 1981). In other words the pressuremeter limit pressure,
PLM, must be more than the horizontal stress that is exerted by the DR column; i.e.:
′ 2‐126
or
1
′ 2‐127
1
Figure 2‐99: Stability condition of a DR column against bulging
It is the author’s experience and internal practice to use a safety factor of 2. As an example
for φ= 37o, Ka= 4, and consequently σ’c‐V ≤ 2 PLM.
As shown in Figure 2‐32b, if the column is short and on firm strata, failure may be by shear
(general bearing). Madhav and Vitkar (1978) have developed a solution for the ultimate
bearing capacity of a footing on a granular trench (plane strain case) based on the general
shear failure mechanism using the upper bound theorem based on kinematic considerations,
shown in Figure 2‐100. Expressions for internal energy were based on Chen (1969) due to
cohesion mobilised along various boundaries of failure surfaces, and the work done by the
141
weight of soil, the surcharge, and the external load. Equating the work done by the external
load, qult, to the energies dissipated by cohesion and the work done on account of soil weight
and surcharge, the final equation can be written as:
Figure 2‐100: Granular trench in weak clay (Madhav and Vitkar, 1978)
1
2‐128
2
where
2‐129
2‐130
and
qult = ultimate external pressure on footing width
cs = soil cohesion
γs = soil unit weight
B = footing (loading) width (see Figure 2‐100)
Df = depth of footing (see Figure 2‐100)
ct = trench cohesion
γt = trench unit weight
142
Parameter Range
D/B 0‐2
t 20‐50o
s 0o
0‐1
1
Table 2‐10: Ranges of parameters
Nc, Nγ and Nq are dimensionless factors that can calculated for the best upper bound value;
i.e., the minimum value of the ultimate bearing capacity of the strip footing. The bearing
factors have been evaluated for the ranges listed in Table 2‐10, and are presented in
Figure 2‐101 to Figure 2‐103. In Table 2‐10, φt and φs are respectively the trench and soil
internal friction angles.
Figure 2‐101: Bearing capacity factor Nc (Madhav and Vitkar, 1978)
When the soil and the trench cohesions and internal friction angles are the same the
soil‐trench system reduces to a homogeneous medium, and the values of the bearing
capacity factors become the same as those of Chen (1975). for an axisymmetrical case,
equation 2‐128 must be modified by incorporating shape factors for the three bearing
capacity factors as suggested by Vesic (1975).
143
Figure 2‐102: Bearing capacity factor NΥ (Madhav and Vitkar, 1978)
Figure 2‐103: Bearing capacity factor Nq (Madhav and Vitkar, 1978)
The values of Nc vary significantly with D/B, φt and ct/cs. When D/B= 0, the problem reduces
to the homogeneous case with soft soil having zero internal friction angle, and the value of
Nc coincides with Prandtl’s value of 5.14. For D/B= ∞, should coincide with Prandtl’s
values corresponding to φt. For an intermediate ratio of ct/cs, Nc values can be interpolated.
The ultimate capacity of a punching (local bearing capacity) failure can be determined by
calculating the end bearing capacity of the column using conventional bearing capacity
theories and adding the skin friction load developed along the side of the column (Barksdale
and Bachus, 1983). Barksdale and Bachus suggest that the bearing capacity of an isolated
144
stone column or column located within a group can be expressed in terms of the ultimate
stress applied to the column:
2‐131
qult = ultimate stress that the column can carry
cu = undrained shear strength of the surrounding cohesive soil
Nc = bearing capacity factor for the stone column
Barskdale and Bachus propose Nc is within the range of 18 to 22, but report that others have
suggested higher values. For example, Mitchell (1981) recommends a value of 25 and Dayte
(1982) proposes 40 for uncased rammed stone columns.
2.6.3 Liquefaction
Research (Hausler, 2002, Mitchell et al., 1995) shows that ground improvement techniques
are indeed able to mitigate liquefaction potential; however, commonly used methods for
evaluation of liquefaction potential such as Youd et al (2001) cannot be used immediately
when treatment is by columnar inclusions, such as dynamic replacement, because the ground
is non‐homogeneous, and the properties of the columns and the soil are different.
Baez and Martin (1993) have proposed a method for the assessment of liquefaction potential
based on shear reduction. Baez and Martin do not take the column’s drainage ability into
account because the time for a positive effect of the drainage is limited to the duration of
the earthquake.
The basic assumption in evaluating the distribution of stresses according to the stiffness of
the individual elements is that shear strains for both loose and stiff material are compatible.
Baez and Martin deem that this assumption is valid because there is no inertial loading from
the superstructure directed to the columns that cause displacements in directions other than
that of the ground motion. Therefore, they assume that:
2‐132
or
145
2‐133
γs = shear strain in the soil
γc = shear strain in the column
τs = shear stress in the soil
τc = shear stress in the column
Gs = shear modulus of the soil
Gc = shear modulus of the column
The shear modulus can be calculated from:
2‐134
2 1
To satisfy equilibrium, the input force from the soil’s inertial loading at the given depth must
be equal to the sum of the forces distributed to each of the elements. As per Seed and Idriss
(1971):
2‐135
τave = input (equivalent average) shear stress and is equal to
0.65 2‐136
Also
amax = peak ground acceleration (PGA)
σ = total stress at depth in question
rd = depth factor
Similar to the area replacement ratio (ac) that was defined in Equation 2‐97, the ratio of the
column to soil shear can be defined as:
146
2‐137
The above equations that were used by Baez and Martin can be combined to find the average
shear stress in the column and in the soil:
1 2‐138
1
2‐139
Equation 2‐139 shows that the stresses will be concentrated in the column proportionally to
Gr. Hence, the shear stress in the soil will be less than when there are no columnar inclusions.
The influence of the columns can be taken into account in the commonly used liquefaction
assessment methods by the introduction of the shear stress reduction factor KG:
2‐140
CSR= cyclic shear ratio
CSRi = reduced cyclic shear ratio
2‐141
′
σ'vo= effective initial overburden pressure
KG can be applied to CSR to calculate the reduced cyclic shear ratio.
2‐142
1 1
147
Figure 2‐104 shows the relation between shear stress reduction factor and area replacement
ratio for different values of Gr.
The stone columns are considered to be able to resist liquefaction if cyclic resistance ratio
exceeds cyclic stress ratio; i.e., CRRMw ≥ CSRi (Youd et al., 2001).
Figure 2‐104: Effect of concentrated shear stresses on the CSR (Baez and Martin, 1993)
Goughnour and Pestana (1998) argue that in general, the shear modulus of the column
(stone) is much greater than that of the soil, and the difference between the two increases
with increasing pore pressure generated as a result of seismic loading. They note that if
unconstrained, wave propagation would be much faster in the stone than in the soil, and
since the cyclic excitation is the same, then the frequencies must also be the same. Therefore,
the wave length has to be much shorter for the waves traveling through the soil. In reality,
the columns are physically constrained by the soil, and unless gaps open up between the
columns and the soil, the two must move more or less together. The velocity of the wave
would result from the compatible movement of the soil and the stone column, and a very
complicated stress state would exist, with heavy stress concentrations at the interface of the
columns and soil.
Therefore, Goughnour and Pestana question the compatibility of shear strains between the
column and soil after these conditions, and develop an equivalent shear modulus, Gceq, for
the column in the form of Equation 2‐143, which accounts for the flexural bending.
148
2‐143
2
d= column diameter
λ= wave length in column
Assuming that ν= 0.3, Equation 2‐143, can be written in terms of Gs:
6.4 2‐144
Dynamic replacement columns are at least 1.5 m and typically 2 to 2.5 m in diameter and at
most 6 m long. It can be observed that the maximum height to diameter ratio will be 4, and
thus based on Goughnour and Pestana’s assessment it is possible to use the method of Baez
and Martin (1993) for this ground improvement technique.
149
2.7 Advances in Equipment
Once it was established that there is a direct relationship between depth of improvement
and pounder weight and drop height (Menard and Broise, 1975), the notion of increasing
these two parameters to achieve greater depths of improvement seemed inevitable.
Early on, when Louis Menard performed dynamic compaction at Mandelieu‐la‐Napoule, he
used an 8 ton pounder that was dropped from 10 m (Communication Department of Menard,
2007). Assuming a coefficient of depth of improvement as low as 0.5 would yield a treatment
depth of about 4.5 m, which at that time was quite an accomplishment compared to what
was achievable using vibratory roller compactors. Soon after, Menard was able to identify
heavy duty cranes that were capable of efficiently lifting pounders weighing up to about 15
tons using single cable lines, and dropping the pounder in pseudo free fall. Implementation
of multiple cable lines was not of interest as such lifting processes would proportionally
reduce impact velocity, increase friction, and consequently reduce impact energy, efficiency
and depth of improvement.
Although true free fall without cable connections provided more impact energy, it introduced
practical difficulties. The release of the pounder caused the crane boom to spring back;
thereby ramming the hook block forcefully into the boom, and potentially damaging it. Shock
absorbers such as truck tyres were frequently used to reduce collision forces. Also, it was
observed that pounder lifting point was slightly offset from pounder impact point; therefore,
crane operators had to compensate the offset based on their experience. Furthermore,
reconnecting the pounder to the hook is a costly activity as it is a slow and time consuming
process that also requires manual labour.
Although dropping the pounder in pseudo free fall with a single cable line proved to be more
efficient and practical than dropping the pounder in true free fall, lift capacity of even heavy
duty cranes was limited to about 15 tons. Hence, depth of improvement remained limited to
about 8 to 10 m. Even though this was a great accomplishment in itself, and satisfied the
needs of numerous projects, there were still other sites that required greater treatment
depths.
150
In order to overcome crane lifting limitations, Louis Menard developed, and manufactured
his own rigs. The 700 tm rig, shown in Figure 2‐105, was able to lift 25 ton pounders. The
ability to assemble and disassemble this rig facilitated its transportation, and made it an
attractive choice for overseas projects. The mega‐machine was capable of delivering 1,600
tm of energy per impact. This machine, shown in Figure 2‐106 was used for the ground
improvement works of Bern to Biel Highway in Switzerland (Communication Department of
Menard, 2007).
Figure 2‐105: Dynamic compaction using the 700 tm rig
Figure 2‐106: Dynamic compaction using 1,600 tm mega machine (Communication Department of
Menard, 2007)
151
The tripod, shown in Figure 2‐107, was an extremely light weight structure that was capable
of delivering 1,600 tm of energy per impact by dropping a 40 ton pounder from 40 m. Similar
to the 700 tm machine, this rig could also be reassembled, and was successfully used in
several countries including the United States, Japan, Dominican Republic, Bangladesh and
Mexico.
Figure 2‐107: Dynamic compaction using the Menard tripod
Figure 2‐108: Dynamic compaction using the giga‐machine
152
Undoubtedly, the giga‐machine, shown in Figure 2‐108, is the most spectacular rig that has
ever been built. This rig had 168 wheels, 7 km of hydraulic hoses, and was specifically built
to treat 20,000,000 m3 of rip rap fill in Nice Airport (Communication Department of Menard,
2007). Impact was provided by dropping pounders weighing 130 to 170 tons from 22 m
(Gambin, 1983). The giga‐machine was decommissioned after the project.
As much as the 700 tm rig, the tripod, the mega‐machine and giga‐machine had their special
applications, they were specifically produced, their numbers were limited and they could not
be manufactured commercially or in great numbers. However, the introduction of a new
generation of high production heavy duty cranes that were able to lift pounders using two
single cable lines that were connected to two hoist drums increased lift capacity from 15 to
25 tons. One of such rigs is shown in Figure 2‐109.
Figure 2‐109: Dynamic compaction using 25 ton pounder
Other than Menard rigs, Lampson also modified a crane that was rented for performing
dynamic compaction using a 30 ton pounder that was dropped from 30 m for the Thermal
Public School project in Palm Spring, USA. This rig is shown in Figure 2‐110. In China, dynamic
compaction has been performed using cranes that have been modified by adding two
additional legs to the boom and turning it into a tripod (Varaksin, 2014).
2.7.2 Pounders
153
(Figure 2‐113); however, it is the personal experience of the author that concrete pounders
are short lived and steel plate pounders can be used over and over again with minimum
maintenance. Dynamic replacement pounders have smaller bases to increase shearing and
punching into the ground.
Figure 2‐110: Lampson crane modified by dynamic compaction
Figure 2‐111: Dynamic compaction in China using cranes that have been modified into tripods
154
Figure 2‐112: Reinforced concrete pounder encased in a steel box
Figure 2‐113: steel pounders composed of steel plates
While studies have recently been conducted on the efficiency of conical bottom pounders
(Arslan et al., 2007, Feng et al., 2000, Feng and Ke, 2005) almost all pounders that are used
in the industry have flat bottoms. As shown in Figure 2‐114, marine dynamic compaction
pounders appear as large graters to decrease water resistance.
More recently, an innovative and patented pounder release and automatic grab system
called MARS has been developed, and will be further described in the case studies presented
in this thesis.
155
Figure 2‐114: Marine dynamic compaction pounder used for seabed improvement at Kuwait Naval
Base (Communication Department of Menard, 2007)
Dynamic compaction is neither the only ground improvement technique that utilises impact
energy to compact the soil nor is it a generic name for soil improvement methods that are
based on impact energy. Dynamic compaction is a term that was chosen by Louis Menard to
specifically define the technology that he had invented. The author notes that in recent years,
as a marketing tool, there is a trend by some contractors to utilise the term dynamic
compaction on its own or with a prefix or suffix to express an alternative impact oriented
ground improvement method. Although this may win those contractors some projects, in the
opinion of the author it will and has brought confusion; thus the author strongly recommends
that terminology be used properly. Rapid impact compaction and impact roller compaction
are two impact oriented techniques that will be considered hereunder.
According to BSP’s website (British Steel Piling (BSP)), the Rapid Impact Compactor or RIC was
developed in the early 1990s in conjunction with the British Military as a means of quickly
repairing damaged aircraft runways. As shown in Figure 2‐115, dynamic energy is imparted
by a weight dropping from a controlled height onto a steel foot that has a diameter of about
1.5 m. Energy is transferred to the ground safely and efficiently as the RIC's foot remains in
contact with the ground and no flying debris is ejected.
156
Compared to dynamic compaction, instead of dropping a heavy weight from a great height
once or twice a minute, the RIC drops a lighter weight, typically weighing 5 to 9 tons, from a
relatively lower height of up to 1.2 m at a rate of 40 to 60 impacts per minute.
Figure 2‐115: Rapid Impact Compactor
The primary usage of RIC in the UK is for shallow compaction of floor slab and roadway
subgrades (Terra Systems). However, the introduction of this technology to other countries
such as the United States and Australia has been associated with a degree of confusion as
the technique is sometimes incorrectly referred to by terms such as dynamic compaction,
light dynamic compaction, controlled dynamic compaction or rapid dynamic compaction.
Disregarding the energy loss due to having to overcome the inertia of the plate resting on
the ground and by applying Equation 2‐41, it can be observed that depth of improvement for
a 9 ton weight that is dropped from 1.2 m will be respectively 2.3 m and 3 m for coefficients
of 0.7 and 0.9. Application of Equation 2‐51 with a pounder radius of 0.75 will not make much
of a difference and will result in the depth of improvement being 3.4 m. It can be observed
that while larger depths of improvement have been reported, the magnitudes of
improvement are considerably less than dynamic compaction, but still quite compatible with
the original purpose of developing the RIC technology, and the improved depths should
generally satisfy FAA (1995) requirements for the compaction of subgrade soil in airports.
Classical circular roller compactors are capable of compacting approximately 0.2 to 0.3 m of
graded soil, but treatment thickness may sometimes be substantially more and a
157
considerable amount of time, money and carbon emissions will be needed to treat the soil
in multiple layers.
The idea of increasing the treatment thickness by implementing non‐cylindrical multi‐sided
geometrical rollers was first recognised and patented in 1935 in Sweden, but it was only 20
years later and in South Africa when impact roller compaction was developed (Avalle, 2004).
Today, impact roller compactors, shown in Figure 2‐116, that are sometimes also called
impact compactors weigh 8 to 16 tons, have 3 to 5 sides, and operate at practical speeds in
the range of 8 to 11km/hr depending on the ground conditions. When travelling speed
exceeds 11km/hr the drum begins to skip as the drum revolution increases (Geoquip). The
drum tends to act with less impact and eventually will run as a circular wheel given sufficient
revolutions per minute.
Figure 2‐116: Impact roller compaction
Impact rollers do not have a motorised form of energy such as the vibratory roller, and derive
their energy by turning on their corner and falling onto the flat side. The number of sides of
an impact roller affects its energy rating. Since an impact roller turns on its corner and falls
onto its flat side, the greater the difference of these radii the greater the lift height. Three
sided rollers produce the maximum lift while five sided rollers produce the least lift. Lifts are
from 0.15 to 0.23 m (Broons Hire, Geoquip).
Impact roller compaction is also sometimes incorrectly referred to as dynamic compaction,
rolling dynamic compaction or even worse, high impact energy dynamic compaction. On its
website, Landpac (Typical Myths About Impact Compaction) notes that:
“… Impact Compaction (IC) and Dynamic Compaction (DC) have some very different
characteristics that make them distinct from one another. DC is performed through a very
large parcel of energy being imparted over a very short time duration by means of a heavy
158
pounder dropped from a great height onto the soil. The repetition of such compaction cycles
is performed at a very low frequency. IC on the other hand is performed through a far smaller
parcel of energy being imparted to the soil over far greater time durations by means of an
eccentrically shaped weight rolling over the soil. The repetition of these compaction cycles is
performed at a far greater frequency that DC.”
Even though impact roller compaction is not dynamic compaction, for comparative purposes
it can be observed that by applying Equation 2‐41 the depth of improvement for an 11 ton
weight dropped from 0.23 m will be respectively 1.1 m and about 1.4 m with coefficients of
0.7 and 0.9. Based on the tests that they had performed, Berry et al. (2004) have proposed
an alternative method for predicting the profile of improvement in unsaturated soils based
on the measurement of ground settlements and a distribution of plastic strains similar to an
adapted Schmertmann strain influence diagram. According to this study the improvement
depth is 2 to 3 times the compactor’s width (900 mm) and the maximum improvement is
achieved at about 0.67 to 1 times the compactor’s width. This will yield an improvement
depth of 1.8 to 2.7 m. Kelly and Gil (2012) also note that ground improvement by impact
roller compaction is typically measured to effective depths of 2 to 3 m (see Figure 2‐117) .
The schematic comparison of depth of treatment between impact roller compaction, rapid
impact compaction and dynamic compaction is shown in Figure 2‐118.
Figure 2‐117: Impact roller compaction depth of improvement (Kelly and Gil, 2012)
159
Figure 2‐118: Schematic comparison of depth of improvement between impact roller compaction,
RIC and DC
160
2.8 Assessment of Dynamic Vibrations
Dynamic compaction's concept is based on generation of waves with sufficiently large
enough energy (and thus magnitude) to be able to re‐arrange the soil in a denser state;
hence, it is of utmost importance to be able to understand safe and acceptable vibration
limits. It has come to the attention of the author that in practice many engineers are
insufficiently familiar with how to properly measure and interpret vibrations. It is the
objective of this section to clarify how safe vibration limits were developed, how to predict
vibration magnitudes, how to properly measure the vibrations, how to compare the
measurements with allowed tolerance, and how to potentially reduce vibration effect on
nearby structures.
From 1930 to 1942 the US Bureaus of Mines (USBM) conducted an extensive research
program to study the seismic effects of quarry blasting on buildings (Nicholls et al., 1971).
Vibration amplitudes ranged from 25 microns to 12.7 mm, frequencies were from 4 to 40 Hz,
and three classifications of damage were proposed based on the degree of plaster failure.
The damage indices were:
No damage
Minor damage: fine plaster cracks, opening of old cracks
Major damage: fall of plaster, serious cracking
There was no sharp distinction between the damage classes. Also, as a note, many other
factors such as ageing, settling, shrinkage, etc. can result in similar damage; hence, it was and
is important to perform a dilapidation survey to assess the structures' soundness prior to
generating vibrations.
USBM’s report (Thoenen and Windes, 1942) of the tests recommended an index of damage
based on acceleration. If accelerations were less than 0.1g, no damage was expected. Minor
damage was expected for accelerations ranging from 0.1g to 1 g, and major damage was
expected for values greater than 1 g. Duvall and Fogelson (1962) statistically showed that
these data gave contradictory results because while minor damage correlated with
acceleration, major damage correlated with particle velocity. Hence, the concept of
161
implementing particle velocity in lieu of acceleration for prediction of (major) damage was
formed.
Langefors, Kilhstrom and Westerberg (1958) also carried out extensive studies to relate
between damage and ground vibrations from nearby blasting, and developed a four degree
damage classification for plaster failure based on particle velocity as follows:
No noticeable damage: 71 mm/s
Fine cracking and fall of plaster: 109 mm/s
Cracking: 160 mm/s
Serious cracking: 231 mm/s
Statistical analyses of these data showed that the degree of both major and minor damage
correlated with particle velocity.
Similarly, Edwards and Northwood (1960) carried out a number of tests and classified
damage as follows:
Threshold: opening of old cracks and formation of new plastic cracks
Minor: Superficial, not affecting the strength of the structure
Major: resulting in serious weakening of the structure
They also concluded that damage was more closely related to particle velocity than to
displacement or acceleration, that damage was likely to occur with a particle velocity of 100
to 125 mm/s, and recommended a safe vibration limit of 50 mm/s.
Figure 2‐119 shows the results of the three tests carried out by USBM, Duvall & Fogelson and
Edwards and Northwood. It can be observed that the regression line with a slope of ‐1 on the
log displacement versus log frequency diagram, which represents a constant particle velocity,
provides a better damage correlation than the 0 and ‐2 slope lines that respectively represent
constant displacement and constant acceleration. The best regression for particle velocity
for predicting major damage, minor damage and safe limit are respectively 193 mm/s, 137
mm/s and 50 mm/s. As shown in Figure 2‐120, the same data can be plotted for the logarithm
of particle velocity versus logarithm of frequency.
162
Figure 2‐119: Displacement versus frequency, combined date with recommended vibration criterion
(Nicholls et al., 1971)
Figure 2‐120: Particle velocity versus frequency with recommended safe limit criterion (Nicholls et
al., 1971)
163
The recommended safe vibration criterion of 50 mm/s particle velocity was a probability type
criterion, and if the observed particle velocity exceeded the safe limit in any of the three
orthogonal components there was a reasonable chance that residential structures could be
damaged. The safe vibration criterion was not a value below which damage would not occur,
and above which it would occur. Many structures could experience vibration levels greatly in
excess of 50 mm/s with no observable damage; however, the probability of damage to a
residential structure would increase or decrease as the vibration level deviates from 50
mm/s.
USBM (Nicholls et al., 1971) recommended that velocity gages should be preferably mounted
on or in the ground rather than the structure because most of the data used in establishing
the damage criterion were obtained in that manner. Mounting the gages in the ground was
understood to alleviate the necessity of considering the response of a large variety of
structures. Also, particle velocity was (and is) be observed in three mutually perpendicular
directions. A safe vibration criterion was based on the measurement of individual
components, and if the peak particle velocity (PPV) of any component exceeded 50 mm/s,
then damage was likely to occur.
Collection of large amounts of data will result in considerable amount of data about the mean
PPV curve; thus most regulations require that without instrumentation maximum probable
velocities be used rather than average values (Dowding, 2000).
According to Siskind et al. (1980), Pennsylvania was the first American state to adopt the 50
mm/s PPV criterion as a safe standard in 1957; however, in 1974 the State was forced to
adopt stricter controls because of citizen pressure and lawsuits involving both annoyance
and alleged damage to residences. Consequently in 1974, USBM began to reanalyse the blast
damage problem, expand the study of Duvall and Fogelson (1962), and overcome its more
serious shortcomings. Part of the new study included emphasis on the frequency dependency
of structure response and damage, and recognising that the response characteristics and
frequency content of the vibrations are critical to response levels and damage probabilities.
Also, an analysis was made of various studies of human tolerance to vibrations, although
most data were from steady state rather than impulsive sources (Siskind et al., 1980).
The measured response of residential structures was a critical indicator of troublesome or
potentially damaging ground vibrations, and corner motion measurements were used to
assess the racking (shearing) motions of the gross structure (see Figure 2‐121). Cracking from
164
vibration can occur where excessive stresses and strains are produced within the planes of
the walls or between walls at the corners. Consequently, the vibration in the corners is
assumed to indicate cracking potential because it corresponds to whole structure response.
Other types of response cause different but consequential results. Mid‐wall motions normal
to the wall surface that are primarily responsible for window sashes rattling, picture frames
tilting and dishes giggling were also measured. Even though structures are designed to resist
normal vertical loads, differential vertical motions can produce high strains in floors and
ceilings.
Figure 2‐121: Types of motion that can cause damage to a building (Siskind et al., 1980)
Although the study recognised that a simple measurement of peak particle velocity was an
oversimplification, it concluded and recommended the continuation of utilisation of peak
particle velocity as the primary measure of ground motion to assess damage.
A simple amplification factor was determined directly from the vibration time histories, and
maximum structure velocities and their times of occurrence were noted. Ground velocities
and frequencies were then picked off the records at the corresponding moments of time or
immediately preceding the time of peak structure vibrations. The ratios of the two velocities
were plotted against the frequency of the corresponding ground motion, and as expected
from the natural resonance frequencies, maximum amplifications were found to be
associated with ground motions between 5 to 12 Hz. The highest amplification factors for
corner motion were approximately 4, with 1.5 being a typical value. Ground motions above
45 Hz produced little or no amplification of the corner measured structure motion.
With many responses occurring at lower frequencies, particularly up to 25 Hz, the maximum
mid‐wall motion amplification factors were found to be greater than for the corners. As with
corner motions, amplification factors for ground motions above 45 Hz were less than unity.
These results suggested that frequencies below 10 Hz are most serious for potential damage
from structure racking. Vibrations with frequencies of less than about 25 Hz can excite high
165
levels of mid‐wall motion, and generate most of the secondary noises, rattling and other
annoyances.
It was understood that for structures not exceeding two stories, constructed on firm
foundation, with dimensions of typical residences and with vibration durations not longer
than a few seconds, the safe vibration level defined as levels unlikely to produce interior
cracking or other damages in residences was determined as presented in Table 2‐11 (Siskind
et al., 1980).
Ground vibration – peak particle
velocity mm/s
Type of structure
At low frequency At high frequency
(<40 Hz) (≥40 Hz)
Modern homes, dry wall interiors 19 50
Older homes, plaster on wood lath
12.7 50
construction for interior walls
Table 2‐11: Safe levels of blasting vibrations for residential type structures, from USBM RI 8507
(Siskind et al., 1980)
Dry walls defined as gypsum wallboards that consist of 10 mm to 15 mm gypsum plaster
panels with paper laminates covering both sides of the wall appeared to be more capable of
withstanding vibrations because the 0.4 mm thick paper contributed greatly to the strength
of the board, and concealed cracking of the plaster core.
Older homes often had interior walls of thick plaster over wood lath support, which was more
susceptible to damage. A minimum safe level of 12.7 mm/s was adopted based on probability
analyses of low frequency shots and the overall study with an assumption of 5% probability
for very superficial cracking. However, this vibration level was also lower than the lowest
level in cases where damage was observed. The 12.7 mm/s level should provide protection
from damage in more than 95% of the cases.
Safe vibration criteria were developed for residential structures, having two frequency
ranges with a sharp discontinuity at 40 Hz, and an intermediate frequency case that was
higher than the structures' resonances (4 to 12 Hz) and lower than 40 Hz. By using a
combination of measured structure amplification and damage summaries, as shown in
Figure 2‐122, a smooth set of criteria were developed (Siskind et al., 1980).
166
Estimation of the predominant frequency was still a problem. Where the wave train is simple,
the period corresponding to the peak level can be directly measured. Otherwise, spectral
analysis will be required (According to Siskind et al. (1980), for frequencies less than 40 Hz,
all spectral peaks within 6 dB (50%) amplitude of the predominant frequency must be
analysed). Occasionally the peak level occurs early in the wave and at a high frequency, with
a long duration wave train of somewhat lesser amplitude following. In the opinion of Siskind
et al. the safest approach was to consider the low frequency part of the time history
separately, and where it was below 40 Hz, to use the 19 mm/s or 12.7 mm/s criteria of
Figure 2‐122.
Figure 2‐122: USBM RI 8507 safe vibration limit criteria (Siskind et al., 1980)
Three years after the publication of USBM RI 8507 (Siskind et al., 1980) the Office of Surface
Mining Reclamation Enforcement (OSM) published its regulation regarding safe vibration
levels (Office of Surface Mining Reclamation Enforcement, 1983). Disregarding the clause
that refers to the scaled distance equation for determining the weight of the explosion
charge, the regulation states that at the location of any dwelling, public building, school,
church or community or institutional building the maximum peak particle velocity shall be
less than the values stated in Table 2‐12 and Figure 2‐123. It can be observed that OSM’s
regulation is in fact a modification of USBM RI 8507’s criteria.
167
Distance from vibration Maximum allowable peak particle
source (m) velocity for ground vibration (mm/s)*
0 to 90 31.75
91 to 1,500 25.4
1,501 and beyond 19
*: Ground vibration shall be measured as particle velocity. Particle velocity
shall be recorded in three mutually perpendicular directions. The
maximum allowable peak particle velocity shall apply to each of the three
measurements.
Table 2‐12: Maximum allowable peak particle velocity for ground vibration (OSM, 1983)
Figure 2‐123: OSM Safe vibration level criteria (Office of Surface Mining Reclamation Enforcement,
1983)
The separate and independent study of vibration damage from blasting, pile driving, and
machine sources by Studer and Suesstrunk (1981) resulted in similar criteria where limiting
particle velocities depend upon frequency, and justified the use of USBM or OSM damage
criteria for dynamic compaction operations.
Nicholls et al. (1971) cite that Windes (1942, 1943) has carried out an investigation for USBM,
and has concluded that window glass failure occurs before any other type of structural failure
due to air blast. In Windes’ investigation some window panes broke by an overpressure of 7
kPa. The condition of the glass in the windows contributed directly to the damage
168
experience. Poorly mounted panes, which have been prestressed by improperly inserted
glazier’s points or other causes, may fail when subjected to overpressures as low as 0.7 kPa.
Based on the research for USBM, Windes concluded that under normal blasting conditions
the problem of damage from air blast was insignificant.
Edwards and Woods (1960) also measured the air blast pressure during their vibration
studies. The measured overpressure ranged from 0.07 to 1.4 kPa at locations outside the
structures that were being monitored. These pressures were considerably below the levels
expected to cause damage, and none of the damage that occurred in any of the structures
was attributed to air blast.
Air vibration, even in blasting operations, is not considered to be a significant factor in
causing damage to residential structures in most operations; however, it can be a nuisance
to humans.
Humans notice and react to vibration at levels that are lower than the damage threshold.
Vibration levels that are completely safe for structures by all standards can be quite
unpleasant when viewed subjectively by people.
Reiher and Meister (1931) studied the subjective human tolerance to 5 minute vertical and
horizontal vibratory motions in a variety of positions. Responses to slightly perceptible
vibrations occurred at 0.25 mm to 0.84 mm/s, and the threshold of strong perceptible
vibrations was 2.55 mm/s. It was noted that over the range of 4 to 25 Hz, the results were
essentially independent of frequency.
Nicholls et al. (1971) adapted Figure 2‐124 from Goldman (1948) to show the subjective
response of the human body to steady state vibratory motion in the frequency range of 2 to
50 Hz. These limits are based on the results for sinusoidal vibrations.
Although most studies of human tolerance to vibrations have been conducted on steady
state sources or those of relatively longer duration than a transient vibration (as what would
be the case for dynamic compaction), it can be expected that the vibration limits required for
a reasonable comfort from a long term vibration source would be more restrictive than for
sources of short duration at infrequent occurrence. Several researchers have recognised that
the duration of the vibration was critical to its undesirability (Siskind et al., 1980). Most
169
evidently, a higher level could be tolerated if the event was short. Consequently, steady state
vibration data cannot be comprehensively applied to vibrations originating from dynamic
compaction.
Figure 2‐124: Subjective response of the human body to steady state vibratory motion (Nicholls et
al., 1971)
As shown in Figure 2‐125, Power (1966) plotted the percentage of complaints versus particle
velocity for the Salmon underground nuclear detonation which was technically a transient
vibration, but lasted 90 seconds. More than 35% of the families located in the zone where 50
mm/s was exceeded filed complaints. This particle velocity is in the intolerable subjective
response zone, and complaints could have been anticipated. In the perceptible zone, less
than 8% of the families complained. This data indicates that a vibration level of 10 mm/s
should not be exceeded if complaints and claims are to be kept below 8%.
Chang (1973) analysed human vibration response literature with particular attention to event
durations, and concluded that Reiher and Meister’s responses could be multiplied by a factor
of 10 for short events.
170
approximately 200 msec duration. The disturbing level mean was 89 to 112 mm/s or over 5
times higher than Goldman’s steady state intolerable level of 19.5 mm/s at 20 Hz.
Figure 2‐125: Complaint history of Salmon underground nuclear detonation with superposed
subjective response (Power, 1966)
As shown in Figure 2‐126, Wiss (1981) has also made a comparison of human response to
transient vibration (Wiss and Parmelee, 1974) and steady state vibration (Reiher and Meister,
1931).
Figure 2‐126: Comparison of Human Response to steady state (Reiher and Meister, 1931) and
transient vibration (Wiss and Parmelee, 1974)
171
2.8.4 Standards
BS 7385‐2:1993 (British Standards Institution, 1993), which is still current, clearly defines
peak particle velocity as the maximum value of any one of three orthogonal component
particle velocities measured during a given time interval. This definition is very important as
it will be seen that sometimes PPV is conservatively redefined using true vector sum (TVS),
as formulated in Equation 2‐146, or erroneously redefined using pseudo vector sum (PVS),
presented in Equation 2‐147.
BS 7385‐2:1993 specifies that measurements should be taken at the base of the building on
the side of the building facing the source of vibration, to define the vibration input to the
building. Where this is not feasible, the measurement should be obtained on the ground,
outside of the building. BS 5228‐2:2009 (British Standards Institution, 2009) further explains
the measuring positions, and specifies that when the purpose of vibration measurement is
to assess the possibility of structural damage, the preferred primary position of monitoring
is in the lowest storey of the building, either on the foundation of the outer wall, in the outer
wall, or in recesses in the outer wall. For buildings having no basement, the point of
measurement should be not more than 0.5 m above ground level. For buildings with more
than one storey, vibration might be amplified within the building. In the case of horizontal
vibration, such amplification might be in proportion to the height of the building, whereas
vertical vibration tends to increase away from walls, towards the mid‐point of suspended
floors. Vibration measurements may also be taken on structures to provide information on
the coupling between the soil and the foundations and amplification effects within a building.
When the building is higher than four floors (approximately 12 m), additional measuring
points should be added every four floors and at the top of the building. Measurements should
be made on the side of the building facing the source. When the purpose of vibration
measurement is to evaluate human exposure to vibration in the building, measurements
should be taken on the structural surface supporting the human body.
BS 5228‐2:2009 also recommends to measure in terms of PPV if the risk of damage to the
building is the primary concern, and there is also an interest in human reaction; however, if
the concern is purely for human tolerance, then weighted acceleration is the preferred
parameter that should be measured. The PPV can be measured in three orthogonal axes at
a point on the ground or inside a property.
172
BS 7385‐1:1990 (British Standards Institution, 1990) and BS ISO 4866:2010 (British Standards
Institution, 2010) that has now replaced BS 7385‐1:1990 give guidance on the measurement
and evaluation of vibration in buildings in the frequency range of 4 Hz to 250 Hz, and define
three categories of vibration‐induced damage (cosmetic, minor and major).
Referring to Siskind et al. (1980), BS 7385‐2:1993 suggest that the probability of damage
tends towards zero at 12.5 mm/s peak component particle velocity, and provides limits for
transient vibration, above which cosmetic damage could occur as given numerically in
Table 2‐13 and graphically in Figure 2‐127. In the lower frequency region where strains
associated with a given vibration velocity magnitude are higher, the guide values for the
building types corresponding to Line 2 are reduced. Below a frequency of 4 Hz, where a high
displacement is associated with a relatively low peak component particle velocity value a
maximum displacement of 0.6 mm (zero to peak) should be used. Minor damage is possible
at vibration magnitudes that are greater than twice those given in Table 2‐13, and major
damage to a building structure may occur at values greater than four times the tabulated
values. This document suggests that where the dynamic loading caused by continuous
vibration is such as to give rise to dynamic magnification due to resonance, especially at the
lower frequencies where lower guide values apply, then the guide values in Table 2‐13 may
need to be reduced by up to 50 %.
BS 6472:1992 (British Standards Institution, 1992), now replaced by BS 6472‐1:2008 (that
was not available to the author at the time of compilation of this thesis), provides guidance
on potential disturbance to persons exposed to vertical building vibration in the frequency
range most applicable to excavation equipment; i.e., 8 to 80 Hz. The levels of vibration that
may be expected to give rise to adverse comment from some people are shown in Table 2‐14.
These values should be regarded as conservative criteria for assessing loss of amenity,
particularly for relatively short term exposure to emissions from construction or excavation
projects.
In residences, adverse comment may arise during daytime as a result of continuous floor
vibration levels in the order of 0.3 to 0.6 mm/s. At night time the residential criterion is
equivalent to a vertical peak vibration level of only 0.2 mm/s. This latter level is barely above
the threshold of human perception.
173
4 to 15 Hz 15 Hz and above
Reinforced or framed structures. 50 mm/s at 4 Hz and
1 Industrial and heavy commercial above
buildings
Unreinforced or light framed 15 mm/s at 4 Hz 20 mm/s at 15 Hz
commercial type buildings increasing to 20 increasing to 50
2
mm/s at 15 Hz mm/s at 40 Hz and
above
Note 1: Values referred to are at the base of the building.
Note 2: For Line 2, at frequencies below 4 Hz, a maximum displacement of 0.6 mm (zero to peak)
should not be exceeded.
Table 2‐13: Transient vibration guide for cosmetic damage (British Standards Institution, 1993)
In offices, vertical vibration levels in the order of 0.6 mm/s are usually acceptable. The
occupants of workshops are likely to be more tolerant, and would normally accept vertical
vibration levels in the order of 1.2 mm/s without adverse comment.
Figure 2‐127: Transient vibration guide for cosmetic damage(British Standards Institution, 1993)
However, as Section 4.1 of BS 6472:1992 states, situations can exist where vibration
magnitudes above those generally corresponding to minimal adverse comment level can be
tolerated, particularly for temporary disturbances and infrequent and intermittent events,
such as those associated with construction projects. While this document provides
reasonable human comfort design goals for long term vibration, significantly higher levels of
short term vibration can be tolerated by many people for construction projects.
174
Vibration levels in mm/s over the frequency range 8
to 80 Hz likely to cause adverse comment
Time Intermittent vibration &
Type of space occupancy of impulsive vibration
Continuous vibration
Day excitation with several
occurrences per day
Vertical Horizontal Vertical Horizontal
Critical working areas
(e.g. some hospital
Day 0.14 0.4 0.14 0.4
operating theatres,
Night 0.14 0.4 0.14 0.4
some precision
laboratories, etc.)
Day 0.3 to 0.6 0.8 to 1.6 8.4 to 12.6 24 to 36
Residential
Night 0.2 0.6 2.8 8
Day 0.6 1.6 18 51
Offices
Night 0.6 1.6 18 51
Day 1.2 3.2 18 51
Workshops
Night 1.2 3.2 18 51
Table 2‐14: Maximum BS 6472 Curve Numbers ‐ Human Comfort (8 to 80 Hz)
Hiller and Hope (1998) have compared the different sections of British Standards in
Figure 2‐128.
While AS 2187.2 (Standards Australia, 2006) recognises both USBM RI8507 and BS 7385‐2,
which are frequency dependant. This Standard also anticipates regulations for users who do
not have the facilities to use frequency dependant assessment method. In such conditions
the Standard recommends ground vibration limits for control of damage to structures and
human comfort to be as shown in Table 2‐15 and Table 2‐16.
175
Figure 2‐128: Summary guidance on vibration criteria given in British Standards. Damage thresholds
are those for domestic buildings (x‐axis is front‐to‐back, y‐axis is side‐to‐side, z‐axis is head‐to‐toe)
(Hiller and Hope, 1998)
Table 2‐15 and Table 2‐16 do not cover high rise buildings, buildings with long span floors,
specialist structures such as reservoirs, dams and hospitals, or buildings housing scientific
equipment sensitive to vibration. These require special considerations, which may
necessitate taking additional measurements on the structure itself, to detect any
magnification of ground vibrations that might occur within the structure. Particular attention
should be given to the response of suspended floors.
Category Peak component particle velocity (mm/s)
Other structures or architectural Table 2 (BS) and referral to USBM RI 8507 and
elements that include masonry, BS 7385‐2
plasterboard in their construction
Unoccupied structures of reinforced 100 mm/s maximum unless agreement is
concrete or steel construction reached with the owner that a higher limit may
apply
Service structures, such as pipelines Limit to be determined by structural design
and cables methodology
Table 2‐15: Recommended ground vibration limits for control of damage (AS 2187.2)
176
Category Type of operation Peak component particle velocity
(mm/s)
Sensitive site* Operations lasting 5 mm/s for 95% blasts per year
longer than 12 10 mm/s maximum unless agreement is
months reached with the occupier that a higher
limit may apply
Sensitive site* Operations lasting 10 mm/s maximum unless agreement is
for less than 12 reached with occupier that a higher limit
months may apply
Occupied non‐sensitive All 25 mm/s maximum unless agreement is
sites, such as factories reached with occupier that a higher limit
and commercial may apply. For sites containing
premises equipment sensitive to vibration, the
vibration should be kept below
manufacturer’s specifications or levels
that can be shown to adversely affect
the equipment operation.
* A sensitive site includes houses and low rise residential buildings, theatres, schools and other similar
buildings occupied by people.
Table 2‐16: Ground vibration limits for human comfort chosen by some regulatory authorities (AS
2187.2)
It can be understood that the codes of practice and standard guidance and recommendations
originate mostly from USBM research (Duvall and Fogelson, 1962, Nicholls et al., 1971,
Siskind et al., 1980, Thoenen and Windes, 1942, Windes, 1942, Windes, 1943) that have
targeted low rise buildings and limits of human annoyance, but that do not cover all
structures and infrastructure. In the absence of guidelines engineering judgement and
experience may become very helpful, and with knowledge that code limits are probabilistic,
sometimes it may be acceptable to exceed code recommendations (that have been
developed for low rise structures). For example, pipelines may be more tolerant to vibrations.
Wiss (1981) indicates that particle velocities of 76 mm/s have not damaged pipes and mains.
He also indicates that high pressure pipelines have withstood particle velocities in the range
177
of 254 to 508 mm/s without experiencing any distress as apparent from dynamic strain gage
measurements.
At the same time it should be noted that at close distances, it may be the permanent physical
movement or ground displacement that leads to damage rather than vibration and particle
velocity. Both field studies (Lukas, 1986), see Figure 2‐63, experimental studies (Hajialilue‐
Bonab and Rezaei, 2009), see Figure 2‐64, and numerical analysis (Poran and Rodriguez,
1992), see Figure 2‐66, have shown that the soil around the pounder impact point undergoes
displacement, and even if a nearby structure or infrastructure is not sensitive to vibration,
ground distortion may become a source of concern. Lukas (1986) notes that as the pounder
strikes the ground a significant portion of the movement is downward, but lateral
displacements occur, and have been reported in literature.
Based on the fact that dynamic compaction will generate high vibration amplitudes and low
vibration frequencies, the US Naval Facilities Engineering Command, NAVFAC, (1997) has
necessitated to maintain the minimum distances to adjacent facilities as follows:
Piles and bridge abutment 4.5 to 6 m
Liquid storage tanks 9 m
Reinforced concrete building 15 m
Dwellings 30 m
Computers (not isolated) 90 m
The author would personally exert extreme caution for any dynamic compaction activity
closer than 10 m, but has personal experience of implementing dynamic compaction at
distances much closer than 90 m to computers, and suspects that NAVFAC may be referring
to main frame computers used in the 1980s that were probably more sensitive to vibration
that the typical computer that is used nowadays.
Field observations indicate that the ground vibration caused by dynamic compaction is in the
range of 2 to 20 Hz (Mayne, 1985, Mitchell, 1981). Van Impe (1989) has mentioned 5 to 12
Hz, and Romana Ruiz and Jurado (1999) have measured vibration frequencies to be 4 to 18
Hz. Hwang and Tu (2003, 2006) have carried out vibration monitoring using a 25 ton pounder
178
that was dropped from 5, 10, 15 and 20 m. Figure 2‐129 shows the vertical, radial and
tangential particle velocity histories at different distances. It can be observed that while the
vertical and radial particle velocity amplitudes are distinctively higher than the tangential
amplitude, the tangential wave form is more complicated. The high frequency peaks in the
front part of the wave trains attenuate quickly with distances in all three directions, but the
low frequency peaks in the rear part of the wave trains attenuate slowly with distance.
Figure 2‐129: Vertical, radial and tangential particle velocities’ histories at different distances
(Hwang and Tu, 2003, 2006)
Figure 2‐130 shows the Fourier amplitude spectra of the vertical, radial and tangential
velocities at various distances from the pounder impact point. The spectra’s shapes at
different distances are similar in the same direction. The spectrum at the distance that is
closer to the impact point has a more noticeable predominant frequency band and a
corresponding peak than the farther distance. Hwang and Tu reason that this is due to energy
focusing at near distance and multiple refractions and reflections of stress waves in between
the soil layers. The primary frequency of vertical vibration was in the range of 10 to 20 Hz,
and the radial vibration had two primary frequencies, from 3 to 4 Hz and from 12 to 13 Hz.
The entire vibration frequency contents were primarily in the range of 0 to 40 Hz for all three
directions.
179
Figure 2‐130: Fourier amplitude spectra of the vertical, radial and tangential velocities at various
distances from the pounder impact point (Hwang and Tu, 2003, 2006)
Hwang and Tu (2003, 2006) did not propose an equation to estimate PPV (referred to as
peak ground velocity in their paper), but have studied the attenuation relationship of PPV
with distance in all three directions for nine consecutive impacts. As shown in Figure 2‐131,
they observed that the attenuation relationships were very close and vibration characteristic
of dynamic compaction seemed to be reliably reproducible.
It is interesting to note that unlike the measurements of Romana Ruiz and Jurado (1999),
Hwang and Tu’s measurements do not indicate that the number of blows on a location will
significantly change PPV. The personal experience of the author suggests that PPV increases
180
as the ground becomes denser. Mayne (1985) has also observed that vibration levels increase
as the treated area is compacted, and notes that a maximum level of particle velocity is
achieved generally after one or two phases of heavy tamping or about 150 tm/m2.
It can be seen in Figure 2‐131 that the vertical and radial particle velocities are greater than
the tangential particle velocity near the impact point. The attenuation rate is highest in the
radial direction and lowest in the tangential direction. At far distances greater than 100 m
the particle velocity in all three directions are almost the same.
Figure 2‐131: Attenuation relationship of PPV with distance in all three directions for nine
consecutive impacts (Hwang and Tu, 2003, 2006)
181
Figure 2‐132 shows the results of Hwang and Tu’s study on PPV attenuation for different
drop heights. As expected, higher drop heights (more energy and momentum) generate
larger PPV. It can also be observed that the attenuation rate is more for the higher drops.
Figure 2‐132: PPV attenuation for different drop heights (Hwang and Tu, 2003, Hwang and Tu, 2006)
An interesting observation in this study can be seen in Figure 2‐133 in which the vertical and
radial Fourier spectrum at 10 m is presented for different drop heights; the spectral shapes
of different drop heights are the same. Higher drops have a more distinct peak spectral value
and clearer predominant frequencies. It seems that the attenuation differences for 15 and
20 m drop heights are small. In the opinion of Hwang and Tu, this seems to indicate that the
ground has a capacity limit for storing the vibration energy. When the drop height is greater
than a certain value the induced vibrations have no significant difference, and Hwang and Tu
understood that the redundant potential energy may have been consumed to create deeper
impact craters.
182
Figure 2‐133: Vertical and radial Fourier spectra at 10 m distance for different falling heights (Hwang
and Tu, 2003, Hwang and Tu, 2006)
Obviously, and as shown in Figure 2‐134, the best way to ensure the accuracy of the
estimated PPV value is to directly measure it on site by the vibration meter. This should be
done definitely if the contract specifies it, or if there are reasons of concern for the parties
involved in the project or the third parties.
Figure 2‐134: Vibration metre for measuring peak particle velocity
183
The relationships among peak values of harmonic waves may be expressed by (Mayne, 1985)
and can be used to estimate the other parameters:
2 2 2‐145
a = acceleration
f = frequency of vibrations
v = particle velocity
s = displacement
Vibration measurements should be taken in three mutually orthogonal directions
simultaneously. Damage criteria developed by USBM, OSM and others have been based on
the maximum single value of the three directional components. This is also in line with
standards such as BS 7385‐2:1993 (British Standards Institution, 1993). Since real waves are
three dimensional and the transducer axes of the 1980s technology may not be exactly in
line with the source of vibrations, Mayne (1985) notes that some engineers prefer to
calculate the true vector sum (TVS) of the triaxial components:
2‐146
where all values are obtained at the same time. Needless to say, TVS will yield a conservative
figure that is always larger than PPV.
Mayne has also noted that several individuals (Dobson and Slocombe, 1982, Skipp and
Buckley, 1977) have mistakenly expressed the vibration levels in terms of the pseudo vector
sum (PVS), formulated in Equation 2‐147. This error is still appearing in more recent works
(Mostafa, 2010).
P 2‐147
Implementation of PVS is unjustified as it is too conservative and not related to a maximum
velocity at a particular time (that has been the basis of all codes of practice). The maximum
184
particle velocity values in the three orthogonal directions rarely, if ever, occur at the same
time, and there does not seem to be any evidence to indicate that they will happen
simultaneously. Mayne has assessed that at most PVS values could be 73% higher than the
maximum single component velocity, and typically, TVS values are about 10 to 40% higher
than the maximum single component velocity.
Wiss (1981) has proposed to express peak particle velocity (PPV) in terms of both distance,
d, and impact energy, E, in a single expression:
2‐148
√
n= vibration attenuation rate. The value of n is generally from 1 to 2, with a relatively
common value of 1.5
K= intercept
/√ = scaled distance
As shown in Figure 2‐135, Mayne et al. (1984) compiled the results of 14 dynamic compaction
sites. Soil types at these sites included silty sands, sandy clay, rubble, coal spoil and debris
fill. For preliminary estimates of ground vibration levels, a conservative upper limit appeared
to be:
.
√
70 2‐149
PPV is in mm/s, d and H are in metres and W is in tons. The variable √ is defined as
inverse scaled distance.
Later and based on a mix of single maximum component and TVS measurements (that should
lead to conservative figures) of 12 sites, Mayne (1985) proposed an upper limit conservative
PPV in the form of:
185
Figure 2‐135: Attenuation of ground vibrations measured on different dynamic compaction projects
(Mayne et al., 1984)
.
√
92 2‐150
Annex E of BS 5228‐2:2009 (British Standards Institution, 2009) that is for informative
purposes has converted Equation 2‐150 to SI units, and predicts dynamic compaction
induced PPV using Equation 2‐151.
.
√
0.037 2‐151
d is between 5 to 100 m, and WH is between 1.0 to 12 MJ.
In another approach and as shown in Figure 2‐136, in order to get a close trend and based
on information accrued by monitoring vibrations produced by different drop heights from a
site in Alexandria, Virginia, Mayne believes that while pounder weight may affect vibration
frequency, the magnitude of vibration particle velocities is slightly more influenced by the
drop height. Therefore, he has proposed to estimate PPV (in the same units as impact
velocity) by normalisation (dividing by the theoretical impact velocity of the falling pounder
186
2 ), and plotting it against the normalised distance to impact (by dividing d by the
pounder radius r):
.
0.2 2‐152
2
Figure 2‐136: Observed attenuation of normalised PPV with normalised distance (Mayne, 1985)
For impact energies in the range of 250 to 300 tm, Varaksin (1981) and Chapot et al. (1981)
estimated PPV to be (Figure 2‐137) as presented in Equation 2‐153.
.
340 mm/s 2‐153
The measurements of Romana Ruiz and Jurado (1999) for an impact energy of 252 tm were
in agreement with Equation 2‐153 only for the first impact and for distances greater than 10
m. The vibration levels increased rapidly with subsequent impacts; in some cases up to 2 or
3 times.
187
Figure 2‐137: Peak Particle Velocity for impact energy in the range of 250‐300 tm (Chapot and et al.,
1981, Varaksin, 1981)
Figure 2‐138: Plot of peak particle velocity versus inverse scaled distance for collapsible soils (Rollins
and Kim, 1994)
188
Based on a number of projects and as shown in Figure 2‐138, Rollins and Kim (1994) have
estimated PPV in collapsible soils to be less than non‐collapsible soils, and as presented in
Equation 2‐154.
.
√ 2‐154
20
It was discussed in Section 2.3.5 that wave amplitudes attenuate with distance; however, this
reduction is sometimes not sufficient to mitigate or eliminate the risk of dynamic compaction
induced damages. In such situations, it may be necessary to reduce the vibrations by
implementing special measures.
It can be understood from the equations of Section 2.8.5.2 that PPV is a function of impact
energy; hence, the simplest way to reduce the vibration amplitude appears to be the
reduction of impact energy itself. This can be merely achieved by reducing either the pounder
weight or the pounder drop height. In such cases, energy deficiency may be compensated for
by increasing the number of pounder drops per print.
This technique is applicable only when lesser impact energies are sufficient to satisfy the
technical specifications and design criteria; otherwise alternative solutions such as
construction of vibration reduction barriers or trenches should be considered.
As shown in Figure 2‐139 the concept of isolation by wave barriers is based on reflection,
scattering and diffraction of wave energy. Wave barriers could be solid, fluid or void zones in
the ground. At a solid to solid interface both P and S‐waves are transmitted, at a solid to fluid
boundary only P‐waves are transmitted, and finally at a solid to void interface no waves are
transmitted. Thus, it appears obvious that the most effective wave barrier is a void trench.
Figure 2‐139: Vibration isolation using an open trench (Shrivastava and Kameswara Rao, 2002)
189
It has been shown (Thau and Pao, 1966) theoretically that in an elastic medium a thin crack
is sufficient to screen vertically polarised SH‐waves which are analogous to the vertical
component of the R‐wave at the surface of the half wave.
An active barrier reduces vibrations at the source. On the other hand, passive barriers screen
vibrations at the location of the receiver. In dynamic compaction the print location is
constantly changing, and it is not possible to construct an active barrier. Hence, only passive
vibration reduction isolators are applicable.
Woods and Richart (1967) and Woods (1968) have carried out broad investigations to
develop guidelines for the design of active and passive barriers. The variables in the tests that
Woods and Richart carried out were the trench depth, HT, the trench length, LT, the trench
width, WT, and the distance from the source of excitation to the trench, RT. The trench sizes
ranged from 0.3 m deep by 0.3 m long by 0.1 m to 1.2 m deep by 2.4 m long by 0.3 m wide,
the vibration excitation was at 1.5 and 3 m from the trench, and the vertical vibration
frequencies were 200, 250, 300 and 350 Hz. The variables were made dimensionless by
dividing them by the wave length LR.
From practical considerations, the critical dimension for trench barriers was the scaled depth
HT/LR; therefore, for each distance from the source, the shallowest trench that satisfied the
criterion of reducing the vertical ground motion amplitude to 25% of its value without
barriers was determined. The minimum scaled depth for the passive trenches was generally
1.2< HT/LR <1.5. To evaluate the total trench area on the screened zone, a quantity HTLT/LR2
was computed for each trench. There was a general trend toward increasing HTLT/LR2 with
increasing RT/LR. Figure 2‐140 shows the amplitude ratio contour diagram for a trench that
satisfied the criterion. It can be observed that the wave amplitude has magnified in front of
and near the end of the trench.
Curves of vertical displacement amplitude versus distance along a line of symmetry from the
vibrator through the trench are shown in Figure 2‐141 for five tests. It can be seen that the
larger trenches are more effective than the smaller ones. Once again, it can be observed that
the amplitudes were magnified in front of the trench.
190
Figure 2‐140: Schematic of test layout for active isolation in the field (Woods, 1968)
Figure 2‐141: Amplitude of vertical displacement versus distance from source for five tests
(Woods, 1968)
191
It can also be noticed from Figure 2‐141 that the trenches are more effective at closer
distances behind the trench than at further distances. In fact it appears that beyond a certain
point the trench will become inefficient. This is in line with BS 5228‐2:2009 that proposes
that for maximum effects the trench should be as close to the source or to the receiver as
possible.
Based on the study of Thau and Pao (1966) it was assumed that the width of an open trench
would not be an important variable. Woods performed a few tests to evaluate this
assumption, and it was found that an increase in width did not cause a significant change in
either the magnitude of reduction or the shape of the screened zone.
Another hypothesis was that open trenches would be more effective than sheet walls as
surface wave barriers. A number of sheet wall barrier tests were performed, and it was found
that in general sheet‐wall barriers are not as effective as open trench barriers for reducing
vibration amplitudes.
Shrivastava and Kameswara Rao (2002) have performed numerical modelling of the same
problem using vertical and horizontal impulse loads. Their study suggests that for vertical
impulses, an increase in the length or depth of the trench reduces the amplitude ratio behind
the trench, but this reduction becomes less as the distance to the trench increases
(Figure 2‐142 and Figure 2‐143).
Figure 2‐142: Effect of length of trench on vibration amplitude (Shrivastava and Kameswara Rao,
2002)
Tsai and Chang (2009) have also studied vibration isolation for trenches with walls by using
the boundary element method. Their research shows that vibration reduction ratio (Ary) is 2
to 10 times more for open trenches with sheet pile or diaphragm walls as compared to when
the trench has no wall (Figure 2‐144).
192
Figure 2‐143: Effect of depth of trench on vibration amplitude (Shrivastava and Kameswara Rao,
2002)
Figure 2‐144: The comparison of screening effectiveness between the open trench barriers: (a) the
amplitude of ground vibration behind the barriers and (b) the ratio of Ary was based on the
amplitude with and without the walls (Tsai and Chang, 2009).
Tsai and Chang also found that changes in either Poisson’s ratio or trench width had
insignificant influence of on vibration amplitude changes for both cases of sheet pile and
193
diaphragm wall trenches. According to Tsai and Chang, Poisson’s ratio does not have a
significant effect because the vibration isolation by a trench is primarily achieved by the
screening of R‐waves rather than P‐waves (refer to Section 0; P‐wave velocity varies greatly
with Poisson’ ratio).
Other research (Gazetas, 1982) indicates that most of the Rayleigh wave energy is contained
in a near surface zone that is about one third to half of a wave length thick. Hence, it can be
understood that the trench depth must be at least a third of the wave length to be rather
efficient.
In their research, Hwang and Tu (2003, 2006) also tested the effectiveness of shallow
isolation trenches, 3 m deep by 1 m wide, which were 10 to 20 m wide and 15 to 150 m away
from the impact point. Their findings indicate that vertical, radial and tangential particle
velocities were basically the same with and without the trenches.
Based on the work of others (Ahmad and Al Hussaini, 1991, Liao and Sangrey, 1978, Watts,
1990) and taking into consideration that most of the Rayleigh wave energy is contained in a
near surface zone about one third to one half of a wavelength thick, Hiller and Hope (1998)
proposed that the isolation barrier depth should be at least one third the wave length:
0.3 2‐155
f is the predominant frequency, and vR is the speed of propagation of a Rayleigh wave in the
ground or, approximately, the shear wave speed, vS.
It can be concluded from the above that:
1. It is most effective is vibration reduction barriers are constructed in the form of open
trenches without walls.
2. The width of an open trench causes small changes to the amount of vibration
reduction. Hence, and from the practical point, it is sufficient to make the trench
width equal to the width of the excavator shovel.
3. The maximum amplitude reduction can be expected to be closest to the back of the
trench; hence, it is best to excavate the trench as close as possible to the zone that
is to be protected. In any case, sufficient distance must be maintained in order not
to endanger the stability of the structures behind the trench.
194
4. Deep trenches are able to reduce the vibration amplitude more than shallow
trenches. The trench depth should be 0.3 times the wave length to be able to
effectively reduce the amplitude of ground motion. In practice the trenches will be
generally 3 to 4 m deep. Practically speaking and if there is a ground water table, the
depth of excavation will have to be limited to the level of the ground water.
5. Trenches with larger areas (length and depth) are more effective. In any case and as
is done in practice, it is best to extend the trench along the boundary where
vibrations must be reduced in order to get the best results.
Although reducing vibrations by excavationof open trenches is common practice in dynamic
compaction, not much has been published on isolation trenches application in DC. Varaksin
(1981) notes that dynamic compaction can be performed very close to rigid structures that
are less than 20 m away, and adds that for normal buildings in good condition a safe distance
of 30 m is necessary. For sensitive buildings in good condition, he recommends safe distance
of 50 m. In case the mentioned distances cannot be honoured, he suggests that vibrations
can be reduced by digging a wave isolation trench that is 1.5 to 2.5 m deep. Varaksin makes
a note that damping appears to be more important than what theory would predict because
even though Rayleigh wave energy is mostly distributed over half the wave length, shallow
trenches such as shown in Figure 2‐145 (reduction of particle velocity of Rayleigh wave versus
distance from pounder impact point) demonstrate effectiveness.
Figure 2‐145: Effectiveness of shallow trenching for reduction of vibration (Varaksin, 1981)
It is the opinion of Varaksin (2014) that trenches are best effective when they are close (5‐6
m) to the impact, and loose efficiency at greater distances (50 m) as the trench depth has to
be one third of wave length, and R wave lengths are around 140 m.
195
2.9 Quality Control
In the earlier sections of this chapter it was shown experimentally (Hajialilue‐Bonab and
Rezaei, 2009) and numerically (Gu and Lee, 2002) that the rate and amount of improvement
reduces as further pounder blows are applied to the ground until improvement increase
tends towards negligible amounts. This would suggest that monitoring the rate of
compaction could be a good indication of the efficiency of dynamic compaction. This may be
achieved by measurement of ground deformation during consecutive pounder blows.
Measurement of crater depth per se does not appear to be an acceptable means of
measuring improvement efficiency because simple crater depth measurement could lead to
erroneous conclusions. The crater depth is not only a function of the induced settlement
under the pounder, but rather the resultant of induced settlement and crater sides collapse
back into the crater floor. This suggests that measurement of the crater volume would be a
more meaningful than the simple measurement of crater depth. Furthermore, measurement
of crater volume on its own would still not be able to accurately predict the efficiency of
compaction as it does not take the deformations of ground around the crater into
consideration. Indeed the efficiency of compaction can be best determined if all volumetric
changes within the crater dimensions and beyond the crater zone are included in the
assessment.
Figure 2‐146: Net effective volume change, modified from Lukas (1986)
196
Figure 2‐146 shows the relationship between the number of pounder blows and net effective
volume change, which is the resultant of the summation of volume change of soil in the crater
and around the crater, and can be used to determine the appropriate number of blows that
achieves an optimum amount of compaction.
While the change in soil volume in the crater is always positive (penetration and reduction of
volume), the peripheral soil can be positive or negative (heave and increase of volume caused
by iso‐volumetric plastic deformations). It is possible that while the pounder continues to
penetrate the ground, and increase the crater depth, the soil is undergoing plastic
deformation, the peripheral ground is heaving, and the net compaction volume is increasing
with a low or negligible rate. In such cases, dynamic compaction is being performed with little
or no efficiency. Figure 2‐147 shows a case encountered during the ground improvement
works for a saturated very silty sand/sandy silt section of Dubai Airport Runway where the
heave and penetration volumes were almost the same (Serridge, 2002) from the very
beginning. This means that application of dynamic compaction in areas with soil behaviour
similar to Figure 2‐147 would not have had any effect, and dynamic replacement should have
been considered.
Figure 2‐147: Inefficiency of pounder blows identifiable from measurement of net volume change
(Serridge, 2002)
To calculate the crater volume, the ground level should be determined before the first blow.
The base area of the crater will be the same as the pounder base area; however, as can be
seen in Figure 2‐148, the crater diameter at ground level is usually larger, and should be
measured in at least 2 directions to obtain an average value. The depth of crater can be
197
measured by placing the staff on the pounder, and adding the pounder height to the
readings.
Figure 2‐148: heave and penetration test
Once the crater dimensions are known the volume can be estimated using any calculation
method. If the top and base of the crater are assumed to be circular with radii RT and RB,
then the volume of a crater with height, Dc, can be estimated by a truncated cone:
1
2‐156
3
Lukas (1986) estimates this volume by taking an average value for the top and base
diameters, and assuming a cylindrical shape. This would yield a volume equal to:
2‐157
2
It can be calculated that estimation of the volume using a truncated cone will yield the same
value as a cylinder when the top and base radii are equal, but will always be slightly greater
when the top radius is larger than the base radius. The difference between the two volumes
is approximately 4% when the top radius is double the base radius and approximately 8%
198
when the ratio of the radii is 3; hence, it is preferable to use a truncated cone volume to
estimate the crater volume.
For estimation of heave volume (which could very well be settlement as well) Lukas (1986)
proposes that within each quadrant, four markers should be set around the print, with the
closest marker being about 0.5 m away from the edge of the print. The furthest marker is
suggested to be about 5 m away from the centreline of the print. The remaining markers
should be equally spaced out between the first and last markers. The elevation of the 16
markers should be determined before the first drop and at two drop intervals. The average
heave (or settlement) for markers located at the same radius from the centre of the print, Si,
should be calculated using Equation 2‐158.
2‐158
4
The heave can then be calculated based on the average heave at various distances. Lukas
notes that heave is generally greatest near the edge of the print, and reduces to small
amounts further away from the print, and suggests that, as shown in Figure 2‐149, heave
calculation can be simplified by assuming a linear heave pattern instead of the actual average
marker heaves. The volume of heave (or settlement) from the simplification will yield:
3 2‐159
L= Distance from edge of crater to where heave reduces to zero
Figure 2‐149: Heaved ground section (Lukas, 1986)
It is the author’s experience that choosing the first set of markers at a distance of 0.5 m from
the print may be too close when the soil is very loose and this ring of markers may fall within
199
the crater volume. Hence, it may be preferable to allow for an extra intermediate ring of
markers or to set the first row of markers at a greater distance from the print. Also, dividing
the ground into 3 parts may be less time consuming, but with sufficient accuracy as setting
the markers at quadrants.
Noting that the pressuremeter test, dynamic compaction and dynamic replacement have all
been the inventions and developments of Louis Menard, it is not surprising that for the 20
years that Menard held the patents of DC and DR, PMT was carried out systematically as the
verification test. Although nowadays Menard Pressuremeter, MPM, is not the only type of
pressuremeter device, all references to PMT in this thesis is meant to refer to the Menard
Pressuremeter Test. Review of literature shows that to date all major standards such as ASTM
(2007) and ENV 1997‐3 (European Standard, 2000) have derived from Menard’s testing
procedure, interpretation and calculation rules (Centre D'Etudes Menard, 1975).
The pressuremeter test is an in situ stress‐strain test that is performed on the wall of a
borehole using a laterally expanding cylindrical probe, and ultimately provides a failure value
in shear mode. PMT is somewhat different from other in‐situ tests, such as SPT or CPT, which
are or were originally used in geotechnical design on the basis of correlations (Gambin and
Frank, 2009). In this test a cylindrical cavity is subjected to pressure increments typically at 1
m intervals of depth, and the resulting expansion is measured in terms of a volume increase.
The result of the test is a failure and a small strain (10‐2) deformation parameter, which make
the test a very suitable device for predicting bearing capacity (shear failure) and
deformations.
The Menard pressuremeter consists of three main elements; i.e., a radially expandable
cylindrical probe that is placed inside the borehole at the desired test level, a control or
monitoring unit that remains on the ground surface, and the tubing that allows the probe to
be pressurised.
200
Figure 2‐150: The main elements of the Menard pressuremeter (Baguelin et al., 1978)
Instead of a long chamber, the probe is made up of three independent cells stacked on top
of each other. Each cell has a rubber membrane, top and bottom discs, and a rigid steel
backbone. All cells are inflated to the same pressure simultaneously. The top and bottom
cells are called guard cells because they protect the middle cell, which is the measuring cell,
from the end effects caused by the finite length of the probe, and allow it to expand only in
the radial direction, as if the probe had an infinite length. This results in plane strain
deformation conditions (Baguelin et al., 1978). The combined height of the measuring and
guard cells should be at least six times the cell diameters (ASTM, 2007).
To protect the probe membranes against the sharp surface of a borehole, the probe can be
initially placed inside a slotted casing, which is a thick steel tube with a number of longitudinal
slots that allow the tube to expand laterally. Commonly, the casing has six slots, each about
1 mm wide. There are a number of standard sizes of slotted casings, but the most widely used
has an outside diameter of 63 mm, an inside diameter of 49 mm, and protects a 44 mm
diameter probe.
The control unit is a container with a front panel on which all the measuring instruments are
installed, and its function is to control and monitor the expansion of the probe. A coaxial
tubing connects the probe to the instruments, and is used to pressurise the probe cells. The
inner tube carries water to the measuring cell, and the annular space between the inner and
outer tubes allows gas, usually nitrogen, to reach the guard cells. Placing the tube supplying
201
the measuring cell inside the outer tube does not allow the inner tube to expand, which
would result in erroneous readings of the amount of injected water and volume change.
2.9.2.2 Calibration
Prior to inserting the probe in the borehole, the equipment should be calibrated to
compensate for pressure losses, Pc, (due to probe rigidity) and volume losses, Vc, (due to
expansion of tubing and compressibility of any part of the testing equipment) by recording
the expansion of the probe for equal increments of pressure (or volume). Pressure and
volume calibration curves are shown in Figure 2‐151.
Pressure correction must be deducted from the pressure readings obtained during the test.
Volume loss correction must also be deducted from the measured volumes during the test.
The latter correction is relatively small in soils, and although ASTM (2007) allows it to be
neglected if it is less than 0.1 % of the nominal volume of the measuring portion of the not
inflated probe per 100 kPa of pressure, it notes that in very hard soils or rock, the correction
is significant and must be applied. In agreement with ASTM, ISSMFE or the International
Society of Soil Mechanics and Foundation Engineering, (Amar et al., 1991) notes that as long
as Menard modulus, EM, is less than 100 MPa volume losses are usually negligible.
Figure 2‐151: Pressure and volume calibration curves (ASTM, 2007)
202
2.9.2.3 Testing Procedure
ASTM (2007) allows pre‐drilled pressuremter tests to be carried out by two procedures; i.e.,
either by equal pressure increments or by equal volume increments; however, as the more
common procedure is the first method, only the first method will be considered here.
The pressuremeter cavity is prepared either by drilling a borehole, or by advancing a sampler.
To ensure adequate volume change capability, the probe must be inserted in a borehole with
a diameter close to that of the probe. If the borehole diameter is too large the probe's
expansion limit may be reached before completion of the test, and if the hole is too tight
then the probe will have to be forced into the hole, and the PMT pressure versus volume
curve will not show a pseudo linear phase. ISSMFE (Amar et al., 1991) recommends that the
hole diameter should not exceed the instrument diameter by more than 10%. Menard
(Centre D'Etudes Menard, 1975) and later ASTM (2007) have specified probe diameters for
typical probe diameters as shown in Table 2‐17. ASTM further stipulates that the borehole
diameter shall be from 1.03 times to 1.2 times the nominal probe diameter. If used, borehole
diameters should also be adapted for slotted casings.
designation* reading V0
(cm3)
Menard ASTM Menard
Minimum Maximum nominal Maximum 535
AX 44 46 52 45 53 535
BX 58 60 66 60 70 535
NX 74 76 80 76 89 790
* Probe diameters are based on the standards of the Diamond Core Drill Manufacturer’s Association
(DCMA)
Table 2‐17: typical probe and borehole diameters (ASTM, 2007, Centre D'Etudes Menard, 1975)
Before the probe is positioned in the hole for testing an accurate determination should be
made of the zero volume reading, V0, which is the volume of the measuring portion of the
uninflated probe at atmospheric pressure. Menard (Centre D'Etudes Menard, 1975) has
already determined the values of V0 for different probes; however, as shown in Figure 2‐151,
V0 can be estimated by fitting a straight line extension of the volume correction curve to zero
pressure (ASTM, 2007):
203
2‐160
4
Vi= intercept of volume calibration line
di= inside diameter of the heavy duty steel casing or pipe
l= length of the measuring cell
The probe should then be lowered to the test depth, which is the midpoint depth of the
probe. In the equal pressure increment method, pressure is applied in approximately equal
increments. Too small steps will result in an excessively long test, and too large steps may
yield results with inadequate accuracy.
Menard (Centre D'Etudes Menard, 1975) suggests testing increments to be preferably 10,
but accepts from 5 to 14 increments as well. ASTM (2007) suggests that pressure steps should
be determined in such a way that about 7 to 10 load increments are obtained. Menard
requires that volume variations be recorded after 15 seconds, 30 seconds and 1 minute;
however, ASTM does not require the 15 seconds reading.
In soft, loose, and sensitive soils, the hole should be predrilled ahead of the testing depth
only far enough so that the cuttings settling at the bottom of the hole will not interfere with
the test; however, in stiff soils and weathered rocks where degradation due to exposure is
not significant, the hole can be predrilled to several test depths When the probe is driven
into the soil, testing can take place continuously (ASTM, 2007).
Menard (Centre D'Etudes Menard, 1975) advises that to obtain as complete a load
deformation curve as possible, the measured volume should reach 700 cm3 if the
pressuremeter limit pressure, PLM, is less than 800 kPa and 600 cm3 if 800< PLM < 1,500 kPa,
but suggests that in other cases the test must be carried out up to pressures of approximately
2,000 to 2,500 kPa in soils and up to 5,000 to 7,000 kPa in rocks. According to ASTM (2007)
when equal pressure increments are used, the test increments should be increased until the
204
expansion of the probe during one load increment exceeds about one fourth of V0 that is
typically about 200 cm3 for a 800 cm3 probe.
Prebored pressuremeter design rules were established historically based on testing without
unload‐reload loops (ASTM, 2007); however, ASTM encourages the performance of
unload‐reload cycle(s).
Lukas (1986) proposes the below guidelines for verification of ground improvement
performance:
1) Regardless of the size of the improved area, there should be at least two borings with in‐
situ testing. A minimum of one boring with in‐situ testing should be carried out for
approximately every 1,000 m2 of area that was compacted for a structure, or 4,000 m2 of
embankment.
2) The borings should extend through the thickness of the deposit that is compacted.
Testing should be carried out at depth intervals not exceeding 1.5 m to determine the
degree and depth of improvement.
3) Additional borings and tests should be carried out in areas where anomalies were
observed during compaction. This would include areas where the crater depths were
significantly greater than average, or where the ground settlement was greater than
normal. The purpose of these tests would be to verify that the magnitude of
improvement in these areas is similar to adjacent areas.
It is the author’s personal experience that performing one boring per every 2,000 m2 of
improved ground in large scale projects with uniform ground conditions will be sufficient and
satisfactory. Also, it is the author’s experience that observation of the prints can assist in
identifying soft or hard spots or zones that react differently to the pounder impact from what
was observed during the heave and penetration test. These spots or zones can then be
further investigated.
205
2.9.2.5 Calculation of PMT parameters
Once pressure and volume readings are corrected, the pressure versus volume curve can be
drawn to produce a curve similar to what is shown in Figure 2‐152. The pressure transmitted
to the soil by the probe is as shown in Equation 2‐161 (ASTM, 2007).
Figure 2‐152: Typical pre‐drilled PMT curve (Amar et al., 1991, ASTM, 2007)
2‐161
P = pressure exerted by the probe on the soil
PR = pressure reading on the control unit
Pδ = hydrostatic pressure between the control unit and the probe, which can be calculated
using Equation 2‐162.
PC = pressure loss correction. The maximum value of PC should be less than 50% of PLM
(ASTM, 2007).
2‐162
Hp = depth of probe below the control unit
δt = unit weight of test liquid in instrument
The corrected volume reading of the probe from the volume readings is (ASTM, 2007):
2‐163
V = corrected increase in volume of the measuring portion of the probe
206
VR = volume of reading on readout device
VC = volume loss correction made at the test pressure readings corresponding to
, and can be calculated from Equation 2‐164.
2‐164
Vr = injected volume at the end of each pressure increment, Pr
a = slope of the volume versus pressure calibration plot
As shown in Figure 2‐152, creep or end of pseudo elastic phase pressure, Pf, is obtained by
plotting the volume differences between the 30 seconds and 1 minute readings against the
applied pressure. Figure 2‐152 also shows that three distinct parts can be found on the test
curve; i.e., the inflation of the probe until the membrane reaches the borehole walls (P ≤ Pi),
the pseudo elastic response of the soil to the probe pressure (Pi ≤ P ≤ Pf), and the plastic
response of the soil (Pf ≤ P) up to ΔV/Vo=1 (volume of probe has doubled to twice the zero
volume). In the pseudo elastic zone the creep curve appears to be a straight line, and a
constant soil deformation modulus called Menard modulus, EM, can be measured to be
(Centre D'Etudes Menard, 1975):
Δ
2 1 2‐165
Δ
ν = Poisson ratio
V0 = zero volume reading
V= corrected volume reading of the measuring portion of the probe
ΔP= corrected pressure increase in the centre part of the straight line portion of the pressure‐
volume curve (see Figure 2‐152)
ΔV= corrected volume increase in the centre part of the straight line portion of the pressure‐
volume curve, corresponding to ΔP pressure increase (see Figure 2‐152)
Vm = corrected volume reading in the centre portion of the ΔV volume increase.
Menard (Centre D'Etudes Menard, 1975) assigns an arbitrary value of 0.33 to the Poisson
ratio. This assumed value has no impact on estimation of settlement as the same value is
assigned in Section 2.9.2.7. ENV 1997‐3 (European Standard, 2000) also accepts this
assumption. Equation 2‐165 can therefore be presented as:
207
Δ
2.66 2‐166
Δ
For the popular AX probe, using V0 as presented in Table 2‐17 (535 cm3), and for Vm
approximately equalling 200 cm3, Menard yields 2(1+ν)( V0 + Vm)≈ 2,000 cm3.
Menard (Centre D'Etudes Menard, 1975) defines limit pressure, PLM, as the abscissa of the
asymptote to the pressuremeter curve, and notes that this parameter can be determined
directly from the curve, but is more conveniently taken as the pressure corresponding to a
volume increase equal to the initial volume of the borehole. Since the initial volume of an AX
or BX probe is in the order of 600 cm3 (V0= 535 cm3 plus the volume injected to contact the
borehole wall), the reading may be assumed for a reading of 700 cm3 (see Figure 2‐153).
Thus:
Pu =pressure reading corresponding to a volume increase equal to the initial volume of the
borehole
Pi = pressure at which the probe makes contact with the borehole.
Figure 2‐153: Pressuremeter pressure versus volume curve (Centre D'Etudes Menard, 1975)
208
If the volumetric increase attained during a test is less than what has been mentioned, it is
still possible to extrapolate the value of PLM with precision, provided that Pf has been
exceeded during the test.
Extrapolating PLM can be done based on the relative or reciprocal volume theories (Centre
D'Etudes Menard, 1975). In the latter method, if the plot of pressure is drawn against the
reciprocal of volume, a straight line can be passed through the last few readings
corresponding to the plastic phase. Menard notes that the point at which the extension of
the straight line intersects the 700 cm3 ordinate corresponds to PLM; however, the borehole
volume does not necessarily have to be 700 cm3, and as shown in Figure 2‐154, can vary from
his rational figure.
Figure 2‐154: Extrapolation of PLM (Amar et al., 1991)
In a similar approach Baguelin et al. (1978), ASTM (2007) note that while theoretical PLM is
defined as the pressure where infinite expansion of the probe occurs, for practical purposes
PLM is defined as the pressure where the probe volume reaches twice the original soil cavity
volume, defined as V0 + Vi (see Figure 2‐155), where Vi is the corrected volume reading at
the pressure at which the probe makes contact with the borehole (corresponding to Pi in
Figure 2‐152).
As shown in Figure 2‐156, ENV 1997‐3 (European Standard, 2000) defines PLM as the pressure
required to double the total volume of the cavity from the point (pr, Vr) at which changes of
209
volume change to pressure change; i.e., d(ΔV)/d(Δp), is a minimum. Vr is the injected volume
corresponding to pressure pr, and V0 is the deflated volume of the probe.
Menard (Centre D'Etudes Menard, 1975) notes that an approximate value of PLM can also be
obtained considering the below statistical results:
Figure 2‐155: Determination PLM from inverse of volume versus pressure (ASTM, 2007)
Figure 2‐156: Interpretation of EM and PLM from PMT from ENV 1997‐3 (European Standard, 2000)
210
Pf is equal to approximately ½ to 2/3 PLM.
For every geological formation, there is a constant EM/PLM value that also depends
on the ground type (see Table 2‐22 and Table 2‐23).
Table 2‐17 presents typical ranges of EM and PLM (Centre D'Etudes Menard, 1975).
Ultimate bearing capacity of a foundation can be calculated using Equation 2‐168 (Centre
D'Etudes Menard, 1975):
2‐168
qu= ultimate bearing capacity
qo= total overburden pressure at the periphery of the foundation level after construction.
This term is defined slightly different in ENV 1997‐3's informative Annex C.1 as the total
(initial) vertical stress at the level of the foundation base. This parameter is calculated
from the unit weight of soil and depth of footing.
k= a bearing factor varying from 0.8 to 9 according to the embedment and the shape of the
foundation level after construction
211
Po= total at rest horizontal earth pressure at the test level (at the time of the test). ENV 1997‐
3 defines Po according to Equation 2‐169.
2‐169
Ko= coefficient of horizontal earth pressure at rest. Menard (Centre D'Etudes Menard, 1975)
notes that experience shows that in very loose soil or compact soil Ko is respectively
approximately 1 or 0.5; however, Appendix C.1 of ENV 1997‐3 assumes conveniently that Ko
is equal to 0.5.
σv= total vertical stress
u= pore water pressure
Instead of qu, Annex C.1 of ENV 1997‐3 presents Equation 2‐168 in the form of
2‐170
′
N= resistance of the foundation against normal loads
A’= effective base area as defined in ENV 1997‐1 (European Standard, 2004)
The difference between PLM and Po is defined as net limit pressure, P*LM (Centre D'Etudes
Menard, 1975):
∗ 2‐171
Likewise, the net ultimate bearing capacity, q*, is:
∗ 2‐172
k can be obtained from Figure 2‐157 or Figure 2‐158 (Centre D'Etudes Menard, 1975)
respectively for isolated (square or circular) footings or strip footings. Soil categories referred
to in these figures are defined in Table 2‐19.
he is the equivalent foundation depth, and hc is the critical depth of embedment or the depth
below which an embedded foundation maintains a constant ultimate net bearing capacity.
Amar et al (1991) cite Menard (1963a) for noting that the ratio of spherical to cylindrical limit
212
pressures remain constant below the critical depth, if the soil properties also remain
constant. This depth is a function of soil category and half width of foundation (R=B/2).
Relative values of hc/R are shown in Table 2‐20. As can be seen in these two Figures, the
minimum value of k is 0.8 when the footing is on the ground surface.
he= equivalent foundation depth, hc= critical depth of embedment, R= half width of footing
Figure 2‐157: Bearing factor for isolated square or circular footings, piers and piles (Centre D'Etudes
Menard, 1975)
Baguelin et al. (1978) recommend that for rectangular footings, with widths and lengths
respectively equal to B and L, that are neither square nor infinitely long, k can be assumed
to vary linearly between the two extremes, and can be calculated using Equation 2‐173.
213
he= equivalent foundation depth, hc= critical depth of embedment, R= half width of footing
Figure 2‐158: Bearing factor for strip footings and diaphragm walls (Centre D'Etudes Menard, 1975)
Foundation type
Soil category
Circular or Square Continuous
I 4 6
II 10 12
III 16 18
IIIA 20 22
Table 2‐20: hc/R for different soil categories and foundation types (Centre D'Etudes Menard, 1975)
214
2‐173
Annex C.1 of ENV 1997‐3 has summarised the bearing factor for spread foundations as shown
in Table 2‐21.
Soil category PLM (kPa) k
Clay and silt A < 700 0.8 1 0.25 0.6 0.4 ⁄ ⁄
B 1,200 to 2,000 0.8 1 0.35 0.6 0.4 ⁄ ⁄
C > 2,500 0.8 1 0.50 0.6 0.4 ⁄ ⁄
Menard (Centre D'Etudes Menard, 1975) notes that it has often been the custom in dealing
with pressuremeter results to tabulate the value PLM – water head. The calculation and
presentation of results is greatly simplified by identifying PLM with PLM – water head as the
error introduced is less than 2 % for most dry sites.
When the footing is on ground with variable strengths at different depths, the equivalent
limit pressure, P*LMe, is defined as the geometric mean of P*LM values obtained near the level
of the foundation:
∗ ∗ ∗ ∗ 2‐174
P*LM1= geometric mean of values measured in levels +1.5B to +0.5B above the founding level
P*LM2= geometric mean of values measured in levels +0.5B to ‐0.5B
P*LM3= geometric mean of values measured in levels ‐0.5B to ‐1.5B
This is equivalent to taking the arithmetic mean of the logarithms (see Equation 2‐175), and
therefore gives less weight to the higher P*LMe values.
215
∗
1 ∗ ∗ ∗
ln ln ln ln 2‐175
3
For shallow foundations, P*LM1 is not introduced, and the equivalent limit pressure becomes:
∗ ∗ ∗ 2‐176
Menard (Centre D'Etudes Menard, 1975) assumes that the variation between the limit
pressures of each elevation band does not exceed ±30% of P*LMe, and advises that should this
limit be exceeded, then the problem should be considered in more detail. Menard
recommends to plot P*LMe against depth, and to smooth out all the peaks in the graph.
Baguelin et al. (1978) assume the variation between the limit pressures does not exceed
±40% of P*LMe.
When the ground properties vary with depth, it is necessary to define the equivalent depth
of embedment he relative to the soil of the founding elevation. This depth can be calculated
by using Equation 2‐177 (Centre D'Etudes Menard, 1975) when limit pressure can be defined
as a function of depth or Equation 2‐178 when limit pressure values are variable in different
ground layers.
1 ∗
∗
d 2‐177
1 ∗
∗
∆ 2‐178
Menard (Centre D'Etudes Menard, 1975) proposes a safety factor of 3 for bearing capacity;
therefore, the allowable bearing capacity, qa, can be expressed as:
2‐179
3
In terms of net allowable bearing capacity, qa, can be expressed as:
216
∗ ∗ 2‐180
3
When loads are eccentric, Menard (Centre D'Etudes Menard, 1975) notes that failure may
occur either by general punching of the footings or by localised failure in the zone of highest
loading due to tilting. He states that general stability will be achieved if:
_ _
2‐181
2
σv_max= maximum vertical stress below footing
σv_min= minimum vertical stress below footing
When relative embedment depth h/R > 1, with a simplification, stability against tilting can
be achieved if:
_ 1.5 2‐182
Baguelin et al. (1978) proposes that effects of eccentricity can be taken into account
conservatively by reducing the footing width to an effective width, B’, which will result in a
concentric load on a fictitious footing with a smaller width. B’ can be calculated from
Equation 2‐183 (Meyerhof, 1953):
2 2‐183
e= eccentricity of the load (see Figure 2‐159)
Figure 2‐159: Eccentric loading on footing, Baguelin (1978) after Meyerhof (1953)
217
Once the footing’s eccentric load has been transformed into a concentric load acting on a
reduced width, then it will be possible to determine the bearing capacity using
Equation 2‐168.
Figure 2‐160: Inclined load on footing (Baguelin et al., 1978)
Baguelin et al. also propose a procedure for determination of bearing capacity for inclined
concentric loads when variation of limit pressure is linear in depth. If the ratio of net limit
pressure at depths of h and h+B (see Figure 2‐160) is expressed by M (see Equation 2‐184),
then a reduction factor iδ should be applied when the load is inclined from the vertical at an
angle of δ degrees. When the inclination angle is between 0 and 30 or 35o, iδ can be calculated
from Equation 2‐185:
∗
∗
2‐184
1 1 1 2‐185
90 20
2‐186
For 0 1 1 2‐187
218
For 1 0 2‐188
For 0 1 1 2‐189
For 1 0 2‐190
For special cases when net limit pressure is constant in depth, or when the footing
embedment depth is at least equal to the footing width:
1 2‐191
90
The bearing capacity of footings resting on sloping ground or near excavations must be
reduced by a coefficient that is a function of β, the angle between the footing and excavation
level. This reduction factor is shown in Figure 2‐161 (Centre D'Etudes Menard, 1975).
Figure 2‐161: Bearing capacity reduction factor for footings resting on sloping ground or near
excavations (Centre D'Etudes Menard, 1975)
Baguelin et al. (1978) propose the same reductive approach that was used for estimation of
bearing capacity of foundations subject to inclined loading to determine the bearing capacity
219
reduction factor near slopes and hillsides. However, in the latter case, the inclination angle,
δ, is replaced, β’, which is shown in Figure 2‐162. Thus Equation 2‐185 becomes:
′ ′
1 1 1 2‐192
90 20
Figure 2‐162: Footing near slope or hillside (Baguelin et al., 1978)
Baguelin et al. note that for steep slopes it is necessary to assess the stability of the slope
itself, and further add that the pressuremeter can only be considered to give a rough
estimate of the soil conditions for slope stability analyses.
When the load is inclined in such a way that the resultant force acts towards the hillside the
combined bearing capacity reduction factor, iδβ, will be used with the same approach as
described above (Baguelin et al., 1978). Hence, Equation 2‐185 will becomes:
′ ′
1 1 1 2‐193
90 20
On the other hand, when the resultant force acts away from the hillside a bearing capacity
reduction factor is calculated with a similar approach for the condition yielding the higher
factor between the two below cases (Baguelin et al., 1978):
When the rupture develops towards the level ground and away from the slope: In
this case the bearing capacity reduction factor is based only on the load inclination
angle, iδ, and Equation 2‐185 will be applicable.
220
When the rupture develops away from the level ground and toward the slope: In
this case the bearing capacity reduction factor, iδβ, will be based on the difference
between β’ and the inclination angle, δ, and Equation 2‐185 will take the form of
Equation 2‐194.
′
1 1 1 2‐194
90 20
If the load is both inclined and eccentric, as shown in Figure 2‐163, then Baguelin et al.
envisage two possible cases:
Figure 2‐163: Footing subject to inclined and eccentric load (Baguelin et al., 1978)
For the case shown in Figure 2‐163(a) the effect of the combination of inclined and
eccentric loading is to magnify the bearing capacity reduction factor, iδe, which, as
shown in Equation 2‐195, is calculated as the product of the reduction factors iδ and
ie. The latter factor, ie, is calculated for footing width B.
2‐195
For the case shown in Figure 2‐163 (b) the effects tend to cancel out one another,
and it will be necessary to consider failure occurring possibly to the left or to the
right.
For failure to the left, ie due to the eccentricity of the load is corrected by dividing it
by iδ because of the inclination of the load (using the actual footing width B for M
and λD). If the resulting value of ie/iδ is less than 1, then the ultimate load, Q1, that
221
the footing can carry if it was loaded vertically and concentrically, is multiplied by this
ratio to give the first estimate of the ultimate inclined load, (Q'1)δ:
If 1 ′ 2‐196
For failure to the right, the ultimate vertical load, Q"1, that is concentric with a
fictitious footing with width of B"= B + 2e is computed. This ultimate load is then
corrected for the load inclination using the actual width, B, for M and λD. Therefore,
iδ is the same as what is calculated for the footing failure towards the left. This will
lead to Equation 2‐197. The inclined, eccentric design load, (Q'1)eδ, to use for the
footing is the lesser of (Q'1)δ and (Q"1)δ.
When the load is inclined and eccentric, and the ground is sloped, then the foregoing rules
are combined as required to suit the particular condition (Baguelin et al., 1978).
Menard and Rousseau (1962) have carried out full scale field tests on footings, and have
observed that contrary to the formulas derived from the elastic theory for homogeneous,
elastic medium, settlement does not increase in direct proportion to the width of the footing.
According to Menard and Rousseau (1962) and Menard (Centre D'Etudes Menard, 1975),
foundation settlement due to external loads is the result of two completely different
phenomena; i.e., volumetric compression and shear deformation. The first phenomenon is
caused by the spherical component of the stress tensor. The increase of bulk pressure
(volumetric compression) causes a reduction in volume of the material in relation to the
modulus of volumetric compression. On the other hand, the latter phenomenon is caused by
the deviatoric components of the stress tensor, and displacements occur without variation
in volume of the material.
The spherical and deviatoric components of the stress tensor are very different at depth; the
spherical component has a maximum value right under the base of the footing; however, the
deviatoric component has a maximum value at a depth that is equal to half the width of the
222
footing. Shear deformation is dominant under footings, shafts and piles, but volumetric
compression predominates under rafts or embankments. As compared to shear deformation,
the relative importance of volumetric compression increases as the safety factor against
failure increases (Centre D'Etudes Menard, 1975). Figure 2‐164 shows the variation of vertical
strain components with depth along the vertical axis below a rigid circular footing (Baguelin
et al., 1978).
Figure 2‐164: Variation of vertical strain components with depth along the vertical axis below a rigid
circular footing (Baguelin et al., 1978)
Total (final) settlement, s, can be estimated for a footing with width and length respectively
equal to B and L, that is embedded (strictly speaking, the equivalent embedment depth) by
at least the footing width (Centre D'Etudes Menard, 1975, Menard and Rousseau, 1962). In
Equations 2‐198 and 2‐199, the first and second terms that are multiplied by the mean
contact added stress respectively represent the influence of the deviatoric and volumetric
stress tensors. Menard (Centre D'Etudes Menard, 1975) notes that there is also a negligible
term corresponding to a purely elastic settlement that has been omitted from the equation.
2
When B > 0.6 m: 2‐198
9 9
223
2
When B <0.6 m: 2‐199
9 9
q= design normal pressure applied on the footing
σvo= total (initial) vertical stress at the level of the footing base
Bo= reference footing width, equal to 0.6 m
B= footing width
α= rheological factor, given in Table 2‐22 and Table 2‐23. It can be noted that this factor can
be different in the two terms of Equations 2‐198 and 2‐199 according to the prevalent
material in each zone of influence.
λd and λc= shape factors, given in Table 2‐24.
Ec= weighted value of EM immediately below the footing; i.e., as shown in Figure 2‐165, EM
from under the footing to a depth equal to half the footing width (Centre D'Etudes Menard,
1975):
2‐200
E1= weighted value of EM from under the footing to the depth of half the footing width
Ed= harmonic mean of EM in all layers down to the depth of 8B (or 16 times half footing width,
R) below the footing
Rock
Extensively fractured 0.33
Unaltered 0.5
Weathered 0.67
Table 2‐23: Rheological factor for rock (Centre D'Etudes Menard, 1975)
224
Menard and Rousseau (1962) note that can be expressed as a function of the ratio of the
standard modulus to the reload modulus:
2‐201
EM+= reload modulus
μ=a small value compared to ½ that depends on secondary factors
Without reference, Baguelin et al. note that Centre D’Etudes Menard sometimes uses the
ratioimprov EM/ EM+ (without the power) for defining ; however, this this may be due to a
typing mistake as Equation 2‐201 appears to be in more agreement with reality.
1
L/B 2 3 5 20
circle square
λd 1 1.12 1.53 1.78 2.14 2.65
λc 1 1.1 1.2 1.3 1.4 1.5
Table 2‐24: Shape factors λd and λc for different footing length to width ratios (Centre D'Etudes
Menard, 1975)
As shown in Figure 2‐165, Ed is calculated by dividing the ground beneath the footing base
into 16 layers, each with a thickness equal to half the footing width. The subscript of each
modulus designates the layer number and location. The harmonic mean of the moduli is
calculated for each layer or group of layers using Equation 2‐202:
1 1
2‐202
225
Figure 2‐165: Layers of ground under the footing that are taken into consideration for the calculation
of Ed (Centre D'Etudes Menard, 1975)
Ei= harmonic mean of moduli of layer i
n= number of moduli layer i
Ej= moduli values in layer i
Ed can be calculated from Equation 2‐203 (Centre D'Etudes Menard, 1975):
4
1 1 1 1 1 2‐203
0.85 , , 2.5 , , 2.5
226
3.6
1 1 1 1 2‐204
0.85 , , 2.5 , ,
Similarly, if E6,7,8 is also not known then:
3.2
1 1 1 2‐205
0.85 , ,
Equations 2‐198 and 2‐199 are applicable to foundations that are embedded at a depth that
is at least equal to the width of the footing. If the depth of embedment is equal to half the
footing width then the settlement should be increased by 10%. When the footing is found on
ground level, the calculated settlement should be increased by 20%.
The above discussion is applicable when the variation of pressuremeter moduli is not very
large with depth.
In the special case when the layer of soft soil is underlain by a layer of better properties at a
depth that is less than one half of the width of the footing, the settlement can be calculated
using Equation 2‐206:
2‐206
α(z)= rheological factor of the soil layer at depth z
EM(z)= Menard modulus at depth z
q(z)= increase of vertical pressure at depth z
β(F)= a coefficient that is a function of the safety factor for bearing capacity, F. This
coefficient can be calculated from Equations 2‐207 and 2‐208 (Centre D'Etudes Menard,
1975):
If 3 1 2‐207
227
2
If 3 2‐208
3 1
When a weak compressible layer of soil is interbedded between two layers, Menard (Centre
D'Etudes Menard, 1975) proposes that the best procedure is to make a calculation for the
settlement (using Equation 2‐198 or Equation 2‐199) of a soil that is assumed to be
homogeneous, and then to add the settlement corresponding to the soft layer. The
settlement corresponding to the soft layer can be estimated from:
1 1
∆ 2‐209
As shown in Figure 2‐166, parameters with subscript c in Equation 2‐209 pertain to the
compressible layer. Hc is the thickness of the compressible layer and Δqc is the vertical stress
increase at the compressible layer.
Figure 2‐166: Compressible layer embedded between two layers (Centre D'Etudes Menard, 1975)
When footings are closely spaced, the stress bulbs overlap, and a more complex method of
settlement calculation is required. As an example, Figure 2‐167 shows a building that is
supporting three basement walls. The shallow footing pressure is qS, and the mean uniform
pressure under the building foot print is qR. The stress distribution due to the loads can be
divided into four parts (Centre D'Etudes Menard, 1975), i.e.:
A spherical component of intensity qR in the zone AR.
A deviatoric component of intensity qR external to the zone AR.
A spherical component of intensity qS ‐ qR under each footing (AS)
A deviatoric component of intensity qS ‐ qR external to zones AS
228
The settlement of the building is then given by:
2‐210
The first and last two terms of Equation 2‐210 are respectively calculated for a bearing area
equal to the building’s foot print and t that of the footing.
Figure 2‐167: Settlement calculation for footings with overlapping stress bulbs (Centre D'Etudes
Menard, 1975)
2.9.2.8 Self‐Bearing
Natural unconsolidated soils and young fills will undergo large settlements with the passage
of time under self‐weight (Centre D'Etudes Menard, 1975). The rate of settlement can be
accelerated by wetting, vibrations, or any phenomenon that tends to temporarily reduce the
shearing resistance between grain contact points, thereby allowing the soil to settle into a
more compact state.
The self‐bearing condition of a soil; i.e., the level of soil parameters that a soil must have so
as not to settle under its own weight, can be related either to physical or mechanical
properties of the soil. Experience suggests that pressuremeter limit pressure is a suitable
characteristic (Centre D'Etudes Menard, 1975). Table 2‐25 presents the required limit
pressure for soils with less than 10 m depth to achieve the condition of self‐bearing. The
influence of external random phenomena such as water table variation, seasonal moisture
content change, road traffic vibrations, etc. have been considered in this table. For depths
greater than 10 m, P*LM values given in Table 2‐25 should be increased.
229
Soil type P*LM (kPa) for self‐bearing condition
Clay 250 to 300
Silt 400
Sand 600
Sand & gravel 800
Soils with organic content 20% increase in P*LM for 1% organic content
Table 2‐25: P*LM (kPa) for self‐bearing condition (Centre D'Etudes Menard, 1975)
According to Menard (Centre D'Etudes Menard, 1975) the self‐weight settlement of a 10 m
thick soft clay layer with , P*LM= 100 kPa will be 1 cm per year. As a first approximation the
one year settlement of any soft layer can be estimated as a function of its limit pressure (in
kPa units) by using Equation 2‐211:
200 cm 1 200
For kPa 2‐211
1000
200
Refer to Table 2‐22 for the values of α. As previously noted, self‐weight settlement rate will
increase due to external causes.
2.9.2.9.1 Moduli
The settlement calculation method described in Section 2.9.2.7 is a valuable tool for
estimating shallow footing settlement; however, there are certain conditions that may
require alternative calculation methods; e.g. application of numerical methods or estimating
settlement of footings with relatively large dimensions compared to the thickness of the
compressible soil. In these two examples, the deformation parameters that are required are
respectively the Young and oedometer moduli.
The relationship between Young and oedometer moduli is:
1
2‐212
1 1 2
230
All experts agree that Menard has determined a relationship as
2‐213
However, the ambiguity appears to be in what E is. In his most renowned publication,
Menard (Centre D'Etudes Menard, 1975) himself notes that EM is a distortion modulus of the
soil measured in a deviatoric stress field, which characterises the pseudo‐elastic phase of the
pressuremeter test. This parameter must not be confused with the oedometer modulus, Eoed,
which is measured in an isotropic or spherical stress condition. Menard further adds that
precise experimental relationships exist between EM and Eoed, and refers the relationship to
later parts of his publication. This would suggest that the correct form of Equation 2‐213 is:
2‐214
Senior engineers, such as Mr Serge Varaksin, who have worked with Louis Menard for years
agree with the form of Equation 2‐214. While implying Menard and Rousseau (1962) as
reference, Amar et al. (1991) have also used the same equation in their publication (The
author notes that he has not been able to find such a relationship in Menard and Rousseau’s
paper, but this could be due to the limited knowledge of the author with the French
language).
On the other hand, Baguelin et al. (1978), who cite Menard (Centre D'Etudes Menard, 1975),
and Leblanc (1982) suggest that the correct form of Equation 2‐213 is as presented in
Equation 2‐215.
2‐215
As mentioned above, the author has not found such a relationship in Menard’s 1975
publication. Baguelin et al. also note that the compression modulus, KM, can be used instead
of EM in Equation 2‐206 by using Equation 2‐216, and define the volume change or
compression (also known as bulk) modulus, K, as indicated in Equation 2‐217
231
2‐216
2‐217
3 1 2
Gambin (1979), also suggest that α relates EM to Ey.
A review of other less known Menard publications is able to clarify the more correct form of
Equation 2‐213 is Equation 2‐214. In his earlier publications (Menard, 1963b), instead of Ed
and Ec, Menard initially formulates settlement using a deviatoric modulus that is measured
by the pressuremeter test and a spherical modulus that is measured by the oedometer or
deducted from the PMT. Later (Menard et al., 1964) he clarifies the spherical modulus, E+,
as:
2‐218
Dauvisis and Menard (1964) then define E+ as the compression modulus, and use
Equation 2‐218 to calculate the second term of Equations 2‐198 and 2‐199. More clearly,
Menard and Lambert (1966) defines α/Ec as the inverse of the oedometer modulus that can
be calculated from the PMT modulus and α in the zone where the constraints tensor is
predominantly spherical or isotropic.
Baguelin et al. (1978) note that theoretical solutions (Bishop et al., 1945, Hill, 1950, Salencon,
1966) based on ideal elastic‐plastic assumptions reduce to Equation 2‐219 for an undrained
condition with Poisson ratio equaling 0.5, and add that Menard (Etude Pressiometrique Louis
Menard, 1967) suggests that the undrained shear strength, cu, can be estimated using
Equation 2‐220.
2‐219
1 ln
3
232
∗
2‐220
1 ln
3
Both Equation 2‐219 and Equation 2‐220 can be written in the form of:
∗
2‐221
Baguelin et al. comment that typical values of E/cu for clay might range from 200 to 2000
(D'Appolonia et al., 1971), which will result in values of 5.2 to 7.5 for β. Briaud (1992)
comments that in fact β depends on the ratio of shear modulus to the undrained shear
strength, and with this ratio being from about 100 to 600, β will be from 5.6 to 7.4 and 6.5
on average. Menard (Etude Pressiometrique Louis Menard, 1970, Menard, 1965) proposes
that the ultimate or residual strength of a clay can be computed using Equation2‐222:
∗
2‐222
5.5
Amar et al. (1991) cite Amar and Jezequel (1972) for limiting the application of Equation2‐222
to conditions in which net limit pressure, P*LM, is less than 300 kPa, and proposing
Equation 2‐223 when net limit pressure is greater than 300 kPa.
∗
25 kPa 2‐223
10
Baguelin et al. (1978) cite Menard (Etude Pressiometrique Louis Menard, 1970) for proposing
Equation 2‐224 and publishing Figure 2‐168 for estimating the drained friction angle, ϕ’, from
the net limit pressure.
∗
2.5 2 2‐224
233
Figure 2‐168: Graph for estimating drained friction angle from net limit pressure (Etude
Pressiometrique Louis Menard, 1970)
Baguelin et al. also note that Muller (1970) quotes the same relationship in a slightly different
form:
∗
2 2‐225
b= 1.8 for homogeneous, wet soil
b= 2.5 on average
b= 3.5 for dry, heterogeneous soil
Baguelin et al. (1978) have reviewed, and interpreted a number of CPT‐PMT correlations such
as those published by Van Wambeke (1962), Cassan (1968, 1969), Jezequel et al. (1968) and
Nazaret (Nazaret, 1972) that were originally printed in French publications. Baguelin et al.
note that while most correlations in technical publications are based on the ratio of qc/PLM,
in spite of introducing uncertainties, the ratio of net values q*c/P*LM would be more
representative. q*c is the net CPT cone resistance, and can be calculated from:
∗
2‐226
234
qo is the total vertical stress. In general q*c/P*LM and qc/PLM are close because qo and Po are
small compared to qc and PLM, but can be quite different at depth in soft clays.
Jezequel et al (1968) studied the influence of depth on q*c/P*LM at the hydraulic fill dikes of
a tidal power project in Rance, France. The fill used was composed of clean sand with dry
density equal to 1.5 t/m3. q*c/P*LM in the upper 1.5 m layer of fill was from 9.11 to 12.03.
Even though qc varied from 2 to 10 MPa, q*c/P*LM was about 6.7 throughout the remainder
of the 20 m thick fill.
Nazaret (1972) did not observe the same independency of q*c/P*LM from q*c in his study on
Loire sand, and reports a tendency of the ratio to increase with the increase of q*c.
Baguelin et al. interpret that the high values of q*c/P*LM near the ground surface are due to
the differences between shallow and deep failure conditions. CPT has a small diameter, and
rapidly reaches its critical depth. However, PMT has to reach an embedment depth of about
1.5 m (1 m in clays, 2 m in sands) before the test is no longer influenced by the surface of the
ground.
According to Baguelin et al. soil type has the greatest effect on q*c/P*LM, and for depths of
about 5 to 20 m there seems to be a narrow correlation between q*c and P*LM. Baguelin et
al. consider that reasonable averages of q*c/P*LM can be considered to be as presented in
Table 2‐26.
Soil Description q*c/P*LM
Very soft to soft clays close to 1 or from 2.5 to 3.5
Firm to very stiff clay from 2.5 to 3.5
Very stiff to hard clay from 3 to 4
Very loose to loose sand and compressible silt from 1 to 1.5 and from 3 to 4
Compact silt from 3 to 5
Sand and gravel from 5 to 12
Table 2‐26: q*c/P*LM for different soil types according to Baguelin et al. (1978)
Baguelin et al. understand that it is very likely that dilatancy is a key factor in sands and
gravels, and q*c/P*LM could prove to be a reliable indicator of the importance of dilatancy in
the resistance of a particular soil. They conceive that a soil is probably non‐dilatant or slightly
dilatant if q*c/P*LM is about 5 to 6, and a ratio of 8 to 12 probably suggests a soil that is
probably dilatant.
235
Campanella et al. (1979) also performed a study on the plastic silt and silty clay fluvial
deposits of the Fraser River delta at Sea Island, Vancouver. Their study showed that qc/PLM is
approximately 2.1 to 4 in the plastic silts, which is of the same magnitude as what Baguelin
et al. had concluded.
Based on theoretical and experimental studies, Van Wieringen (1982) proposed that qc can
be correlated to PLM using Equations 2‐227 and 2‐228.
For clays 3 2‐227
For sands . 2‐228
15 tan ′
Briaud et al. (1985) collected 82 PMT borings data from various projects from 1978 to 1985,
and proposed the correlations of Table 2‐27 (Briaud, 1992).
Based on research (Cassan, 1968, Cassan, 1969, Hobbs and Dixon, 1969, Waschkowski, 1974,
Waschkowski, 1976), carried out on Leucate and Dunkirk sands, Devonian marl of
Monmouthshire, and silty sands of Blois region, Baguelin et al. (1978) observed a large scatter
in the ratio of SPT blow counts, N, to PLM, and described a major reason for this scatter being
the fact that reaming was not used. They concluded that N/PLM, with limit pressure in kPa,
varies from 1/50 to 1/20, and as a provisional recommendation propose:
1
For sands in kPa 2‐229
50
236
Using Meyerhoff’s correlation qc/N= 400, with cone resistance in kPa, Equation 2‐229 will
yield qc/PLM= 8, which is within the range proposed in Table 2‐26. In view of the large scatter
of N in clay, Baguelin et al. do not propose any correlation.
Briaud (1992) proposes that in sands:
1
in kPa 2‐230
47.9
1
in kPa 2‐231
383
More recently, Bozbey and Togrul (2010) have proposed Equations to for correlating PMT
parameters to SPT blow counts corrected to 60% efficiency, N60, during a case study in
Istanbul.
For sands . 2‐232
330 in kPa , 0.74
For sands . 2‐233
1330 in kPa , 0.82
For clays . 2‐234
260 in kPa , 0.67
For clays . 2‐235
1610 in kPa , 0.7
Currently, dynamic compaction quality control and verification is performed as spot checks
at a certain frequency (refer to Section 2.9.2.4). Observation of ground behaviour during
impact, and the crater sizes can also be of great assistance in identifying soft or hard spots or
zones. It could be expected that quality control would greatly improve, if the observational
procedure was to be integrated into the verification process using a continuous monitoring
mechanism.
Based on roller integrated compaction metres that continuously measure the acceleration of
the roller drum and calculate a roller measuring value from the acceleration signal,
237
Continuous Compaction Control (CCC) measuring technology was initiated in Europe in the
1970s for use on vibratory rollers compacting granular material (Briaud and Saez, 2012,
Forssblad, 1980, Thurner and Sandström, 1980, White and Vennapusa, 2010). For vibratory
rollers, an accelerometer that is mounted to the roller drum provides spatial records of
compaction quality when linked to position measurements and a documentation system. An
alternative to accelerometer‐based vibratory measurements is the measurement of rolling
resistance or machine drive power, and can be applied to both vibratory and non‐vibratory
roller operations. When the measurement system provides Automatic Feedback Control
(AFC) for roller vibration amplitude or frequency, it is referred to as Intelligent Compaction
(IC).
Roller measurement values calculated based on accelerometer measurements use one of
two different approaches; i.e., calculation of a ratio of selected frequency harmonics for a
set time interval, or calculation of ground stiffness or elastic modulus based on a
drum‐ground interaction model and some assumptions. Regardless of the technology, the
basis of CCC and IC is that the measurement values are related to traditional compaction
measurements, and will be useful as part of effective earthwork compaction operations and
quality control and quality assurance practices (White and Vennapusa, 2010).
White and Vennapusa (2010) note that the below measurement technologies have been
developed for roller compaction methods:
Compaction Meter Value (CMV) (Sandström, 1994)
Oscillometer Value (OMV) (Thurner and Sandström, 2000)
Compaction Control Value (CCV) (Scherocman et al., 2007)
Roller‐Integrated Stiffness (ks) (Anderegg and Kaufmann, 2004)
Omega Value (Kröber et al., 2001)
Vibratory Modulus (EVIB) Value (Kröber et al., 2001)
Similar measurement systems and technologies are not available for dynamic compaction
CCC, but research in this field is under way, and Kopf and Paulmichl (2004) have developed a
concept (published in German) that has been cited in several publications (Adam and Brandl,
2009, Adam et al., 2007, Kopf et al., 2010).
Pounder acceleration during impact can be measured to determine the reaction forces of the
ground; however, these reaction forces are influenced by soil compaction, replacement,
238
liquefaction, excessive pore water pressures, local ground failure, plastic and elastic
deformations, etc. Consequently, Kopf and Paulmichl consider that the reaction force is not
suitable as a clear characteristic value required for a reliable compaction control. However,
in their opinion, the decay of free soil vibrations caused by the falling pounder after each
impact is characteristic of the soil‐pounder interaction, and enables a site‐specific
optimisation and quality control of dynamic compaction. The decaying pounder acceleration
after impact can be seen in the measured example of Figure 2‐169.
Figure 2‐169: Acceleration of pounder after impact (Adam et al., 2007)
The idea of Kopf and Pualmichl’s (2004) concept is to create a correlation between measured
vibration parameters and soil parameters of an idealised linear‐elastic half space with the
assumption that an elastic decay of free soil vibrations under still (possibly meant as constant
and without change) increased pore water pressures can be modelled and solved using a
theoretical analysis similar to a viscously damped single degree of freedom system (SDOF).
Consequently, measuring the acceleration of the falling pounder during the decay of free soil
vibrations will provide the damped frequency, ωd, and Lehr’s damping coefficient, ζ, if a
viscously damped SDOF system is assumed.
During dynamic compaction the pounder penetrates deeper and deeper into the ground
(refer to Section 2.4.6 to 2.4.9). This changes the dynamic conditions of the system
depending. In order to quantify the specific effects, extensive parameter studies were
performed using the boundary element method, which only requires the discretisation of the
boundary (Banerjee, 1994), and is particularly suitable for linear‐elastic half space problems.
Kopf and Pualmichl note that the assumption of linear‐elastic soil behaviour is applicable and
with sufficient accuracy because the decay of the vibration after impact behaves as an elastic
239
problem. The pounder‐half space system was simulated in the frequency domain using a
rotational‐symmetric model. The penetration depth, Dc, is variable, and using the SDOF
analogy, an analytic approximate solution exists for the case in which the pounder has not
penetrated the ground. Approximations are also available for the spring and damping
coefficient when the depth of penetration is greater than zero, but they are limited to
geometrical properties and the penetration depth. Therefore, boundary element method
numerical simulations that did not consider material damping were carried out to analyse
the decay of free soil vibrations.
Numerical calculations suggested that Lehr’s damping coefficient derived from the SDOF
analogy depended only on Poisson’s ratio, was practically independent of the deformation
modulus of the half space, and varied with depth for all penetration depths. The influence of
the pounder penetration depth, i.e., the crater depth, can be calculated by a reduction factor
κ, which is a function of Poisson’s ratio.
In summary, the steps for deriving the soil parameters from the measured vibration
parameters are:
1. Determination of the damped natural frequency, ωd/2π, and the Lehr’s damping
coefficient, ζ, from the post impact decaying free vibration parameters (period, T,
and consecutive maximum wave amplitudes Zn and Zn+1, refer to Figure 2‐170 ). ωd
the logarithmic damping coefficient, δ, are respectively calculated from the decaying
free wave using Equations 2‐1 and 2‐237. As can be seen be seen in Figure 2‐171, in
reality the ground is not a linear‐elastic half space with negligible material damping,
and the actual decaying acceleration wave period and logarithmic damping
coefficient may vary in the cycles. ζ and the (undamped) natural frequency, ω, are
respectively calculated from Equations 2‐238 and 2‐239.
2 / 2‐236
ln 2‐237
240
Figure 2‐170: (Idealised) decaying wave (Adam et al., 2007)
Figure 2‐171: Decaying acceleration wave measured after pounder impact (Adam et al., 2007)
2‐238
2 1
2
2‐239
1
2. Estimation of Poisson’s ratio, ν, from the chart of Figure 2‐172, and by using ζ and
the actual pounder penetration depth. It is noted that this chart has been developed
based on a pounder with diameter and weight being respectively 1.8 m and 16.5 t.
241
The Young modulus and soil density used were respectively 16 MPa and 2 t/m3, and
no damping was included in the calculations.
Figure 2‐172: Chart for estimation of Poisson’s ratio (Adam et al., 2007)
3. The (uncorrected) half space modulus is determined from the chart of Figure 2‐173
based on the natural frequency (reciprocal of ω) and Poisson’s ratio for the case of
no pounder penetration into the ground (Dc= 0). It is noted that, as before, this chart
has been developed based on a pounder with diameter and weight being
respectively 1.8 m and 16.5 t, and no damping.
Figure 2‐173: Chart for estimation of Young modulus for the case of Dc= 0 (Adam et al., 2007)
242
4. The corrected Young modulus is calculated by multiplying the uncorrected modulus
derived from of the previous step by κ, which is estimated from the chart of
Figure 2‐174. This chart has been developed based on a pounder with diameter and
weight being respectively 1.8 m and 16.5 t. The Young modulus and soil density used
were respectively 16 MPa and 2 t/m3, and no damping was included in the
calculations.
Figure 2‐174: Chart for estimation of κ (Adam et al., 2007)
The method presented above appears to be very promising; however, the impact of some
assumptions used in this analogy, such as assuming zero material damping, requires further
consideration. Also, the ground is assumed to be a linear elastic half space, and the
deformation modulus that is derived applies to the whole of the half space mass. This is
obviously not compatible with reality, and it is well established that the effect of
improvement diminishes with depth.
243
2.10 Special Topics
It is the experience of the author that a number of questions are frequently asked by clients,
consultants, associates and engineers who are considering the application of ground
improvement in general and dynamic compaction or dynamic replacement in particular.
These topics are reviewed in this section.
Ground improvement is generally taken into consideration after it has been established by
some means that the in‐situ ground conditions are not able to satisfy the project
requirements. It will then be necessary to develop the ground treatment specification and
consequently the acceptance criteria, which can have numerous forms.
It is the objective of this section to review and to compare a number of possible and common
methods of how soil improvement specifications are prepared and stipulated. Some
specifications and acceptance criteria potentially inherit more risk and cost than others, and
the outcome of a project may be determined by the quality of specifications and criteria
rather than the success of the works.
Sometimes, prior to the tender and the award of the contract to a tenderer, ground
improvement specifications are developed in full detail by the party who has prepared the
tender documents. In such a case, based on the geotechnical advisor’s design, which is
usually not made available, a ground improvement technique is specified, the scope of work
to be performed is described in detail, and the construction method is outlined.
The responsibility of the contractor will usually include the procurement of the working team,
equipment, material and execution of the works according to the specification’s instructions.
In this type of project acceptance criteria is generally non‐technical and based on performing
the works according to the specified dimensions, spacing, quality and quantity of materials
that have been used.
Testing may be specified, but it will not be (or at least in a fair contract it should not be) the
contractor’s responsibility to meet any specific value as other parties have been engaged in
244
the development of the methodology and design of ground improvement. Should test results
fail to meet expectations, then the contractor will be required to perform additional work
using the same technology or an alternative ground improvement method, and would be
fully paid for any extra work if the works were carried out properly in the first place,.
As an example, the ground improvement method in the specification may be stipulated to
be dynamic replacement. The column spacing, number of phases, number of blows per print,
pounder weight and drop height, and execution method will all be specified, and the
contractor will have to provide the working team, DR rig, pounders, other equipment, and
material, and perform the works according to what has been described in the specifications.
Even if tests carried out after the works demonstrate that the technical requirements of the
project have not been satisfied; e.g. achieving the required DR column diameter or
penetration depth, should the contractor have performed the works correctly, he will be
exempted from all other responsibilities and consequences. Correction of the works to meet
the project’s requirements will be the responsibility of others.
At first glance this approach and development of acceptance criteria may appear to be
favourable. Indeed this is the way that many projects are performed in most engineering
fields, and under the influence of other disciplines, ground improvement is sometimes
carried out in a similar manner.
Although it is acknowledgeable that developing a design by an experienced geotechnical
engineer will provide certain amounts of confidence, especially in the absence of specialised
ground improvement contractors, due to the history of development of ground improvement
technologies and the heterogeneity of the soil, in the presence of specialist ground
improvement contractors, this method possesses a number of major draw backs that
inevitably deviate the project’s path from an optimised approach.
It should not be forgotten that academia and consulting engineers have been the front
runners of many engineering disciplines, and consequently it is not surprising that they drive
the concepts of those fields. However, if we look back at the histories of ground improvement
techniques, it can be observed that almost all have been developed, advanced or patented
by contracting companies. For example, further to the invention of jet grouting by the
Japanese in the early 1970s (Ichese et al., 1971), almost all major further developments and
supplementary innovations that have been patented in the United States are by specialist
contracting companies (Shibazaki and Yoshida, 1995, Tokoro et al., 1984, Yahiro and Yoshida,
245
1977). Later, research by academia has provided the theoretical background and
understanding of the method; e.g. the theoretical modelling of jet grouting (Modoni et al.,
2006). The same concept is true with other ground improvement techniques; i.e., dynamic
compaction, dynamic replacement, Menard vacuum, controlled modulus columns
(Communication Department of Menard, 2007), vibro compaction and stone columns (Better
Ground), and many other techniques.
In most cases, eliminating the input of specialist contractors saves the project neither time
nor money. As much as an academician or geotechnical consultant may be up to date with
the latest ground improvement developments, it is always possible that specialist contractors
have further advanced a technique or developed a new idea that can bring savings to the
project.
While most tenders accept alternatives, it still appears to be a waste to allocate time and
money to prepare a detailed design and specification of ground improvement, only to replace
it with an alternative in the final stage. It is the author’s belief, in the presence of specialist
ground improvement contractors, that identifying the ground’s problems, developing a
concept based on budget, time, environmental and other constraints, and allowing the
industry to propose an optimised solution will more readily meet the interests of the project.
Stipulating non‐technical acceptance criteria results in success only if the anticipated
methodology works perfectly and when there are no problems. However, serious problems
may arise if things do not work out well. Remedial solutions may not be possible with on‐site
capacity, and the contractor may not have the technology or resources to execute an
alternative solution. Furthermore, even if there are no latent conditions, any rectification of
the ground improvement solution beyond the design will immediately become a cost to the
owner and most probably a delay in the construction programme.
It is definitely better for acceptance criteria to be a technical requirement rather than tied to
the quality and quantity of performed work. There is truly no justification to allocate the
project’s financial resources and time to developing a detailed solution that may be subject
to major and fundamental changes and derivations during later stages. It seems best for the
project’s consultant to identify the need for ground improvement, to identify feasible
solutions that would be able to satisfy the technical requirements, to provide a realistic
budget for planning purposes, to prepare a specification for ground improvement works that
246
is open to solutions that are able to satisfy technical requirements and to supervise the works
in a manner that will verify that technical acceptance criteria have been satisfied.
A better approach for developing ground improvement specification and acceptance criteria
would be when the criteria are based on technical performance rather than complying with
a rigid methodology.
In such cases, acceptance criteria are sometimes stipulated to be a minimum test value (or
envelope) or a correlation to the test value. Sometimes the specified testing method is
impractical. For example, it may be stipulated to carry out in‐situ density tests for the ground
improvement works of a road that is to be constructed on a thick reclamation. Although this
testing method may be appropriate for thin layers of roller compacted soil in a typical road
construction project and it is not impossible to obtain undisturbed samples from deep
depths, questions can and should be raised that is the envisaged test method the appropriate
procedure for the specific project conditions, and will it be able to suitably, technically,
practically and economically demonstrate that design requirements have been satisfied.
Possibly being aware of the above discussion, sometimes the specifications supposedly
facilitate the application of an impractical testing method by correlating it to a practical and
more suited testing procedure. For example the specifications may stipulate a minimum in‐
situ density for a thick reclamation project, but as it is impractical to perform such tests, the
density acceptance criterion may be correlated to CPT cone resistance. One should wonder
about the objective of going through all this trouble to develop a correlation that is simply a
best fit curve with some values for the standard deviation and coefficient of determination
while it would seem more rational to develop the acceptance criteria simply based on the
cone resistance without correlating to density at all. This example demonstrates how the
background and mentality of engineers who have gained expertise in other fields may deviate
ground improvement acceptance criteria to what is more acceptable in the other field.
Although quite often referenced in many publications, relative density is also a very
unsuitable acceptance criterion that has found its way into ground improvement projects,
especially in reclaimed sites. Due to the importance of this topic, Section 2.10.2 “Unreliability
and Unsuitability of Relative Density as a Ground Improvement Criterion” has been allocated
to this subject.
247
Quite commonly, ground improvement acceptance criteria are based on minimum (or an
envelope of minimum) passing values of practical, efficient and economical field tests such
as SPT, CPT or PMT. This, itself, is an important step forward as it recognises that establishing
acceptance criteria based on direct measurement of parameters is much more rational and
beneficial than relying on redundant correlations. However, there is probably still another
step that must be taken forward.
The question that should now be posed is on the mechanism of determining the minimum
passing values. Those figures are most probably based on a calculation that has been carried
out by the geotechnical engineer to ensure certain design requirements such as bearing
capacity, total and differential settlements, liquefaction mitigation or long term
consolidation have been satisfied. However, we must note that a condition in which the soil
layers in the test just reach the minimum passing value that have been determined by the
geotechnical engineer is only one possibility of the hundreds of results that can still satisfy
the design criterion.
Certainly in ground improvement techniques such as dynamic replacement, stone columns,
jet grouting, deep soil mixing and controlled modulus columns, in which inclusions are
introduced into the soil and vertical loads are distributed between the in‐situ soil and the
inclusions by arching the concept of minimum passing value becomes meaningless. Indeed,
the in‐situ soil improves negligibly, and it is only the added elements that improve the soil
mass.
It is evident that this type of acceptance criteria rules out the techniques that may be most
applicable. Of course, the specification may be written to stipulate average minimum passing
values, but that does not cover all aspects either. For example, an average minimum value
will still not be able to address the issue of column bulging in very soft saturated clays. The
cycle of adding clauses to the specification can go on and on.
Even if the ground improvement technique; e.g. dynamic compaction or vibro compaction,
does not incorporate inclusions, most probably there will be test results that are well above
the minimum passing value. A minimum passing criterion neglects the benefits of the
additional improvement, and does not take it into consideration. However, a recalculation
may very well be able to demonstrate that design requirements can still be satisfied what all
layers actually passing the specified minimum acceptance value.
248
One may argue that in the minimum passing value criterion provides a higher safety factor
and is on the conservative side, and ignores the benefits of layers with better than required
properties. As true as this argument is, adding more and more safety factors well beyond
what is considered as sound engineering practices is poor engineering practice that will
inevitably lead to more effort, energy, construction time and costs. Highly conservative
engineering is not an achievement. Today, we do have enough knowledge about ground
improvement technologies to be able to design soil treatment without resorting to such out
dated procedures.
Acceptance criteria based on minimum passing values may unjustifiably and unnecessarily
lead to non‐conformance while still meeting the foundation design requirements, and it is
the opinion of the author that acceptance must be developed in a more appropriate manner.
Having discussed acceptance criteria that are usually unable provide the best performance
with minimum cost and time requirements, it can now be proposed that optimal acceptance
criteria should be directly based on design criteria. Indeed, there is no better way of making
sure that a certain aspect of design has been fulfilled than directly verifying that specific
criterion itself.
Relevant tests should be carried out simply to verify that ground conditions after soil
improvement are able to satisfy the design requirements. In reality there is no need to add
proxies to the specifications by introducing minimum passing values or correlations. The
foundation design requires certain conditions to be met; e.g. achieving a defined bearing
capacity subject to a limit on the total and differential settlements.
Of course, tests and measurements must be made to demonstrate that design criteria have
been satisfied, so the testing programme and interpretation method should be clearly
defined.
With this information, the specialist contractor will know exactly what is expected, and will
be able to offer his best proposal. If the contractor believes that alternatives to the
specifications can benefit the project, he should clearly propose the alternative offer based
on alternative design criteria or verification testing. The designer can then assess the
proposals, and select the best proposal with minimal complications during construction.
249
2.10.2 Unreliability and Unsuitability of Relative Density as a Ground
Improvement Criterion
2.10.2.1 History
The concept of relative density (Dd) was first introduced by Terzaghi (1925) to bring the
behaviour characteristics of soils together on a common basis in consistent and practically
useful relations and to provide a tool for communications between engineers (Burmister,
1948). It was suggested that this parameter would be an appropriate means to define the
looseness and denseness of sand or sand‐gravel soils in a meaningful way because important
properties were assumed to correlate quite well by this means.
Relative density is a definition rather than an inherent property of the soil. Per se, it has no
significance and influence on performance and compared it may be possible to satisfy design
criteria with a higher safety factor without even complying with the relative density criterion.
In fact, confusion in the use of relative density began as soon as engineers started utilising it
as a soil parameter because there was no common definition or set of standards to work
from (Holtz, 1973b). Recognising that some kind of standard would be necessary, Section D,
Subcommittee 3 of Committee D‐18 of ASTM was established in 1954 for determining the
minimum and maximum densities of sand and gravel soils. The work of the Subcommittee
resulted in ASTM D2049‐69 (ASTM, 1969), which was approved by Committee D‐18 in 1964.
It was the opinion of the majority of Subcommittee members that the benefits of a
reasonably good standard method far out weighted any disadvantages of the particular test
that was then proposed (Holtz, 1973b). In addition, it was felt that by getting a tentative
standard published, many persons and organisations would work with it, evaluate its
suitability, and suggest beneficial modifications and improvements.
The introduction of relative density as a criterion for acceptance whereas it is compared with
a critical value to give a yes or no type of answer has led to considerable amount of research,
most of which point towards the insufficiency of its accuracy and reliability, and it is the
opinion of the author that although still popular and systematically used by those engineers
who have never looked deeply into this concept, the truth is that after years of having had
its suitability evaluated, relative density has not fulfilled expectations and should be
abandoned as a criterion altogether. That seems to be already under way. There was a time
when liquefaction analysis was found on the concept of relative density (Seed and Idriss,
250
1967, Seed and Idriss, 1971); however, today while the same methodology has been
retained, relative density has lost colour and is no longer part of liquefaction evaluation
procedure (Youd et al., 2001), and in the process of re‐evaluating the concept of relative
density, as proposed by Selig and Ladd (1973), ASTM withdrew its D2049‐69 standard for
measuring relative density as early as 1984, and has replaced it with the more meaningful
standards for measuring maximum and minimum index densities (ASTM, 2006a, ASTM,
2006b). In its new standards, ASTM (2006a) cautiously states that it is generally recognised
that either relative density or per cent compaction is a good indicator of the state of
compactness of a given soil mass. However, the engineering properties, such as strength,
compressibility, and permeability of a given soil, compacted by various methods to a given
state of compactness can vary considerably. Therefore, considerable engineering judgment
must be used in relating the engineering properties of soil to the state of compactness.
Mayne (2006) cautiously notes that in his opinion relative density is a rather weak parameter.
Likewise, Bowles (1996) states that in his opinion the relative density test is not of much value
since it is difficult to obtain maximum and minimum unit weight values within a range of
about ±0.5 kN/m3.
2.10.2.2.1 The Definition of Relative Density and its Intended Range of Application
ASTM (2006a, 2006b) defines relative density a ratio, expressed as a percentage, of the
difference between the maximum index void ratio and any given void ratio of a cohesionless,
free‐draining soil; to the difference between its maximum and minimum index void ratios,
or:
100 2‐240
emax= maximum index void ratio or the reference void ratio of a soil at the minimum index
density/unit weight.
emin= minimum index void ratio or the reference void ratio of a soil at the maximum index
density/unit weight.
e= the in situ or stated void ratio of a soil deposit or fill.
251
Equation 2‐8 can also be expressed in terms of maximum and minimum indexes and dry
densities or dry unit weights (ASTM, 2006a, ASTM, 2006b) as formulated in Equation 2‐241
and Equation 2‐242.
100 2‐241
100 2‐242
dmax or dmax = the reference dry density/unit weight of a soil in the densest state of
compactness that can be attained using a standard laboratory compaction procedure that
minimizes particle segregation and breakdown.
dmin or dmin = the reference dry density/unit weight of a soil in a standard state of
compactness at which it can be placed using a standard laboratory procedure which prevents
bulking and minimizes particle segregation.
d or d = dry density/unit weight of a soil deposit or fill at the given void ratio.
ASTM makes a number of statements that deserve review and consideration. Firstly, the
applications of ASTM testing methods are conditioned to a range of selected soils, or in other
words, the concept of relative density is not applicable to all other soils. According to ASTM
(2006a, 2006b), for its testing methods to be applicable, the soil can contain up to 15%, by
dry mass, of soil particles passing a 75‐μm sieve, provided they still have cohesionless, free‐
draining characteristics. Also, for determination of minimum index density and unit weight
of soils, the three accepted methods are applicable to soil in which 100%, by dry mass, of soil
particles pass respectively a 75 mm, 19 mm and 9.5 mm sieves. In the first method the soil
may contain up to 30%, by dry mass, of soil particles retained on a 37.5 mm sieve, and the
third method is applicable only to fine and medium sands which may contain up to 10%, by
dry mass, of soil particles retained on a 2 mm sieve (ASTM, 2006b). To summarize, as shown
in Figure 2‐175, at best, relative density is applicable only to soils with less than 15% of the
material being silt or clay and can have gravel up to a diameter of 75 mm provided that 70%
of the grain size are less than 37.5 mm and the soil is still cohesionless and free draining.
252
Figure 2‐175: Applicable range of relative density (non‐shaded areas) as defined by ASTM
ASTM has rightfully introduced the term free draining even though the applicable zone of
relative density is for sands and gravels. As much as sands and gravels are assumed to be free
draining, the author have come across non‐free draining sands with less than 15% fines
content, but containing high amounts of clay. For example in the dynamic compaction project
of Salam Resort (unpublished), on the seashore of Bahrain, sieve analyses showed that fines
content of the soil was less than 20% and the soil was identified as sand; however clay
content was relatively higher than normally encountered and approximately 10%. When
ground improvement was carried out on the site it was observed that the soil was not
draining, as would be expected in sand, and reduction of pore water pressure to initial values
required a period of more than one week.
Engineers who have been involved in reclamation projects know very well that even if
specifications limit the soil’s grain sizes to the limits defined by ASTM, there is still a good
chance that due to segregation of material or change of material source, layers or bands of
soil with higher fines content will be realised here and there. On the other end, there are
many situations where the soil contains larger cobbles and boulders.
When developing a specification it would seem rational to stipulate acceptance criteria that
will most likely not require reassessment and revision during the project. In the author’s
opinion relative density does not comply with this statement, as in the event that further
supplementary investigations prove that the soil is non‐free draining, then the specifications
253
will not apply to the soil any longer and alternative criteria must be sought. It may be simply
a more reasonable approach to stipulate the alternative specification in the first place.
If the over and undersize percentage of materials are identified in advance, the oversize may
be separated or crushed or the material may be mixed to reduce the fines content, but there
are techniques, such as dynamic compaction, that can treat soils containing more fines and
oversized cobbles than allowed by ASTM (for example, refer to the treatment of rock fill by
dynamic compaction at Udevalla, Sweden in Section 2.10.4) and the extra activities and costs
could have been avoided by stipulation of an appropriate acceptance criterion.
If over or undersize materials are detected after reclamation, one ill‐advised solution may be
to remove the soil and to replace it with soil that meets the required grading; however, this
is more easily said than done. Alternatively, there are others who may re‐define the grading
range defined in relative density standards to fit their needs. Of course, these scenarios could
be totally avoided simply by defining a reliable criterion.
For many types of free‐draining, cohesionless soils, ASTM specified test methods cause a
moderate amount of degradation (particle breakdown) of the soil (ASTM, 2006a). When
degradation occurs, typically there is an increase in the maximum index density/unit weight
obtained, and comparable test results may not be obtained when different size moulds are
used to test a given soil. This raises the question of comparability of test results, and as will
be later discussed, any small changes to the limit indexes will result in dramatic changes in
relative density.
Referring to Holtz (1973b), ASTM (2006a, 2006b) also acknowledges that there are published
data to indicate that these test methods have a high degree of variability, but proposes that
the variability can be greatly reduced by careful calibration of equipment. Of course,
calibration of equipment can only reduce systematic errors caused by equipment conditions,
but it will have no effect on random errors. The sources of errors and variations of relative
density testing require further review.
Relative density testing is subject to a combination of systematic errors, random errors and
mistakes. Systematic error is a measure of accuracy and the difference between correct value
and the measured average of a set of repeated tests. Random error is the precision of a
quantity and is measured by the scatter in the results of a group of repeated tests (Selig and
254
Ladd, 1973). The distributions of these errors around the true value are shown in
Figure 2‐176.
Figure 2‐176: Systematic and random errors (Selig and Ladd, 1973)
The reliability of relative density measurements has, for the most part, been taken for
granted by most engineers, but not by all. Yoshimi and Tohno (1973) have assumed a
statistical approach, and note that since relative density is proportional to (d ‐ dmin)/ d,
even a small variation in d or dmin can cause a considerable variation in relative density
when d ‐ dmin is small; i.e., when relative density is low. For example, when dmin= 13.5
kN/m3, dmax= 16.37 kN/m3, and d= 14.25 kN/m3, relative density will be 30%. If dmin is
increased by 1% to 13.635 kN/m3, Dd reduces to 25.8%, which is 14% less than the initial
value. In other words, the relative deviation in relative density is 14 times that of dmin.
Yoshimi and Tohno express the influence of dry unit weight random errors on relative density
using Equation 2‐243:
2‐
243
The terms in the parentheses are coefficients of variation. SDd, Sdmax , Sdmin and Sd are
respectively the standard deviations for Dd, dmax, dmin and d, and Cdmax, Cdmin and Cd are
error propagation factors that are expressed in the forms of Equations 2‐244 to 2‐246.
2‐
244
255
1 1 2‐
1
245
1 2‐
1 1
246
As illustrated in Figure 2‐177, for the case of Cdmax= 4.7 which is a representative value for
clean sand with low uniformity, these equations show that while Cdmax is independent of
relative density, Cdmin and Cd increase as relative density decreases until they reach infinity
when relative density reduces to zero.
Figure 2‐177: Error propagation factors for relative density, redrawn from Yoshimi and Tohno (1973)
Random errors can be reduced to any desired degree by repeating the test and averaging the
results (Selig and Ladd, 1973).For the case when the coefficients of variation are equal, for a
defined coefficient of variation, the number of measurements, m, for each dry unit weight
was calculated by Yoshimi and Tohno (1973) to be as expressed in Equation 2‐247.
2‐247
256
For example when Cdmax=4.7, if Dd= 50%, (SDd/Dd)=5% and coefficient of variation equals to
1% then 7 tests must be made for each of the maximum index, minimum index and specimen
unit weights. For Dd= 22%, with the previous other parameters, 40 tests for each unit weight
will be required. Unfortunately, and even to the acknowledgement of ASTM (2006a, 2006b)
it seems that in practice this recommendation of Yoshimi and Tohno has gone unheard of
and most commonly, regardless of the value of relative density, only one test for each
parameter is carried out.
When studying the influence of dry unit weight systematic errors on relative density, Yoshimi
and Tohno (1973) formulate relative deviation in relative density, ΔDd/Dd, due to a deviation
in any one of the dry unit weights; i.e., Δdmax, Δdmin and Δd, and express them in the forms
of Equations 2‐248 to 2‐250.
∆ ∆ 2‐
∆
∆ 248
∆
∆ 1 ∆ 2‐
∆
∆ 249
∆
∆ ∆ 2‐
∆
∆ 250
∆
For simultaneous errors in dmax,dmin, andd the resulting values of ΔDd/Dd may be added,
keeping in mind the correct signs. In view of relatively large values of Δdmax or Δdmin due to
the systematic errors of the unit weight measurements, it can be understood that ΔDd/Dd
may reach tens of per cent.
Referring to Equations 2‐8 to 2‐242, there are three parameters that must be determined for
calculation of relative density; i.e., emax, min or min that describes the loosest state of the
soil, emin, max or max that describes the densest state of the soil and e, d or d that describes
the in‐situ state of the soil.
Due to the relative magnitude of the maximum index, minimum index and in‐situ dry unit
weights, relative density is computed from the ratio of small differences between large
numbers (Tavenas et al., 1973). This implies that small variations of large numbers will be
257
magnified to produce a great variability in the computed result. The simple application of the
theory of errors led Tavenas and La Rochelle (1970) to conclude that any laboratory
determination of relative density would be affected by a large variability, even if the ASTM
standard method was used.
As part of the study of Tavenas et al. (1973) 87 laboratories in the United States and Canada
participated in grading tests, minimum and maximum index density tests. Some very
important, if not dramatic, conclusions that they were able to draw from their findings
included:
Even though ASTM standard tests for determining the minimum and maximum
(index) density of cohesionless materials was considered as "normally accurate" soil
mechanics tests with observed coefficient of variability of the order of 2.5 per cent
and coefficient of reproducibility of the order of 0.8 per cent, the use of these
parameters in the relative density formula leads to a result of poor quality, since it is
characterised by coefficients of variability of the order of 15 to 40% and by
coefficients of reproducibility of the order of 3 to 15% in most of the usual cases.
Thus, and simply due to the formulation of relative density, the variability is
multiplied by a factor of 10.
The variability of the relative density increases as the maximum grain sizes of the
tested materials increases. Standard deviations were found to be 60 to 100% larger
for gravelly sand than for fine sand. Since the fine sand tested was close to the ideal
material with a small maximum grain size and a coefficient of uniformity of the order
of 3, and no particles passing sieve 0.075 mm, the variability and reproducibility
observed on this material were the best possible with the testing technique. Thus, it
cannot be expected to determine any relative density with a width of the 95%
interval, less than 10% if the results obtained by one technician only are considered,
and less than 40% if the results obtained by different laboratories are analysed. The
results obtained by different operators in the same laboratory fell in between.
A third basic parameter influencing the determination of the relative density is the
actual dry unit weight itself. This parameter is also affected by a certain error. In the
most important case of the measurement of the in situ unit weight, the methods
were such that any value of dry unit weight could not have been defined with an
error less than ±30 N/m3. The error on the in situ dry unit weight had to be combined
258
to the previously discussed variability to give the final variability of the relative
density. As shown in Figure 2‐178, Tavenas et al. demonstrate that the width of the
95% interval for the correlation between the in situ dry density and relative density
of fine sand and gravelly and is respectively 64% and 94%, which is close to the full
range of possible values for the relative density, and assess that under such
circumstances the probability of evaluating the correct relative density by a wild
guess is at least equal to that of measuring it by the standard method.
Figure 2‐178: Error on the in‐situ relative density, redrawn in SI units from Tavenas et al. (1973)
Due to the very large variability of the relative density between laboratories, the
comparison of relative densities measured by different laboratories was totally non‐
significant. There were important practical implications of this fact: all established
correlations between relative density and various properties of cohesionless soils
such as standard penetration index, point resistance in a static penetration test,
friction angle, modulus of compressibility, shear wave velocity, etc., are useless to
anyone but the operator who has established them, since that person is the only one
who can reproduce the relative density of the considered soil with sufficient
accuracy.
259
It appeared that due not so much to the variability of the minimum and maximum
unit weights but essentially to the formulation of relative density itself, the resulting
accuracy of this parameter was so poor that its use was related to major
uncertainties (the best case was of ideal material such as the tested fine sand, and
was deemed to be practically meaningless in most of the other cases).
Similarly, Tiedemann (1973) has conducted research on the results of minimum and
maximum index unit weight tests performed by 14 US Bureau of Reclamation soil
laboratories, and has concluded that based on results obtained from the cooperative test
program and other studies, it appeared that the variations associated with the minimum and
maximum unit weight tests investigated were approximately the same as, or less than, those
associated with the impact‐type compaction test. However, when the results are used to
compute relative density, large variations can occur. For the extreme variations in both the
minimum and maximum density, a dry unit weight of 17.3 kN/m3 could be reported as a
relative density varying from 40 to 76% on a between‐laboratory basis or from 52 to 66% if
the limiting unit weights were determined by a single operator.
One may argue that in reality the chances of the occurrence of the worst case scenarios as
envisaged by Tavenas et al. (1973) and Tiedemann (1973) are quite low. As true as that may
be, these worst case scenarios are just an indication that as unlikely as may be, relative
density criterion could seriously misrepresent the soil conditions to the point of rejecting a
well compacted soil or leading to a total disaster by accepting a loose soil. Of course, there
are all the other non‐worst case scenarios that nevertheless also lead to misrepresented
ground conditions. It would indeed be much more appropriate not to tie a project’s destiny
to a parameter that is referred to as being as accurate as a guess.
Tiedemann also reports that variatons in duplicate tests were less than for duplicate
specimens, indicating that there was considerable variation in the specimen. According to
Holtz (1973b) it is usually more difficult to produce uniform samples and specimens of the
coarse sand and gravel soils than fine silty and clayey soils. The difficulty usually becomes
greater when the soils are cohesionless and as the range of particle sizes increase. This is
primarily due to segregation of the material. Segregation can even effect the uniformity of
sampling when closely controlled quartering and splitting methods are used. The scooping
and placing of cohesionless, coarse‐grained soils in moulds for the standard minimum and
maximum index density tests can cause segregation, which can produce nonhomogeniety of
260
gradation within a test specimen and gradation variations between the so‐called duplicate
specimens.
Similar findings have been also reported by others, such as Gupta and McKeown (1973) who
encountered problems for applying relative density criterion for Kettle Generating Station in
Canada. In the end they concluded that the effects of variations in minimum unit weight on
relative density were startling. They noted that although the variation was decreasing with
the increase of relative density, nevertheless, it still created dilemma for effective quality
control in the field in terms of enforcing the requirements of 75% relative density as
stipulated in the contract.
Gupta and McKeown used four technicians for each test to carry out 10 tests on each soil
sample; thus resulting in 40 tests per sample. For the analysis of the tests, they only used
those tests that fell within the middle 90% spread. Among the numerous batches of 40 tests,
in the best case, relative density at 50% and 75% respectively had errors of ±5% and ±1.6%.
In the worst case, a 50% and 75% relative density respectively had errors of ±26% and ±6.6%.
This means that a sample with relative density of 50% could have been reported as 64% and
a sample with relative density equal to 75% could be reported as 69%. While the true
condition of one sample is loose and the other dense, errors may lead to a condition where
the material may be practically assumed to have similar test values. Obviously, this may lead
to a non‐compliance with project specifications solely due to the inherent defaults of the
choice of specification. Had Gupta and McKeown not limited the tests to the middle 90%
spread, the variations would have been even more concerning.
Reviewing the works of others during the ASTM Symposium for Evaluation of Relative Density
and its Role in Geotechnical Projects Involving Cohesionless Soils at the 75th Annual Meeting,
Selig and Ladd (1973) assess that the error in dmax has the largest effect for high Dd while
error in dmin has the greatest effect for low Dd and add that, in general, relative density has
associated with it a random error of about ±10 to 15 standard deviation and a systematic
error of 25 to 30 range. They conclude that in their opinion relative density has value, but
that it has frequently been overextended with a false sense of reliability, or improperly used.
Ladd and Selig were probably using a mild tone not to upset any supporters of relative
density. If there is any value in relative density, with all that has been discussed, the value is
not in being an acceptance criterion of ground improvement and it should be sought
elsewhere.
261
2.10.2.2.3 4. A Second Glance at the Standards
After this discussion, it may be interesting to review the ASTM standards (2006a, 2006b) once
again. ASTM allows four alternative methods for determination of maximum index
density/unit weight of which two methods are on oven dried soil and two are on wet soil.
The wet method may be conducted on either oven dried soil to which sufficient water is
added or, if preferred, on wet soil from the field.
Although it may be expected that the application of any of the methods should yield the same
result, ASTM notes that the wet method can yield significantly higher values of maximum
index density/unit weight for some soils, and adds that such a higher maximum index density,
when considered along with the minimum index density/unit weight, will be found to
significantly affect the value of relative density. ASTM does not define wetness, and it is not
clear what would be considered as the minimum water content to be considered as wet.
ASTM notes that although the dry method of testing is often preferred because results can
usually be obtained more quickly, as a general rule the wet method should be used if it is
established that it produces maximum index densities/unit weights that would significantly
affect the use of the value of relative density. However, ASTM does not define what it
considers as being significant. Noting that any small changes in limit indexes are magnified
several times in changes of relative density, any difference may in fact be significant.
Likewise, ASTM (2006b) specifies three testing methods for determination of minimum index
density/unit weight, but allows the individual assigning the test to specify the method to be
used. If no method is specified, the provisions of first method shall govern. The two other
methods are provided for guidance of testing used in conjunction with special studies,
especially where there is not enough material. Once again, surely the application of these
three testing methods may yield different results that could result in values of relative density
with significant differences.
ASTM (2006a, 2006b) is well aware of relative density deviations and has performed a series
of three replicate tests per type and single test per type for poorly graded sand and specifies
acceptable ranges of results for it. For single tests, the acceptable range between two results
is 1.15 kN/m3. It has already been discussed that for a much smaller value of 0.135 kN/m3
Yoshimi and Tohno (1973) calculated 14% change in relative density; hence, it could be
expected that the acceptable range of ASTM could result in larger errors.
262
2.10.2.3 Relative Density Correlations
The unreliability of relative density due to its formulation has complicated its usage as a
ground improvement acceptance criterion, but correlation of relative density to other
commonly used field tests may appear as an attractive means of applying the philosophy of
relative density without the actual implementation of the problematic formulation. Thus, it
can be occasionally seen that a project’s specification stipulates that the improved ground
must exceed a minimum relative density value or curve based on SPT or CPT correlations.
Correlations, themselves, are as questionable as the concept of relative density, and have
been the subject of discussions by the experts since many years ago. In this section a number
of well‐established correlations that are commonly referred to will be reviewed, and the
unreliability of relative density correlations as acceptance criteria for ground improvement
shall be demonstrated.
A fundamental question that crosses the mind is that, irrespective of the limitations and
errors that are associated with relative density concept and formulation, would it be true to
say that if two soils had the same relative density then they would possess the same physical
characteristics and will behave the same? It would have been very satisfying if the reply to
this question was positive; however, this is not the case, and the use of relative density
correlations based on average sand to predict soil behaviour without considering other
parameters can result in poor or misleading predictions.
Youd (1973) conducted a study on the maximum and minimum indexes of 13 specimens of
different clean sands with less than 5% fines passing the 0.075 mm sieve, and was able to
identify clear relationships between emax, and emin, with particle roundness, grain shape,
range of particle sizes defined by the coefficient of uniformity, and the type of gradational
curve. Figure 2‐179 and Figure 2‐180 show the relationship between limit void ratios with
roundness and coefficient of uniformity. Youd did not identify a unique relationship between
mean particle diameter (D50) and limit void ratios; however, others (Cubrinovski and Ishihara,
1999, Skempton, 1986, Tokimatsu and Yoshimi, 1983) have noted the importance of the
effect of particle diameter on relative density‐field test relationships.
263
Figure 2‐179: Density limits as a function of grain shape for laboratory fractions with Cu=1.4 (Youd,
1973)
Holubec and D'Appolonia (1973) focused their study on the sphericity and the shape of four
types of medium to fine sands with the same gradation but varying particle shapes. What
they have found is that granular soils at the same relative density can have drastically
different engineering properties, and have assessed that the use of relative density criteria
in design, without considering particle shape, can result in poor or misleading predictions of
soil behaviour. As can be seen in Figure 2‐181, Holubec and D’Appolonia show that while
both emin and emax increase with the increase of the coefficient of angularity, the increase in
emax is considerably more than that for emin and that the difference between these two limits
increases with the angularity of the soil particles.
264
Figure 2‐180: Generalized curve for estimating emax and emin from gradational and particle shape
characteristics. Curves are only valid for clean sands with normal to moderately skewed grain size
distributions. (Youd, 1973)
Figure 2‐181: Effect of particle shape on minimum and maximum void ratios (Holubec and
D'Appolonia, 1973)
265
It can also been seen from the work of Holubec and D’Appolonia that each of the four sands
that they tested had separate and distinct relative density versus friction angle relationships
whereas whilst the roundest particles had the least friction angles and exhibited the
minimum increase in friction angle with increasing relative density, the sands with angular
particles had the highest friction angles with an intermediate increase with increasing
relative density. What can be assessed is that friction angle is a function of both relative
density and particle shape. In fact, the research suggested that equally large differences in
friction angle are possible with variations of particle shape as with changes in relative density.
Stress‐strain relations are also influenced by particle shape. Holubec and D’Appolonia show
that the more angular the particles are, the greater the failure strain is for a given relative
density. They performed a miniature penetration test using a 12.5 mm diameter rod that was
driven to a depth of 20 cm in sands that were compacted to selected densities. The data
showed a marked increase in the penetration resistance with the increase of angularity of
the sand particles. For example, they observed that the required number of blows for sand
with angular particles is nearly double the values for sub‐rounded sand when relative density
is 70%. Conversely, a blow count of 20 indicated a spread in relative densities from 66 to 86%.
More recently, Liu and Lehane (2012) have studied the behaviour of four uniformly graded
silica soils with similar mineralogical compositions but with distinctly different particle
shapes. The soils were subjected to centrifuge CPT and direct shear tests at different relative
densities in dry and saturated states. In these tests gradation and material were controlled
to be similar; hence, any differences in test results could be attributed to the grain’s
sphericity, roundness and roughness.
Figure 2‐182 shows the differences between effective peak and critical state friction angles
at different effective vertical stresses for the four soils when relative densities were 20%,
50% and 80%. It can be seen that while the general trend is the reduction of friction angle
differences at higher effective vertical stresses, the lines fitted to the plotted points of the
four soils are distinctly different for any of the three relative densities.
It can be understood from the work of Youd (1973), Holubec and D'Appolonia (1973), and Liu
and Lehane (2012) that applying relative density correlations from one type of sand to the
other, even if the gradation and mineralogy are similar, could lead to very misleading results.
266
Figure 2‐182: Difference between effective peak and critical state friction angles in four different
uniformly graded silica soils with similar mineralogical compositions at different relative densities;
reconstructed from Liu and Lehane (2012)
By definition correlation is a statistical relation between two or more variables such that
systematic changes in the value of one variable are accompanied by systematic changes in
the other. Correlations are not physical laws or theorems; they are simply statistical relations
and only meaningful once their scatters, deviations and variances are known. In general,
empirical correlations are derived from a set of specified data under special conditions that
are not necessarily applicable to other data, soil and conditions. In addition to the inherent
drawbacks and limitations of relative density, when consideration of a correlation’s
applicable domain is not taken into account, inaccurate and unrepresentative outcomes
should be expected.
This section will limit its review and discussion to some of the better known relative
density ‐ field test relationships as discussing all research is beyond the scope of this thesis;
however, the trend of this discussion and its conclusions is equally applicable to any other
such correlations. Also, this discussion and review will not include any limitations and
drawbacks that are associated with any of the field testing methods as that is a different issue
and not directly associated with the unreliability of relative density.
267
2.10.2.3.2.1 Relative density – Standard Penetration Test Correlations
While the first correlations between relative density and SPT were qualitatively realised by
Terzaghi and Peck (1948), probably the best known and most referenced estimation method
has been developed by Gibbs and Holtz (1957) based on data obtained from calibration
chamber tests performed at the US Bureau of Reclamation. Gibbs and Holtz performed tests
on a fine grained and a coarse grained sand by placing them at controlled densities and
moistures in a heavy steel tank, 0.9 m in diameter and 1.2 m in height. Overburden pressure
was realised by load plates and loading springs. Maximum density was determined by
vibrating the saturated material to constant density or by using extreme compaction hammer
blows, whichever gave higher values, in a container of known volume. Minimum density was
found by lightly pouring the dry material into the container of known volume.
Gibbs and Holtz carried out this research in 1957; i.e., 12 years before ASTM published its
first standard on relative density (ASTM, 1969). Obviously the testing procedure for
measuring limit densities could not have been as per a non‐existent standard of its time, and
even small differences in limit indexes originating from Gibbs and Holtz’s testing method
could have had significant impacts on the calculated relative densities.
Unlike the rather well known Gibbs‐Holtz relative density versus overburden pressure
diagram that has been referenced in numerous publications, those who have actually seen
Gibbs and Holtz’s (1957) paper know that they published a relative density versus
penetration resistance chart, as reconstructed in Figure 2‐183, not the better known chart of
Figure 2‐184 that first appeared in Bureau of Reclamation’s (1960) Earth Manual in
accordance to Gibbs and Holtz. It is noted that the Bureau of Reclamation (1998) does not
use the term correlation, but rather refers to the chart as the criterion for predicting relative
density of sand. In addition to the limitations that ASTM (2006a, 2006b) has in place for the
application of relative density, Bureau of Reclamation also limits the use of its chart to sands
containing less than 10% fines and no gravels.
Osterberg and Varaksin (1973) extracted frozen soil samples from Lake Michigan and
compared them with what was estimated by Gibbs and Holtz. Predictions were quite
different from reality, and it was concluded that relative densities obtained from SPT using
Gibbs and Holtz chart had no relationship to the actual relative densities. This study was an
indication that relative density correlation of one sand cannot be extended to all other sands.
268
Figure 2‐183: Relative density versus penetration resistance curves for fine and coarse sands,
reconstructed from Gibbs and Holtz (1957)
Figure 2‐184: Criterion for predicting relative density of sand from the penetration resistance test
(Bureau of Reclamation, 1998)
Contrary to the general over confidence by others, Holtz (1973a) does not share the same
unconditional and unlimited trust that others have for their chart. In the discussion on the
results of Osterberg and Varaksin (1973), Holtz notes that everyone should recognise that
the Standard Penetration Test is a relatively crude test, and no one should expect to
determine the relative density of sands to the nearest one per cent or anything like that, and
further adds that he and Gibbs developed a set of correlations to take into account the effect
of overburden pressures, and that they never indicated that the sets of curves developed at
269
that time were necessarily applicable to all cohesionless soils under all conditions. More
importantly he clarifies that they had always stressed on relative density trends indicated by
SPT values rather than the specific individual values. This discussion suggests that the
research of Gibbs and Holtz may have been blown out of proportion, and rather than
acknowledging the trend of relative density versus overburden pressure it is used
systematically for something that was never the intention.
Meyerhof (1957) has formulated the coarse sand graph of Gibbs and Holtz (1957) and
expressed it in the form of Equations 2‐251 or 2‐252 (after conversion to SI units); however,
in practice this equation is often extended to most types of sands, regardless of the soil
particle size and shape, gradation and mineralogy. Obviously, this will further cast doubt on
the reliability of the correlation as it is very clear in the original Gibbs and Holtz research that
the fine and coarse sand curves do not coincide.
17 0.25 ′ 2‐251
2‐252
17 0.25 ′
N= SPT blow count
Dd= relative density expressed as a ratio (not percentage)
’v= effective vertical stress (kPa)
In lieu of Equation 1‐12, Haldar and Tang (1979) have approximated Gibbs and Holtz’s relative
density correlation using Equation 2‐253 (after conversion to SI units).
. 2‐253
20 0.21 ′
Haldar and Tang who note the difficulty of obtaining data that actually includes
measurements of Dd, N values and ’v have plotted measured versus calculated (using Gibbs
and Holtz’s relationship) relative density of what they could gather (Gibbs and Holtz, 1957,
Marcuson III and Bieganousky, 1977, Moretto, 1954, Varaksin, 1970, Wu, 1957) onto the
graph of Figure 2‐185. It can be seen that there is considerable spread about the 45o bisecting
parity line. The majority of data is above the 45o line, suggesting that Gibbs and Holtz’s
270
relationship may be non‐conservative. Figure 7 also shows a mean line, an upper bound (line
B) and a lower bound (line E). For a given value of relative density calculated from Gibbs and
Holtz’s relationship, the measured relative density may be assumed to follow a triangular
distribution between the upper and lower bounds and be symmetrical about the mean line.
On this basis, the mean measured relative density is only 75% of the calculated value and the
scatter around the mean has a coefficient of variation of 27% which is constant at any relative
density calculated. Haldar and Tang conclude that except for sand exhibiting a small
difference between the limit indexes, the uncertainty in direct laboratory determination of
in‐situ relative density (which is prone to large errors) is expected to be less than estimations
using Gibbs and Holtz’s relationship for normally consolidated deposits, and that a systematic
bias appears to exist in Gibbs and Holtz’s prediction relationship.
Figure 2‐185: Comparison of measured versus calculated relative density using Gibbs and Holtz’s
relationship (Haldar and Tang, 1979)
Peck and Bazaraa (1969) proposed Equations 2‐254 and 2‐255 (after conversion to SI Units)
for predicting relative density. Comparison of these relations will show that for equal N
values, these equations will consistently yield higher relative density estimates than Gibbs
and Holtz’s relationship.
271
for ′ 72kPa 2‐255
65 0.21 ′
Lacroix and Horns (1973) note that it is their experience Gibbs and Holtz’s method of
predicting relative density yields results that are too high for heavily compacted fill. As noted
by Peck and Bazaraa (1969), they also agree that while Gibbs and Holtz greatly underestimate
N values corresponding to 100% relative density, Bazaraa’s (1967) relationship was in better
agreement with the blow count data.
Marcuson III (1978) who refers to a study on four sands at various relative densities under
overburden pressures in the laboratory of the US Army Waterways Experiment Station (WES)
concluded that a simple family of curves relating N values, overburden pressure and relative
density for all sands under all conditions is not valid. He also concluded that based on
comparisons between the relationships presented by Gibbs and Holtz, Bazaraa and WES, SPT
is not sufficiently accurate to be recommended for final evaluation of the density or relative
density, unless site specific correlations are developed.
Skempton (1986) proposes the relationship between relative density and N values in the
general form of Equation 2‐256, with a and b as two parameters. Skempton tested five
different types of sand and proposed values for a and b ranging respectively from 15 to 30
and from 17 to 24 in the imperial system (ton/ft2). It can be seen that the values for these
parameters can be respectively 100% and 40% more than the least values. Inclusion of
Bazaraa (1967) would even further increase the figures respectively to 300% and 47%. This
clearly suggests that the specific studies of one site cannot be used on other sites so simply.
2‐256
′
Among his observations, Skempton (1986) notes that at a given relative density and
overburden pressure, N values are higher for sands with larger grain sizes (D50). He also
assesses that there is direct evidence that aging of sand will increase the SPT blow counts
(also refer to Section 2.5.6). This suggests that not only is relative density influenced by
numerous other parameters such as gradation, particle size, overburden pressure,
mineralogy and particle shape, and hence its correlation in one sand is not reliable for any
272
other sand, but even the correlation for a soil at a specific time or state may be not valid and
applicable at other times and conditions. Skempton also identifies a relationship between
the effects of over consolidation and relative density, and introduces an over consolidation
coefficient, Coc, into Equation 2‐256 to derive Equation 2‐257. Coc itself can be calculated from
Equation 2‐258.
2‐257
′
1 2
2‐258
1 2
KONC and KO are the ratio of effective horizontal to vertical stresses respectively when the soil
is normally consolidated and over consolidated. Skempton refers to Mayne and Kulhawy
(1982) for determining these coefficients as a first approximation according to Equations 2‐
259 and 2‐260.
1 sin ′ 2‐259
2‐260
OCR= over consolidation ratio
φ’= effective internal friction angle
Tokimatsu and Yoshimi (1983) who were studying soil liquefaction modified Equation 2‐252
to Equation 2‐261 by taking into account the effect of fines content, Fc, and introducing ΔNf
as a correction term (refer to Table 2‐28). However, they themselves did not demonstrate
confidence in their proposed equation, and note that its application had yet to be proven.
∆
0.21 2‐261
′ 1.7
0.7
98
Fines content (%) ΔNf
273
0‐5 0
5‐10 interpolate
10‐ 0.1Fc+4
Table 2‐28: ΔNf – fines content correlation (Tokimatsu and Yoshimi, 1983)
More recently, Hatanaka and Feng (2006) have carried out a comparative study on high
quality undisturbed samples recovered by the in‐situ freezing method. The material used in
this study was less than 4.75 mm in size and D50 was less than 1 mm. They then compared
the measured and calculated values of relative density using Meyerhof (1957), Bazaraa
(1967) and Tokimatsu and Yoshimi (1983). As shown in Figure 8, the estimated values of
relative density based on Meyerhof’s method were in the range of +15% to ‐45% of the
measured values. Similar to Haldar and Tang (1979), this research also shows a large scatter
of results about the prediction equation; however, contrary to Haldar and Tang, the scatter
is mostly concentrated on the lower side of the bisecting parity line, suggesting that
Meyerhof’s equation is generally underestimating relative density. Similar results were also
obtained when the correlation of Tokimatsu and Yoshimi (1983) was used. In this case, the
range of estimated to measured relative density was from +25% to ‐20%.
Tatsuoka et al. (1978) who examined the accuracy of Meyerhof’s (1957) expression by
studying the results of SPT on normally consolidated sandy deposits and conventional
undisturbed samples made a similar assessment, and concluded that Meyerhof’s formulation
tends to underestimate the relative density of fine sands and silty sands.
For estimating relative density, Cubrinovski and Ishihara (1999) note the importance of grain
size and propose Equation 2‐262.
. 98
2‐262
9 ′
274
Figure 2‐186: Figure 8: Comparison of measured versus calculated relative density using Meyerhof’s
equation (Hatanaka and Feng, 2006)
N78= SPT blow count corresponding to an energy rod ratio of about 78% from the theoretical
free‐fall energy.
Cubrinovski and Ishihara report that by using Equation 2‐262, in 84% of instances they were
able to calculate relative density within a deviation of 20% from the measured values, and
more than half the calculations were within 10% deviation. Although this amount of accuracy
is by no means sufficient enough to reliably establish an acceptance criterion upon, in any
case, here we are dealing with three parameters, two (limit void ratios) of which are included
in the original definition of relative density, and the third (N value) is a function of the in‐situ
density. It may be simply more appropriate to use the definition of relative density (ASTM,
2006a and 2006b) to calculate a value which is prone to error rather than going to the trouble
of estimating a value which may be even more erroneous and misleading.
We can review or introduce many other SPT‐Dd correlations and continue this discussion
endlessly; however, what is evident and established is that due to the numerous other
parameters that influence the outcome, SPT‐relative density correlations will have an
unreliably large amount of scatter, and cannot be trusted as ground improvement criteria.
Many, for example Jamiolkowski et al. (2001), note that the first CPT versus relative density
correlation was published by Schmertmann (1976). This is not entirely accurate as the
275
authors are aware of at least one previous publication by Schmertmann (1975) and a
reference by Schmertmann (1978) to Mitchell and Gardner (1975).
Although the author agrees that Schmertmann’s 1976 publication should be recognised as
an early research on the subject, it has unfortunately been published in the form of a
contract report for Waterways Experimental Station and today, this document is not even
available in WES’s own library (Dolan, 2011). The author has unsuccessfully attempted to
obtain a copy of the report through numerous sources, and to date has not met anyone who
has actually read the publication. We can only speculate that many researchers who are more
recently referring to this publication have also never read the original publication, and are
simply citing information from other publications. This can be quite concerning due to
possible publication errors, misinterpretations or the fact that each of the intermediate
authors only extract that scope of work that is relevant to their own study. By reviewing
Schmertmann (1978) and other research such as Villet and Mitchell (1981) the author
understands that the updated correlation of Schmertmann (1976) is the same as what was
later published in Schmertmann (1978).
Schmertmann (1978) who himself is aware that cone resistance is significantly affected by
grain size distribution, cementing, lateral stresses, depth of overburden, compressibility,
pore pressure and thin layer effects modifies and updates his earlier work (Schmertmann,
1975) after performing 80 more tests in a calibration chamber with a 1.2 m diameter on two
artificial sands with opposite extreme crushabilities, two natural fine sands, and one natural
and one artificial medium sand, and proposes a correlation chart which only takes into
account the effect of vertical effective stress and is for normally consolidated, recent,
uncemented fine SP sands. He also proposes a correction factor for converting over
consolidated sands’ cone resistances, qcOC, to normally consolidated sands’ cone resistances,
qcNC. Once simplified, the expression will take the form of Equation 2‐263.
3 .
1 4 1 2‐263
4
As shown in Figure 2‐187, there are significant differences between the relative density
estimations of Schmertmann (1975) and Schmertmann (1978). Although the latter
publication does not show the scatter of testing results, Schmertmann does note that for
clean quartz sands his chart is able to estimate relative density with a standard deviation of
276
about 10%. The standard deviation of Schmertmann (1975) was 7%. This suggests that
irrespective of the scatter that one may experience in one study, there is a strong possibility
that the correlation would be almost meaningless for another soil.
Figure 2‐187: Comparison of relative density – CPT cone resistance relationships for Schmertmann
(1975), Schmertmann (1978) and Villet and Mitchell (1981)
This has also been observed by Villet and Mitchell who performed a series of tests on four
gradations of a commercially available windblown dune sand in a calibration chamber with a
diameter of 0.76 m and height of 0.8m. The sands were mainly of quartz and feldspar grains
and sub‐rounded to sub‐angular in shape.
The relationship of Villet and Mitchell is also shown in Figure 2‐187. The differences between
Villet and Mitchell’s measurements and Schmertmann (1978) are significant and from as low
as 20% in dense sand with lesser vertical effective stress to 140% for loose sand subject to
more vertical effective stress. This is shown in Table 2‐29. After comparing their calibration
chambers and chamber boundary conditions, Villet and Mitchell conclude that these large
differences are due to soil type, thus suggesting that cone resistance, vertical stress, relative
277
density relationships are not unique and universal for all sands, but rather a function of the
sand being penetrated. This conclusion is applicable to any other research that defines a
relationship between these parameters.
In fact, had Villet and Mitchell used the same calibration chamber as Schmertmann then the
differences would have been even larger. A number of researchers such as Parkin (1977) and
Parkin and Lunne (1982) have studied the effects of calibration chamber size and boundary
conditions. Parkin’s results indicate that calibration chamber diameter ratio to the CPT cone
diameter and boundary conditions affect the test results whereas either decreasing diameter
ratio or maintaining constant stress rather than constant volume boundary conditions results
in lower cone resistance.
In addition to the corrections that may be necessary due to the geometry and boundary
conditions of the calibration chamber, Bolton and Gui (1993) mention the simulation method
of overburden pressure by surcharging as another limitation of the calibration chamber as
the effect of stress gradient due to the self‐weight of the soil cannot be simulated in this way.
Bolton and Gui conclude that the results obtained from calibration chambers always leave
room for questioning, and propose the use of centrifuge testing.
Irrespective of Bolton and Gui’s view, calibration chamber testing has remained the
dominant approach for developing relative density – CPT correlations. Baldi et al. (1986) who
have proposed one of the most popular correlations performed calibration chamber tests on
Ticino sand and Hokksund sand. They proposed an expression in the form of Equation 2‐264.
1
ln 2‐264
′
Co, C1, and C2 = experimental coefficients
278
Baldi et al. then refer to earlier work by themselves, and define ’ as effective vertical stress
if the sand is normally consolidated or as the effective horizontal stress or effective mean
stress if the soil is over consolidated. This was based on the fact that using effective vertical
stresses in over consolidated sands led to an overestimation of relative density, as also
observed for SPT.
Although Baldi et al. propose experimental coefficients for 10 cases of normally consolidated,
over consolidated and normally/over consolidated conditions for both Ticino sand and
Hokksund sand, for some unknown reason, the normally consolidated Ticino sand has
become the better known and popular correlation as expressed in Equation 2‐265 and what
will generally appear in one form or the other when relative density correlations are used as
ground improvement acceptance criteria.
1
ln . 2‐265
2.41 157 ′
Had normally consolidated Hokksund sand gained fame, the expression would have had been
as shown in Equation 2‐266. The difference between predicted relative density values using
Equations 2‐265 and 2‐266 increases with effective vertical stress and the reciprocal of cone
resistance and can be more than 20%. This is yet another example of the fact that relative
density correlations are not unique and are dependent on soil type.
1
ln . 2‐266
3.29 86 ′
In line with the above, Jamiolkowski et al. (2001) have proposed Equation 1‐28 using Ticino
sand, Hukksund sand and Toyoura sand. Here, the data scatter is more for each of the Ticino
sand or Hukksund sand equations. Equation 1‐28 yields the lowest estimate of relative
density as compared to Equations 2‐265 and 2‐266 in most cases, except in loose sand at high
effective vertical stresses.
279
1 98.1
ln . 2‐267
3.10 ′
17.68 98.1
It should be noted that Jamiolkowski is well aware of the unreliability of relative density as a
criterion, and is merely proposing expressions to better the estimate of relative density.
Jamiolkowski and Pasqualini (1992) note that quality control only on values of relative density
can be insufficient to evaluate the ground modifications achieved by compaction. Even if
effective horizontal stresses are used in the estimation expression, they conclude that
important factors such as compressibility, aging, and the presence of fines limits the use of
correlations as a guide for evaluating in‐situ density of clean, predominantly silica sands of
recent, uncemented deposits. Evaluation of relative density in over consolidated sands
becomes less reliable because of the inherent difficulties in proper assessment of effective
horizontal stresses for improved soils.
Relative density correlations are even more unreliable when it comes to calcareous sands.
Almeida et al. (1992 ) who carry out calibration chamber tests on the calcareous Quiou sand
conclude that for the same relative density, cone resistance in the calcareous sand is well up
to half the value of qc measured in the silica Ticino sand. The observed trend in the
differences was greater for higher relative densities.
More recently, Al Hamoud and Wehr (2006) report that based on correlation charts of
Robertson and Campanella (1985), relative density of 60% was required for land reclamation
projects (Palm Jumeira) in Dubai (The authors note that there are no such charts in Robertson
and Campallena's cited publication and Al Hamoud and Wehr have probably made a mistake
in their reference). Due to the difficulties in achieving the requested penetration resistance
in some zones of the compacted fill, a need was felt to verify whether silica sand based
correlation was equally applicable to calcareous sand.
Al Hamoud and Wehr refer to the unpublished work of Gudehus and Cudmani who had
performed calibration chamber tests on Dubai’s calcareous sand and Karlruhe’s quartz sand.
According to Al Hamoud and Wehr the calibration chamber diameter and height were
respectively 0.95 m and 1.5 m. The diameter of the CPT rod used in the test was 36 mm.
Without entering into a detailed discussion Al Hamoud and Wehr state that a shell correlation
(correction) factor of 1.5 for depths greater than 8 m, 1.6 for depths of 4 to 8 m, and 1.7 for
280
depths less than 4 m must be applied to Dubai sand. According to this study, these correlation
factors should be seen as lower limit and conservative for calcareous sand if the material in
the field is much coarser than the soil fractions used in the experiments since larger shells
crush easier than very small ones. They also note that penetration resistance for Dubai sand
was reported to be about 37% lower than that of Karlsruhe sand for a medium dense state.
They also divide the best fit exponential curves of Dubai sand to a number of other silica
sands to derive a shell correlation factor fshell which is expressed in Equation 2‐268.
Comparing Dubai sand with Karlsruhe sand demonstrates that, as previously mentioned,
there is no unique relationship between relative density and cone resistance. At the same
time, the cone resistance difference between Dubai sand and Karlsruhe sand does not mean
that the same difference would have been observed if the same sand (referenced incorrectly
by Al Hamoud and Wehr) as the project’s acceptance criteria was used for. The same is
equally true for the expression that was derived in Equation 1‐29. In fact, entering a relative
density of 0.6 into Equation 1‐29 will yield a correlation factor of 1.6389 which, if used, would
have resulted in about 10% overestimation of Dubai sand relative density at depths deeper
than 8 m.
Furthermore, the calibration chamber diameter to cone diameter ratio in this testing
programme was 26.4. This is almost half of what Parkin and Lunne (1982) propose for
boundary effects to become negligible in normally consolidated dense sands. With the small
ratio that has been used by Al Hamoud and Wehr, the difference between the calibration
chamber test and the project’s acceptance criteria may have simply been due to the
boundary effects.
2.10.3 Reclamations
It is the author’s observation that while most engineers well understand that dumped fills on
land are non‐engineered fills, and subject to numerous geotechnical problems such as low
bearing capacity, creep and excessive settlements under external loads, many do not have
an equal comprehension when (sand) filling is done in the sea as reclamation. Perhaps this is
due to the fact that those engineers are not acquainted with ground improvement
techniques, and hoping that constructing an engineered fill above ground water level will
281
alleviate future foundation problems. Unfortunately, reclamations systematically realise
loose to very loose fills that, as a least, bear the same problems as land based non‐engineered
fills.
Reclamation can either be by dump trucks tipping fill into the sea or by hydraulic placement
from the sea. Sladen and Hewitt (1989), Lee et al. (1999), Lee et al. (2000), Lee (2001) and Na
et al. (2005) have studied the effects of placement methods on the geotechnical behavior of
sand fills.
Figure 2‐188: Dump truck tipping fill into the sea
The density of sand that is dumped by trucks (Figure 2‐188), and then pushed into the sea by
a bulldozer is usually low, with relative density of about 20%. Exceptions can be thin layers
that have been compacted by the traffic of earthmoving equipment. Hydraulic placement
can be subaqueous by hoppers or bottom dump barges. When possible sand is discharged
by means of a big door located on the bottom of the hull, but when the water is shallow
alternative methods; i.e., pipeline discharge or subaerial rainbow discharge will be used. In
pipeline discharge low velocity water‐sand slurry is pumped; however, in the rainbow
method the dredger sprays a high velocity water‐sand mixture onto the reclamation. These
processes are schematically shown in Figure 2‐189.
282
(a)
(b)
(c)
Figure 2‐189: (a) subaqueous discharge by hopper or bottom dump barge (b) subaerial rainbow
discharge (c) pipeline discharge, redrawn from Lee et al. (1999)
The variation in fill densities achievable by hydraulic placement is large and closely related to
the placement method. Hopper placed sand is denser than pipeline placed sand. Sand
deposited by hydraulic filling below water level generally has a low to medium relative
density of about 20 to 60% due to the loose packing from self‐weight sedimentation of sand
particles under water. The zone with the least strength could be expected to be just beneath
water level if fill is placed by subaqueous discharge through hydraulic pumping. Sand placed
283
above water table by hydraulic filling tends to have a higher relative density in the range of
60 to 80% because of dense packing from downward seepage and reduction in void ratio as
a result of sliding and rolling of the sand particles mixture.
Hopper or bottom dumping achieves a higher density than pipeline discharging for a number
of reasons. Firstly, the sand mass stored in a hopper has a higher bulk density than the sand
slurry that is discharged from a pipe. Also, dumping a large quantity of sand from a hopper
in a short period will result in the sand mass falling as a slug rather than as individual particles.
Furthermore, the simultaneous opening of all bottom doors prohibits the entrapment of
fresh water into the slug that would reduce the fall velocity and expand the slug size. The fall
energy of the slug is likely to be dissipated in compaction of berm through impact and
shearing. The loosest possible state would likely be achieved if the pipe discharge was placed
near the water surface in such a way to allow maximum fresh water entrapment. In such a
case the slurry becomes a clod with falling velocity being close to the falling velocity of
individual grains. Each particle will basically come to rest in the position that it makes contact
with the previously placed fill. Impact may result in some pushing around of the grains, but
the impact velocities and forces can be expected to be small. Subaerial rainbow dredging can
be expected to yield similar results to pipe discharging.
Once the process of reclamation is understood, it will not be difficult to be able to foresee
that reclaimed sand fills will most probably be loose and potentially subject to a number of
geotechnical problems. Hence, project developers and designers must be aware that they
may require specific geotechnical solutions such as ground improvement in reclaimed
projects.
Menard (1971) was quick to notice the benefits of applying dynamic compaction to
reclamation projects, where the fill is usually composed of good quality material such as rock
waste, rubble, sand, etc., and the but the reclamation’s apparent poor behaviour is merely
due to its loose state and low compaction. Dynamic compaction has been used in many
noteworthy reclamations; e.g. Changi International Airport (Bo et al., 2009, Choa, 1980, Choa
et al., 1979), Nice International Airport (Gambin, 1983), Macau International Airport
(Spaulding and Zanier, 1997), Tsing Yi Oil Terminal (Hendy and Muir, 1997), or Port Botany
Expansion (Berthier et al., 2009).
284
2.10.4 Offshore Dynamic Compaction and Dynamic Replacement
It can be observed that the notion of improving the ground for engineering purposes initially
developed implicitly to resolve land based foundation problems because the quantity of
foundations on land is by far more than what is encountered in marine environments and
offshore. However, the 20th century was witness to a number of marine and onshore
geotechnical failures such as the 1916 collapse of Gothenburg Harbour’s Stigberg Quay in
Sweden (Massarsch and Fellenius, 2012), and the 1979 failure of Nice Harbour in France (Dan
et al., 2007). Hence, it was inevitable that sooner or later attention would be drawn towards
modifying or adjusting ground improvement techniques for application to subaqueous near
shore and offshore projects.
The first applications of marine ground improvement can be traced back to the 1970s.
Menard carried out the first offshore dynamic compaction project in 1973 as part of the
construction of Brest Naval Port’s prefabricated dry dock in France. In this project a specially
designed 11 ton pounder that is shown in Figure 2‐190 was used to compact about 3 m of
loose alluvium seabed over an area of 4,500 m2 (Boulard, 1974, Gambin, 1982, Menard, 1974,
Renault and Tourneur, 1974). Dynamic compaction was applied with variable energies
ranging from 240 tm/m2 in one pass, 480 tm/m2 in two passes, up to 700 tm/m2 in special
cases, and on average applied energy was 400 tm/m2. PMT parameters before and after
dynamic compaction are shown in Figure 2‐191.
Figure 2‐190: Application of dynamic compaction at Brest Naval Base using a specially designed 11
ton marine pounder (Boulard, 1974)
285
Figure 2‐191: EM and PLM before and after offshore dynamic compaction at Brest Naval Base
(Boulard, 1974)
In 1975, offshore dynamic compaction was used at Pointe Noire in Gabon to improve the
passive resistance of loose sand in front of a cellular sheet piled wall. The specially designed
pounder used in this project weighed 12 tons. Prior to ground improvement a 0.4 m thick
blanket of gravelly sand was placed over the seabed, located 13.5 m below sea level.
Generally, in this 14,000 m2 project two phases of dynamic compaction with compaction
energy intensity of 240 tm/m2 were applied to provide an allowable bearing capacity of 400
kPa; however, in one sandy area one phase, and in one very silty sand area three phases of
compaction were utilised. PLM before improvement ranged from 250 to 600 kPa in the upper
5 m of seabed. Post improvement tests indicated that the limit pressure had increased by
about 100 to 300% (Menard, 1978).
As part of the 50,000 m2 dynamic compaction project at Udevalla, Sweden in 1975, offshore
DC was applied for the treatment of a granite rock fill that was placed below the
prefabricated caissons. The rock size was up to 1 m in diameter, the top of the fill was 12 m
below sea level, and its thickness was variable from 17 to 20 m. Due to the size of the rock
fill, dynamic compaction was performed by using a 40 t pounder that was dropped from 40
m using the Menard’s Mega Machine (refer to Section 2.7.1). In this area 20 tm/m3 of energy
was applied to increase EM in the upper 11 m of rock fill below the casing from 4.85 MPa
before DC to about 110 MPa after dynamic compaction (Menard, 1978). The cross section of
the project is shown in Figure 2‐192.
286
Figure 2‐192: Cross section of rock fill at Udevella, Sweden (Menard, 1978)
As shown in Figure 2‐193, in 1977 a 32 ton pounder was used to compact a 5 m thick layer of
silty sand and a 1.5 to 2 m thick rock fill blanket at the depth of 10 m below seawater level
to mitigate the risk of liquefaction of a breakwater foundation at Kuwait Naval Base (Chu et
al., 2009, Gambin, 1982, Menard, 1978). Marine ground improvement was carried out over
an area of approximately 36,000 m2.
Figure 2‐193: Application of offshore dynamic compaction at Kuwait Naval Base
Also in 1977, offshore dynamic compaction was applied for the treatment of seabed at Sfax
Fishing Quay in Tunisia (Gambin, 1982, Menard, 1981). In this project water depth was 1 m,
and the seabed was composed of 5 m of very soft clay followed by 2 m of silty sand and 3 m
of denser clayey sand, and firm clay.
Conventionally, in such projects the solution would have been to remove the soft clay,
replace it with sand, drive in the sheet pile wall, and deepen the quay by dredging in front of
the quay wall. However, as sand was not available in the required quantity, and the project
287
owner intended to construct a gravity quay wall, the foundation construction technique
consisted of removing the soft clay and replacing it with a rock fill blanket, and improving the
silty and clayey sands to provide a an allowable bearing capacity of 300 kPa. Dynamic
compaction was performed using a 17 ton pounder with a square grid spacing of 2 m.
Compaction was carried out in 3 to 5 phases with 2 to 10 blows per print.
Figure 2‐194: Offshore DC at Kuwait Naval Base using a 32 t specially designed pounder
Prior to placement of the prefabricated sections of a dry dock in Lagos, Nigeria, dynamic
compaction was carried over an area of 13,800 m2 in 1979. Initially, the seabed was
excavated to ‐15 m RL (reduced level), and backfilled with 1 m of rock fill. As shown in
Figure 2‐195, soil improvement was carried out using a 40 t pounder with the intention of
treating 15 m of soil. Up to 5 passes of dynamic compaction were applied in some locations.
The square compaction grid spacing varied from 2 to 4.5 m. As can be seen in Figure 2‐196,
large improvement was achieved in the upper 10 m, and a mean soil reaction modulus of
0.82 kPa/m was obtained for a 20x44 m2 slab (Gambin, 1982).
Figure 2‐195: Dynamic compaction at Lagos dry docks, Nigeria
288
Figure 2‐196: Comparison of EM before and after dynamic compaction at Lagos Dry Docks (Gambin,
1982)
In 1980, a desalination plant was assembled in Japan, towed to Yanbu in Saudi Arabia, and
sunk on a seabed foundation prepared by dynamic compaction. The seabed was composed
of 8 m of loose silty sand, with the upper 4 m being very heterogeneous. Hence, the top 4 m
of seabed was removed, and replaced by crushed coral, and dynamic compaction was applied
using a 17 ton pounder that was dropped on a square grid with 8 m spacing. The number of
blows per print in the compaction phases varied from 12 to 10. By this means the allowable
bearing capacity of the seabed was increased from 30 to 120 kPa (Gambin, 1982). The treated
area in this project was 4,000 m2.
289
3 Distinguished Dynamic Compaction and
Dynamic Replacement Projects
290
3.1 Choice of Projects
Millions of square metres of ground improvement have been undertaken by dynamic
compaction and dynamic replacement since Menard invented and developed these
techniques more than 40 years ago. These techniques have generally been used to improve
the soil strength, to increase bearing capacity and to reduce external load induced
settlements, but many projects have also been undertaken to mitigate soil liquefaction or to
prevent settlements originating from the self‐weight of the soil.
The author has had the unique opportunity to be involved in numerous state‐of‐the‐art
dynamic compaction and dynamic replacement projects that can be of great value for
improving our understanding of what to expect by performing these techniques, to optimise
design and management procedures, and to develop new methods for quantifying ground
and soil behaviour. Hence, a number of such projects with various aspects have been chosen
for this purpose, and will be documented, analysed and compared with existing publications
in this thesis.
A summary of the projects that are documented and analysed in this research are presented
in Table 1. To the knowledge of the author, a number of projects that have been chosen in
this research have been world records or first time applications at their time. These include:
291
Project Name Location Area (m2) Technique Ground Genre
Blue City Oman 225,000 DC Saturated Buildings
loose silty
sand
Marjan Island Ras Al 198,000 DC Reclamation Road
Road Corridor Khaimah
UAE
Al Nakhilat Ras Laffan 175,000 DC Reclamation Quay, dry
Ship Repair Qatar dock
Yard
Ritz‐Carlton Abu Dhabi 90,000 DR Saturated soft Building,
Hotel UAE very silty Embankment,
sands wall
foundation
Umm Quwain Umm 86,000 DR Saturated soft Buildings
Marina Quwain very silty sand
UAE
Al Jazira Steel Mussafah, 39,000 DR Saturated soft Industrial
Pipe Factory Abu Dhabi very silty building
UAE sands
Fujairah Fujairah 28,000 DC Loose sand Industrial
Desalination UAE
Plant Phase II
Reem Island Abu Dhabi 8,000 DC Reclamation Road
Causeway & Reem
UAE
Heavy Fuel Oil Ras Laffan ~ 12,200 DC Loose sand Tanks
Tanks Qatar and sabkha
Palm Jumeira Dubai ~ 2,000 DC Reclamation Tanks
Sewage UAE
Treatment
Plant Tanks
Al Khaleej Raw Dubai, UAE ~ 3,300 DR Loose sand Silos
Sugar Silos
292
Project Name Location Area (m2) Technique Ground Genre
Palm Jumeira Dubai, UAE ~ 900 DC Reclamation Building
Trial
Quay Southeast ~ 550 DR Disturbed Offshore
Expansion Asia clayey seabed
Table 3‐1: Case studies used for documentation, analysis and comparison with previous
publications
Abu Dhabi New Corniche, UAE: In this project 900,000 m2 of ground was reclaimed
by hydraulic filling. This project was 6 km long, which makes it the longest dynamic
compaction project that has been carried out in a metropolitan area. Figure 3‐1
shows the New Corniche against Abu Dhabi’s skyline after the completion of the
project.
Figure 3‐1: Abu Dhabi Corniche: 6 km long project in a metropolitan area
Al Quo’a New Township, UAE: In this project 1.13 million m2 of levelled dune sands
were treated using dynamic compaction. While, at most, the ground had to support
two storey villas, the maximum treatment depth for creep or self‐bearing was 28 m.
The MARS (Menard Accelerated Release System) was invented and used to drop a
35 t pounder in free fall and automatically reconnect it to the DC rig in this project
for the first time. Figure 3‐2 shows the pounder lifting and dropping phases.
King Abdulla University of Science and Technology (KAUST), Saudi Arabia: In this
project more than 2.6 million m2 of ground was treated using dynamic compaction,
dynamic replacement, and dynamic surcharging. At its time of construction this
project was the largest (in size) DC or DR project that was carried out in one contract
293
by a ground improvement specialist contractor. KAUST also scored the world record
for the project with the most number of DC and DR rigs as it implemented 13 rigs
working two shifts a day. Consequently, at its peak, production reached
approximately 600,000 m2 per month, which was also a world record at its time.
Extreme heterogeneity of the ground conditions over short distances, a very tight
programme requiring treatment before completion of architectural and structural
designs, and late changes in design have all made this project one of the most
challenging projects that has ever been undertaken.
Figure 3‐2: First application of MARS pounder release and reconnect mechanism in the world
Figure 3‐3: Application of DC and DR at KAUST using 13 (9 shown in the photograph) DC‐DR rigs
working in two shifts per day.
294
Al Falah Community, UAE: This development required application of dynamic
compaction for the treatment of loose desert sands over an area of 4.84 million m2,
making it also the largest (in size) DC project that has been performed in one contract
by a ground improvement specialist contractor at its time of construction. Although
the number of DC rigs used in this project was 11 (working in two shifts) production
rate reached 966,000 m2 per month, which surpassed KAUST’s record, and became
the new world record. Figure 3‐4 shows the application of DC at Al Falah Community.
2
Figure 3‐4: Treatment of 4.84 million m of loose sands at Al Falah Community.
Quay in Southeast Asia: In this project dynamic replacement was applied to the
seabed at the depth of 30 m below sea level, which makes it the world’s deepest
application of DC or DC in an offshore project. The double sided DC‐DR grated
pounder that is shown in Figure 3‐5 and that was used in this project was also an
innovation in pounder construction.
Abu Dhabi New Corniche, Al Quo’a, KAUST, and Al Falah Community are all mega DC
or DR projects, but it is the author’s experience that projects larger than 30,000 m2
can also be considered as large size projects. Thus, Blue City, Marjan Island Road
Corridor, Al Nakhilat Ship Repair Yard, Abu Dhabi Ritz‐Carlton Hotel, Umm Quwain
Marina, Al Jazira Steel Pipe Factory, and Fujairah Desalination Plant can be
understood to be large or very large size projects as well.
295
Figure 3‐5: First application of a double sided DC‐DR grater pounder in the world
296
3.2 Abu Dhabi New Corniche Road
As shown in Figure 3‐6 Abu Dhabi New Corniche (Beach Road) is a 6 km long (see Figure 3‐6)
reclamation with an area of 900,000 m2 that has been hydraulically reclaimed from the
Persian Gulf using the pipeline discharge method (Figure 3‐7). The reclamation began at the
face of the original beach road, varied in width from zero to 300 m, and on average extended
160 m into the sea. Maximum reclamation thickness was about 12 m at the sea facing, which
was to be retained by sheet piles, and average thickness of the fill was 6.5 m. In addition to
the leisure areas, pedestrian pathways and bicycle lanes on each side of the road, the road
itself was designed to have 4 lanes in each direction that were to be separated in the centre
by a variable width median.
Figure 3‐6: Plan of Abu Dhabi New Corniche
Figure 3‐7: Reclamation of New Corniche by pipeline discharged hydraulic fill
The preliminary geotechnical investigation revealed that the seabed was composed of
medium dense fine grained sand followed by a dense layer of sand, shells and ultimately the
limestone bedrock.
297
Project specifications stipulated that fines content in the reclamation material had to be less
than 10%, and the fill’s relative density had to be at least 80% with an SPT blow count
correlation that is shown in Table 3‐2.
Depth (m) NSPT
0‐ 2 15
2‐ 5 18
5‐ 8 20
8‐11 22
Table 3‐2: SPT blow counts acceptance criteria based on correlation to 80% relative density
Post reclamation testing indicated that the fill, with a maximum thickness of 12 m, was loose
to very loose with SPT blow counts in the range of 1 to 10, and that acceptance criteria were
not satisfied. Hence, the ground improvement works were tendered, and the project was
awarded to a ground improvement specialist contractor who had proposed the
implementation of dynamic compaction.
Further testing upon award of the ground improvement package and during the course of
the works revealed that the hydraulic fill had segregated, and in some locations a silty layer
with at least 0.5 m thickness was covering the seabed.
This project is an example of one of many problems associated with application of relative
density (refer to Section 2.10.2). ASTM (2006a), (2006b) specifies that the concept of relative
density is applicable to soils with less than 15% fines; however, as this was not the case
throughout the soil profiles of the project, the concept of relative density and thus any
correlation with it would become inapplicable.
Additionally, sampling with the split spoon of the SPT device proved to be difficult, and
occasionally produced unreliable results. Theoretically, the split should be able to extract
0.45 m long samples from the ground; however, this is not always the case in practice, and
in some boreholes the samples were just a few centimetres and sometimes the sampler had
total loss of sample.
Thus, in addition to hte criteria of relative density and SPT, pressuremeter testing was also
carried out for the design of sheet piles and further verification of the work.
298
The 900,000 m2 work area was divided into 22 sequential sub areas along the length of the
project. Energy intensity, pounder weight, drop height and number of phases were varied
based on the treatment thickness and confirmed by a calibration programme. Pounder
weights varied from 12.5 to 25 tons and a maximum drop height of 20 m was implemented.
In this project dynamic compaction was carried out only as deep treatment phases without
ironing because it was possible to meet project requirements without performing the latter
phase.
In areas with less than 6 m thickness, two phases of deep treatment using 12.5 and (14 to)
16 ton pounders were utilised. Depending on the fill thickness Phase 1 had a grid spacing of
6 m, and included 10 to 12 blows per print. Phase two prints were offset in two directions,
and were located in the middle of Phase 1’s grid. In this phase 8 to 10 blows were applied to
Phase 2 prints and an additional 2 blows were also applied to Phase 1 prints.
In deeper areas four phases of dynamic compaction comprising of three phases using 25 ton
pounders and one phase using 14 or 16 ton pounders were carried out. When fill thickness
was limited to 8 m or was 8 to 12 m it was envisaged that the number of blows per print in
the first three phases would be respectively 10 or 16 blows. Grid spacing for Phases 1, 2 and
3 were respectively set at 12 m, 12 m, and 8.5 m. Phase 4 had a grid spacing of 6 m, and each
print received 12 blows. Additionally, during this phase, prints of Phases 1 to 3 received
another 3 blows.
Figure 3‐8: Typical crater size and depth in the deep treatment areas of Abu Dhabi New Corniche
299
Sometimes due to the buildup of pore pressure the first three phases of areas with more
than 6 m had to be performed in sub phases, and sufficient time was permitted for the
excessive pore pressure to dissipate to tolerable values. The maximum breakdown of phases
occurred in one area where Phase 1, 2 and 3 were respectively performed in 4, 3 and 2 sub
phases. After each sub phase the craters were backfilled with sand, time was allowed for the
pore water pressure to dissipate, and then the next sub phase was carried out. A typical
crater size and depth in the deep treatment areas of the project is shown in Figure 3‐8.
While the suitability of dynamic compaction application is well established for saturated
granular soils, the author is frequently asked if this ground improvement method can be used
for treating saturated sands, and about the minimum distance that should be maintained
between the works and the shoreline. As shown in Figure 3‐9 the works can be carried out
basically at the shoreline itself; however, this must be done with consideration of work safety
practices.
Figure 3‐9: Application of dynamic compaction at the shoreline
Generally, a point of concern for application of dynamic compaction in residential areas is
noise and vibration disturbance (see Section 2.8.3) due to pounder impact. Although the
results of vibration monitoring is not available to the author, as can be seen in Figure 3‐10, it
was possible to perform dynamic compaction in two shifts per day parallel to the skyline of
Abu Dhabi over a length of 6 km. To minimise disturbance during nights, works were
scheduled in such a way that the areas closest to the towers were treated only during day
time.
300
Figure 3‐10: Application of dynamic compaction in front of Abu Dhabi skyline
The works were carried out with a maximum number of 7 DC rigs in two working shifts over
a period of 7 months. The maximum production rate during the course of this project was
200,000 m2 of improved ground per month.
SPTs and PMTs were carried out in this project as part of quality control, quality assurance,
and verification of works. PLM and EM at three points before and after dynamic compaction
have been shown in Figure 3‐11 to Figure 3‐13. For clarity, average PLM and EM of the same
points have also been shown in Figure 3‐14 (only Point No 3 in shown for depth of 1 m) .
It can be observed that due to the earthmoving and construction equipment traffic before
ground improvement the upper crust of the reclamation was quite dense with PLM exceeding
1,000 kPa, but the soil rapidly became very loose, generally with PLM being less than 300 kPa,
and subject to creep under self‐weight (see Section 2.9.2.8). After dynamic compaction PLM
increased up to 3,000 kPa in the upper sand layers and to at least 750 kPa at depth. Similarly,
EM increased up to 30,000 kPa in the upper sand layers, and to at least 7,000 kPa at depth. It
should be noted that improvement of the upper ground layers could have been even better
if the ironing phase of dynamic compaction was also performed.
For convenience, the ratio of pressuremeter parameters after dynamic compaction to before
dynamic compaction is defined as improvement ratio; i.e.
301
Improvement Ratio 3‐1
Figure 3‐15 to Figure 3‐17 show the improvement ratios of PLM and EM for the three test
points shown in Figure 3‐11 to Figure 3‐13. Also, the average improvement ratios of PLM and
EM for the same three test points are shown in Figure 3‐18. As can be observed that while it
is general perception that the most dynamic compaction improvement would be achieved at
the upper ground levels, in this project specifications were satisfied without the ironing
phase, and consequently improvement at the shallow depths was less than what would be
typically encountered if ironing was also performed. It is observed that the improvement
ratio of PLM was about 1.5 at the depth of 1 m, but on average approximately 7 and at most
nearly 10 at the depth of 3 m. As expected, the ratio then gradually reduced at further depths.
In this project the average peak improvement ratio was 7 (i.e., 600% increase) and the
maximum peak in any of the tests under study was approximately 10 (i.e., 900% increase).
Comparison of PLM improvement ratios in this project with previous research (Lukas, 1986)
that suggests increases would be in the order of 100 to 400% (see Section 2.5.5) indicates
that greater peak improvement ratios can be considered. However, the average PLM
improvement ratio for all depths in this project was 4.1, which is in the order of magnitude
that Lukas has noted. It should be noted that in general the pattern of foundation stress and
improvement ratio reduction in depth is similar; hence, assuming an average improvement
ratio will probably yield conservative values.
The improvement ratio of EM appears to have been bounded by a lower value of
approximately 4, which matches the order of magnitude that Lukas has observed.
302
Figure 3‐11: PLM and EM before and after dynamic compaction at test point No. 1
Figure 3‐12: PLM and EM before and after dynamic compaction at test point No. 2
303
Figure 3‐13: PLM and EM before and after dynamic compaction at test point No. 3
Figure 3‐14: Average PLM and EM before and after dynamic compaction for three points
304
Figure 3‐15: PLM and EM improvement ratios for Point No. 1
Figure 3‐16: PLM and EM improvement ratios for Point No. 2
305
Figure 3‐17: PLM and EM improvement ratios for Point No. 3
Figure 3‐18: Average PLM and EM improvement ratios
306
3.2.5 Lessons and Conclusion
1. If project size justifies, in an optimal dynamic compaction design the treatment area
should be broken down to sub areas based on design and acceptance criteria,
loading, and ground conditions. Pounder weight and drop height, compaction
intensity, grid size, and the other design parameters can consequently be
determined.
2. Relative density is an unreliable acceptance criterion.
3. It is possible to perform dynamic compaction in populated areas.
4. It is possible to perform dynamic compaction at close vicinity to the shoreline.
5. Average PLM improvement ratios after dynamic compaction were from 1.5 to 7. This
suggests that maximum improvement ratio can be more than what has been
reported in previous publications.
Figure 3‐19 shows Abu Dhabi Corniche after completion.
Figure 3‐19: Abu Dhabi Corniche
307
3.3 Al Quo’a New Township
Al Quo’a is a remote desert township that is located about 100 km from Al Ain and on the
UEA side of the United Arab Emirates ‐ Oman border (Figure 3‐20). The first phase of this
town was constructed by creating a levelled platform for the town’s foot print by cutting and
filling the dune sands. The town structures were built directly on the platform, and although
the buildings were only two stories high, most that were built on the fill areas suffered from
severe damages and substantial amounts of cracking. Consequently, the town developers
became well aware of problems associated with construction on young non‐engineered fills,
and stipulated that specific measures had to be implemented as part of the construction of
later phases of the project to avoid more damage.
Figure 3‐20: Location of Al Quo’a
The second phase of Al Quo’a, consisting of 450 two floor villas and the associated
infrastructure was planned to be constructed on a site with a rectangular shape that was 2.3
km long and 1.65 km wide, and that had an area of more than 3.8 million m2. This phase was
also anticipated to be constructed on a relatively flat platform. Thus, as shown in Figure 3‐21,
the dune hills were once again cut and dumped into the lower level areas as non‐engineered
fill.
308
Figure 3‐21: Grand size cutting and dump filling of dune sands for the construction of the second
phase of Al Quo’a Township.
The sieve analysis of the dune sand that was used for constructing the town’s platform is
shown in Figure 3‐22. It can be observed that the soil was poorly graded fine clean sand. As
could be expected, groundwater was not observed or recorded in any of the tests that were
carried down to a maximum depth of about 30 m below platform level.
Figure 3‐22: Sieve analysis of the dune sand
The fill areas were measured, and it was calculated that they covered an area of 1,135,000
m2. More detailed assessments revealed that 44% of the fill material was placed at depths
less than 6 m, 35% were from 6 to 12 m deep, 13% were from 12 to 16 m deep and the
remaining 8% were located at depths greater than 16 m below the finished working platform
(±0.0 m RL, reduced level). The maximum depth of the fill was 28 m.
309
CPT test results indicated that while the ground condition in the cut areas was satisfactory
enough to support the structures and infrastructures, as could have been predicted, the
dumped dune sands in the fill areas were in a rather loose state and except for the upper 2
m, qc was constantly in the low range of 2 to 4 MPa. The higher values of the soil strength (qc
in the range of 10 to 15 MPa) in the upper crust was contributed to the effects of the
earthmoving equipment traffic. Likewise, pressuremeter tests that were later carried out as
part of the post ground improvement programme indicated that PLM was in the range of 100
to 500 kPa, indicating that the young uncontrolled fill would be subject to creep under self‐
weight (refer to Section 2.9.2.8).
The low strength of the fill indicated that the ground would not be able to safely support the
footing loads of the villas. It was expected and calculated that the facilities of this phase of
the project would not only have problems with settlement and bearing of structural loads,
but would also suffer from creep settlements, cracks and other damages if no specific remedy
was anticipated for the loose fill.
Piling was deemed as an unfavourable solution namely because it was only feasible for
supporting the structural loads of buildings, and was not applicable for providing a
geotechnical solution for creep of the town’s platform and non‐building areas. Furthermore,
although calculations were not made at that phase, the piles had to be longer than the fill
thickness to avoid the settlements of the piles themselves in the creeping soil. This would
have incurred huge costs and a very long construction time. Consequently, Ground
improvement was deemed as a possible solution for the treatment of the very loose fill.
Noting that the maximum depth of ground improvement was 28 m, a figure that appears to
be out of dynamic compaction’s reach with typical equipment, the ground improvement
tender envisaged vibro compaction as the ground improvement method, and acceptance
was based on CPT results. While vibro compaction could be viewed as a preferred method of
soil improvement when depth of treatment of loose sands is in the order of Al Quo’a project,
its application on this jobsite would have been very challenging and difficult due to dry state
of the dune sands and the logistics for providing water for jetting. Although vibro compaction
can be done without water jetting by utilising compressed air, it is the author’s experience
that this may also turn into a challenging endeavour if the insertion cavity does not collapse,
and the vibroflot vibrations are not transferred to the soil mass. Lukas (1995) has carried out
310
a cost comparison between different ground improvement solutions. His research shows that
when applicable, dynamic compaction is the most affordable ground improvement
technique.
If applicable (due to depth constraints), dynamic compaction could have been a more
preferable solution with consideration of the below:
Dynamic compaction does not require water (or compressed air) for the works.
Dynamic compaction requires the minimal amount of equipment; i.e., the DC rig and
the pounder, and a loader for levelling the ground after each phase of compaction.
On the contrary, vibro compaction requires a rig, the vibroflot, water tanks, water
pumps, a generator, and a compressor. Handling a larger pool of equipment in an
isolated and remote site such as Al Quo’a was more challenging.
Implementation of dynamic compaction with PMT for verification was proposed as an
alternative ground improvement solution by one of the specialist soil improvement
contractors, and ultimately accepted. In addition to being the most affordable solution, the
alternative proposal had to address the technical objectives and concerns of the designers as
well.
The structures in Al Quo’a were only 2 floors, so it could have been envisaged that the stress
bulb generated by the structural loads would not extend deeper than about 6 m below
ground level. However, the ground was a thick uncontrolled and non‐engineered fill that
could have further settled under its own weight, external vibrations, and water ingress. These
conditions suggested that the specifications could specifically address external and structural
loadings on the upper ground layer when there were loads to be considered and self‐bearing
(refer to Section 2.9.2.8) at other depths.
In villa areas, single footing loads were known to be less than 100 tons, so it was possible to
stipulate an allowable bearing capacity (200 kPa) and thus a maximum footing size (2.25 x
2.25 m2). Total settlements were to be limited to 25 mm, and angular distortions were
311
required to be less than a very stringent value of 1/1,250 due to the special psychological
concerns and the previous disappointing results of the first phase of the project. It was also
possible to determine the stress bulb influence depth and to specify self‐bearing treatment
for further depths. In non‐villa areas a nominal bearing capacity of 100 kPa was stipulated for
light and unanticipated loads. The bearing capacity safety factor was 3. The approved design
criteria of the project are summarised in Table 3‐3.
By implementation of PMT calculation methods (Centre D'Etudes Menard, 1975) that have
been described in Sections 2.9.2.6 and 2.9.2.7, acceptance criteria based on design criteria
was developed and tabulated as presented in Table 3‐4 and Table 3‐5.
The depth of improvement is mainly a function of ground conditions, pounder weight and
drop height (refer to Equation 2‐41):
312
√ 2‐41
The percentage of each treatment area was defined in the project’s ground conditions
(Section 3.3.2). Assuming the coefficient of depth of improvement for dry sand is 0.7 to 0.8,
the required impact energy to reach the bottom of fill for each zone will be as shown in
Table 3‐6:
It could be understood that it was possible to treat almost 90% of the site using conventional
DC rigs that were capable of lifting pounders weighing up to 15 tons using a single cable line
or 25 tons using two single cable lines; however, the deepest region required more energy,
at least in the order of what could be provided by Menard’s 700 tm rig and in the order of
what the mega‐machine and tripods could provide for the deepest areas (refer to
Section 2.7.1).
While Menard’s 700 tm rig was available, the mega‐machine and tripods had long been
decommissioned, and an alternative solution was developed and patented specifically for
this project, and later used in other projects, such as Federal Highway BAB A 71 in Germany
(Chaumeny et al., 2008).
MARS (Menard Accelerated Release System) is an innovative automated pounder free fall
release and grab mechanism. As was discussed in Section 2.7.1, although free fall drops
provide more impact energy than cable connected drops, the former is associated with
practical difficulties as the disengagement of the pounder from the hook would cause the
crane boom to spring back; thereby ramming the hook block forcefully into the boom and
potentially damaging it in the absence of reliable shock absorbers. Furthermore,
reconnecting the pounder to the hook is a costly activity as it is a slow and time consuming
process that also requires manual labour.
313
In MARS, the pounder is released and allowed to drop in true free fall during its process of
pseudo free fall and as it is falling while connected to the hoist line. This unique release
mechanism prevents the boom to snap back, and thus, eliminates the risks of the hook block
striking the boom. Also, the pounder drop location remains in line with the vertical alignment
of the cable; hence, making it possible for the system’s mechanism to grab the pounder
without the assistance of time consuming labour. Figure 3‐23 and Figure 3‐24 show the
implementation of MARS in Al Quo’a for dropping a 35 ton pounder in true free fall and
automatically grabbing it without manual labour.
Figure 3‐23: Free fall of 35 ton pounder using MARS
In all, a total of 6 rigs were used for performing the ground improvement works at Al Quo’a
over a contract period of 10 months. Of these, 3 were single cable line heavy duty DC rigs
with lifting capacities of 15 tons (Figure 3‐25), one was a two single line heavy duty DC rig
with lifting capacity of 25 tons (Figure 3‐26), one was the Menard 700 tm rig (Figure 3‐27),
and one was a 250 ton rig for lifting the 35 ton pounder using MARS (Figure 3‐28).
Dynamic compaction energy intensity was optimised based on the treatment thickness and
acceptance criteria. The heavier 35 (free fall using MARS) and 25 ton pounders were used for
the first phase of treatment of the deep fill areas, and heavy duty rigs dropping up to 15 ton
pounders were utilised for the improvement of subsequent phases of those deep treatment
314
areas and the remaining areas of the project. By this means, the project was completed
within the contracted period of 10 months.
Figure 3‐24: Automatic grabbing of a 35 ton pounder by MARS
Figure 3‐25: DC rig for lifting 15 ton pounders at Al Quo’a
315
Figure 3‐26: DC rig for lifting 25 ton pounders at Al Quo’a
Figure 3‐27: Menard 700 tm rig at Al Quo’a
As the minimum distance between the first and second phases of Al Quo’a developments
was 200 m it was not expected that dynamic compaction induced vibrations would become
a problem. From Equation 2‐150, it can be calculated that PPV generated by a 35 ton pounder
that is dropped from 23 m will be about 3 mm/s, which is barely perceptible (see
Figure 2‐126) and well below the limit that could be damaging to the villas (see Figure 2‐122).
Actual site measurements confirmed that PPV was less than the estimated value.
316
Figure 3‐28: 250 ton rig for lifting the 35 ton pounder using MARS
The tender had stipulated that CPT tests be carried out in advance of the ground
improvement works to confirm the treatment depths. This verification programme, itself,
would have become a critical activity, and would have impacted the scheduling and progress
of works. Noting that the intent of this testing was strictly limited to determining the
boundary between loose and dense sand, the specialist ground improvement contractor
proposed an innovative solution in which a static penetration machine was developed to
perform a quick probing. Similar to a wick drain rig, as shown in Figure 3‐29, the main
components of this machine included an excavator base, a mast with sufficient height, and a
tubular rod to be pushed into the ground in lieu of a wick drain mandrel.
By this means more than 25,000 m of quick probings were carried out at 1,600 points over a
3 week period. The same activity could have required between 25 to 30 weeks using a single
CPT rig.
A total of 250 PMTs were carried out at the site. For comparative purposes, 50 tests were
carried out before ground improvement, and the remaining 200 tests were performed after
dynamic compaction. Depth of testing was based on the fill thickness and depth of dense in‐
situ soil.
317
Figure 3‐29: Quick probing machine composed of an excavator base, a mast and a penetrating
tubular rod.
Figure 3‐30 to Figure 3‐33 show PLM values before and after dynamic compaction and PLM
improvement ratios in four locations. As can be observed that except the upper layer of
ground that was relatively dense due to earthmoving equipment traffic, the soil was in a very
loose state, with PLM consistently between 200 to 500 kPa. These low values suggest that the
soil was indeed subject to creep under self‐weight (see Section 2.9.2.8). It can be observed
that post dynamic compaction PLM values have substantially increased in an almost classical
sickle shaped curve to nearly 2,500 kPa and at least to about 1,400 kPa in the upper 6 m. PLM
values then gradually reduced to figures in the order of 700 to 900 kPa. The test results
indicate that acceptance criteria defined in Table 3‐4 and Table 3‐5 have been satisfied. Peak
PLM improvement ratios in these tests were as high as 20.5 in Test No. 1 (refer to Figure 3‐30)
and from 5.2 to 6.6 in other tests. The improvement ratio at depth ranged from 1 to 3.
Average PLM values and PLM improvement ratios for the four tests are shown in Figure 3‐34.
Averaging the values at each level smoothens out any unusually high or low values and allows
a more representative view. It can be seen that before dynamic compaction the upper
ground level is relatively dense due to earthmoving equipment, but the soil then becomes
very loose and subject to creep under self‐weight. The limit pressure values appear to
increase linearly and gently with depth. After dynamic compaction PLM values in the upper
layers of ground increase substantially, and are about 2,000 kPa over a thickness of about 5
m, and consistently around 800 kPa at the deepest layers. Peak PLM improvement ratio is 6.7
318
at 4 m, and more than 5 (400% improvement) over a thickness of about 3.5 m). After the
peak value, improvement ratio initially declines with a greater slope to 2.6 at the depth of 9
m, but then continues to reduce in value to 1.4 at the depth of 15 m.
As can be observed that all post improvement test results basically follow the variation with
depth that Lukas (1986) has suggested. PLM values increase to a peak value at about ⅓ to ½
of the influence depth and then deceases to a point where improvement would be negligible.
Lukas (1986) also assumes an upper bound post dynamic compaction PLM value of 1.9 to 2.4
MPa for sands and gravels (refer to Section 2.5.5), which is in line with the results of the
works carried out in this project.
Lukas (1986) also suggests an upper bound improvement value of 400% for PLM due to
dynamic compaction. The results of Figure 3‐30 to Figure 3‐34 indicate that it is possible to
improve PLM of very loose dune sands by up to 1950% at least in one instance and by up to
570% as an average for peak improvements. In fact, test results show that in the study area
of this project, on average, approximately 20% of the treated thickness improved by at least
400%.
In addition to the PMT, 3 zone load tests and 5 plate load tests were also carried out in this
project. Plate load tests were performed on a 900 millimeter diameter plate at ‐0.75 m RL to
a maximum pressure of 200 kPa. The objective of the zone load tests was to measure the
settlements under an applied load of 25 kN/m2, similar to the total loading of an average villa,
for 7 days. The loading dimensions were equal to the villa dimensions.
Measured site settlements after dynamic compaction were in the range of 0.60 to 0.80 m.
The review of this project can provide the geotechnical engineer with a number of lessons to
be incorporated in future projects, including:
1. It is more likely that non‐engineered backfilling will be loose and potentially subject
to low bearing and excessive settlements. It is recommended that ground
improvement be envisaged during planning stage to avoid any surprises and
disappointments during development.
319
Figure 3‐30: PLM and ratio of PLM before and after dynamic compaction at test point No. 1
Figure 3‐31: PLM and ratio of PLM before and after dynamic compaction at test point No. 2
320
Figure 3‐32: PLM and ratio of PLM before and after dynamic compaction at test point No. 3
Figure 3‐33: PLM and ratio of PLM before and after dynamic compaction at test point No. 4
321
Figure 3‐34: Average PLM and ratio of average PLM before and after dynamic compaction
2. Dynamic compaction can be understood to be an affordable and effective ground
improvement for thick dry desert dune sands in isolated locations.
3. Proper determination of design criteria is very important, and failure to adopt a
suitable specification can lead to unnecessary treatment, additional costs and delay.
4. The most suitable acceptance criteria are based on design criteria. It is not necessary
to specify minimum test results as proper testing should be able to verify that design
criteria have been satisfied. Also, one criterion does not have to govern throughout
the depth of treatment. In Al Quo’a the upper layers were treated for bearing
capacity and settlement under structural loads while the deeper soils were treated
for self‐bearing.
5. In large projects, it is preferable to mobilise dynamic compaction rigs with different
capacities and to optimise treatment by applying various energy intensities based on
the requirements of each zone.
6. It is possible to efficiently improve the depth of influence in dynamic compaction by
implementation of the free falling and automatic grabbing MARS technology.
322
7. It is possible to improve maximum PLM values of dune sands on average by up to
570% and at least in one instance by up to 1950%. Improvement beyond 400% was
observed over a thickness of 20% of treatment depth.
323
3.4 King Abdulla University of Science and Technology
The 5.6 million m2 King Abdulla University of Science and Technology (KAUST) is located in
Rabigh, on the Red Sea coast and near the city of Jeddah, in Saudi Arabia. KAUST, originally
anticipated to have buildings with at most two to three stories, includes the university
campus and academic administration core, a desalination plant, wind turbines, residential
neighbourhoods, a research park, a commercial centre, a waste water treatment plant and a
beach club. The university plan is shown in Figure 3‐35.
Figure 3‐35: Plan of KAUST
The concept of the project was developed in 2006. According to the schedule, master
planning, architectural and structural design and construction had to be completed in less
than three years and handover date was set at September 2009.
The preliminary geotechnical investigation that was carried out at a relatively wide grid
indicated that the ground was very heterogeneous and composed of a combination of loose
324
or soft soils with rapid variations within short distances. For example, qc of two test locations
that were 30 m apart are shown in Figure 3‐36. It can be observed that in the first test qc was
approximately 5 MPa in the upper 1.5 m, but then dropped to almost null for the next 4 m
until the ground strength increased and ultimately reached more than 20 MPa at the depth
of 8.5 m. However, in the second test qc in the upper generally sandy 2.5 m of ground is about
7 MPa, but drops to about 4.5 MPa (which is much more than the values of the first test at
the same depths) in the finer soil that extends down to the depth of about 5.5 m, and then
further reduces to about 2 MPa (which is much less than the values of the first test at the
same depths).
Figure 3‐36: Variation of ground conditions for two test locations that were 30 m apart.
This investigation and further testing during the works revealed that more than 2.6 million
m2 of the construction area was to be built on soil consisting of up to 9 m of loose silty sand
or soft sandy silt, which is locally called sabkhah, and originates from storms or eolien actions
that have resulted in tidal lagoon deposits, and contain high concentrations of salt.
A generalised schematic profile of the ground and a summary of the test results for each soil
layer are respectively shown in Figure 3‐37 and Table 3‐7.
325
Figure 3‐37: Ground profile at KAUST
Figure 3‐38: CPT‐19
326
A number of CPTs that demonstrate the variation of ground conditions are shown in
Figure 3‐38 to Figure 3‐43.
Figure 3‐39: CPT‐178
Figure 3‐40: CPT‐27
327
Figure 3‐41: CPT‐12
Figure 3‐42: CPT‐138
328
Figure 3‐43: CPT‐206
As groundwater level was less than 1 m below in‐situ ground level it was decided to raise the
ground by about 3 m to be safely above high tide. Approximately 1 to 1.5 m of this granular
fill was placed before ground improvement.
The unsuitable ground conditions was a warning to the project team that they should be
expecting foundation problems, and with the objective to complete the project in due time,
it was decided to develop and apply a ground improvement plan while design was
proceeding.
329
3.4.3 Development of the Ground Improvement Solution
While it was evident that only a well worked out schedule that incorporated design and
construction could meet the project’s deadline, the problematic soil that covered an area of
approximately 2.6 million m2 posed a serious threat to this programme as it was not possible
to design the foundations without the finalisation of the buildings’ locations and architectural
drawings. The size of the area containing the problematic soil was so large that if ground
improvement was to be undertaken in a single package (which became the case), then to the
knowledge of the author, it would have become the world’s largest ground improvement
project of its time.
All possible foundation solutions were considered. Piling was immediately deemed as
infeasible because without architectural and structural design it was not possible to define
the pile locations. Thus, only ground improvement remained as a potential solution.
Ultimately, the record breaking (to the knowledge of the author) ground improvement
design and construction package was awarded to a specialist contractor that had proposed
the application of dynamic compaction and dynamic replacement.
Design criteria were stipulated as:
Footing location: Any place within the treatment area
Maximum footing load: 1,500 kN
Allowable bearing capacity: 200 kPa
Maximum total settlement: 25 mm
Maximum differential settlement between two adjacent footings: 1/500
Liquefaction mitigation for an earthquake with (peak ground acceleration) PGA=
0.07g
Foundation level: 0.8 m below final ground level, but in any case at least 2 m above
sabkhah level.
While it was possible to construct a footing at any location without any concerns in the sandy
zones that were to be treated by dynamic compaction, per se the same was not true after
treatment of the zones consisting of sabkhah by dynamic replacement. In the latter case the
structural loads would have had to distribute between the in‐situ soft soils and dynamic
replacement columns by arching (refer to Sections 2.4.9.12.4.9 and 2.6.1). Consequently and
as shown in Figure 3‐44 to Figure 3‐46, finite element analyses was undertaken to determine
the DR grid size for safely distributing the footings’ loads without punching into the ground
330
in between the DR columns. Based on a minimum PLM value of 180 kPa in the sabkhah layer,
the optimised column grid was determined to be a square grid of 3.8 m with minimum DR
column diameters of 2.2m.
(a)
(b)
Figure 3‐44: Finite element modelling for DR columns at (a) 5.5 m grid and (b) 3.8 m grid
(a) (b)
(c) (d)
Figure 3‐45: Vertical effective stresses for 5.5 m spaced DR grid elevations (a) 0.0 m RL, (b) ‐1 m RL,
(c) ‐2 m RL and (d) ‐3 m RL
331
(a) (b)
(c) (d)
Figure 3‐46: Vertical effective stresses for 3.8 m spaced DR grid elevations (a) 0.0 m RL, (b) ‐1 m RL,
(c) ‐2 m RL and (d) ‐3 m RL
The foundation concept is shown schematically in Figure 3‐47.
Figure 3‐47: Foundation concept
Prior to application of ground improvement by a combination of dynamic compaction and
dynamic replacement techniques the working platform was constructed by dumping and
levelling 1 to 1.5 m of granular fill. Roller compaction was not necessary as the platform
332
material was to be compacted in the process of soil improvement works. This was
advantageous because not only was the cost of levelling in layers, watering, removal of
oversize material, mixing and roller compaction eliminated, but more importantly precious
construction time was saved.
In this project pounders weighing up to 25 tons were used to compact soils with fines content
up to about 30 to 35% using dynamic compaction. Dynamic replacement was used in areas
where the maximum depth of sabkhah was 5 m. When the depth of sabkhah was more than
5 m, high energy dynamic replacement (HDR) was used in conjunction surcharging the areas
for a period of 6 weeks. The surcharge was placed to 3 m above final grade level.
Figure 3‐48: Application of 3 m surcharge above final grade level after dynamic replacement in areas
with sabkhah being deeper than 5 m.
After completion of ground treatment in specific areas, it became known that the revised
master plan incorporated 36 buildings with more floors and heavier loads. Consequently, a
derivative of dynamic compaction called dynamic surcharging was also used to consolidate
the deep sabkhah layers. In this technique the combination of preloading and vibration is
used to re‐introduce pore pressure in the soil‐water system and consequently to accelerate
settlement rates. It is the author’s experience (Varaksin, 2014) that dynamic surcharging is
applicable to silts, but not to clays, and additionally the degree of consolidation must be
roughly in the range of 50 to70%.
Hence, in addition to the engineered fill required for reaching final grade level, a 3 m high
surcharge was placed, and as shown in Figure 3‐49, dynamic compaction was performed on
top of the surcharge. The surcharge was left in place for 3 weeks.
333
Figure 3‐49: Dynamic Surcharging in KAUST
As previously noted, variations in ground conditions occurred quite rapidly, within short
distances and the field geotechnical tests were performed in a relatively wide grid; hence, it
was necessary to develop a method for determining the ground condition, and applying the
relevant ground treatment method. Although performing additional field tests would have
been the first logical choice, the introduction of hundreds of additional tests would have
delayed the project. Therefore, a highly accurate observational approach was implemented.
In this approach the trend of sabkhah depth was initially identified by the field tests.
Differences in ground behaviour due to pounder impact enabled the site supervisors to
assess the rapidly varying ground conditions, and to apply the appropriate ground
improvement technique as needed. It was observed that while the first DC pounder impact
penetrated the ground by about 0.25 m, the DR pounder penetration was substantially more
and in the range of about 1 m. Also, performing DC frequently resulted in the seepage of
groundwater to the surface, but this phenomenon was rarely encountered in the DR areas.
Ground rest periods in between DC phases were 1 to 3 days. However, this period was
considerably longer and from 7 to 21 days when DR was utilised. Furthermore, ground heave
due to pounding was not observed in DC areas, but was evident in DR zones.
A pilot test was realised with pressuremeter testing and SPT (for grain size analysis) to define
the boundaries of application of the dynamic compaction and dynamic replacement
techniques as a function of grain size, PLM and applied energy. Furthermore, a spread sheet
334
based on PMT interpretation rules (Centre D'Etudes Menard, 1975) was prepared for the
quick estimation of the bearing capacity and settlement by the site engineer.
Figure 3‐50 shows the flow chart for determining the applicable ground improvement
technique.
Figure 3‐50: Flow chart for determining applicable ground improvement technique
The allocated time frame for mobilisation, execution and testing of ground improvement
works was 10 months. Ground improvement was carried out over a period of 8 months using
a total of 13 rigs working two shifts per day. To the knowledge of the author, this is the world
record for the number of DC rigs working at the same time on a single project. A review of
production data shows that at its peak, the rate of ground improvement was about 600,000
m2 per month, which to the knowledge of the author was the world record at that time.
Figure 3‐51 shows 12 rigs symbolically lifting their pounders. The 13th rig was used for
providing the necessary elevation for taking the photograph.
th
Figure 3‐51: Utilisation of 13 DC‐DR rigs in two shifts (photograph taken from the 13 rig)
335
In addition to the 13 rigs, other equipment that were implemented in the project included
15 pounders weighing from 12 to 25 tons, 30 vehicles, 1 truck equipped with a crane, 1
forklift, 3 CPT rigs, 3 SPT rigs, and 3 PMT rigs.
During the course of ground improvement works a total of 76 test pits, 122 SPTs boreholes,
672 CPTs, and 403 PMTs were carried out for determination of the appropriate ground
improvement method, fines content, liquefaction mitigation, bearing capacity and
settlements.
PLM before and after dynamic compaction in one of the test locations is shown in Figure 3‐52.
It can be observed that while PLM was as low as 200 kPa at some depths before dynamic
compaction, PLM increased by 3 to more than 4 times after dynamic compaction. PLM
improvement ratio can be seen to be variable from as low as 1.9 to 4.5 with an average value
of 2.9.
Figure 3‐52: PLM before and after dynamic compaction and PLM improvement ratio
Figure 3‐53 shows PLM before and after (between DR columns and inside one of the DR
column) dynamic replacement in one of the test locations. It can be observed that dynamic
336
replacement has been able to the improve the sabkhah layer in between the DR columns
throughout the testing depth to at least 190 kPa and that average PLM improvement ratio of
the in‐situ is approximately 2.3; however, this amount of improvement would not have been
sufficient to satisfy the design requirements without the DR columns. The peak and average
PLM improvement ratio of the DR column were respectively 18.56 and 10.2, which
demonstrate a massive improvement in the ground conditions.
Figure 3‐53: PLM before and after dynamic replacement and PLM improvement ratio
Due to the wide range of ground conditions, including soil classification, strength, and
thicknesses of layers, test results showed a wide range of results. Although the results clearly
demonstrated that design criteria had been satisfied, it is interesting to note that it is possible
to classify the test results according to the amount of dynamic energy applied to the ground
and the amount of improvement. Figure 3‐54 shows the relation between the net
pressuremeter limit pressure (PLM‐Po) and the energy per volume of improved soil. As
expected, it can be observed that the improvement factor (I) or the ratio of after to before
soil improvement net limit pressure (only the in situ soil without the DR columns) is higher
and more efficient when the soil is more granular and contains lesser fines. Furthermore, it
can be noted that the DR columns play a vital role in providing the required ground properties
in the sabkhah soils, and their exclusion will result in failure of satisfying the design criteria.
337
Figure 3‐54, which has been realised by Serge Varaksin, also indicates that the boundary
between the application zones of DC and DR is at about 30% fines content.
Figure 3‐54: The relationship between net limit pressure, fines content and improvement energy
The effect of surcharging in one of the areas where the depth of sabkhah exceeded 5 m is
shown in Figure 3‐55. As can be observed, even though dynamic replacement columns had
strengthened the ground in this area, settlement of the deep soft soils would have potentially
remained a geotechnical concern, and could have resulted in damaging settlements without
the application of surcharging.
Figure 3‐55: Settlement induced by surcharging after implementation of DR in areas where sabkhah
depth exceeded 5 m.
338
More interesting than inducing settlement of deep sabkhah layers in areas that have been
treated by dynamic replacement is the pattern of induced settlements realised by dynamic
surcharging. As can be seen in Figure 3‐56, prior to the commencement of the first phase of
dynamic surcharging the ground was gently settling due to the static preload that had been
placed on the ground. However, contrary to classical consolidation theories that predict a
reduction of rate of consolidation as time progresses, application of the first phase and
second phases of dynamic surcharging suddenly increased both the magnitude and rate of
settlement.
Figure 3‐56: Settlement induced by static and dynamic surcharging
This project demonstrates that classical solutions are not able to meet the requirements of
special projects in which the site is unusually large, preliminary field tests are most probably
insufficient, design and construction period is relatively short, and construction must
commence in parallel to the preparation of the architectural and structural drawings.
However, it is possible to construct such a project successfully if:
1. Design and construction phases are merged smartly and efficiently.
2. Design and acceptance criteria are specified based on the actual project
requirements.
3. A combination of ground improvement techniques are used for treating different
ground conditions rather than forcefully trying to implement one technique to a
339
variety of different conditions. Furthermore, it may be more efficient to provide
different amounts of treatment for each technique.
4. The limits of each technique are well understood. The efficiency of dynamic
compaction rapidly deteriorates with the increase of fines content. It appears that
the limit of efficiency is when fines content reaches approximately 30%.
5. It is the author’s experience that dynamic surcharging is applicable to silts, but not
to clays, and additionally the degree of consolidation must be roughly in the range
of 50 to70%.
6. Possible production rates are well understood and sufficient amounts of plant and
equipment are mobilised.
7. Site observations are given value and changes in actual ground conditions are taken
into account during the process of the works.
Figure 3‐57 shows King Abdulla University of Science and Technology after construction.
Figure 3‐57: King Abudulla University of Science and Technology
340
3.5 Al Falah Community
As part of the development of Abu Dhabi, Aldar Properties launched the 12.7 million m2 Al
Falah Community Development project in 2008 in the outskirts of the capital city of UAE. The
project was anticipated to include 5,000 villas, 2,300 townhouses, 2,100 apartment houses,
14 schools, a hospital, a shopping mall, a number of hotels, restaurants, and health clubs.
The town’s master plan is shown in Figure 3‐58.
Figure 3‐58: Master plan of Al Falah Community
The geotechnical investigation of the project indicated that the site was covered with a
superficial layer of silty sand that had a variable thickness of only several centimetres to more
than 18 m, and which was underlain by sandstone or siltstone bedrock.
As illustrated in Figure 3‐59 (white areas), the major portion of ground at the site was very
dense, and it was possible to construct shallow foundations without any geotechnical issues.
The SPT blow counts in these areas were consistently more than 50, and CPT penetration
would generally reach refusal within the first metre.
341
Figure 3‐59: Site plan and limit of loose silty sands (hatched areas)
However, the ground conditions were not suitable for foundation construction throughout
the entire site, and the test results indicated the presence of loose soil, shown as hatched
zones in Figure 3‐59, that approximately covered a total area of 4.84 million m2. SPT blow
counts in the superficial layer of these areas were generally from 7 to 12 and occasionally as
low as 2. Similarly, qc was mostly 2 to 3 MPa. Fines content of these loose soils were usually
less than 25% but occasionally higher. CPT friction ratio was generally less than 1%, but
occasionally increased to as high as 3%. Typical SPT and CPT results are shown in Figure 3‐60.
(a) (b)
Figure 3‐60: Two typical test results of initial ground conditions (a) CPT, (b) SPT
342
Further study and analyses of the data indicated that the thicknesses of loose soil deposits
were less than 3 m in 66% of the site. 20% of the loose soils had thicknesses of 3 to 6 m, 8%
had thicknesses of 6 to 10 m and 6% had thicknesses of more than 10 m and exceptionally
up to 18 m.
It was observed that while the groundwater level was relatively deep and from 11 to 18 m
below ground level, moisture content of the ground varied from 8 to 35% in the non‐
saturated layers.
The study, analyses and assessment of the geotechnical investigation and preliminary
calculations demonstrated that the areas with poor ground conditions could not meet the
project’s foundation design criteria, which was defined as:
Allowable bearing capacity: 150 kPa for conventional strip or pad foundations with
maximum dimensions of 1.5m1.5m2 (villa areas) or 33 m2 (heavy loads).
Maximum total settlement: 25 mm for a maximum pressure of 150 kPa applied to
the above mentioned footings
Differential settlement: 1:500 measured between surface points not closer than 8 m
apart.
Liquefaction mitigation: for an earthquake with magnitude 6 and PGA= 0.15g
It was assessed that piling could not be included among the optimised foundation solutions
as it was not only costly, but could not be carried out within the allocated time frame for
foundation works; hence, ground improvement was identified as the preferred method for
foundation construction.
Based on the experiences of a number of mega size ground improvement projects local to
the Middle East, such as the 2.6 million m2 King Abdulla University of Science and Technology
(Section 3.4), the 1.13 million m2 Al Quoa New Township (Section 3.3), and the 0.9 million m2
Abu Dhabi New Corniche (Section 3.2), one of the ground improvement specialist contractors
proposed the implementation of dynamic compaction, and was awarded a 240 day design
and construct contract to carry out the ground improvement works.
To the knowledge of the author, this project was the largest (in size) single soil improvement
contract that had ever been undertaken at its time. Furthermore, the schedule was very tight,
343
and it was expected that the average ground improvement production rate during the
contract period had to exceed 700,000 m2 per month for the project to reach completion on
time. As production rate was expected to increase gradually with the mobilisation of
additional DC rigs, peak production rate was expected to exceed 900,000 m2 per month for
reaching the defined milestones. This, itself, was a challenge as the peak production rate
world record was understood to be 600,000 m2 per month (refer to Section 3.4) at that time.
It was evident that meeting the design specifications for the world’s most demanding
dynamic compaction project required the optimisation of impact energy by selecting the best
combination of DC parameters. As a first step, it was decided to divide the treatment zones
according to treatment thicknesses (0 to 3 m, 3 to 6 m, 6 to 10 m, and more than 10 m) and
loading intensity (low rise villas or more heavily loaded apartments).
A total of 11 DC rigs were mobilised to complete the project within the contract period.
Figure 3‐61 shows six DC rigs during execution of the works. Heavy steel pounders weighing
up to 23 tons were implemented to reduce the required number of drops per print.
Figure 3‐61: Implementation of 11 special cranes for DC (not all shown in this photograph)
To facilitate site verification of tes results without detailed calculations and comparison of
results with design criteria, a quick verification method that is shown in Table 3‐8 was
developed. It can be calculated by using Menard’s method (Centre D'Etudes Menard, 1975)
that safe bearing capacity and settlement of villas with the PMT parameters of Table 3‐8 will
344
be respectively 360 kPa and 5 mm. For the heavily loaded structures safe bearing capacity
and settlement can be calculated to be respectively 310 kPa and 7 mm.
If test values exceeded the defined parameters, then acceptance would have definitely been
achieved; otherwise detailed calculations would have been necessary.
The optimisation of the DC design and ability to provide sufficient number of rigs allowed the
completion of the project before the due handover date; i.e., within a period of 7 months. It
can be observed from Figure 3‐62 that the maximum ground improvement production rate
per month was 966,000 m2, which to the knowledge of the author is the world’s current
ground improvement production rate record.
Figure 3‐62: Accumulative treated area
3.5.4.1 Calibration
Due to the scale of the project and benefits of optimising the DC parameters as much as
possible, prior to commencement of dynamic compaction a number of calibration
programmes were carried out in areas with different thicknesses. The results of one of these
345
studies is discussed and analysed hereunder. In this programme the weight of the pounder
was 20 tons, its base dimensions were 2x2 m2, and the drop height was 20 m.
Figure 3‐63 shows the preliminary stage of the calibration programme where the pounder
was dropped 4 to 8 times on 5 prints. The distance between the centres of the prints were
selected at 14 m to eliminate the effect of compacting one print on the other. In this part of
the calibration programme a CPT was carried out next to (7 m away from the centre of) each
print, and one CPT in the centre of each print after compaction. Preliminary calibration as
undertaken in this project is, if ever, seldom, performed on typical projects as the potential
benefits do not justify the required time and effort to perform it; however, the size of this
project was an opportunity to perform this additional activity.
Figure 3‐63: Calibration of phase 1 number of blows
CPT cone resistance before and after dynamic compaction for 6, 7 and 8 blows are compared
in Figure 3‐64. It can be observed that prior to dynamically compacting the prints the upper
surface of the ground was very dense, but the sand then became loose to very loose.
Application of the minimum number of 3 blows has met the quick criteria of Table 3‐8, and
this calibration suggests that there is no need to apply more than 3 blows to satisfy the
project requirements at depth in the print location. Although the magnitude of qc suggests
that the quick criteria may have been also satisfied in between the prints, the results of this
preliminary calibration per se do not definitively make such a conclusion and further testing
during the calibration programme will be required.
It can be seen that DC blows have improved the soil strength at depths lower than about 1.2
to 1.5 m in the order of 3 to 5 times. The thicknesses of the superficial loosened sands were
somewhat more than the pounder penetration depths. This is not very unexpected as the
impact points were backfilled with sand from the platform, and ironing was not carried out.
346
(a) (b) (c)
Figure 3‐64: CPT cone resistance before and after DC (a) 6 blows, (b) 7 blows, (c) 8 blows
Cone resistance of CPTs performed in the prints after dynamic compaction is shown in
Figure 3‐65. Although it appears that the sequential increase of blows at some depths has
led to progressive improvement, the variation of the CPT profiles does not lead to such a
conclusive observation.
Figure 3‐65: CPT cone resistance after compaction of prints
The layout of one of the calibrations is shown in Figure 3‐66. In this part of the programme
the same pounder (with a base area of 2x2 m2 and weighing 20 tons) was also dropped from
20 m in 4 compaction patterns that have been shown in Figure 3‐66. In the first phase of
compaction the pounder was dropped 3 to 6 times on 4.95x4.95 m2 grids. In the second phase
of compaction the in between prints that were on a 3x3 m2 grid were each subjected to 3
347
blows, and each of the prints of phase 1 received one more blow. A penetration test was also
carried out using 20 blows.
Figure 3‐66: Dynamic compaction calibration layout
Prior to dynamic compaction a CPT was carried out at the calibration area, and 3 CPTs were
also carried out after execution of dynamic compaction in each of the compaction patterns
shown in Figure 3‐66. These latter CPTs were carried out inside phase 1 print centres (with
varying blow counts depending on the pattern), inside phase 2 print centres (with 3 blows)
and in between the prints. For each pattern, one PMT was also performed in between the
prints.
Figure 3‐67 shows qc values of the preliminary calibration (3 tests), the patterns (1 test) and
the average qc of all 4 tests before dynamic compaction. As before, it can be observed that
the superficial ground layer is very dense, but the soil then becomes loose with qc in the range
of 2.5 to 3.5 MPa.
Cone resistances and average qc improvement ratio, calculated as the ratio of average qc after
dynamic compaction to average qc before DC, are shown in Figure 3‐68 to Figure 3‐71. It can
348
be seen that even the pattern with the least amount of energy intensity (3 blows in phase 1)
is able to satisfy the quick criteria. More interestingly, the variation of qc in phase 1 prints,
phase 2 prints and in between prints is small, in particular below the depth of about 1.5 to
2m, and can be assumed to be practically the same as the average. At depth above
approximately 1.5 to 2 m, although the quick criteria are still satisfied in all CPT profiles, it
seems that the lowest qc occurs in between the prints. No specific relationship can be
observed in this superficial layer between phase 1 and phase 2 prints as qc values are
sometimes higher in phase 1 prints, sometimes higher in phase 2 prints, and sometimes
appear to be practically equal.
In all patterns qc improvement ratio is greatest at depths of about 3 to 4 m, and in the range
of about 4.5 to 5.5, which is in the order that has been proposed by Lukas (1986).
Comparison of the average qc and average qc improvement ratio in Figure 3‐72 indicates that
the maximum amount of improvement has been achieved in the pattern with the maximum
energy intensity (6 blows in phase 1), but the overall qc improvement ratio of the four
patterns seem to be very close, and it cannot be immediately concluded that applying
additional blows has improved the ground strength.
Figure 3‐67: CPT cone resistances before dynamic compaction
349
Figure 3‐68: qc and average qc improvement ratio for pattern with 3 blows per print in phase 1
Figure 3‐69: qc and average qc improvement ratio for pattern with 4 blows per print in phase 1
350
Figure 3‐70: qc and average qc improvement ratio for pattern with 5 blows per print in phase 1
Figure 3‐71: qc and average qc improvement ratio for pattern with 6 blows per print in phase 1
351
Figure 3‐72: average qc and average qc improvement ratios of all four patterns
In addition to the PMT that was performed before dynamic compaction, one PMT was carried
out in between the prints after DC in each pattern. Due to project requirements and gage
limitations tests were terminated for PLM of 2,500 kPa and EM of 20,000 kPa (assuming EM
/PLM= 8, refer to Table 2‐22). Figure 3‐73 to Figure 3‐76 compare PLM and EM values in the
four patterns before and after dynamic compaction. As before, it can be observed that even
the pattern with the least amount of energy intensity is able to satisfy the quick criteria with
a margin of almost 100%. Comparison of PLM and EM values of the four patterns in Figure 3‐77
is more oriented than the CPT results, and the trend of increasing soil strength with
application of additional energy appears to be more obvious. However, even in the PMT
results there is a fluctuation, and the pattern with the most blows does not yield the highest
PLM and EM values.
PLM and EM improvement ratios of the patterns are shown in Figure 3‐78 to Figure 3‐81. While
at first glance it appears that the magnitude of PMT improvement ratio is less than CPT
improvement ratio, it is reminded that PMTs were terminated at 2,500 kPa, and higher
improvement ratios would have been achieved if the tests were continued. PLM and EM
improvement ratios of the patterns are compared in Figure 3‐82. As with the PLM and EM
values themselves, while the general trend of improvement indicates that additional blows
352
further increase PMT parameters, the pattern with the most energy intensity does not yield
the highest parameters.
Figure 3‐73: PLM and EM for pattern with 3 blows per print in phase 1
Figure 3‐74: PLM and EM for pattern with 4 blows per print in phase 1
353
Figure 3‐75: PLM and EM for pattern with 5 blows per print in phase 1
Figure 3‐76: PLM and EM for pattern with 6 blows per print in phase 1
354
Figure 3‐77: Comparison of PLM and EM before and after dynamic compaction
Figure 3‐78: PLM and EM improvement ratios for pattern with 3 blows per print in phase 1
355
Figure 3‐79: PLM and EM improvement ratios for pattern with 4 blows per print in phase 1
Figure 3‐80: PLM and EM improvement ratios for pattern with 5 blows per print in phase 1
356
Figure 3‐81: PLM and EM improvement ratios for pattern with 6 blows per print in phase 1
Figure 3‐82: Comparison of PLM and EM improvement ratios of all 4 patterns
357
After dynamic compaction, the 4 patterns were levelled, and the amount of settlement
measured. It was observed that the ground settlement in the first (with the least energy
intensity of 3 blows in phase 1) to fourth pattern (with the most energy intensity of 6 blows
in phase 1) were respectively 0.16, 0.25, 0.26, and 0.32 m.
The crater diameter, pounder penetration, and crater volume for every pounder blow during
the heave and penetration test is shown in Table 3‐9. Pounder penetration was measured by
both surveying four points on the pounder (the average penetration is presented here only)
358
and implementing the drop line measuring device of the rig. It can be observed that the
measurements using these two methods are fairly similar in the practical sense.
The accumulative penetrations that were measured by surveying and the crater volume are
plotted against the number of blows in Figure 3‐83. It can be observed that the shapes of
accumulative penetration and crater volume curves are very similar, and the slope of both
curves reduce as the number of blows increases. This can be better observed in Figure 3‐84,
which shows the percentages of final accumulative penetration and final crater volume
against the number of blows. It can be seen in this plot that 80% of the final pounder
penetration was achieved in the first 10 blows while the pounder only penetrated the last
20% of the total penetration in the second 10 blows. Likewise 80% of the crater volume was
achieved during the first 10 blows, and the second 10 blows only contributed to 20% of the
crater volume.
The project specification required that bearing and settlements to be verified by performing
PMTs. Also, it was anticipated to carry out CPTs for verification of liquefaction mitigation and
as an additional control measure. Consequently, in addition to the 18 PMT and 32 CPT that
were carried out as part of the full scale DC calibration program, a total of 282 PMT and 1,029
CPT were also carried out to confirm that the project requirements had been met. As the
final test results are comparable and compatible with the calibration results further tests will
not be presented.
Figure 3‐83: Accumulative penetration and crater volume against number of blows
359
Figure 3‐84: Percentages of accumulative penetration and crater volume against number of blows
To the knowledge of the author, at the time Al Falah Community Development was under
construction, it held the world records for largest (in size) single ground improvement
contract at 4.84 million m2 and highest ground improvement production rate of 966,000 m2
per month.
Worthwhile points in this project are:
1. The site was broken down to sub‐areas, and treatment was performed based on
ground conditions and structural loads.
2. It is possible to meet the project specifications in a shorter time by using high heavy
pounders with high drops rather than using lighter pounders with the same energy
intensity.
3. Early planning will allow mobilisation of sufficient numbers of plant and equipment
to perform grand size projects in relatively short periods.
4. The improvement ratio in the calibration programme was in the order that has been
suggested by Lukas (1986). However, it could be cautiously expected that higher
energy intensities will result in better results.
Figure 3‐85 shows Al Falah Community Development after dynamic compaction and
completion of villas and roads.
360
Figure 3‐85: Al Falah Community
361
3.6 Marjan Island Road Corridor
Marjan Island is the first manmade group of island(s) of its kind that has ever been
undertaken in the northern emirate of Ras Al Khaimah in the United Arab Emirates. This
project is located approximately 27 km southwest of Ras Al Khaimah city and 54 km
northwest of Dubai city.
Marjan is the Arabic word for coral, and as can be seen in the project’s master plan that is
shown in Figure 3‐86, the reclamation is composed of a peninsula followed by four coral
shaped islands that are connected together via bridges. Upon completion Marjan Island will
include various structures including low rise villas, townhouses, apartment buildings, retail
and leisure areas, utilities, hotels, resorts, and open spaces.
Figure 3‐86: Master plan of Marjan Island (main road corridor shown in white)
Unlike most manmade islands in the UAE where land has been reclaimed from the sea by
hydraulic filling, in Marjan Island trucks were used to cart and dump soil more than 3 km into
the Persian Gulf to an average elevation of +4 m ACD (Admiralty Chart Datum) that is 2 m
above average groundwater level.
362
The preliminary geotechnical investigation on the Island’s main road (shown as a red strip in
Figure 3‐87) that passes through the Peninsula, Island 1 and Island 2 consisted of 37 SPT
boreholes drilled down to depths of 2.5 to 16 m. These tests indicated the presence of a
heterogeneous fill. On average, in the top 7 m that extended down to ‐3 m ACD the fill was
composed of very loose to medium dense sand that was occasionally interbedded with
boulders at different depths. Fines content was variable from 13% to 30%, and SPT blow
counts were generally low, sometimes as low as 4, but rarely exceeding 50. The second layer
of soil, extending down to ‐12 m ACD, appeared to possess better mechanical properties.
This layer was composed of medium dense to very dense silty sand with occasional
interbedded pockets of sandy silt. Fines content was generally from 5% to 30%, and SPT blow
counts were from 10 to more than 50.
Figure 3‐87: Marjan Island satellite image (taken in 2009) with the main road identified in red
Marjan Island’s SPT borehole layout plan of the Peninsula and Island 1 are shown in
Figure 3‐88. Figure 3‐89 shows the blow counts of the upper 8.5 m of the Peninsula’s first 5
SPT boreholes that were spread out at distances of approximately 70 to 80 m from one
another. It can be observed that not only was the soil generally in a loose state, but also that
the scatter of blow counts at any one depth was considerable, which indicated that in
addition to bearing and total settlement issues, the ground could have also been subject to
excessive differential settlements.
363
Figure 3‐88: SPT borehole layout
Figure 3‐89: SPT blow counts of several boreholes at Marjan Island before dynamic compaction
Later, 16 PMTs were also carried out as part of a supplementary geotechnical investigation.
PLM in these tests also indicated that the ground was sometimes very loose in the upper 7 m,
in such a way that the lowest measured PLM was as low as 50 kPa. Although due to the traffic
of the earthmoving equipment the upper 1 to 2 m of ground was generally very dense and
PLM of more than 1,000 kPa was commonly observed, the limit pressure of the deeper layers
364
was usually less than 600 kPa. Similarly, while EM was generally 4 to 8 MPa at depths of 2 to
7 m, values of less than 1 MPa were also occasionally encountered. The layout plan of the
PMTs on the Peninsula, PLM and EM of the first three tests of the Peninsula are respectively
shown in Figure 3‐90 and Figure 3‐91. The distances between these three test locations were
about 200 to 250 m.
Figure 3‐90: PMT layout plan on the Peninsula
Figure 3‐91: Several PMT results at Marjan Island before dynamic compaction
365
As explained in Section 2.9.2.8, natural unconsolidated soil or young fills will undergo large
settlements with time, even if they are very lightly loaded. The self‐bearing condition of a
soil; i.e., the minimum value of the soil parameters that a soil must have so as not to settle
under its own weight, can be related to physical or mechanical properties of the soil.
Menard’s experience shows that the limit pressure is a suitable characteristic, and for sand
and sand with gravel the P*LM at depths less than 10 m must be at least respectively 600 kPa
and 800 kPa to prevent creep settlement (refer to Section 2.9.2.8). Menard proposes to
estimate the self‐weight settlement of soft or loose soils during a period of one year as a first
approximation using Equation 2‐211.
200 cm 1 200
For kPa 2‐211
1000
200
Assuming creep due to the soil’s self‐weight will reach stabilisation after t years according to
a logarithmically decreasing curve, self‐weight settlement, sn, in year n will be:
1
ln
3‐3
ln 1
Thus, it can be estimated that a one metre thick layer of sand with an initial limit pressure of
400 kPa will undergo 18 to 25 mm of creep settlement depending on whether the
stabilisation period is assumed to be 10 or up to 20 years.
The heterogeneity in soil strength and the presence of loose spots near dense spots indicated
that it was possible for the ground to undergo large differential settlements due to the self‐
weight of the soil without external loading. Structural and traffic loads further increased the
risks of excessive ground deformations.
The geotechnical concerns were considered to be of high priority for the main road corridor
that passed through the Peninsula, Island 1 and Island 2 (refer to Figure 3‐87). Hence, the
project management team approached geotechnical specialist contractors to propose
solutions for mitigating the geotechnical problems.
366
3.6.3 Development of Solution and Application of Dynamic Compaction
After reviewing the proposals the project’s management team and consultant awarded the
198,000 m2 ground improvement works of the main road corridor to a specialist ground
improvement contractor who had proposed the application of dynamic compaction. 123,000
m2 of the works was in the Peninsula and Island 1, and 75,000 m2 was located in Island 2.
Ground improvement design criteria were specified to be:
Maximum total settlement: 25 mm under a uniform load of 20 kPa on the road area
Maximum differential settlement: 1:500 between any two points on the road with a
distance of 10 m under a uniform load of 20 kPa.
Verification and acceptance of the works was defined to be by PMT.
As can be observed bearing capacity was not a requirement of this project as it was expected
that the ground loaded over a large area would cease to function by excessive settlement
rather than by bearing failure.
Dynamic compaction was carried out by 3 DC rigs using 17 and 20 ton pounders. Each print
location was subjected to 5 to 8 blows dropped from 20 m using a final grid size of 5x5 m2. In
the ironing phase the 17 and 20 ton pounders were dropped from 12 m or a 12 ton pounder
was dropped from 17 m.
The author is frequently asked about the effect of dynamic compaction and vibrations
generated by DC on non‐sensitive structures and facilities. It is important to note that while
application of dynamic compaction at close distances creates displacements and vibrations
that can crack buildings and damage structures, the level of tolerance for different structures
and facilities varies (refer to Section 2.8.1and 2.8.4). Figure 3‐92 shows the application of
dynamic compaction at the vicinity of the reclamation slope’s rock armour. As can be
observed the rock armour is able to remain intact even when works are performed at close
proximity.
Mobilisation of equipment, ground improvement works and testing by PMT were carried out
during a period of 5 months.
367
Figure 3‐92: Application of dynamic compaction at the vicinity of slope armour protection
3.6.4.1 Calibration
At the beginning of the works a calibration dynamic compaction programme was performed
in two zones to verify and to optimise the ground improvement parameters. The results of
the first zone, which included two patterns is discussed hereunder.
Figure 3‐93 and Figure 3‐94 respectively show the location and layout of the dynamic
compaction calibration programme in which a 20 ton pounder was used to compact the
prints of phases 1 and 2 from 20 m height, and dropped once from 12 m height on each print
of the ironing phase. The grids of the first and second phases of Pattern 1 and Pattern 2 were
respectively 7x7 m2 and 5x5 m2. In addition to the 4 HPTs, one PMT was performed before
dynamic compaction and 6 PMTs were carried out in phases 1 and 2, and in between the
prints of the two patterns (refer to Figure 3‐94). Figure 3‐95 and Figure 3‐96 respectively
show the measurement of the crater’s depth and upper diameter during the heave and
penetration test. Pounder diameter, crater diameter and amount of heave (or subsidence)
were not available to the author, and volumes reported herein are from project reports.
The results of the heave and penetration tests in the first and second phases of Patterns 1
and 2 are shown in Figure 3‐97 to Figure 3‐100. In each pattern, the first HPT was carried out
in the first print of the pattern’s first compaction phase. After completion of compaction of
the first phase prints, the second HPTs were performed in the first print of the second phase
368
prints of the two patterns. 11 blows were applied in the first HPT of Pattern 1, but all other
HPTs were subjected to 10 blows.
Figure 3‐93: Location of calibration area
Figure 3‐94: Dynamic compaction calibration layout
It can be observed that in both patterns the accumulative pounder penetration and
compaction volume of the first phase HPT is greater than the second phase HPT. This
indicates that, similar to the laboratory scale tests of Hajialilue‐Bonab and Zare (2014), which
were discussed in Section 2.5.4, the soil in the zone of influence of phase 2 was already
somewhat improved due to the compaction carried out during phase 1; hence, the amounts
of penetration, reduction in void ratio and volume reduction were less in the second DC
phase.
369
Figure 3‐95: Measurement of crater depth in the heave and penetration test
Figure 3‐96: Measurement of the crater’s upper diameter in the heave and penetration test
Figure 3‐101 shows the ratio of phase 2 to phase 1 accumulative penetrations and
compaction volumes. It can be seen that the ratios are fluctuating within a limited band that,
except for one data set, is less than one for both penetration and compaction volume ratios.
While it could be expected that the tighter grid should yield lower penetration and
compaction volume ratios (because the ground strength in the zone of influence of the
second phase has been improved more due to the compaction of the tighter first phase grid),
this cannot be concluded for the HPT results of this project. This may be explained when the
penetration of the first phases of the two patterns are compared. While both HPTs were
performed on virgin ground that were not influenced by any ground improvement activities,
and the measured penetrations should have logically been approximately the same figures
370
for both patterns, the penetrations of the first pattern are greater than the second pattern.
This suggests that the original ground conditions of the latter pattern could have been better,
which consequently, led to a larger decrease of phase 1 penetrations compared to phase 2
crater depths.
2
Figure 3‐97: Heave and penetration test in phase 1 print of 7x7 m grid pattern
2
Figure 3‐98: Heave and penetration test in phase 2 print of 7x7 m grid pattern
371
2
Figure 3‐99: Heave and penetration test in phase 1 print of 5x5 m grid pattern
2
Figure 3‐100: Heave and penetration test in phase 2 print of 5x5 m grid pattern
The ratios of the compaction volume to accumulative heave of the four HPTs are shown in
Figure 3‐102. It can be observed that the ratios increase with the number of blows. Noting
that the dimension of the ratios are length squared, the ratios of the four HPTs were fitted
to second degree polynomials, and it was observed that, within the range of calibration blow
counts, compaction volume to accumulative heave ratios of all four HPTs can be formulated
by second degree polynomials with very high coefficients of determination.
372
Figure 3‐101: Ratio of Phase 2 to Phase 1 penetration and volume reduction in the two patterns
Figure 3‐102: Ratio of compaction volume to accumulative penetration in HPTs
Replotting the same data of Figure 3‐102 with the ratio of the square root of the compaction
volume to accumulative penetration yields Figure 3‐103, which shows that after the first
several (3) blows the ratio is generally between 2.15 to 2.4. Thus:
√
2.15 2.4 3‐4
373
V= compaction volume (m3)
Dc= crater depth (m)
Figure 3‐103: Ratio of square root of compaction volume to accumulative penetration in HPTs
The range shown in Equation 2‐40 should be dealt with caution at this phase as compaction
volume is a function of the pounder base’s area (bottom of crater), and it is not known if the
basis of calculations in this project was the customary assumption that crater base diameter
was equal to the pounder’s diagonal length or the equivalent pounder diameter (which in
the author’s opinion is the more reliable assumption).
Based on the result of the HPT 6 blows were applied from the height of 20 m using a 20 t
pounder in both the first and second DC phases of pattern 1 with 7x7 m2 grid. For the latter
pattern with 5x5 m2 grid the blows in the second phase was reduced to 5. Ironing was
performed by dropping the same pounder from 12 m height.
PLM and EM before and after dynamic compaction inside two prints of compaction phases 1,
in between the prints, and their average values for Pattern 1 are shown in Figure 3‐104. It
can be observed that the variation of the results inside phase 1 prints and in between the
prints are practically insignificant, and in fact the difference between the two inside print PLM
values is greater than the difference between them and the in between print test. The
average test values that are also shown in Figure 3‐104 are suitably representative of the
compacted ground condition.
374
Figure 3‐105 shows PLM and EM improvement ratios for inside phase 1 prints, in between the
prints, and their averages. While the peak PLM improvement ratio of all tests is approximately
6, the peak average PLM improvement ratio was about 5, which is compliant with 400%
improvement that has been proposed by Lukas (1986). Peak average EM improvement ratio
appears to be slightly more and approximately 6, which is still in the magnitude that has been
noted by Lukas. Peak improvement ratios were observed at depth of about 2 m to 3 m, and
as expected, the amount of improvement then decreased with depth.
Ground settlement after the final levelling in the area with 5x5 m2 grid pattern was measured
to be 0.29 m. Although final settlement is the accumulative effect of phases 1 and 2 of deep
compaction and ironing, using Equation 2‐40, the crater depth can be estimated to have been
1.12 m to 1.25 m, which reasonably agrees with Figure 3‐99 and Figure 3‐100.
2
Figure 3‐104: PLM and EM values before and after dynamic compaction in pattern with 7x7 m grid
In the second pattern with 5x5 m2 grid, one PMT was performed in each of the deep phases’
prints, and one PMT was carried out in between the prints. PLM and EM values of these tests
and the average test values are shown in Figure 3‐106. It can be observed that in this pattern
the improvement of the ground in phase 1, phase 2 and in between the prints is reasonably
close and the average test values are representative of the ground’s mass improvement.
375
2
Figure 3‐105: PLM and EM improvement ratios in pattern with 7x7 m grid
Figure 3‐107 shows PLM and EM improvement ratios for inside phase 1 and phase 2 prints, in
between the prints, and their averages. The peak PLM improvement ratio of all tests is about
5.2, and the peak average PLM improvement ratio is about 4.2, which is somewhat less than
an improvement of 400% that has been proposed by Lukas (1986).
Figure 3‐108 and Figure 3‐109 respectively show the ratio of Phase 1 (average of two prints
for Pattern 1) to Phase 2 limit pressure values for Patterns 1 and 2 and the same ratio and
the ratio of Phase 1 to in between the prints PLM for Pattern 2. It can be observed that, except
of the variations on the most upper testing level, Phase 1 to Phase 2 ratios for both patterns
are very similar, and it appears that in general the ground strength is up to 30% more in Phase
1 prints. Interestingly, Figure 3‐109 suggests that at some depths the strength of ground is
more in between the prints that in Phase 1 prints.
376
2
Figure 3‐106: PLM and EM values before and after dynamic compaction in pattern with 5x5 m grid
Figure 3‐107: PLM and EM improvement ratios in pattern with 5x5 m2 grid
377
Figure 3‐108: Ratio of Phase 1 to Phase 2 PLM values for Patterns 1 and 2
Figure 3‐109: Ratio of Phase 1 to Phase 2 and in between prints PLM values for Pattern 2
378
3.6.4.2 Final Testing
16 PMTs were carried out before dynamic compaction. A further 32 PMTs were also carried
out after ground improvement to verify the ground conditions and to confirm that
acceptance had been achieved. Of these, 14 tests were on the Peninsula, 5 were on Island 1
and 13 were on Island 2.
For comparative purposes EM values of two test locations that were substantially different
are shown in Figure 3‐110. In these tests EM values were assumed to have a maximum value
of 25 MPa due to equipment gauge limitations. Although in reality the distance between
these two tests was much more than 10 m, it is of interest to study the total settlements.
Settlements can be calculated using Equation 2‐198:
2
2‐198
9 9
Figure 3‐110: Two post dynamic compaction EM logs
379
Assuming a loading area of 100 m by 10.5 m subject to a uniform load of 20 kPa, and
conservatively assuming EM= 20 MPa for all layers below testing levels, it can be seen that
the total settlement of the testing point with lower moduli will result in only 2.2 mm of
settlement. This figure itself is much less than the acceptable differential settlement between
two points, and demonstrates that acceptance has been achieved.
Figure 3‐111 shows average EM before and after dynamic compaction. It can be seen that
while it is generally expected for the improved test profile to be sickle shaped, in this case
because average EM before dynamic compaction was least where the improvement was
most, the improvement profile appears to be more linear and only slightly curved.
Figure 3‐111 also shows the average EM improvement ratio, and it can be observed that the
ratio appears to be in the shape of a sickle with maximum improvement at about half the
depth of improvement. The maximum average EM improvement ratio was 5.31, which is quite
compatible with the indicative upper bound figure of 400% that has been proposed by Lukas
(1986) as a guideline.
Figure 3‐111: Average EM before and after dynamic compaction and average EM improvement ratio
380
3.6.5 Lessons and Conclusions
Marjan Island is a reclaimed site that has been realised by dumping sand into the sea by
trucks. The geotechnical investigation showed that the soil was of variable strength and
generally very loose down to the depth of about 7 m, and preliminary analysis indicated that
the ground was subject to creep settlement due to self‐weight and differential settlements.
Dynamic compaction was implemented to improve the main road of the Peninsula, Island 1
and Island 2.
Post ground improvement PMTs were carried out, and were able to demonstrate that ground
settlements due to the project’s uniformly distributed design load was negligible and much
less than the acceptance criteria.
In summary, it was observed that:
1. Ground reclaimed by dumping sand into the sea was in a loose state.
2. Should the grid pattern be close enough, compaction of first phase prints will also
somewhat densify the ground influenced by the second phase prints. Consequently,
it may be possible to apply lesser amounts of energy in the second phase of
compaction.
3. In this project, the PMT parameters’ improvement ratios were in the order suggested
by Lukas (1986).
4. In this project (reclamation by dumping sand), the square root of the compacted
volume (in m3) of ground is about 2.15 to 2.4 times the crater depth (in m).
Figure 3‐112 shows Marjan Island Road Corridor after completion of dynamic compaction
works.
381
Figure 3‐112: Marjan Island road after completion of dynamic compaction works
382
3.7 Al Nakhilat Ship Repair Yard
Ras Laffan, located on the southern coast of the Persian Gulf and approximately 70 km north
of Qatar’s capital city, Doha, houses the onshore facilities of the world’s largest gas field.
Nakhilat Ship Repair Yard that is shown in Figure 3‐113 is part of Port of Ras Laffan’s
expansion programme, and has been constructed on land that was hydraulically reclaimed.
Figure 3‐113: Plan of Nakilat Ship Repair Yard
Seabed level at the location of the project was variable from ‐9.1 m to ‐13.2 m CD (chart
datum). Design (final platform) level was set at +3.5 m CD; however, it was envisaged that
the hydraulic fill would be placed in a loose state, ground improvement would be required,
and the platform level would consequently reduce. Ground subsidence due to treatment was
estimated to be in the range of 0.6 to 0.8 m; hence, the working platform was reclaimed
approximately to +4.1 to +4.3 m CD.
Reclamation was carried out using carbonate sand and gravel that were dredged from the
sea for deepening the port. The fill’s grain size was generally less than 75 mm, but stones as
large as 500 mm in diameter were also present. The maximum fines content of the fill was
mostly less than 10% in the upper elevations, but there were occasional lenses of silt at depth
with thicknesses varying from 0.2 to 0.4 m. Carbonate content of the reclamation material
was approximately 90% as CaCO3.
383
CPTs were carried out as part of the geotechnical investigation after reclamation. In areas
DDR4 (57,064 m2), DDR5 (35,643 m2) and DDR6 (82,962 m2) of the dry dockyards (see
Figure 3‐113) the soil in the upper 3 m to 5 m was medium to very dense with qc ranging from
as low as 5 MPa to more than 20 MPa. The soil then became loose to medium dense with qc
fluctuating between 1 MPa to 7 MPa. Dense seabed was encountered at depths of 13 m to
17 m, and CPT friction ratio was understood to be generally well below 1%. Three typical CPT
qc logs are shown in Figure 3‐114.
Figure 3‐114: Comparison of CPT cone resistance before ground improvement with target relative
density
While it was understood that less sensitive areas of the project would require lesser ground
treatment, areas DDR4, DDR5 and DDR 6 with a total area of more than 175,000 m2 were
deemed to be sensitive, and project specifications stipulated that minimum relative density,
Dd, in these areas had to be 60% based on the correlation of Baldi et al. (1986), which has
been derived for normally consolidated Ticino sand (see Equation 2‐265).
384
1
ln . 2‐265
2.41 157 ′
As discussed in Section 2.10.2.3.2.2, the penetration resistance of calcareous sands is lower
than quartz sands with similar grain size distributions. Thus, in this project a correction factor
of 1.94 was defined in the specifications for application to the cone resistances of the
calcareous soil. For the purpose of calculations it was specified that the saturated and
unsaturated densities of the soil were to be assumed to be respectively 18.7 kN/m3 and 15.2
kN/m3. Average groundwater level was assumed to be at +0.5 m CD.
By also plotting Baldi’s 60% relative density curve (red line) in the CPT logs of Figure 3‐114 it
can be seen that the hydraulically reclaimed fill was not able to satisfy the project’s
acceptance criterion, and it was confirmed that ground improvement would be required.
Figure 3‐114 also shows the relative density design curve, which takes the soil’s better
characteristics in the upper layers into account, and assumes a number of line segments in
lieu of the actual relative density correlation curve.
The ground improvement works of the project was awarded to a specialist geotechnical
contractor who had proposed the application of dynamic compaction. With consideration of
the numerous draw backs of relative density as a ground improvement criterion (refer to
Section 2.10.2), and with the knowledge that foundation requirements could be satisfied
more affordably by using alternative criteria that were directly based on design
requirements, (refer to Section 2.10.1.12.10.1), the specialist contractor proposed
alternative criteria based on a worst case isolated footing scenario:
Footing load: 4000 kN
Allowable bearing capacity: 200 kPa
Maximum settlement: 50 mm
The project’s schedule stipulated that mobilisation, ground improvement and testing to be
completed according to the below milestones:
DDR4: 154 days after issuance of notice to proceed
DDR5: 63 days after issuance of notice to proceed
385
DDR6: 91 days after issuance of notice to proceed
Based on the fill thickness and the phase of dynamic compaction, soil improvement was
carried out in areas DDR4, DDR5 and DDR6 using a combination of 15, 25, 28 and 35 ton
pounders. As shown in Figure 3‐115, the 35 ton pounder was dropped in free fall from 25 m
using MARS that had previously been used successfully in Al Quo’a (refer to Section 3.3.4).
Figure 3‐115: Free fall drop of a 35 ton pounder using MARS technology
Prior to the commencement of dynamic compaction production a number of heave and
penetration tests were carried out, of which one that was carried out in DDR6 has been
selected for review, analyses and discussion. In this HPT a 25 ton pounder was dropped 20
times from 20 m. The pounder was square with each side being 2 m. The assumption that
the pounder’s diagonal length of 2.82 m should be taken as the crater’s bottom diameter
does not seem justified in this project because firstly the pounder was lifted using two
parallel single cable lines; hence, the pounder could not have rotated as it was lifted and
dropped onto the print with different orientations, and secondly the crater’s upper diameter
in the first few blows was less than the pounder’s diagonal length. Therefore, for calculation
purposes, it is assumed that the diameter of the crater’s base is the equivalent pounder
diameter of 2.25 m.
386
Crater depth, which is equal to the accumulative pounder penetrations, crater, heave and
the net compaction volumes are shown in Figure 3‐116. As also observed in the HPT of Al
Falah Community (Section 3.5.4.1) and Marjan Island Road Corridor calibrations
(Section 3.6.4.1) the rate of crater depth increase is initially higher, but then reduces.
Except for the first several blows where heave volumes are negligible, additional blows
generated negative heave volumes; i.e., additional settlement and compaction around the
print. The rate of negative heave volume appears to be constant and not reducing noticeably
with the increase of the number of blows. However, the net compaction volume rate seems
to be declining. Thus, it can be envisaged that while the soil directly within the pounder print
is reaching an asymptote value, the soil in the periphery is still able to undergo further
compaction. Comparison of crater, heave (subsidence in this case) and net volume changes
suggest that approximately 80% of the compaction volume originates from the crater and
20% can be contributed to the peripheral subsidence. Of course, this should not be
necessarily interpreted as the crater volume being the result of the sole compaction of the
soil directly beneath the crater.
Figure 3‐116: Heave and penetration test results
Figure 3‐117 shows the relationship between the number of blows and the ratio of the square
root of the compaction volumes to the crater depth. It appears that after the first several
blows the ratio becomes independent of the number of blows. The ratio is approximately
2.32 when the net compaction volume, being the resultant of the crater and periphery
volume changes, is considered, and is about 2.11 when only the crater volume is considered.
387
This suggests that the ratio of the square root of the crater volume to penetration is
approximately 90% of square root of the net volume to penetration.
Figure 3‐117: Ratio of square root of compaction volumes to penetration
A similar conclusion can be made when the compaction volume of the crater and each of the
concentric rings around the crater are compared with the total volume of compaction for 6,
10 and 20 blows in Figure 3‐118. It can be seen that the ratio of crater compaction volume to
total compaction volume reduces with the number of blows from approximately 90% for 6
blows to about 80% for 20 blows. Compaction volume rapidly reduces with radius whereas
98% of the compaction volume occurs within a radius of 5 m for 6 blows, and within a radius
of 10 m for 20 blows. Also, it can be noted that approximately 8% of the compaction has been
realised within the concentric circle with 3 m radius. Noting that final grid sizes are usually
about 4 to 6 m, an educated guess would suggest that compaction of second phase prints
could increase the compaction volume of the soil within the zone of the first phase prints by
about another 3 to 5%.
Figure 3‐119 shows the ratio of the difference between the crater top diameter, DT, and base
diameter, DB, to the crater depth. It can be observed that for the first 7 blows the ratio
increases almost linearly with a gradient of approximately 14%; however, it then almost
suddenly reaches a constant value of about 1.05. This suggests that the crater sides are
initially steeper, possibly and partially due to the pounder side retaining the soil, but reach
equilibrium at an angle of about 28o from the vertical after several blows when the crater
becomes deep (about 1.5 m at 7 blows).
388
Figure 3‐118: Percentage of compaction share of the crater and the peripheral concentric rings
Figure 3‐119: Variation of (DT‐DB)/Dc and crater side angle with number of blows
A penetration test (only crater measurements without recording peripheral heave) was also
carried out in DDR‐5 using a 35 ton pounder that was dropped from 25 m in free fall using
MARS. The pounder was 2.4x2.4 m2, and had an equivalent diameter of 2.7 m.
Figure 3‐120 shows the ratio of the square root of the crater compaction volume to crater
depth. Once again, it can be observed that after the first several blows the ratio levels off at
approximately 2.36. Noting from Figure 3‐117 that about 90% of the compaction volume
389
originates from the crater itself, it could be envisaged that the ratio for the total compaction
volume would have been in about 2.6, which would have been higher than what was
observed in Al Falah Community, Marjan Island, and for the 25 ton pounder calibration in this
project. Thus, Equation 2‐40 can be modified to become:
Figure 3‐120: Ratio of square root of crater compaction volume to penetration
√
2.15 2.6 3‐5
Similar to Figure 3‐119 and as shown in Figure 3‐121, changes in (DT‐DB)/Dc and crater side
angle values are higher for the first several blows, but then the rate reduces with more blows.
However, in this calibration there does not seem to be an asymptote value, and it appears
that (DT‐DB)/Dc and crater side angle diagrams follow logarithmic curves.
Changes of (DT‐DB)/Dc become more meaningful when they are studied against the changes
of impact energy. Figure 3‐122 shows the changes of (DT‐DB)/Dc against impact energy for
two calibrations that were presented in this Section, and suggests that the variations of
(DT‐DB)/Dc are almost linear against the logarithm of the impact energy. The comparison of
only two curves is not sufficient to conclude anything concrete and definite, but one of the
below scenarios is likely:
390
Figure 3‐121: Variation of (DT‐DB)/Dc and crater side angle with number of blows
1. These two curves should be the same, there is only a single curve that explains the
relationship between (DT‐DB)/Dc and impact energy, and the two curves in
Figure 3‐122 have not fallen on top of each other due to testing accuracy. This
scenario appears reasonable as it is possible that crater proportions are independent
of the energy of a single impact.
Figure 3‐122: Variation of (DT‐DB)/Dc with impact energy
391
2. These two curves that are for two different impact energies per blow are different,
and are part of a family of parallel curves that define the relationships between (DT‐
DB)/Dc and impact energy. This scenario appears to be equally reasonable as it is
possible that crater proportions are dependent on the energy of each drop.
3.7.5.2 CPT
In order to verify that the project requirements had been satisfied one CPT was carried out
per every 600 m2 of improved ground. A number of PMT were also carried out for
comparative purposes.
Two commonly encountered post ground improvement soil profiles (using a 28 ton pounder)
in area DDR4 and four commonly encountered post ground treatment soil profiles in area
DDR6 are respectively shown in Figure 3‐123 and Figure 3‐124. The soil compositions of these
profiles are summarised in Table 3‐10. It can be well seen that while the fines content of the
upper layers in all areas and the entire fill thickness in some areas is composed of clean sand
as per the project specifications, the post treatment tests were able to identify substantially
siltier soil (fines content greater than 10%) at depth in some areas. Fines content in some
layers were considerably higher than expected, and even reached 100%. Similarly, the CPT
friction ratio in some locations was measured to be up to 7%.
For design purposes, a worst case scenario of the qc curve was chosen from DDR4 Profile 1
(the black line segments in Figure 3‐123a), and simplified into a number of segments. The
same line has been copied onto Figure 3‐123b and Figure 3‐124 for comparison purposes. It
can be observed that the actual ground improvement is considerably better in the other
profiles.
Irrespective of all the serious drawbacks and ambiguities that the application of relative
density can have on a project (refer to Section 2.10.2), it is still possible to draw the relative
density curve (knowing that it does not apply when fines content exceeds 15%) and to
compare it with the worst case scenario design curve. Figure 3‐123c compares the qc design
curve, developed in Figure 3‐123a, with the target relative density curve and relative density
curve, developed in Figure 3‐114. It can be seen that while the worst case scenario design
curve has higher values in the upper soil layers, the relative density curve has greater values
at depth.
392
Soil Profile Ground Levels Layer information Friction Fines
(m CD) Description bottom level Ratio (%) Content
(m CD) (%)
DDR4‐ Profile 1 +3.5 to ‐13.5 clean sand ‐13.5 0.2 0 to 5
+4
DDR4‐ Profile 2 +3 ‐13 clean sand ‐4.5 to ‐7 0.2 0 to 5
silty sand ‐7 to ‐9.5 1 to 3 15 to 50
clean sand ‐13 0.2 0 to 5
DDR6‐ Profile 1 +2.5 ‐11 clean sand ‐6 to ‐9 0.2 0 to 5
silty sand ‐11 0.2 to 0.5 5 to 10
DDR6‐ Profile 2 +3 ‐10.5 clean sand ‐8.5 to ‐9.5 0.2 0 to 5
silty sand 10.4 0.2 to 0.5 5 to 10
silt (0.5 m 4 to 5 50 to 70
thick)
inter‐bedded
in silty sand
layer
DDR6‐ Profile 3 +2.5 to ‐10.5 clean sand ‐5 to ‐6.5 0.2 0 to 5
+3 silty sand ‐10.5 0.5 to 1 20 to 30
silty sand/ 1.2 to 5 30 to 70
silt
three bands,
each 0.2 to
0.4 m thick
DDR6‐ Profile 4 +1.5 to ‐10.5 clean sand ‐3.5 to ‐8 0.2 0 to 5
+3.2 silty sand ‐10.5 0.5 to 1 20 to 30
silty sand/ 4 to 7 70 to
silt 100
four bands,
each 0.2 to
0.6 m thick
Table 3‐10: Soil profiles in areas DDR4 and DDR6
393
(a) (b) (c)
Figure 3‐123: Post ground improvement soil profiles in DDR4
(a) (b) (c) (d)
Figure 3‐124: Post ground improvement profiles in DDR6
394
Ultimate bearing capacity, qu, for a square footing on a cohesionless soil can be estimated
from Bowles (1996):
. 3‐6
4800 0.9 300 kPa
Jullienne (2008) calculated the allowable bearing capacity for each of the qc curves with a
factor of safety of 3. The allowable bearing capacity for the relative density design curve and
worse scenario design curve are respectively 400 and 414 kPa, which are both much more
than 200 kPa.
Likewise, using the method proposed by Schmertmann (1970) and Schmertmann et al.
(1978), Jullienne also calculated the settlements for the relative density design curve and
each of the soil profiles that have been shown in Figure 3‐123 and Figure 3‐124. In his
calculation Jullienne assumed that the square footing is subject to 4,000 kN and subject to
200 kPa of pressure (a 4.5 x 4.5 m2 footing). The result of the calculations is summarised in
Table 3‐11. It can be seen that it is in fact the relative density design curve that will result in
the maximum amount of settlement.
Profile settlement (mm)
DDR4‐ Profile 1 35.52
DDR4‐ Profile 2 30.67
DDR4‐ Dd design curve 50.16
DDR6‐ Profile 1 28.33
DDR6‐ Profile 2 33.66
DDR6‐ Profile 3 33.66
DDR6‐ Profile 4 35.06
DDR6‐ Dd design curve 50.16
Table 3‐11: Settlements of the Dd design curve and the soil profiles subject to a load of 4,000 kN
The above calculations show that qc values after dynamic compaction at depth may have
been less than the relative density curve; however, all soil profiles have performed better
than the relative density specification, and regardless of the drawbacks and unreliability of
relative density as a ground improvement criterion, it can be seen that application of
specifications that unnecessarily limit the acceptance zone by defining a hypothetical
minimum passing value can result in failure to meet acceptance without failure to meet
395
design requirements. Indeed, as discussed in Section 2.10.1.3, acceptance criteria that are
based on design criteria without proxy will yield the most efficient result and consequently
the shortest construction programmes with the least construction costs.
After execution of dynamic compaction in DDR5 using a maximum pounder weight of 28 tons
(without ironing) it was decided to perform a dynamic compaction trial to study the
improvement effects using a 35 ton pounder that was dropped by MARS. This process
included 3 deep compaction phases and an ironing phase.
As shown in Figure 3‐125, 3 PMTs were carried out next to 3 CPTs in the below order:
Before phase 1: PMT‐007 and CPT‐551 (in between impact points)
After phase 1: PMT‐009 and CPT‐576 (in between impact points)
After phase 3: PMT‐010 and CPT‐595 (in impact point)
Figure 3‐125: CPT and PMT layout
CPT qc values are shown in Figure 3‐126. Likewise, PMT PLM and EM are respectively shown in
Figure 3‐127 and Figure 3‐128.
396
Figure 3‐126: CPT cone qc used in the correlation
Figure 3‐127: PMT PLM values used in the correlation
397
Figure 3‐128: PMT EM values used in the correlation
The ratios of qc/PLM are shown in Figure 3‐129. It can be observed that the average qc/PLM
values for 21 tests points, which exclude the uppermost test points of PMT007 and PMT010
due to the differences in between the shallow and deep failure modes, are equal to 4.54 (the
average qc/PLM value is 5.20 when the top two ratios of PMT007 and PMT010 are also
included). The average qc/PLM value for the three correlations on Ras Laffan carbonate sand
are 4.1, 5 and 5.3 excluding the mentioned uppermost points. Minimum and maximum
qc/PLM values were respectively 2.9 and 9.1 for the 21 points. Briaud et al. (1985), see
Table 2‐27, have proposed the correlation to be PLM =0.11qc, which is equivalent to qc/PLM=
9. This is almost twice the average qc/PLM value that was derived in this project1.
As shown in Figure 3‐130, q*c/P*LM plots are identical in shape and very close in value to the
qc/PLM ratios, and indicate that implementation of qc/PLM ratios has yielded the same results
as q*c/P*LM in this saturated sand. The average q*c/P*LM value for the 21 tests points is equal
to 4.82, which is just below the range of 5 to 12 that has been proposed by Baguelin et al.
(1978) that was presented in Table 2‐26. The average q*c/P*LM values for the three
1
The author has previously published a paper (Hamidi, B., Varaksin, S. & Nikraz, H. (2010) Correlations
between CPT and PMT at a Dynamic Compaction Project. 2nd International Symposium on Cone
Penetration Testing (CPT10), Huntington Beach, California, 9‐11 May, paper 2‐04.) The interpretation
of that paper is erroneous as there has been a mistake in the comparison of results. The interpretation
in the thesis is correct.
398
correlations are 4.3, 5.4 and 5.2 excluding the mentioned uppermost points. Minimum and
maximum q*c/P*LM values for the 21 points were respectively 3 and 9.3.
Figure 3‐129: qc/PLM for Ras Laffan carbonate sand
Figure 3‐130: qc*/P*LM for Ras Laffan carbonate sand
As shown in Figure 3‐131, EM/qc values of the two uppermost shallow points do not seem to
correlate differently with the deeper points as the result of differences between the shallow
399
and deep failure modes. The average EM/qc value for the 23 test points is 1.35. The average
EM/qc values for the three tested locations on Ras Laffan carbonate sand were 1.5, 1.3 and
1.1. Minimum and maximum EM/qc values for the test points were respectively 0.3 and 1.91.
Briaud et al. (1985) have proposed a correlation of EM=1.15qc (see Table 2‐27), which is
almost 85% of what has been measured in this project.
Figure 3‐131: EM/qc for Ras Laffan carbonate sand
With consideration of the two uppermost points, it can be concluded that for Ras Laffan
saturated carbonate sand, after dynamic compaction:
∗
∗ 5 3‐7
1.35 3‐8
Sieve analyses and hydrometer tests carried out on samples extracted from boreholes that
were drilled as part of the geotechnical investigation during ground improvement indicated
the presence of bands of fine soils in the area of the skid beams in DDR6. Fines content in
these layers were considerably higher than the specified 10% limit and up to 97%. As shown
in Figure 3‐132, clay content of some layers was up to 38%.
400
Hole No: BH‐DDR6‐08
Sample No: 12
Depth (m): 10.90 – 11.35
Figure 3‐132: Fines content of reclaimed soil with very high fines content
The CPT profiles of a test that was carried out before and a test that was performed after
dynamic compaction in the same area is shown in Figure 3‐133. The CPT friction ratio also
confirmed the presence of the same fine layers.
A PMT that was also performed in the area with fine soil layers is shown in Figure 3‐134.
While this test does not identify fine soil layers, a noticeable reduction of PMT parameters is
seen at the depth of the fine soil layers.
The presence of a potentially soft clayey soil at depth raised concerns about the long term
settlement of the ground under external loads. Therefore, a full scale zone load test was
performed at the location of tests shown in Figure 3‐132 to Figure 3‐134 that appeared to
have the most amount of silty or clayey layers with a thickness of 4 m from approximately ‐5
m to ‐9 m CD, and which has been marked with a circle in Figure 3‐135.
As shown in Figure 3‐136 (relative to the DC grid and test locations) and Figure 3‐137(a)
(relative to one another, the loading units and settlement measuring locations), similar
loading to the skid beams was carried out using 1.5 m3 concrete blocks (3.75 tons) placed in
two parallel strips, each composed of 3 rows and up to 5 columns. For the first layer, 18
blocks, spaced at 0.2 m distances from edge to edge, were placed over a length of 25.7 m.
The second to fifth layers each had from 17 to 14 blocks. Maximum height of the blocks (from
ground level to the top of the fifth layer) was 5.7 m. The centres of the strips were 6 m apart.
401
It can be assumed that the boundary influence will be eliminated between block rows
number 5 to 13. Hence, on an area of 53.6m2 (14.3x3.75 m), the load intensity was assumed
to be 105 kPa.
Figure 3‐133: results of CPT‐615 (before DC) and CPT‐828 (after DC)
Figure 3‐134: Results of PMT‐025
402
Figure 3‐135: Location of zone load test in area with the most amount of fine soil
Zone loading strips
Figure 3‐136: Location of zone load test and field tests shown in Figure
After levelling the areas to +3.5 m CD, four settlement monitoring beacons (the blue
monitoring points in Figure 3‐137) were installed as in‐survey points. These were measured
immediately before and right after the test loading. After the first layer of blocks was placed
and all monitoring points (red points at the edge of the blocks) were established and
measured, the remaining four block layers were also placed on the strips. The placement was
done layer by layer. Upon completion of loading, the red monitoring points shown in
Figure 3‐137 were immediately measured. The blocks were removed after the testing period.
403
Figure 3‐137: Skid beam zone load test details (a) plan view, (b) side view, (c) front view
The non‐processed (red) monitoring point settlements for a duration of 41 days are shown
in a semi logarithmic chart in Figure 3‐138. It can be observed that while the settlements
have undergone some oscillations due to the accuracy of the surveying, maximum ground
settlement at the end of the monitoring ranged from 9 mm to 16 mm. It can be extrapolated
404
that the average settlement of the ground 10 years after monitoring will remain less than 18
mm, which is still significantly less than the design criterion of 50 mm.
A similar surcharging was repeated in a second area. Results of this monitoring were quite
compatible with the first area, with final settlements after 40 days being in the range of 6
mm to 20 mm. The maximum settlement probably occurred due to a movement of the blocks
as the first day settlement of this point itself was 14 mm, which was more than double the
of the next largest measured settlement. Further settlement rates then followed the same
pattern as all other points.
Figure 3‐138: Settlement monitoring of settlement monitoring points
The review of this project can provide the geotechnical engineer with a number of lessons to
be incorporated in future projects, including:
1. It can be envisaged before reclamation that young hydraulic fills will most likely be
placed in a loose state and potentially subject to low bearing, excessive settlements
and creep.
2. Project specifications should be defined appropriately.
a. Relative density is an unsuitable quality control criterion (also refer to
Section 2.10.2), and this parameter should not be used as an acceptance
criterion.
405
b. It is possible to reach better results without meeting relative density
specifications. This means that while it is likely that ground improvement
works can satisfy a project’s technical and design requirements, it may
appear as non‐conforming with relative density specifications due to the
poor choice of defining acceptance criteria.
c. Further to the in‐situ tests that showed a reduction of strength and
deformation modulus at depth, the monitoring of settlements under a full
scale zone load test showed that it is not necessary to have the same amount
of improvement in all soil layers. Ground behaves as a mass with each layer
contributing to the settlement according to its physical and mechanical
properties. It is the summation of the settlement increments that must be
less than design and acceptance settlements, not the settlement of each
individual layer.
3. This project has demonstrated the ability of dynamic compaction to treat loose
hydraulic fills as deep as 16 m. The technique has been successful even when fines
content was more than what was originally expected.
4. It was observed that for Ras Laffan saturated carbonate sand the relationship
between the CPT cone resistance and PMT parameters are:
a. qc / PLM= q*c /P*LM= 5
b. EM/qc= 1.35
The above correlations did not appear to be influenced by depth
5. In this project (hydraulic fill), the ratio of the square root of the compacted volume
of ground (in m3) to the crater depth (in m) is about 2.11 (when only the crater
volume is considered) to 2.32 (when the net compaction volume, being the resultant
of the crater and periphery volume changes is considered). This suggests that
approximately 90% of the compaction volume has originated from the crater and
10% from the ground around the crater.
6. There appears to be a linear relationship between the logarithm of impact energy
and (DT‐DB)/Dc. More research is required to determine if the relationship is unique
for a given soil type or whether it is also a function of the drop energy.
406
3.8 Abu Dhabi Ritz‐Carlton Hotel
Ritz‐Carlton Hotel is a luxurious resort constructed on the southern coastlands of Abu Dhabi,
UAE. It is constructed on a plot of land with an area of 210,000 m2 that includes the previous
site of the now demolished Gulf Hotel and the virgin site, which was situated on its southern
border. An 11 storey hotel with an underground parking is located in the middle of the resort.
This structure is encompassed on the north and south by two man‐made hills gradually
elevating to about 8 m above street level. 100 single storey chalets were designed to be
spread out throughout the hills and the original ground. The hills are retained on the outer
perimeter with mechanically stabilised earth (MSE) walls and slope inwards towards a large
pond (originally envisaged to be a man‐made lagoon) that is excavated in front of the hotel.
The project was originally commenced as J W Marriott Hotel, but was later rebranded as the
Ritz‐Carlton. Figure 3‐139 shows the project’s layout.
Figure 3‐139: Layout of Abu Dhabi Ritz‐Carlton Hotel
Noting that the site was located in a part of the city with a known history of soft soils and
with the recognition that the two floor buildings of the original Gulf Hotel were built on piles,
it was expected that the project would also be situated on problematic soils. This speculation
was confirmed by the preliminary soil investigation.
407
The original ground level in the undeveloped southern part of the site was mostly from +1.6
m to +2.1 m RL (reduced level= Abu Dhabi Datum); however, the level in the eastern portion
of this area was even lower and at +0.9 m RL. The original ground level in the location of the
previous Gulf Hotel was about +2 m RL. Groundwater depth was recorded to be from 0.5 m
to 1.8 m below borehole levels, and groundwater level was from +0.4 to +1.30 m RL.
During the preliminary geotechnical investigation 23 SPT boreholes were drilled down to
depths of 15 m and 30 m. The testing layout is shown in Figure 3‐140. It was observed that
the upper 1 to 2 m of soil was loose silty sand with fines content ranging from 5 to 25% and
generally less than 15%. The upper sandy layer was underlain by very soft silty to clayey sand
to sandy silt that was 3 to 4 m thick, and had fines content ranging from 35 to 70%. The
bottom of that layer was generally at the depth of 4 to 5 m and exceptionally as deep as 6 m.
SPT blow counts were consistently low, often as low as 1. The very soft layer was followed by
a medium to very dense silty to very silty sand layer with SPT blow counts commencing from
at least 18 and rapidly increasing to 30 and even more than 50. The bottom of this layer was
at the depth of 12 m. Mudstone with occasional layers of crystalline gypsum or sandstone
was encountered in all boreholes. Figure 3‐141 shows the profiles of four SPT boreholes in a
section through the site.
Figure 3‐140: Layout of SPT tests
408
Figure 3‐141: The profiles of four SPT boreholes in a section through the site
PMT carried out during later phases of the project as part of a supplementary geotechnical
investigation also indicated the presence of a very soft layer with PLM as low as 110 kPa. The
ground profile is summarised in Table 1.
Further to the geotechnical investigation, it became even more evident that the ground
conditions would be challenging for construction when the equipment mobilised for
demolition of Gulf Hotel began sinking in areas where the top of the soft soil was at shallow
depth. Figure 3‐142 shows an excavator that sunk into the clay during the works, and was
ultimately pulled out with a crane.
409
Figure 3‐142: Construction equipment sinking into the ground due to poor ground conditions
The construction of the original Gulf Hotel on piles, the local failures of the ground under the
load of the construction equipment, the poor results of the soil investigation, and the
preliminary calculations all indicated that the site was situated on soft compressible soil and
unable to support the loads that the new project.
The project team decided that very heavy structures such as the 11 storey hotel and the
underground car parks (that were in fact buried under the hills) were to be built on piles;
however, piling did not seem to be a solution of interest for an area of approximately 90,000
m2 that covered the hills, MSE walls, chalets and roads as it would have taken too much time
to execute, and would have been very costly. Thus, ground improvement was concieved as
an economical solution for treating the ground under the mentioned above.
As a first step it was required to develop appropriate design criteria based on the special
requirements of the embankments, walls, and chalets. The stipulated design criteria were:
1. Allowable Bearing capacity:
a. Chalets: 100 kPa under raft footings located 1 m below finished floor level
(variable) of the structure.
b. Man‐made landscape hills: To support fill compacted to 95% proctor dry
density (assumed to have a unit density equal to 18 kN/m2) for a slope less
than 2V:3H.
c. Retaining walls: To support 1.3 times the fill weight within a strip behind the
wall with a width equal to 70% of the wall height.
2. Total settlement
410
a. Chalets: 25 mm under actual footing loads (on average 20 kPa).
3. Differential settlement
a. Chalets: 1/500 between adjacent columns under footing loads
b. Added fill: 1/500 after 90% consolidation for two points located 10 m apart
under fill load
4. Consolidation ratio: 90% after 90 days of constructing the full fill height.
As can be observed that although the design criteria could have been further optimised, they
have still been able to concisely address each of the designer’s concerns. The required
allowable bearing capacity of the chalet does not appear to be in line with the actual 20 kPa
load of the chalet; however, as this bearing can be usually achieved under a raft footing
without extra work and cost, it was accepted. On the other hand stipulating settlement
criterion for 100 kPa uniform load over the chalet foot print would be very irrational because
that load is several times larger than the actual load, will never be realised in a one storey
chalet, and is in the magnitude of multi‐storey buildings. Relating settlements to a load
equivalent to the bearing capacity would needlessly increase construction energy and cost,
to a point where ground improvement may no longer be a feasible solution due to poor and
unsuitable specifications. Bearing or shear failure for the walls has been addressed, but as
the wall design was not completed during design of ground improvement it was not possible
to optimise design criteria by comprehensively stipulating wall stability as well. Total hill
settlement was rightfully deemed as unimportant during construction phase, and was not
incorporated in the criteria; however, total and differential settlements had to be defined
after a certain amount of time to ensure that the walls, chalets, utilities and roads on the
embankments would not be subject to damage.
The design and construct ground improvement tender was based on achieving the
mentioned design criteria, and any ground improvement method was deemed as acceptable.
In the opinion of the author this is the preferred method of performing ground improvement
projects as specialist contractors are given an equal opportunity to propose any technique
that will meet the requirements with the lowest cost. Furthermore, soil improvement is a
technology that is usually driven by specialist contractors who are continuously developing
enhancements to their methods to improve performance.
The project was awarded to a specialist contractor who had based ground treatment on the
dynamic replacement method.
411
If necessary and possible, soft soil may be excavated at the location of the print and backfilled
with granular material prior to implementation of dynamic replacement. This may be due to
the existence of a stiff superficial layer or to facilitate deeper penetrations. In this project
due to the availability of redundant sand in the lagoon area and affordable sand resources in
the vicinity of the project it was decided to pre‐excavate the DR prints and to backfill them
with sand. The pre‐excavation of a DR print is shown in Figure 3‐143.
Figure 3‐143: Pre‐excavation of DR prints
In order to provide a minimum thickness of granular material as the working platform that
was sufficiently above groundwater level, the lowlands and areas with thin layers of
superficial sand were backfilled with sand in such a way that the thickness of the granular
layer was at least 1.5 m and the platform level was at least 1.3 m above groundwater level.
High energy impacts such as what is experienced in dynamic replacement increase the pore
water pressure. In cohesive soils that do not allow the rapid dissipation of pore water
pressure, each consecutive blow increases the pressure to the point where the soil liquefies.
In some soils only one blow is enough to liquefy the soil, and to hinder the application of
additional blows. In such a case, additional blows can (or should) only be applied when the
pore water pressure has sufficiently decreased. As can be seen in Figure 3‐144, built up of
pore pressure and its release in the form of sand boiling was observed in prints surrounding
the impact location during the soil improvement works.
412
Figure 3‐144: Release of excess pore water pressure in the form of sand boiling
Application of pre‐excavation and backfilling of DR prints with sand was able to accelerate
the pore water pressure distribution in the soil; however, even the implementation of this
technique was not able to prevent the liquefaction of soil during the first few poundings, and
work had to be continued in sub‐phases once the pore water pressure had reduced
sufficiently.
Experience gained during the works suggests that the liquefaction of the soil can be retarded
by minimising the time interval between backfilling pre‐excavated prints and pounding. This
may be somewhat expected as increasing the time interval between the different stages of
the work process will allow the backfill material’s water content to increase.
While design calculations indicated that execution of dynamic replacement columns would
be able to satisfy the design criteria in most areas, the wall stability analyses suggested that
additional measures may be required. MSE walls were 3 m, 5 m, 5.5 m and 8 m above finished
road level (+2 m RL). Calculations using numerical analyses showed that it was necessary to
install DR trenches in lieu of DR columns to reach a safety factor of 1.5 (see Figure 3‐145). For
walls higher than 7.5 m, the trench had to extend 5.25 m and 2 m respectively behind and in
front of the MSE wall. For shorter walls, the treatment zone required only 50% replacement.
Hence, while maintaining the original band width, the excavation was done as alternative
segments within the trench setout path with a total coverage of 50%.
413
Figure 3‐145: Stability analysis of MSE Wall on pre‐excavated DR trench
In order to ensure that the bottom of the trenches or segments would have a high friction
angle for stability assessment, the excavations were originally backfilled with demolished
concrete pieces and then with sand that was similar to what was utilised in the DR columns.
Excavation of trenches and execution of dynamic replacement for the walls are shown in
Figure 3‐146.
Figure 3‐146: Backfilling of the lower part of the trenches beneath the wall with demolished
concrete debris
414
Two DR rigs were allocated to the project to complete the works within 150. DR pounders
weighing 12 to 14 tons were used for the works. Based on the ground conditions and project
requirements, drop heights were variable from 5 to 15 m, and grid spacing was 5 m or 6 m.
22 pressuremeter tests (PMT) were carried out after dynamic replacement to verify that
acceptance had been achieved. PLM and EM values before ground improvement, after ground
improvement in between the prints and after ground improvement in a DR column are shown
in Figure 3‐147. It can be observed that while PLM in between the DR columns increased by
about 100% from 110 kPa to 220 kP after dynamic replacement, the magnitude of
improvement has still remained insufficient, and calculations can demonstrate that this
amount of treatment on its own would not have been enough to satisfy acceptance, and it
was in fact the substantial increase of strength in the DR columns that made acceptance
achievable. The same discussion is valid for EM, and it can be seen that the moduli in between
the prints have improved marginally, but the improvement inside the DR columns is
substantial. Figure 3‐147 shows a reduction in parameters at depth, but this is interpreted as
the better properties of the in‐situ ground at the testing location before DR compared to the
post DR testing locations, and is not considered to be the result of applying dynamic
replacement.
PLM and EM improvement ratios in between and inside prints are shown in Figure 3‐148. It
can be observed that while limit pressure improvement in between the columns was at best
100%, the maximum amount of improvement inside the columns exceeded 900%, and was
5.15 times more than improvement in between the columns. The same approximate order
of difference is also noticeable for the modulus.
Bearing capacity acceptance was confirmed by using the method proposed by Centre
D'Etudes Menard (1975) (refer to Section 2.9.2.6).
While calculations confirmed that settlement criteria had been satisfied, 15 settlement plates
were additionally installed to measure the ground subsidence due to construction of the hills
(see Figure 3‐149). Fill height and the associated settlements of one of the settlement plates
are shown in Figure 3‐150. The diagram has not been corrected for possible errors in
surveying, vibrations caused by construction equipment, etc., and it appears that the true
settlement at the 90th day may have been more than what was recorded.
415
Figure 3‐147: Comparison of PMT parameters before and after DR (in between and inside DR
columns)
Figure 3‐148: Comparison of PMT parameters ratios in between and inside DR columns
416
Figure 3‐149: Construction of hills supported by MSE walls on one side
Figure 3‐150: The measurements of fill height and ground settlement under the fill during a time
interval
Consolidation ratio and the maximum ground settlement under the weight of the hill at the
location of each plate was estimated using the method of Asaoka (1978) in which ground
settlement at a specific point is measured at constant time intervals, and settlements are
plotted on both coordinates. The abscissa of each point on the graph is the settlement, Sn,
and the ordinate is the next settlement, Sn+1. At 100% consolidation the settlements plot
must intersect the bisector. As shown in Figure 3‐151 the consolidation ratio of the
settlement plate that was shown in Figure 3‐150 reached 99% after 90 days, and the
stipulated 90% criterion was achieved about 6 weeks before 90 days.
417
Figure 3‐151: Estimation of settlement and consolidation ratio using Asaoka’s method
1. Pre‐excavation was used to accelerate pore water pressure dissipation and to allow
the application of more pounder blows during each sub‐phase of DR. At critical
locations pre‐excavation was further adopted in the form of DR trenches.
2. By pre‐excavation and allowing accelerated dissipation of pore water pressure it is
possible to use sand for backfill material.
3. While PMT parameters of the soft layer increased after DR by up to 100%, the
improvement was not sufficient to satisfy acceptance criteria without the
introduction of compacted granular material in the form of DR columns. PLM
improvement inside the DR column was more than 900% the value of the original
ground, and 5.15 times more than the improvement in between the columns.
Figure 3‐152 shows the completed Abu Dhabi Ritz‐Carlton Hotel.
418
Figure 3‐152: Completed chalets at Abu Dhabi Ritz‐Carlton Hotel
419
3.9 Al Jazira Steel Pipe Factory
Al Jazirah Steel Pipe Factory (AJSPF) is constructed on Plot 203ER6 of Industrial City Abu Dhabi
(ICAD) Phase 1 Extension in the UAE. The plot is approximately a chamfered square with an
area of 397,889 m2. As can be seen in Figure 3‐153, six buildings were considered for the first
phase of AJSPF. These included the 31,200 m2 main and annexed building, the 2,700 m2
workshop, the 3,300 m2 hot bending building, the 1,700 m2 administration building, a
mosque with associated washrooms, and the fire water tank and pump room.
Figure 3‐153: Layout of Al Jazira Steel Pipe Factory structures
Different activities including plate stacking, preparing, rolling and bending, welding,
inspecting and controlling are undertaken in the main building, and the maximum vertical
load on a stanchion of this building was estimated to be 2,500 kN. As shown in Figure 3‐154
floor loads ranged from as low as 40 kPa to as high as 200 kPa. The annexed building mainly
consists of storage areas, laboratories and changing rooms, and the maximum vertical
column load was estimated to be 300 kN. The workshop and hot bending buildings each have
50 columns that carry crane loads and support vertical loads of up to 1,000 kN. The
administration building was the only two storey structure on site with column loads of up to
1,500 kN.
420
Figure 3‐154: Loading patterns in the main building
Ground level was generally 1 m below final finished levels of the buildings, and groundwater
level was 1.8 to 1.6 m below site level.
SPT results suggested that the geotechnical profile of the site was composed 11 to 12 m of
soil followed by bedrock. SPT blow counts in the upper 5 m were generally as low as 10 and
up to 15, but then the soil became very dense with blow counts occasionally exceeding 50. It
appeared that up to 2 m of the compressible soil at the depth of 1.5 to 2 m was high in fines.
CPT results illustrated a more problematic condition with the presence of a soft layer of silty
or clayey material that commenced at the depth of about 1.5 to 2 m, was approximately 2 m
thick, generally had qc in the range of 1.5 to 4 MPa but sometimes as low as of 0 MPa.
Figure 3‐155 shows the typical soil profile.
Figure 3‐155: Soil profile
421
Supplementary PMTs that were later carried out proved that the CPT was more
representative of the ground conditions. PLM in the soft layer was generally in the range of
0.2 MPa and EM was as low as 15 kPa.
The history of almost all projects in ICAD using piles as their foundation system and the poor
ground conditions was enough evidence for the project’s developer to conclude that the
project required the implementation of special measures for safely transferring the loads to
the ground.
The developer of AJSPF considered the alternatives of piling or performing ground
improvement and utilising shallow footings. Piling was well established in the region and was
a method that was technically acceptable; however, the cost study revealed that installation
of 2,500 piles was more expensive than ground improvement. Among possible soil
improvement technologies stone columns and dynamic replacement were proposed, but
after weighing the technical and financial aspects of each method, it was the latter that was
deemed as the more appropriate of the two techniques, and a design and construct contract
was awarded to the specialist geotechnical contractor who had proposed DR.
Due to the variation of loading and the deformation tolerances of the different components
of the project the design and acceptance criteria consisted of a number of bearing capacity
and settlement requirements that have been summarised in Table 3‐13, and quality control
was based on PMT. Footings in the main building were from 1.5 x 1.5 m2 to 4.25 x 5.5 m2.
Maximum Settlement
Type Allowable Bearing Capacity (kPa)
Total (mm) Differential
Building footings 200 kPa, 1.5 to 3.5 m below finished 25 1/1,000 to
floor levels 1/500
Equipment footings generally 75, locally up to 200 25 1/1,000
Ground slabs
Steel plate storage 100 75 1/500
Water tank 150 50 1/500
Other industrial 40 35 to 50 1/500
Other non‐industrial 15 to 25 25 1/500
Table 3‐13: Summary of design criteria
422
Initial calculations suggested that heavy point loads would require a number of DR columns
to safely transfer the loads to the foundations, and a cost study suggested that it would be
more economical to pre‐excavate the soft soils, backfill the foundation with granular material
and to compact the backfill dynamically rather than to create a group of individual DR
columns under the load. It should be noted that this decision was made attractive because
the excavated material did not contain any contaminants that would have required
treatment, and it was possible to distribute the excavated material throughout the vacant
factory boundaries at very low cost (ironically, while the excavated material was deemed as
waste material, the project’s developer was later approached by other organisations who
were interested in buying the excavated material). Thus the soft soil was pre‐excavated and
replaced with granular material under the stanchion and equipment footings. In the storage
areas with 100 and 150 kPa of uniform loading and smaller footing locations the excavation
and removal was limited to the DR column locations. The schematic illustration of this
process is shown in Figure 3‐156.
Figure 3‐156: Schematic illustration of the dynamic replacement process
Although dynamic replacement is frequently carried out using crushed stone because it can
provide better properties, any granular material including construction debris and
demolished concrete that will not rapidly liquefy under dynamic loading conditions can be
used for creating DR columns. As the closest stone quarries were about 300 km away from
the project, in addition to the high price of crushed stone, transportation of the material
would have also been costly. Calculations indicated that it would be possible to meet the
423
design and acceptance criteria by using locally sourced sand. The emirate of Abu Dhabi is
basically flat ground and suitable sand quarries were still about 100 km away; thus, as a
further step towards optimisation, sand was supplied from excavation sites within the
metropolitan areas. The portions of the upper sandy soil that were not mixed with the fine
soils during pre‐excavation of the DR columns were also used for backfilling.
(a) (b)
Figure 3‐157: (a) Pre‐excavation, (b) backfilling with sand
Mobilisation, calibration, completion of 37,000 m2 of dynamic replacement and testing was
performed during a period of less than 4 months.
3.9.4.1 DR Calibration
Prior to the commencement of ground improvement works two heave and penetration tests
were carried out of which one test was for dynamic replacement and the other was for
pre‐excavated dynamic replacement. The results of the first test is described, reviewed and
analysed in this section.
The testing layout is shown in Figure 3‐158. For the purpose of the calibration programme a
13 ton pounder was dropped from 10 to 20 m.
424
Figure 3‐158: Dynamic replacement HPT test layout
Pounder penetration and volumetric changes versus the number of pounder blows are
shown for HPT‐1001 in Figure 3‐159. During the HPT the pounder was dropped consequtively
8 times when crater depth reached 1.74 m, and the crater was backfilled with sand before
additional blows were applied. Both crater depth (penetration reset at backfill) and
accumulative penetration depth are shown in Figure 3‐159. It can be observed that
introduction of backfill in a loose state after the 8th blow has resulted in a sharp increase in
the rate of penetration at the 9th blow, but the penetration curve returns back to the previous
logarithmic trend at the 10th blow.
Contrary to HPT in sand (e.g. Marjan Island Road Corridor Project described in Sectoin 3.6.4.1)
where successive blows created settlement (negative heave) around the pounder’s impact
point, it can be observed in Figure 3‐159 that pounder blows are creating heave at the
periphery of the impact point. The compaction, heave and net compaction volumes
(difference between compaction and heave volumes) rates appear to be relatively linear, and
compaction has not reached a stage where the soil has liquefied, and additional blows simply
result in volumetric displacement with little or no compaction, such as was observed during
the ground improvement works of the saturated very silty sand to sandy silt section of Dubai
Airport Runway that is shown in Figure 3‐160 (Serridge, 2002).
425
Figure 3‐159: Pounder penetration and volumetric changes during the DR HPT
Figure 3‐160: Heave and penetration test at Dubai Airport (Serridge, 2002)
The result of the HPT becomes more meaningful when before and after dynamic replacement
PMTs are compared in Figure 3‐161 and Figure 3‐162. It can be observed that PMT
parameters have increased both in between and inside DR prints down to the depth where
the ground becomes very dense; however, the amount of improvement inside the prints are
larger. The initial PMT parameters at the depth of soft soil, the amount of improvement and
that depth, and the amount of ground heave suggests that supposedly soft cohesive layer at
the testing location was probably higher in sand content than the typical location
encountered at site; hence, the amount of improvement in the soft layer was comparable to
other layers.
426
Figure 3‐161: DR calibration PMT results
Figure 3‐162: DR calibration PMT improvement ratios
427
3.9.4.2 Final Testing
In addition to the six PMTs that was performed during calibration, 41 PMTs were also carried
out during the works. 37 of these tests were performed after ground improvement for
verification purposes.
The results of three PMTs are shown in Figure 3‐163. For comparative purposes a PMT was
performed before dynamic replacement. The test was located in between pre‐excavated DR
prints, and was re‐tested after ground improvement, and a PMT was also performed in one
of the print locations next to the location that had been tested earlier. Ground level at the
testing points was +4.4 m ADD (Abu Dhabi Datum). It can be seen that the soft layer was
approximately from elevation +1.5 to +2.5 m ADD in the test area. After dynamic
replacement PLM and EM increased both inside and outside of the DR columns in the first
several upper metres until very dense sand was encountered. Improvement was also
observed in the soft layer that would typically be assumed to remain unimproved in
calculations; however, the magnitude of improvement in the soft soil was noticeably less
than the sand and what was observed in Figure 3‐161 and Figure 3‐162. The soft layer has
somewhat improved because in reality it was not pure silt and clay, and it did contain an
amount of sand in its grading. Hence, asa also observed in KAUST (see Figure 3‐54 in
Section 3.4.5) the impact did not only lead to plastic displacement and heave, and the impact
energy was able to improve this layer to an extent as well.
The effectiveness of dynamic replacement is more observable when PLM and EM
improvement ratios in between and inside DR prints are compared with each other. It can be
seen that improvement ratios in the sand layers are of the same order, but there is a
pronounced difference at the level of the soft soil. In this band the PLM and EM improvement
ratios of in‐situ ground are respectively 3.9 and 2.6, but the replacement sand’s PLM and EM
improvement ratios are respectively 12.2 and 12.4; i.e., respectively 3.1 and 4.8 times more
than that of the in‐situ cohesive soil.
The harmonic mean of PLM in the pre‐excavated DR column is 2,817 kPa. Assuming that the
lower layers also have the same limit pressure, it can be calculated using the method
proposed by Centre D'Etudes Menard (1975) (refer to Section 2.9.2.6) that a sandy soil with
a footing that is on ground surface will be able to provide a safe bearing of 800 kPa with a
safety factor of 3. Bearing will be even higher for embedded footings; however, in that case
it will be the settlement criteria that will govern. For example, calculation for a 3.5x4.5 m2
428
stanchion footing demonstrates that settlement will be 9 mm when the foundation is
subjected to a load of 2,500 kN. Clearly, settlement under a load of 12,600 kN (800 kPa x3.5
m x4.5 m) will greatly exceed the settlement criterion.
Figure 3‐163: Comparison of PMT parameters before and after DR
Figure 3‐164: Comparison of improvement ratios in between and inside DR columns
429
The author notes to have observed that some engineers are in the habit of using formulas
for predicting bearing capacity that are based on a 25 mm settlement. This project, where
settlement criteria are variable and not necessarily equal to 25 mm, is a very good example
of how this approach could provide erroneous results. Bearing capacity and settlement are
two different phenomena. Each should be individually assessed; however, in the end it will
be one of them that will govern the design.
Detailed calculations were carried out using the PMT results to verify that all design criteria
had been satisfied.
Al Jazira Steel Pipe Factory was the first of a series of ground improvement projects in the
Industrial City of Abu Dhabi. In this project dynamic replacement and pre‐excavated DR were
successfully used to provide foundations for footings and slabs on a site that contained a
layer of saturated soft soil.
It was noted that:
1. Acceptance criteria should be based on design criteria.
2. Sand can be used in lieu of commonly used crushed stone for DR column material to
reduce material and transportation costs.
3. Pounder impacts create heave when the ground contains saturated cohesive soils.
4. Although DR was able to improve the PMT parameters of the soft soil layer, the
amount of improvement in the soft layer was noticeably less than the other layers.
However, inside the DR column, the replacement backfill improved as much as the
other layers.
Figure 3‐165 shows the completed main building.
430
Figure 3‐165: Al Jazira Steel Pipe Factory
431
3.10 Reem Island Causeway
Reem Island, previously called Abu Shaoum, is a small island that is located approximately
0.4 km north of Abu Dhabi, United Arab Emirates. The island was basically vacant until 2005
when it was decided to develop it into a modern suburb of UAE’s capital city.
One of the first requirements of the new development was the construction of a causeway
to link the island to the mainland. According to the design, the causeway was to be composed
of approach roads on the two sides and a bridge structure in the centre. The approaches on
each side were to be approximately 150 m in length. As can be seen in Figure 3‐166, the
reclamation was anticipated to be approximately 135 m and 50 m long respectively on Abu
Dhabi and Reem Island sides.
Figure 3‐166: Site plan of Reem Island Causeway
432
The approach road was to be constructed on coastal grounds and on reclamation. The road
level was anticipated to be from +2.0 m RL (Reduced Level= mean sea level) to +7.0 m RL at
bridge level. The maximum elevation difference between the low and high points of the
approach road was 5 m, and the road slope was 3.25%. The longitudinal profile of the project
(Abu Dhabi side) is shown in Figure 3‐167.
Figure 3‐167: Longitudinal profile of the approach road
The approach road and bridge were designed to have four lanes in each direction. The width
of the approach road leading to the bridge was 28 m. An additional lane was envisaged on
each side for drivers who intended to turn back without crossing the bridge. In order to limit
the total width of the road to 38 m the stability of the two sides of the bridge’s access road
was to be provided by an MSE (mechanically stabilised earth) wall. The cross section of the
project at bridge level is shown in Figure 3‐168.
Figure 3‐168: Project cross section at bridge level
Natural ground levels (NGL) in Abu Dhabi and Reem Island approaches were respectively at
about ‐0.5 m and +1 m RL, but rapidly dropped to about ‐7 m and ‐5.5 m RL on the sides of
the bridge. Groundwater level in the boreholes varied from +0.7 to ‐0.7 m RL.
Although NGL in the marine boreholes differed, as summarised in Table 3‐14, the in‐situ
ground profile was generally the same within the project’s area. The upper 0.8 to 1.5 m of
soil was soft sandy silty clay. This layer was followed by a very loose to very dense sandy layer
with a variable thickness of 0 to 2 m that contained less than 20% fines. This latter layer
433
overlaid bedrock. The bottom elevation of the loose sandy layer was from about ‐6 m to ‐8
m RL.
Fill (reclaimed by up to 9 ‐ < 20% 250 to PMT values after
dumping) 400 reclamation and before
ground improvement
marine mud (sandy 0.8 to 1.5 0‐2 50 to 80% ‐ removed before ground
silty clay) improvement
loose in‐situ sand 0 to 2 4 to < 20% 500 to
30 700
bedrock ‐ ‐ ‐ ‐ encountered at
elevation ‐6 to ‐8 m RL
Table 3‐14: Ground profile of reclamation area before ground improvement
Although the marine mud thickness was at most 1.5 m, it was understood that the
consolidation of this layer during the life time of the project could cause excessive
settlements; hence, it was removed by dredging the seabed prior to filling and reclamation.
With similar intent to minimise foundation settlements due to poor ground conditions and
the marine environment in which the bridge piers had to be constructed in, the foundations
of the piers were designed as bored piles.
As shown in Figure 3‐169, reclamation was anticipated to be done by dumping sand and
pushing it into the sea. Soil layers above the sea level were to be compacted using vibratory
rollers; however, the submerged fill mass and existing loose sand would have remained in a
loose and uncompacted state. Although it was expected that the sand fill would consolidate
in a relatively short period under the embankment loads, it was also recognised that the fill
could pose a number of problems such as insufficient bearing capacity, creep settlement
under self‐weight, excessive differential settlements of the MSE walls, and excessive total
and differential settlements under vibratory traffic loads. These problems could be most
evident in the form of unpleasant bumps at the interface of the approach road and the bridge
abutment (Li, 2005).
434
Figure 3‐169: Reclamation by dumping sand and pushing it into the sea
Based on the above concerns, ground improvement of the submerged fill was envisaged to
be carried out in the form of a design and construct package, and the contract for soil
treatment was awarded to a specialist contractor who had proposed the application of
optimised design criteria, and implementation of dynamic compaction along with an
objective testing method based on PMT.
The criteria that were defined to address the geotechnical concerns of the project were:
1. Safe bearing capacity under the approach road: 120 kPa with a safety factor of 3.
2. Total settlement of the fill in the approach road with a uniform loading of 20 kPa: 30
mm
3. Differential settlement of the fill in the approach road with a uniform loading of 20
kPa: 1:500
The maximum required bearing capacity of the fill is at the location where the approach road
reaches the bridge elevation at +7 m RL, which is realised by constructing 5 m of embankment
in between the MSE walls. Conservatively assuming that the unit weight of the engineered
fill is 20 kN/m3 and adding an further 20 kN/m2 for traffic loads, the required bearing capacity
will be 120 kPa. At the same time the pavement of the road was designed to be able to sustain
435
a maximum total settlement with a condition that differential settlements did not exceed
1:500.
It is noted that in this project the settlement criterion was defined for the actual post
construction loading, not for a load that was equal to the bearing capacity. Once the ground
has been stabilised, and the project has been constructed, the road will be subject only to
traffic loads that at most will be 20 kPa, and will never reach an extraordinary road loading
of 120 kPa that is approximately equivalent to a 12 m high oil tank or 10 storey building
spread out over the project’s area. Indeed, specifying the same load for the bearing capacity
and settlement criteria would have been poor engineering and meaningless.
Due to the requirements of the main contractor, ground improvement works were carried
out with one mobilisation and demobilisation for each end of the bridge, and each part was
executed within two weeks.
Based on the recommendations of the ground improvement specialist contractor, the fines
content of the fill that was used for reclamation was limited to 20% to remain compatible
with the in‐situ sand and within the range of dynamic compaction treatment.
The fill was tipped into the sea to an average elevation of +1.35 m RL using dump trucks. In
addition to the road’s 38 m width, an extra 5 m was also initially reclaimed on each side of
the approach road to ensure that the slopes of the embankment would also receive sufficient
compaction. The extra width was later further increased to allow temporary passage of traffic
between Abu Dhabi mainland and Reem Island. The extra width was removed after ground
improvement at a later stage of the project, and replaced with geotextiles and rock armour
to protect the reclamation against wave action and erosion.
A 15 ton steel pounder with an area of 2x2 m2 and a drop height of 20 m was used for dynamic
compaction. Other compaction parameters; i.e., grid size, number of impacts per print
location and number of phases were optimised during a calibration program that was carried
out at the beginning of the works. Based on the HPT and PMT results, a total energy of 240
tm/m2 was applied in three phases. Figure 3‐170 shows the application of dynamic
compaction at the second end (Reem Island side) of the project. Also visible in the left hand
side of Figure 3‐170 is the rig used for installing the bored piles.
436
Figure 3‐170: Application of dynamic compaction at Reem Island Causeway
3.10.4.1 DC Calibration
As part of the calibration programme two PMTs (before and after improvement), a HPT and
a penetration test were carried out during phase 1 of dynamic compaction (see Figure 3‐166).
HPT‐1, shown in Figure 3‐171, and PT‐2, shown in Figure 3‐172, were carried out respectively
with 12 blows and 20 blows. It can be observed that implementation of consecutive blows
not only created a crater under the impact point but caused the ground at the periphery of
the print to subside (negative heave) in a progressive manner as well. It can be observed that
initially the amount of pounder accumulative penetration (crater depth) and compaction
volume per blow follows a higher rate, but after about 6 blows the rate decreases, and it can
be extrapolated and expected that asymptotes will be reached at about 35 to 40 blows.
It is generally neither necessary nor justifiable to apply the number of blows for reaching the
asymptote. It is much more preferable to be able to implement the feasible heaviest pounder
and highest drop height to make use of the higher rates of improvement with lesser blows.
Further blows realise less achievement with the same impact energy.
437
Figure 3‐171: Heave and Penetration Test results
Figure 3‐172: Penetration Test results
It should be noted that optimisation of dynamic compaction is based on both the HPT results
and the verification testing that follows. It is possible to optimise the number of blows by
reviewing the penetration and volume changes of HPT, but the amount of penetration, even
if considerable, does not necessarily imply that design criteria have been satisfied. Design
requirements can be confirmed only by proper testing and measuring the parameters that
are able to demonstrate that specifications have been satisfied. In this project PMT was used
for this purpose.
438
The square roots of the crater volume and net compaction volume (summation of crater and
periphery subsidence volumes) are shown in Figure 3‐173. It can be observed that similar to
the results achieved in Marjan Island Road Corridor (refer to Section 3.6.4.1) and Al Nakhilat
Ship Repair Island (refer to Section 3.7.5.1) the square root of the net volume to penetration
curve initially commences with larger values, but slopes down and levels off after the first
several blows to a value that is about 2.6. Therefore, Equation 1‐5 is still able to formulate
the relationship; however, it seems that the contribution of the peripheral subsidence is less
(10% in Al Nakhilat Ship Repair Yard) and about 5%.
Figure 3‐173: The relationship between the square root of volume to penetration and number of
blows
The curves of (DT‐DB)/Dc and crater side angle versus number of blows are shown in
Figure 3‐174. The form and values of these curves appear similar to what was observed in
Nakhilat Ship Repair Yard (refer to Section 3.7.5.1).
Figure 3‐166 shows the location of PMT‐001 and PMT‐004 that were carried out before
dynamic compaction and PMT‐i05 and PMT‐i06 that were performed respectively in DC
phase 1 and phase 2 prints. PLM and EM of these tests are shown in Figure 3‐175, and the PLM
and EM improvement ratios are shown in Figure 3‐176. It can be observed that while
construction equipment traffic had improved the PMT parameters of the upper meter of soil
439
crust, deeper layers were initially very loose with PLM values generally less than 600 kPa and
as low as 220 kPa.
Figure 3‐174: The relationship between (DT‐DB)/Dc and number of blows
Figure 3‐175: PLM and EM values before and after dynamic compaction
440
Figure 3‐176: PLM and EM improvement ratios
Post treatment testing demonstrates that the soil parameters improved significantly
throughout the fill layer whereas maximum PLM and EM improvement ratios were respectively
6 (500% increase) and 7.8 (690% increase). Comparison of PLM improvement ratio in this
project with previous research (Lukas, 1986) that proposes an increase in the order of 100 to
400% (see Section 2.5.5) suggests that greater peak improvement ratios can be considered.
Lukas has reported that the upper bound for PLM after dynamic compaction in sands and
gravels will be 1.9 to 2.4 MPa. It can be seen that the maximum PLM value achieved in this
project has exceeded that expectation in one of the test locations. The PLM values are also
higher than what Mayne et al. (1984) have reported for a number of sites. However, tests
carried out in other projects that have been studied in this thesis, e.g. Al Quo’a New Township
(refer to Section 3.3.5) are in line with the results of the PMT carried out in this project and
indicate that it is possible to achieve larger values than the mentioned range. Case studies by
others such as Spaulding and Zanier (1997) also indicate that higher PLM values of up to 4
MPa can be achieved after dynamic compaction.
It is interesting to note that improvement at depth and at the bottom of the subgrade fill is
still substantial and in the range of 80 to 130%. This massive improvement may have been
441
due to the fact that the very young fill was placed only a short period before ground
improvement works and in a very loose state.
The ultimate bearing capacity of a foundation can be conservatively calculated using
Equation 2‐168 by assuming a value of 0.8 for bearing factor (k) and not considering qo and
Po. It can be calculated that the geometric mean of the PLM the two post treatment PMTs are
respectively 2,640 kPa and 1,890 kPa; thus, with a safety factor of 3, the allowable bearing
capacity will be respectively 704 kPa and 505 kPa, which exceeds the design criterion of 120
kPa by far.
Similarly, the harmonic mean of the EM for the two post treatment tests can be calculated to
be respectively 10.7 MPa and 9.3 MPa. Eoed can be calculated from EM with α= 1/3 for sands
(refer to Equation 2‐214). Thus, the harmonic Eoed for two test locations can be calculated to
be respectively 32.1 MPa and 27.9 MPa.
Once again, the settlement can be conservatively calculated with the assumption that load
intensity does not reduce as stresses distribute in the ground. It can be then readily
calculated with Hooke’s law that the settlement will be respectively 5 mm and 6 mm.
The above figures show that while the settlements are much less than the design criterion in
reality, had the settlement requirements been inappropriately tied to the required bearing
capacity, which is 6 times more than the actual design load, then the settlements calculated
from the second post treatment test would have exceeded acceptance. This comparison
demonstrates the importance of developing proper design criteria.
As a result of the ground improvement works by dynamic compaction the fill settled about
40 cm on average. Should this amount of dynamic compaction induced ground subsidence
be plotted against the applied energy of 240 tm/m2, and compared with previous research
(Mayne et al., 1984) that is shown in Figure 3‐177, it can be observed that the amount of
subsidence in this project falls approximately in the middle of other sites for the same
amount of energy. Thus, it can be anticipated that PLM values of this site are not extraordinary
either, and it may be possible to obtain higher post DC PLM values than what Lukas (1995)
had conceived.
442
Figure 3‐177: Comparison of ground subsidence versus energy with previous research (Mayne et al.,
1984)
The proper stipulation of specifications benefited the project by allowing the engineers to
introduce an affordable ground improvement method with a short construction schedule.
The application of dynamic compaction for the treatment of a submerged fill has proven to
be very successful, and PMTs were able to demonstrate that the design criteria were readily
satisfied. PLM improvement ratio in the upper several metres of soil was 500 to 700%. The
parameters’ improvements was still substantial and in the range of 80 to 130% at depth.
Useful lessons that can be noted include:
1. Stipulating specifications and project criteria based on requirements is a superior and
an optimised approach. This allows works to be carried out in the shortest possible
duration and with the most affordable cost while satisfying the project’s
requirements without introducing technical drawbacks or sacrifices.
2. It would be realistic to assume that reclaimed fills will be in a loose state, and most
will require of ground treatment for most construction projects. Although the soil
above groundwater level may be dense due to construction equipment traffic or
443
other reasons, it is highly likely that there will be loose layers of soil below
groundwater level.
3. There are numerous methods for improving saturated fills; it is the responsibility of
the engineer to identify the best feasible solutions and to implement the one or ones
that can provide the most benefits to the project. In this project the thin layer of
saturated clayey soil was removed by dredging. Dynamic compaction was used to
improve a relatively thick submerged subgrade with fines content up to 20%.
4. It is not necessary to compact the ground until reaching the asymptote of ground
settlement (or compaction) versus number of drops (or accumulative energy). What
is relevant is to demonstrate that design criteria (settlements, bearing capacity, etc)
have been satisfied.
5. It is possible to achieve PLM values greater than 2.4 MPa after dynamic compaction
in saturated sands; however, the peak value decreases with depth.
6. Noting that the amount of improvement is a function of impact energy, it is possible
to improve PLM of saturated sands by 500%. However, the peak amount of
improvement decreases with depth. The maximum amount of improvement appears
to be in the upper half of the depth of improvement.
7. The square root of net compaction to crater depth was measured to be
approximately 2.6 m after the first several blows.
8. Measurements suggest that 95% of the net compaction volume originated from the
crater.
As shown in Figure 3‐178, Abu Dhabi – Reem Island Causeway has been completed, and is
operational. As a result of ground improvement the dumped fill is performing as per the
project requirements and no settlements or indications of cracking due to poor foundation
or bumping has been reported.
444
Figure 3‐178: Abu Dhabi – Reem Island Causeway after completion
445
3.11 Ras Laffan Heavy Fuel Oil Bunkering Facility
3.11.1 Introduction
The Qatari industrial city of Ras Laffan, located on the southern shores of the Persian Gulf
and 80 km from Doha, is one of the world’s largest oil and gas hubs. This industrial complex
is continuously and rapidly expanding to increase the production of gas from the North Field.
One of the projects that has been constructed in Ras Laffan is the Heavy Fuel Oil (HFO)
Bunkering Facility for Qatar Fuel (WOQOD). The project included the design and construction
of an import line, the storage and process area and an export line, along with the necessary
utilities and civil works. The storage and process area includes three HFO tanks, a sub‐station,
a pump shed and other equipment, piping, shelters and other allied facilities. The project
location and site plan can be seen and Figure 3‐179.
Figure 3‐179: HFO Bunkering Facility site in Ras Laffan
The tanks were all designed as steel tanks with fixed roofs. The diameters of Tanks T‐01 and
T‐02 were 60 m while Tank T‐03 had a smaller diameter of 42 m. The bottoms of all tanks
were to be at +3.2 m QNHD (Qatar National Height Datum), and the foundations were to be
subjected to a uniform pressure of 170 kPa.
446
3.11.2 Ground Conditions
As can be seen in Figure 3‐180, the HFO Bunkering Facility is located in a reclaimed area of
Port of Ras Laffan. Site survey and initial tests indicated that ground level was generally flat
and from +1.2 to +1.8 m QNHD. Groundwater was recorded to be 0.95 to 2.3 m below ground
level.
Figure 3‐180: Location of HFO Bunkering Facility in Ras Laffan
13 boreholes, 4 per tank and one in the pump station, were drilled and SPTs were carried
out. The upper 12 m of ground appeared to be silty sand and gravel with cobbles. The limited
information suggested that fines content was from 8 to 23%. While the SPT blow counts in
all boreholes were generally high and in the range of 25 to 50, layers of 1 to 2 m thick with
lower blow counts of 11 to 14 were encountered from depths of 5 to 8 m. The SPT blow
counts of borehole BH2B is shown in Figure 3‐181. It can be noted that testing was aborted
for unrecorded reasons at depths of 2.5, 3 and 5 m. The ground then became limestone with
UCS values mostly recorded to be from 10 to 30 MPa.
At later stages of the project, supplementary PMTs were carried out that appeared to be in
disagreement with the results of the initial SPTs. While SPT blow counts were quite high and
suggested the presence of very dense layers, PMTs indicated that the soil composition and
the presence of cobbles had resulted in misleading and unrepresentative blow counts.
447
Figure 3‐181: SPT blow counts in borehole BH2B
For comparative purposes, PMT T2‐01 was carried out in the same location as BH2B. PLM and
EM values of this test along with PMT T1‐01 and PMT T1‐03 that were performed in the other
tanks are shown in Figure 3‐182. As can be seen, while the SPT blow counts were indicative
of dense to very dense soil, PLM readings were high only in the upper most soil layer, possibly
due to traffic and movement of construction equipment. PLM in deeper layers of the
reclamation was from approximately 100 to 250 kPa, which indicated the presence of loose
sand subject to creep under self‐weight (refer to Section 2.9.2.8).
The comparison of the results of the SPT and PMT can be an oblique reminder of the fact that
SPT was originally developed for sampling soil (Rogers, 2004) rather than assessing its
strength, and that higher blow counts in gravel and cobble strewn formation grounds are
recorded when gravel or cobbles plug the end of the split‐spoon (Abramson et al., 2002). This
can lead to false and very misleading interpretations and can result in dangerously under
designed foundations.
448
Figure 3‐182: PLM and EM values in PMT T2‐01
Before the supplementary PMTs were carried out and the presence of loose soils throughout
the reclaimed site was established, the medium dense sandy layer at the depth of
approximately 5 to 8 m was sufficient reason for applying applying specific geotechnical
measures to ensure the safe transfer of loads from the structures to the ground. Piling,
although applicable, was deemed as an expensive solution and ground improvement was
preferred as a more competitive alternative.
The ground improvement package was awarded to a specialist geotechnical contractor who
had proposed the application of dynamic compaction. For the tanks, bearing capacity was
determined based on the tank loading. Total and differential settlement acceptance criteria
were proposed based on Mobil (1990) specifications. Thus the criteria that had to be sastified
became:
Bearing capacity: 170 kPa
Total settlement: 300 mm under a uniformly distributed load of 170 kPa
Differential settlement: 1:180 under a uniformly distributed load of 170 kPa
449
PMT and finite element analyses were to be used for the verification of design criteria.
Interpretation of PMT was carried out using the method of Menard (Centre D'Etudes
Menard, 1975).
The total ground improvement works are summarised in Table 3‐15.
Noting that the risk of a large diameter tank failing in bearing is very unlikely, especially when
deeper layers are limestone bed rock, failure could be expected to be either by local shear
beneath the tank wall or due to excessive settlements. Therefore, it was defined that bearing
capacity as calculated using Menard’s (Centre D'Etudes Menard, 1975) equation (refer to
Equation 2‐168), had to be satisfied in the superficial upper 5 m that could be subject to shear
failure.
2‐168
450
With a safety factor of 3, assuming that the base of the tank is at ground level and with k=0.8,
it can be calculated that the required geometric mean value of PLM will be 638 kPa. At greater
depths it was defined that Pf had to be greater than the extra stress; i.e.
∆ ′ 3‐9
Equation 2‐168 yields: Pf ≥ 170 kPa
It can be conservatively assumed that PLM ≥ 170 kPa with consideration that creep pressure
is one half to two thirds of the limit pressure (Centre D'Etudes Menard, 1975); i.e.
1 2
3‐10
2 3
Settlements were calculated using numerical analysis software. In the model whose
geometry is shown in Figure 3‐183, it was assumed that the bottom plates of the tank were
represented by a beam with flexural rigidity EyI and axial stiffness EyA. The Young modulus
of steel was assumed to be 210,000 MPa, and the thickness of the bottom plates was
considered to be 2 cm. The radii of the tank and dynamic compaction were respectively
assumed to be 30 m and 35 m, and the uniform load applied to the tank was 170 kPa. The
parameters used in the model are shown in Table 3‐16.
The numerical analysis whose results are shown in Figure 3‐184 indicated that maximum
settlements at the centre and edge of the tank were respectively 14.9 cm and 9.0 cm that
are less than 300 mm. The centre to edge differential settlement can be calculated to be 5.9
cm, which is equivalent to 0.1475% (0.26/180) and is less than 1:180.
Prior to dynamic compaction the ground was raised to +2.3 m QND by backfilling the site with
sandy soil.
Based on the depth of treatment it was decided to apply heavy dynamic compaction using a
23 ton pounder that was dropped from 22 m. Compaction was applied in three deep phases
for the tank foundations. Based on the results of the calibration, it was originally decided to
apply 15, 11 and 8 blows respectively in deep compaction phases 1, 2 and 3; however, to
451
further increase the safety factor and to achieve even more improvement than necessary to
meet design the number of blows in phases 1, 2 and 3 were respectively increased to 25, 15
and 8. A further 7 blows were then applied to phase 3 prints to homogenise the compaction
of phases 2 and 3. As shown in Figure 3‐185 the first and second phases of deep compaction
were carried out in a rectangular pattern of 8.3 m by 9.6 m, and the third deep compaction
phase’s grid was designed in such a way that the overall deep compaction grid became an
equilateral triangle with 4.8 m sides. Ground improvement was carried out with on offset
width of about 5 m beyond the tanks’ boundaries.
Figure 3‐183: Numerical analysis model geometry
(kN/m3) 18
0.33
Table 3‐16: parameters used in the numerical analysis model
Dynamic compaction of the buildings was performed using 2 deep compaction phases, with
15 blows in each phase.
Upon mobilisation and completion of calibration, dynamic compaction works for the three
tanks and three buildings were completed in less than 4 weeks.
452
(a)
(b)
Figure 3‐184: (a) Vertical displacement contours, (b) vertical displacement at the ground surface
3.11.4.1 DC Calibration
Prior to commencing ground improvement a calibration programme that included 3 heave
and penetration tests was carried out respectively in the prints of deep compaction phases
1, 2 and 3 of Tank T‐02. Although it is common practice to carry out the calibration outside
the project’s treatment area, due to space limitations, in this project the calibration was
performed within the tank’s boundary.
453
Figure 3‐185: Layout of Dynamic Compaction works and PMTs
Ground surface settlements measured during HPT‐01 in deep compaction phase 1 due to 22
pounder drops is shown in Figure 3‐186. The 23 ton pounder that was dropped from 22 m
had a base area of 1.9x1.9 m2. In addition to measuring the crater’s upper diameter and
depth in four corners, changes in ground elevation were also measured up to 6 m away from
the print centre in three directions.
It can be observed that while the ground around the pounder impact point has also settled,
most of the deformation has occurred either directly around the pounder or at its vicinity.
Yet, irrespective of the vibration magnitude, the ground deformation at even the farthest
measured distance is sufficient to cause damage to most structures. Therefore, it can be
concluded that at very close distances, both vibrations and ground deformation can result in
cracking and damage to buildings.
Figure 3‐186: Measured ground settlement during HPT‐01
454
Figure 3‐187 shows the increase in crater depth, crater volume, heave volume and net
volume of compaction due to the increase of pounder blows in HPT‐01. It can be observed
that similar to the previous projects that have been studied in this thesis, the ground around
the print has also undergone settlement (negative heave) due to pounding. However, it
appears that while about 90% of net compaction volume originated from the crater in
Nakhilat Ship Repair Yard (Refer to Section 3.7.5) that was also in Ras Laffan, and about 95%
of net compaction volume in Reem Island Project was realised from the crater, in this project
only about 60 to 65% of the net compaction volume was measured as crater volume.
Figure 3‐187: crater depth, crater, heave and net compaction volume changes due to consecutive
pounder blows during HPT‐01
Figure 3‐188 shows that the square root of the crater volume to penetration was measured
to be approximately 2.2 after the first several blows, which is similar to the other projects
that have been reported in this thesis, but the same ratio for net volume is higher than the
previous projects and about 2.75. Therefore, the range of Equation 3‐5 can be extended to
2.75; i.e.
√
2.15 2.75 3‐11
455
Figure 3‐188: Variation of the square root of compaction volumes with consecutive pounding
Measurement of ground levels after dynamic compaction in Tank T‐02 indicated that the
ground level had dropped by about 0.85 m from +2.3 m QND to +1.45 m QND. With
consideration that phase 1 prints amounted to 27% of all deep phase prints, by applying a
factor of 0.8 based on the observation that the ratio of phase 2 to phase 1 compaction in
Marjan Island Road Corridor Project was about 0.8 (refer to Section 3.6.4.1), it can be
estimated that the average blow count per print (25 blows per phase 1 print and 0.8 times
15 blows per phase 2 and 3 prints) would be 16. The crater depth for 16 blows is 1.63 m in
Figure 3‐187. By assuming that the ratio of the square root of net compaction volume to
crater depth is 2.75, it can be estimated that the volume of compaction per print is
approximately 20 m3. The unit cell for a triangular grid of 4.8 m is 19.94 m2, which suggests
that the ground should have settled about 1 m. This estimate is about 17% more than reality
(0.85 m), and suggests that the ratio of 0.8 for the ratio of phase 2 to phase 1 compaction
volume may have been too high when the ratio of crater compaction volume to net
compaction volume is about 0.6 to 0.65; however, it may still be a suitable way to make a
first estimate of the amount of ground subsidence due to dynamic compaction.
PMT parameters in Tank T‐02 before and after dynamic compaction are shown in
Figure 3‐189. T2‐01 demonstrates the ground condition before application of dynamic
compaction. T2‐010 was carried out in a phase 1 print during the calibration programme after
456
14 pounder blows, T2‐011 and T2‐012 were performed in between the prints during
calibration (T2‐011 after 22 blows in phase 1 and 21 blows in phase 2), T2‐013 was conducted
in a phase 3 print after 15 blows, and T2‐014 was executed in a phase 1 print.
Figure 3‐189: PLM and EM values before and after dynamic compaction
The effect of increasing the impact energy intensity is observable in Figure 3‐189. It can be
seen that the PMT parameters in a phase 1 print after 14 blows is similar to when higher
energy was implemented in the upper and lower depths of the soil. However, there is a
noticeable difference in the intermediate treatment depth. The scatter of results between
tests carried out in the compaction prints and in between the prints suggests that, on
average, in between prints have improved as much as the prints.
It is interesting to note that while Lukas (1995) suggests that dynamic compaction can
increase PLM values to 2,400 kPa and improved limit pressure values in previously studied
projects in this thesis were generally less than 3,000 kPa, it seems that in this project the
maximum value of PLM exceeds the more common observations and, on average, is about
5,000 kPa.
457
PMT parameters’ improvement ratios based on T2‐01 and the average values after dynamic
compaction are shown in Figure 3‐190. As can be observed PLM and EM improvement ratios
are both extremely high (possibly due to the very low PMT parameters before ground
improvement), and by far larger than what has been observed in other projects. For example,
peak values for both PLM and EM improvement ratios are in the order of 40 while the peak
figures in the previously studied projects of this thesis were in the order of 6 to 12.
Figure 3‐190: Ratio of PLM and EM before and after dynamic compaction
It can be observed from the test results that bearing capacity and settlement criteria have
been well satisfied.
Hydro testing was performed after completion of ground improvement and the fabrication
of the tanks. Of the 32 points on the shell of Tank T‐02, maximum and minimum total
settlements were 13 and 26 mm, and differential settlement between any two consecutive
points that were 5.89 m apart was less than 4 mm that equated to less than 1 in 1,472, and
was considerably less than the design criterion of 1:180.
458
3.11.5 Lessons and Conclusion
In this project it was observed that:
1. SPT can result in misleading interpretations when the ground is composed of large
size particles.
2. Compaction results very impressive and it was seen that even at the deepest layers,
at 11 m, improvement was still significant.
3. Maximum percentage of improvement was much more than what is suggested by
earlier literature and observed values for other projects that have been studied in
this thesis. PMT parameters’ improvement ratios were in the order of 40 and
maximum PLM was in the order of 5,000 MPa.
4. In addition to the PMTs and finite element calculations, a full scale hydro test was
performed for each tank. These tests were able to demonstrate that both total and
differential settlements were considerably less than design criteria.
5. Buildings that are at very close distances from dynamic compaction works may
sustain damage not only due to vibrations, but also due to DC induced ground
deformations.
6. Crater volume was about 60 to 65% of net compaction volume. This is much less than
about 90% that was recorded in Al Nakhilat Ship Repair Yard Project.
7. The ratio of the square root of net compaction to crater depth was measured to
approximately 2.75 after the first several blows. This figure is higher than the other
projects that have been studied in this thesis prior to this project.
8. It may be possible to make a first estimation of the amount of ground subsidence
due to dynamic compaction by measuring crater depth and applying the value to the
ratio of the square root of compaction volume to crater depth.
Figure 3‐191 shows the HFO tanks during construction.
459
Figure 3‐191: HFO tanks under construction
460
3.12 Palm Jumeira Sewage Treatment Plant Tanks
Palm Jumeira is a group of man‐made reclaimed islands off the coast of Dubai in the United
Arab Emirates and the first of the three famous and world renowned palm shaped
reclamations that lead to the notion of figure shaped reclamation projects. The Palm consists
of a tree trunk, a crown with 17 fronds, three surrounding crescent islands that form an 11
km long breakwater and two identical smaller islands on the sides of the trunk that are in the
shape of the logo of The Palm. The island itself is 5 km by 5 km, and has added about 78 km
to Dubai’s original 72 km coastline.
In total, 94 million m3 of sand and 7 million m3 of rock have been used in the construction of
Palm Jumeira. Calcareous sand was dredged from the Persian Gulf using trailing suction
hopper dredgers (Dowdall Stapleton, 2008). During the dredging process the dredger initially
lowered the suction pipes on both sides of the ship all the way to the seabed. Sand pumps
transferred the sand dredged up by the suction head into the hold or hopper, and the excess
water was drained off via the overflow pipes. When the hopper was full, the ship sailed off
to the reclamation area. When possible, the hopper was discharged by means of a big door
located on the bottom of the hull, but when the water was shallow, that was not possible
and the dredger sprayed the sand and water mixture onto the reclamation by rainbow
discharge. The process of rainbow discharge that was utilised on Palm Jumeira is shown in
Figure 3‐192.
Figure 3‐192: Reclamation by rainbow discharge at Palm Jumeira
In general, the reclamation at Palm Jumeira was about 12 to 14 m thick of which about 3 to
4 m was above sea level. It was observed that CPT cone resistance of the deposited
calcareous sand above water level was very high and in the range of 20 to 40 MPa. The soil
then became very loose in the rainbow discharged sand layer below water level with qc as
461
low as 1 MPa in the next 4 to 5 m of soil. Loose to medium dense sand with qc varying from
4 to 8 MPa was encountered down to depths of about 12 to 14 m where the soil became very
dense. Carbonate content of the sand, measured as CaCO3, varied from as approximately
60% to more than 90%.
Due to the low strength and high compressibility of the soil, almost the entire reclamation
was treated using vibro compaction. Heavily loaded structures were additionally constructed
on piles. The two sewerage treatment plant (STP) tanks were the only heavily loaded
structures that were neither treated with vibro compaction nor were supported on piles.
As shown in Figure 3‐193, a sewerage treatment plant has been constructed at the tip of each
of the lower crescents of Palm Jumeira. The SPT lots were denoted A‐A and G‐G, and each
plant includes one reinforced concrete tank that has a diameter of 35.1 m.
Figure 3‐193: Location of the two sewerage treatment plant on Palm Jumeira
Each tank was subject to the below loads:
Dead load corresponding to the reinforced concrete structure: 41 kPa
Dead load corresponding to steel structures: 8 kPa
Live load corresponding to the liquid inside the tank: 71 kPa
Total dead and live load: 120 kPa
462
3.12.2 Preliminary Geotechnical Investigation
While no geotechnical investigation was available for Lot G‐G, two SPT boreholes and two
CPTs were carried out not very far from Lot A‐A’s tank location.
The boreholes indicated that the upper crust of the soil was generally very dense with SPT
blow counts up to 28; however, the deeper layers of soil were less dense, with minimum
blow count in the upper 8 m of the reclamation dropping to as low as 5. The soil then
appeared to become denser with a minimum blow count value of 18 below 8 m depth and
exceeding 50 at the depth of 13 m. The fines content of the soil in these two boreholes was
from 2 to 10% in the upper 13 m of soil, but increased to 22% at the depth of 14.5 m. Ground
level was at +4 m RL (reduced level) and groundwater was at the depth of about 3 m.
CPT readings also suggested that the upper 2 m of soil was composed of very dense sand
with qc as high as 25 MPa. The soil then became loose with qc as low as 3 to 4 MPa down to
the depth of approximately 13 m where great resistance was encountered and testing was
terminated.
Although the SPT and CPT results suggested that the soil was clean sand, fines content as
high as 30% was observed in a number of boreholes that were not very far from the project.
In the absence of conclusive and definite geotechnical information, the limited and
somewhat contradicting geotechnical data from nearby boreholes and the general pre‐
ground improvement conditions of Palm Jumeira was enough evidence to indicate that
specific foundation measures were required.
Although piling was deemed as a feasible method the costs associated with the execution of
deep foundations made this solution unattractive. Steel piles were both uncommon and
expensive in the UAE. Cast in place concrete piles were very common; however, providing
concrete to the crescent islands by marine transportation made this method very inefficient
and costly.
In the event that a suitable ground improvement solution was developed, acceptance criteria
at tank foundation level (+2.5 m RL) were specified to be:
Bearing capacity: 160 kPa with a safety factor of 3
463
Differential settlement: 1/750 for a uniformly distributed load of 120 kPa
During that period vibro compaction was the commonly practiced method of ground
improvement on Palm Jumeira, but the possible presence of silty sand placed doubt on the
applicability of this technique. Stone columns (vibro replacement) were another feasible
solution, but also proved to be costly.
One of the ground improvement specialist contractors proposed the application of dynamic
compaction.
Using Equation 2‐41 with c= 0.7 and a drop height of 20 m, it can be estimated that treating
a loose sand layer that is about 13 m thick would require a pounder that weighs 18 tons;
however, at that time, and the heaviest pounder that was available to the specialist
contractor weighed 15 tons.
As a supplement to dynamic compaction it was decided to implement dynamic surcharging
(also refer to Section 3.4.4) to improve the results by combining the effect of static loading
and high energy impacts to generate acceleration in the soil under static loading in such a
way as to produce a shearing process around the surcharge fill. This process was to reduce
the spreading of the load that was initially caused by the high strength of the upper layers.
Generation of vibrations and increasing the pore pressure under the tank was to result in a
reduction in friction between the granular particles of the soil and to ultimately lead to the
collapse of the foundation soil under the influence of dynamic surcharging.
To realise this process, a surcharge was to be initially placed onto the treatment area and
dynamic compaction was to be performed. Although granular materials settle under static
loads, as dynamic shear modulus has been found to decrease significantly with increasing
values of shear strain amplitude (Silver and Seed, 1971), it can be expected that introducing
vibration will increase the amount of settlement under the surcharge. Furthermore, the rate
of consolidation of fine soils is greatest when pore water pressure is high, and as was
observed in King Abdulla University of Technology (refer to Section 3.4) it is possible to
increase the rate of consolidation back to previously high values by inducing pore water
pressure through vibration.
464
3.12.4 Supplementary Geotechnical Investigation in Lot A‐A
In order to verify the ground conditions before finalising the treatment scheme, four SPT
boreholes were drilled, and tested in the centre and three sides of Lot A‐A’s tank. These
boreholes also indicated that the upper 3 m of sand was very dense, but the soil then became
very loose to medium dense at groundwater level. SPT blow counts at depths of 3 to 8 m
varied from as low as 4 to as high as 14, recorded N values were then from 11 to 20 to depths
of approximately 12 to 13 m where the ground became very dense, and blow counts
exceeded 50. Fines content of the 38 samples that were extracted from the four boreholes
ranged from 16% to 21%. This was much more than the average 5% that was indicated by
the preliminary geotechnical investigation. Also, although no silt pockets were identified
under the tank, as fines content was observed to be more than 20% in almost half of the
samples and as high as 30%, it was understood that the tank’s location was probably one of
the siltiest areas of the reclamation. SPT blow counts and fines content in the preliminary
and supplementary boreholes are shown in Table 3‐17.
Furthermore, two PMTs were also carried out in the tank area. As it was already established
that the ground was very dense above water table level, testing was done at 1 m intervals
below sea level. These tests also reconfirmed that the submerged soil was in a loose state. In
this zone PLM was less than 100 kPa to about 700 kPa, and EM was measured to be from less
than 1 MPa to 6 MPa.
As shown in Figure 3‐194 and prior to the placement of any surcharge material four
settlement monitoring plates were installed at the tank location. One plate was installed in
the centre of the tank, and the remaining three were installed at 120 degrees angles from
one another on a ring with a radius of 17.5 m.
The surcharge was provided by scraping sand from the vicinity of the tank and placing it
evenly over the loading area with a base diameter of 44.2 m. The fill height was 4 m with a
lateral slope of 1V:1.5H; thus the top diameter of surcharge was 32.2 m. The surcharge is
schematically shown in Figure 3‐195.
465
Depth Supplementary SPT Preliminary SPT
BH‐1 BH‐2 BH‐3 BH‐4 BH‐13 BH‐14
SPT fines SPT fines SPT fines SPT fines SPT fines SPT fines
0.0 28
0.5 22
1.0 18 17 4%
1.5 8 8
2.0 15 17
2.5 13 21
3.0 14 29
3.5 19 5
4.0 4 11% 13 14% 13 12% 9 5% 7
4.5 6 9% 7 11
5.0 7 15% 7 11 20% 15 12
5.5 10 30%
6.0 7 15% 15 17% 14 16% 8 6% 12
6.5 13 30%
7.0 12 10% 22 30% 14 20% 18 11 6%
7.5 9 30%
8.0 5 22% 14 30% 10 15% 38 10% 16
8.5 5 27%
9.0 14 21% 12 21 22% 55 36
9.5 18 17%
10.0 11 6% 13 17% 13 22% 26 41
10.5 11 17%
11.0 24 14% 17 17% 23 22% 18 20 2%
11.5 11 17%
12.0 37 29% 20 23% 64 20 3% 200
12.5 11 22%
13.0 36 17% 24 33 22% 133 26 5%
13.5 38 15%
14.0 38 14% 53 52 20% 25 47
14.5 22%
15.0 200
Table 3‐17: SPT blow counts and fines content in the preliminary and supplementary boreholes
466
Figure 3‐194: Layout of settlement monitoring plates
Figure 3‐195: Schematic illustration of the surcharge
It can be estimated that the approximate volume of the surcharge was 4,700m3. Assuming
that the in‐situ unit weight of the surcharge was 17 kN/m3, the total weight of the surcharge
material can be estimated to be 80 MN. Noting that ground and tank levels were respectively
at +4 m RL and +2.5 m RL, that overburden will be the equivalent of an additional surcharge
load of approximately 25 MN. Although the uniformly distributed load acting specifically on
the tank’s area was 93.5 kPa (4 m of surcharge plus 1.5 m of over burden pressure), which is
equivalent to 78% of the tank’s total load, the 105 MN of load applied over the tank’s area
and its periphery equates to approximately 90% of the tank’s total load. Hence, it would be
467
fair to assume that this amount of surcharge would have been able to well consolidate the
deep layers of silty sand satisfactorily.
As graphically shown in Figure 3‐196 and numerically presented in Table 3‐18, the monitoring
plates’ settlements were recorded during the stages of the surcharge. Ignoring the errors in
recording, it can be seen that the plates settled prorate with surcharge placement. Once
backfilling was completed, the surcharge was left in place for 5 additional days. It can be
predicted from the settlement trend that the ground could have settled an additional
maximum settlement of 5 mm in the long run.
Figure 3‐196: Ground settlement in Tank A‐A during static and dynamic surcharging
Before commencement of dynamic surcharging, as shown in Figure 3‐197, 26 DC print
locations were pre‐excavated by approximately 1 m at 5.95 m intervals on a ring around the
468
tank with a diameter of 49.2 m. The excavation was carried out to increase the pounder’s
depth of influence by lowering the impact level and also reducing the amount of energy
absorption in the very dense superficial layer. Then, as shown in Figure 3‐198, a 15 ton
pounder was dropped a total of 30 times per print in 6 consecutive cycles.
Figure 3‐197: Pre‐excavation of dynamic surcharging prints
Figure 3‐198: Dynamic surcharging
It can be assessed from Figure 3‐196 and Table 3‐18 that dynamic surcharging has increased
ground settlements by 1.6 to 5.2 times the value of the static settlements. As could have
been anticipated, the maximum effect and increase was on the periphery where impact wave
amplitudes were greatest. It can also be observed that by the sixth cycle, the efficiency of
469
dynamic surcharging seems to be diminishing and additional cycles would have yielded a non‐
linearly lower rate of induced settlement.
Although the maximum differential settlement of outer monitoring plates were 7 mm at the
end of static surcharging, the maximum differential settlements during dynamic surcharging
increased by 4 times to 40 mm. This suggests the possibility of large differential settlements
under seismic and vibratory loads in untreated areas of the same ground.
After completion of dynamic surcharging the surcharge was removed, and the ground was
excavated to working platform level at +2.8 m RL. This elevation was defined with the
intention of allowing for 0.3 m of induced settlement due to dynamic compaction,
approximately reaching the tank’s base level of +2.5 m RL at the end of the ground
improvement works, and sufficiently and safely remaining above groundwater level without
destabilising the platform or delaying the works due to ground liquefaction and water
seepage to the surface.
The diameter of the working platform level after excavation and top of excavation were
respectively 41.2 m and 46 m, and the excavation cross section is shown in Figure 3‐199.
Figure 3‐199: Cross section of excavation for reaching working platform level
Dynamic compaction was carried out on prints located in the centre of the working platform
and on 4 concentric rings around the central print. The description of the print rings is
summarised in Table 3‐19. As with the dynamic surcharging prints, each DC print was pre‐
excavated by approximately 1 m to facilitate pounder penetration and to increase the depth
of influence.
Approximately 150 m3 of crushed rock and cobbles were added to the total of 58 DC prints.
This amount equates to about 2.6 m3 of rock per print or an equivalent of approximately 0.13
m of rock per every metre of ground within the treatment zone. This amount of stone is
470
insufficient to effectively increase the ground strength, and was rather used to increase soil
permeability.
Due to the small size of the project and consequently the limited number of prints, in the first
phase of compaction each print received between 28 to 35 blows from a drop height of 20
m. In the second phase of compaction, each print received an additional 3 to 5 blows. During
ironing stage the pounder was dropped with low energy once onto the points of a of 2x2 m2
grid. A photograph of the works is shown in Figure 3‐200(a). As shown in Figure 3‐200(b), due
to soil liquefaction and an increase in pore water pressure during dynamic compaction, water
was observed to boil out from the craters.
(a) (b)
Figure 3‐200: (a) Dynamic compaction of the tank’s foundation, (b) water boiling into the craters
As shown in Figure 3‐201, pounder penetration and the outer crater diameter for each print
location were measured during the first two phases of dynamic compaction. In the first
phase, the average pounder penetration depth and average upper diameter of crater were
respectively 1.7 m and 5 m. In the second phase, these numbers significantly reduced and
respectively dropped down to 0.4 m and 2.3 m. The diameter of the crater at the base could
be assumed to be equal to the pounder’s dimension; i.e., 1.7 m.
471
(a)
(b)
Figure 3‐201: Crater top diameter and depth in (a) the first and (b) the second phase of DC
At the end of dynamic compaction, the ground level dropped to +2.25 m RL. Noting that 150
m3 of stone that equated to a thickness of 0.13 m had been added, it can be calculated that
472
the ground had settled 0.68 m in addition to the settlements induced by dynamic
surcharging.
Noting that the treatment diameter was 41.2 m, it can be calculated that the compaction
volume was 906 m3. Assuming √ 2.26 (refer to Equation 3‐11) for each of the two
phases of the 58 prints will yield the same volume. Assuming the ratio to be equal to 2.15
and 2.75 will respectively result in total compaction volumes of 775 and 1268 m3, which
respectively underestimate the compaction volume by 15% and overestimate the
compaction volume by 40%. √ 2.5 will result in 16% overestimation in this case.
Comparison of the above figures suggests that Equation 3‐11 has sufficient accuracy for a
first estimate or at least as a rule of the thumb estimation of the compaction volume, and
thus the average induced strain in the soil.
Although the magnitude of this settlement is much larger than what was measured during
dynamic surcharging, it should be reminded that dynamic surcharging was aimed at reducing
the settlement of the deeper and siltier layers that may have been too deep to effectively
reach with dynamic compaction using a 15 ton pounder or too silty for dynamic compaction
to be efficiently effective.
Upon completion of dynamic compaction and levelling of the site, 4 PMTs were carried out
within the treatment area to assess the ground conditions. PLM before and after ground
improvement are shown in Figure 3‐202. As could have been predicted, maximum
improvement occurred to depths of about 8 where, excluding the highest values of one of
the tests(A2), PLM ranged from approximately 2 to 4 MPa that is quite more than 1.9 to 2.4
MPa that Lukas (1986) has expected. EM ranged from 23 to 30 MPa at depths of about 4 m
to 8 m. Due to the combination of dynamic surcharging and pre‐excavated dynamic
compaction with high number of blows improvement can still be observed at greater depths.
Furthermore, the minimum PLM value after improvement is greater than 600 kPa, which
demonstrated that the young hydraulic fill was no longer subject to creep due to self‐weight
(see Section 2.9.2.8).
473
Figure 3‐202: Pre and post ground improvement PMT PLM and EM values
Figure 3‐203 shows the average PLM and EM improvement ratios. It can be seen that while the
lower bound of improvement ratio at depths of 4 to 8 m is about 4 to 5, which is compatible
with the upper bound values suggested by Lukas (1986), upper bound values are quite higher
and in the range of 10 to 18.
Bearing capacity can be calculated by using Equation 2‐168. The geometric mean of the
average of the 4 post ground improvement PMTs is 1,684 kPa. Conservatively assuming that
the deeper layers also have the same geometric value and that the foundation is on ground
level, with k=0.8 and safety factor= 3, the allowable bearing capacity can be calculated to be
449 kPa, which is more than the required 160 kPa.
Although it could have been possible to calculate the tank settlements using hand
calculations, it was decided to study the total and differential settlements by using finite
element analysis. Therefore, a three dimensional model consisting of 41,920 elements and
45,979 nodes was realised (refer to Figure 3‐204(a)). Loads consisted of the weight of tank
and concrete raft that were assumed to be about 1800 kN and a uniform pressure of 63.8
kN/m² on the raft that represented 6.5 m of water height in the tank.
474
Figure 3‐203: Average PLM and EM improvement ratios
The model was developed in half space due to symmetry. In this model the tank was placed
on a 0.5 m thick concrete raft and two layers of sand of which the upper layer had higher
values for its geotechnical parameters. Also, in order to assess differential settlements, the
ground in the model was made more stiff on one side of the tank by assuming that the
thickness of the upper and lower sand layers in the left and right sides of the tank were
respectively 7.5 m and 3.5 m. On the right side, the upper and lower sand layers were
respectively assumed to be 6.5 m and 6.5 m. An enlarged image of the model is shown in
Figure 3‐204(b).
(a) (b)
Figure 3‐204: (a) Finite element model, (b) zoom of model details
475
The principal characteristics of the material (see Figure 3‐204(a)) and dimensions for the
model are shown in Table 3‐20.
Material ID 1 2 3
Ey (MPa) 11 000 59.7 22.5
Thickness (m) 0.50 6.5 – 7.5 3.5 ‐ 6.5
Table 3‐20: The principal characteristics and dimensions for the model
The result of the finite element analysis is shown in Figure 3‐205 and Figure 3‐206. As can be
observed the maximum settlement at the tank’s centre was calculated to be 21.35 mm.
Minimum tank settlement at the shell was 10.91 mm. Thus, differential settlement over the
radial length of 17.55 m was 10.44 mm or less than 1/1,681, which was much smaller than
the allowed value of 1/750. Differential settlement from one side to the other side of the
tank can be calculated to be 3.13 mm or less than 1/11,200.
Figure 3‐205: Settlement contours
476
Figure 3‐206: Settlement variation under Tank A‐A
Based on the general understanding of the project and the experience gained in Lot A‐A,
similar works were carried out in Lot G‐G. Here, only a summary of Tank G‐G will be presented
as the results are very similar to what has been already reported in Lot A‐A.
Geotechnical tests comprising of 5 SPTs and 3 PMTs were carried out before ground
improvement. SPTs were carried out from the depth of 4 m. The soil was identified to be
loose to medium dense sand with blow counts in the range of 6 to 16 down to depths of
about 9 to 12 m. The sand then became denser with blow counts falling within the range of
about 16 to 40. Fines content ranged from 5 to 25% and with an average of 13%, indicating
that the soil in this Lot contained lesser fines than Lot A‐A.
PLM from the depth of 4 to 13 m ranged from as low as 100 kPa to 1,500 kPa. PLM and EM are
shown in Figure 3‐208.
Once the ground conditions were confirmed, the surcharge was placed in the same manner
as has already been explained for Lot A‐A. However, in this Lot, the surcharge was placed
over a longer period due to logistics difficulties, and only 5 cycles of dynamic surcharging was
performed. The static and dynamic settlements are shown in Figure 3‐207. It can be observed
that the magnitudes of both static and dynamic settlements are larger in Tank G‐G. The ratios
between last measured static settlements and final settlements were from 1.3 to 3 that are
less than both the maximum and minimum ratios of Tank A‐A.
477
After removal of surcharge and excavation to platform level dynamic compaction was carried
out. Settlement induced by dynamic compaction in the foundation of Tank G‐G was 0.64 m,
which is very similar to the 0.68 m figure of Tank A‐A. Post ground improvement PMT results
are shown in Figure 3‐208.
Figure 3‐207: Ground settlement in Tank G‐G during static and dynamic surcharging
Figure 3‐208: Pre and post ground improvement PMT PLM and EM values
478
Similar to Tank A‐A, the ground in this Lot also had the most amount of improvement down
to depths of about 8 to 9 m; however, improvement is observable and rather uniform
throughout the remainder of the testing depth. The deeper layers may have improved
uniformly due to the effects of dynamic surcharging. As can be seen, PLM and EM values have
improved rather similarl to Lot A‐A whereas they are respectively from approximately 2 to 5
MPa and from 15 to 50 MPa at depths of about 3.5 to 7.5 m.
The geometric mean of the average PLM of the 3 post ground improvement PMT is 2,199 kPa,
which is 30% higher than the geometric mean of Tank A‐A, and bearing capacity criterion can
be deemed to have been satisfied without further calculation.
Finite element for the calculation of Tank G‐G was also done using the same principles as
Tank A‐A; however, in this tank the upper and lower sand layers were assumed to be
respectively 5 m and 6 m with equal thickness throughout the treatment zone. Here, Ey for
the upper and lower sand layers were assumed to be respectively 69.6 MPa and 24 MPa. As
shown in Figure 3‐209, maximum settlement in the centre of the tank was 18.39 mm. Shell
settlement was calculated to be 11.61 mm; thus differential settlement from the centre to
the shell of the tank can be calculated to be 1/5,177 that is much less than 1/750.
Figure 3‐209: Settlement profile under Tank G‐G
This project has been able to demonstrate the effectiveness of combining two techniques;
i.e., dynamic surcharging and dynamic compaction, to improve the ground and achieve
results that would have been otherwise infeasible with the available equipment.
479
Dynamic surcharging was able to induce additional settlement compared to what was
realised under static loading conditions. This has not only shown the value of dynamic
surcharging for increasing induced settlements and reducing soil porosity, but is also a
reminder that even if settlements are acceptable under static loading conditions, vibration
of the ground due to earthquakes or any other sources can generate more settlement.
Some lessons that can be concluded from this project include:
1. Dynamic surcharging can be used to increase induced foundation settlement under
static surcharge by 1.3 to 5 times, depending on the distance from the pounder
impact point, to treat siltier material that would normally not be treatable by
dynamic compaction, and to increase the depth of treatment.
3. Equation 3‐11 was able to estimate the compaction volume in this project the
equation equaled 2.26; i.e., when √ 2.26. The upper and lower bound values
of 2.15 and 2.75 in Equation 3‐11 resulted in 15% underestimation and 40%
overestimation. √ 2.5 will result in 16% overestimation in this case.
4. It may be possible to make a first estimation or at least a rule of thumb estimation of
the induced strain in the treated sand using crater depth.
5. Excluding the highest values, average PLM and EM respectively ranged from
approximately 20 to 40 MPa and from 23 to 30 MPa at depths of about 4 m to 8 m.
These values are significantly higher than what has been suggested by Lukas (1986).
6. Due to the combination of dynamic surcharging and pre‐excavated dynamic
compaction with high number of blows improvement can still be observed at greater
depths.
7. Maximum improvement ratios were in the range of 10 to 18 that are significantly
higher than the range that has been suggested by Lukas (1986).
Figure 3‐210 shows Tank A‐A after construction.
480
Figure 3‐210: Tank A‐A after construction
481
3.13 Al Khaleej Raw Sugar Silos
3.13.1 Introduction
Al Khaleej Sugar Factory is located on the southern coast of the Persian Gulf in Jebel Ali Free
Zone, Dubai, UAE (see Figure 3‐211). As part of the company’s development programme, it
was decided to construct two identical 72 m high reinforced concrete sugar silos, each 34 m
in diameter and 3.5 m apart from one another.
Figure 3‐211: Satellite image of location of Al Khaleej Sugar Factory
The geotechnical investigation report indicated that ground level on site was variable from
+4.2 to +4.8 m RL (reduced level). Mean High High Water (MHHW) and Mean Low Low Water
(MLLW) were respectively at +1.7 m and +0.5 m RL, indicating a relatively high groundwater
level at all times.
As part of the soil investigation, six boreholes were drilled and SPTs were carried out in them.
These boreholes indicated that the first layer of soil, extending down to approximately +2 m
RL, was loose to medium dense silty fine sand fill. Occasional high fines content in this layer
suggested the presence of bands or pockets of silts and clay. The top soil layer was followed
by 4 m of medium dense to dense slightly silty sand that terminated at ‐2 m RL. Fines content
in this layer was less than 20%.
The subsequent layer was 6 m of weak sandstone or cemented sand extending down to
approximately ‐8 m RL. Then, a very weak to weak layer of sandstone that was 4 m thick and
extended to ‐12 m RL was encountered. The next layer that was only 1 m thick and
482
terminated at ‐13 m RL was very weak to very stiff calcisilite. Then, very weak calcisiltite or
conglomerate, 4 m thick, and extending down to ‐17 m RL was identified. Finally, weak to
moderately weak calcisiltite or conglomerate was observed down to the end of boring at ‐35
m RL. Ground layers and their assumed design properties are summarised in Table 3‐21.
Calculations showed that due to the presence of the superficial loose layers of soil, it was not
possible to construct the silos without the implementation of specific geotechnical solutions.
Preliminary studies suggested that safe foundations for the silos could be provided by using
piles and concrete rafts. Depending on the design pile diameter, pile lengths were calculated
to be from 29 to 37 m.
Although piles and raft footings were technically acceptable solutions, the cost of installing
long piles in rock was estimated to be very high, and the project engineers began considering
the option of ground improvement. Thus, several specialist contractors were approached,
and subsequently a design and construct contract was awarded to the company that had
proposed the implementation of dynamic replacement at approximately one fifth of the price
of piling.
483
As shown in Figure 3‐212, in order to optimise the ground improvement solution, the
geometry of the project was reviewed, and silo dimensions and distances were revised. In
the new design silo heights were reduced to 55 m, diameters were increased to 40 m, and
the distance between the silos was increased to 25 m to reduce the side loading effect of one
silo on the other.
Figure 3‐212: Schematic cross section of the sugar silos
1 2 3 4 5 6
Load Compacted Rock Weaker Rock Substratum Replacement
Description
transfer Sand
platform
Thickness (m) 1.50 4.00 6.00 9.00 13.00 2.00
First loading Ey 80 60 540 132 360 100
(MPa)
Reloading Ey 104 78 700 264 720 130
(MPa)
Cohesion (kPa) 0 0 0 0 0 0
Friction angle (°) 35 38 40 40 40 40
Unit weight 18 18 18 18 18 18
(kN/m3)
Table 3‐22: Design parameters for numerical analyses
484
The concept of applying dynamic replacement was realised based on excavating the
unsuitable man‐made top soil layer down to working platform level at +2.5 m RL, carrying
out ground treatment and improving the properties of the sandy layer down to ‐2 m RL and
finally constructing an engineered fill from finished soil improvement level to bottom of raft
level. The ground improvement concept was based on application of dynamic compaction
for compacting the sandy layer and reinforcing it with crushed stone inclusions that were
backfilled into excavated pits within the tank area, and installing a crushed stone shear ring
trench for supporting the shell. The depth of the ring was assumed to be 2 m. Top and
bottom widths of the trench were designed to be respectively 8 and 6 m. Design parameters
that were used in the numerical analyses of Figure 3‐213 using finite element software are
presented in Table 3‐22.
Figure 3‐213: Model used for the numerical analysis of the sugar silo
In the numerical analyses the thickness and unit weight of the raft were respectively assumed
to be 1.5 m and 25 kN/m3. The 0.45 m thick shell was modelled by a 1,222 kPa load spread
over the shell thickness, and an additional 25 kPa of uniform load was used to model the
tunnel and slab below the raft. Sugar heights in the centre and at the shell were assumed to
be respectively 52.7 m and 40.2 m, and unit weight of sugar was taken as 9 kN/m3. The
Sugar’s dynamic load during the unloading phase was taken as 2,444 kPa spread over the
wall and with a reduction factor of 2 with consideration of the dynamic modulus of soil being
higher than the static modulus.
485
In the calculations 3 load cases were analysed; i.e., after construction (empty silo), after full
loading and upon unloading (inclusive of sugar dynamic loading). Vertical displacements for
these load cases are shown in Figure 3‐214 to Figure 3‐216.
Figure 3‐214: Total settlement after construction
Vertical displacements along the ground surface of the model are plotted in Figure 3‐217. As
can be seen, calculations showed that maximum settlement of approximately 71 mm was
realised in the centre of the silo when it was fully loaded. While it was calculated that
maximum settlement would increase negligibly to 71.5 mm in the centre at the beginning of
sugar unloading, it can be observed that settlement at the shell increases by about 70%
during that phase. It can also be seen that the effect of one silo loading on the other is
negligible at 25 m.
The finite element analysis also showed that approximately 70% of the ground settlements
were due to the compression of the engineered fill and the improved ground below it.
Figure 3‐218 to Figure 3‐220 show horizontal stresses (tensile stresses are positive) in the
concrete raft. The values in the boxes are the maximum tensile stresses. It can be seen that
the maximum tensile stresses reach 1.8 MPa, and occur on the top fibre of the concrete raft.
486
Figure 3‐215: Total settlement after first loading
Figure 3‐216: Total settlement after beginning of first unloading
487
Figure 3‐217: Vertical displacement of the ground due to different load cases
Figure 3‐218: Maximum horizontal stresses after construction
Figure 3‐219: Maximum horizontal stresses after first loading
488
Figure 3‐220: Maximum horizontal stresses after beginning of first unloading
A similar set of calculations were also undertaken for the reloading of the silo using the reload
moduli that have been presented in Table 3‐22. Figure 3‐221 shows that the maximum
displacement is 46 mm during reloading. Similarly, the maximum tensile stresses after
unloading with reload moduli will reach 2.3 MPa on the top fibre of the concrete raft due to
the dynamic effect of unloading.
0
0 10 20 30 40 50
Vertical displacement (cm)
-1
-2 Loading
Beginning of unloading load
Complete unloading
-3
-4
-5
Distance from centre (m)
Figure 3‐221: Vertical displacement of the ground due to different reloading cases
The curves of Figure 3‐222 allow the assessment of the behaviour of the soil under the cyclic
loadings. The last curve corresponds to the vertical displacements of complete unloading
with reload moduli minus beginning of unloading with reload moduli plus beginning of first
unloading with initial moduli. These curves show that maximum differential settlement
489
occurs during the first loading and is equal to approximately 1/500. During complete
unloading, differential settlement does not exceed 0.94/1000.
0
0 10 20 30 40 50
-1
Vertical displacement (cm)
-2
-3 Construction (Einit)
First loading (Einit)
Beginning of first unloading (Einit)
-4 Complete unloading (Ereload)
-5
-6
-7
-8
Distance from centre (m)
Figure 3‐222: Vertical displacement of the ground due to different loading and reloading cases
Taking into consideration the ground deformation shapes and the bending moments that
would be realised in the reinforced concrete raft, the application of dynamic replacement
was approved with target values of Ey being 60 MPa and 100 MPa respectively under the silo
and the shell.
In practice, the location of the DR columns in each silo was excavated and backfilled with
crushed stone with a replacement ratio of 13% using a grid that was approximately 4.9x4.9
m2. On average, each print was subjected to 12 blows. Dynamic compaction of the shear ring
is shown in Figure 3‐223.
The presence of an already existing concrete silo at a minimum distance of 10 m from the
dynamic replacement works (see Figure 3‐224) presented a challenge as it was expected that
at such distances peak particle velocity would exceed what is recommended by USBM
(Siskind et al., 1980). When distances were at a minimum a constant dilapidation survey was
performed during dynamic compaction, to observe the behaviour of the existing silo, and if
necessary, to devise an alternative work procedure. However, the survey demonstrated that
no specific procedures had to be introduced.
490
Figure 3‐223: Dynamic compaction of the shear ring
Figure 3‐224: Dynamic replacement next to an existing concrete silo
During the works it was observed that, as shown inFigure 3‐225 , a very dense layer that was
0.5 to 1 m thick had capped a portion of the loose sand in the second silo. This layer damped
the energy transmitted to the loose layer beneath. Hence, in this zone the pounder was
dropped in total 25 to 30 times to break through the very dense layer and to achieve the
desired compaction.
491
Figure 3‐225: Presence of very hard damping layer
After completion of soil improvement works, at the end of dynamic replacement, the
project’s client and consultant decided to reduce foundation level to +2.5 m RL to eliminate
the cost of constructing the engineered fill; thus calculated settlements were further reduced
due to the elimination of this layer.
PMTs were performed inside the DR columns, in between the columns and in the shear ring
trench before and after the ironing phase of dynamic replacement. The summary of test
locations for the first silo has been tabulated in Table 3‐23, and EM values are shown in
Figure 3‐226. All tests before ironing were intended to be carried out to the depth of 4 m
below working platform level; however, some tests reached refusal before that depth.
It can be understood from the PMTs that ironing has been able to increase EM values at the
depth of 1 m by 2.6 to 3.5 times. However, calculations show that it would have been
possible to meet acceptance criteria even without application of the ironing phase.
The harmonic mean values of Ey for each test location can be calculated from EM values using
Equations 2‐212 and 2‐214 with rheological factors of 1/3 for sand and 1/4 for gravel to
demonstrate that acceptance EM values have been achieved.
492
PMT Location Time
1 in DR column before ironing
1B in DR column before ironing
2 in between DR columns before ironing
3 in between DR columns before ironing
4 in shear trench, in DR print before ironing
4B in shear trench, in DR print after ironing
5 in between DR columns before ironing
6 in shear trench, in DR print before ironing
6B in shear trench, in DR print after ironing
7 in shear trench, in between DR prints before ironing
8 in shear trench, in between DR prints before ironing
9 in DR column after ironing
Table 3‐23: PMT Schedule
Figure 3‐226: PMT modulus before and after ironing in the first silo
493
3.13.5 Lessons and Conclusion
In this project two sugar silos, each 55 m high, have been constructed on improved soil. This
project faced numerous challenges including the very heavy loading of the silos, the
insufficient distance between the silos, the presence of an existing nearby silo, a high
groundwater level, and several metres of loose soil. Lessons learned in this project include:
1. By reviewing and redesigning the silos’ layout, adopting appropriate design
procedures and acceptance criteria, and implementing state‐of‐the‐art ground
improvement techniques it may be possible to construct very heavy structures on
improved ground.
2. The mass mechanical properties of sand layers can be further improved by
introducing dynamically compacted crushed stone inclusions.
3. Application of ironing phase of DC will improve the superficial ground parameters.
Figure 3‐227 shows Al Khaleej raw sugar silos after completion of construction.
Figure 3‐227: Al Khaleej raw sugar silos after completion of construction
494
3.14 Trail for Quay Expansion in Southeast Asia
3.14.1 Introduction
Recently, dynamic replacement was carried out in Southeast Asia to treat soft marine
deposits more than 30 m below seawater level for the construction of a container terminal
using caisson seawalls. According the original design the soft marine clay at the seabed was
to be dredged to the depth of 30 m below sea level where the shear strength of the stiff clay
exceeded 250 kPa. The excavated key was to be then backfilled with sand and compacted
using vibro compaction under 3 m of additional overburden sand fill. Next, the surcharge had
to be removed, a rubble mound was to be placed over the sand key, and as shown in
Figure 3‐228, caissons were to be then sunk onto the mound.
Figure 3‐228: Cross section of container terminal based on original foundation concept
As SPT blow counts exceeded 50 and the assumed clay shear strength of 250 kPa was
achieved at dredge level, works progressed by backfilling sand and compacting the fill using
vibro compaction.
While the clay at dredge level was initially very stiff, dredging works and cutting into the clay
softened the upper 1 to 1.5 m of the exposed clay surface and post dredging CPT tests
performed before the removal of the overburden sand fill indicated that the clay’s shear
strength had dropped to about one third of its original value; i.e., to approximately 80 kPa.
PMTs carried out at later stages suggested that the shear strength had even further reduced
at some points to a mere 16 kPa.
495
3.14.3 The Solution: Offshore Dynamic Replacement
Further dredging of the softened clay and replacing it with more sand fill did not appear to
be an effective method because it was expected that this would lead to the disturbance of
deeper clay layers and the persistence of the problem.
Due to the nature of the soft soil and its thickness, marine dynamic replacement was
envisaged as a possible treatment solution. Based on previous experiences (refer to
Section 2.10.4), it was anticipated that if proper equipment; i.e., a large stable barge, a
specialised crane with a sufficiently powerful winch system for lifting a heavy pounder and
resisting tidal action, and a special pounder for transmitting sufficient impact energy to the
seabed were available, it would then be possible to drive granular material into the soft clay
and improve its properties.
Unlike land based dynamic replacement where suitable material can be pushed into the
crater by a loader, this method is not possible in offshore dynamic replacement, and material
can only be punched in from the load transfer platform. Hence, a stone blanket was used to
feed the DR columns and to create the load transfer platform. This layer also prevented the
contamination of seawater by the flow and dispersion of suspended clay particles produced
by the pounder’s impacts.
In the proposed dynamic replacement methodology it was assumed that a 1.8 m thick granite
rock fill layer would be placed over the soft clay layer. The blanket material was chosen in
such a way that 30% of the stone diameters were from 150 to 200 mm and the remaining
70% were from 200 to 300 m. The DR rock columns were designed to be 2 m in diameter, in
a 4.5 m grid and with a replacement ratio of 15%.
As shown in Figure 3‐229, in this project a specially designed grater shaped marine pounder
weighing 38.5 tons was used to drive the rock into the ground to form the DR columns, and
to dynamically compact the rock blanket. The pounder’s dimensions were 1.7 m by 1.7 m on
the DR side and 2.3 m by 2.3 m on the DC side. Figure 3‐230 and Figure 3‐231 show the
pounder being used respectively for dynamic replacement and dynamic compaction works
from a barge that was 15x50 m2 in platform area.
Previous experiences by the working team suggested that water resistance could greatly
reduce the effect of significantly high drops. Hence, the drop heights during the trial were
set to 5 m above seabed level. Records of the crane’s winch speed during the works indicate
496
that the maximum drop speeds were in the order of 430 m per minute. This speed is
equivalent to a free fall with a drop height of 2.6 m (in air), and verifies the original
assumption that a great portion of the drops’ kinematic energies would have been lost due
to water resistance.
Figure 3‐229: Marine DR (bottom side) ‐ DC (top side) pounder
Figure 3‐230: Marine pounder being utilised for dynamic replacement
Each dynamic replacement print location was subjected to 30 blows. Furthermore, 3 to 6
blows were applied as ironing using the larger end of the pounder. As shown in Figure 3‐232
the penetration of the pounder into the ground was measured for every blow. It can be
497
observed that while the pounder penetrated the ground at a more pronounced rate during
the first four blows, the penetration rate then rapidly decreased to the point where it appears
that no penetration was practically observed after the 15th blow. The amount of pounder
penetration was variable from 1.1 to 1.7 m. Comparing these figures with the thickness of
the soft soil prior to dynamic replacement, it can be interpreted that the pounder impact was
able to effectively drive the granular material of the blanket to the end of the soft soil layer
with the first 4 to 12 blows and then to further compact the granular rock fill. It can also be
observed that the maximum penetration values per print were sometimes more than the
assumed soft soil layer’s thickness. This indicates that either the DR columns have penetrated
into the stiffer clay or that the actual soft layer’s thickness was more than originally
anticipated at some locations.
Figure 3‐231: Marine pounder used for dynamic compaction
The total ground settlement was measured by echo sounding and the survey showed that
the top of the blanket dropped by 0.38 m as a result of the ground improvement works.
Divers were used to visually inspect the impact results at seabed level. Based on the larger
amounts of crushed rock at the DR column location, it was determined that the columns were
2.4 m in diameter, which equates to the diagonal length of the pounder’s base on the DR
side. It can be interpreted that the larger DR columns’ diameter as compared to the
pounder’s base may have been formed by a combination of soft soil being pushed away
498
laterally due to the high horizontal stresses exceeding the soil’s strength at impact location
and possible rotations of the pounder during the impacts (although the latter is less likely
due to the implementation of synchronised double winches in the drop mechanism). Thus,
the actual DR replacement ratio was 22.3% in lieu of the originally assumed 15%. Target rock
friction angle was 45 degrees.
Figure 3‐232: Pounder penetration at several DR print locations
Due to the large water depths and open sea working conditions 63 mm slotted casing PMTs
were carried out using 100 mm guide tubes followed by the 60 mm PMT tube. During testing
visual observation on the return of drilling fluid was recorded. No return of drilling fluid
indicated that the test was carried out in the free‐draining rock material whereas testing in
impervious clay was identified by the return of the drilling fluid.
Figure 3‐233 shows the trial zone and PMT locations. Two PMTs (Pre‐2 and Pre‐8) were
carried out prior to dynamic replacement and six (Post‐2, 2a, 2b, 2c, 8 and 9) were carried
out after treatment. EM and PLM before and after treatment values are tabulated in
Table 3‐24. It can be observed that Post‐8 registered a non‐yielding curve with a high value
of PLM, probably due to a localised closer matrix of rock pieces in the vicinity of the slotted
casing, and as such, was deemed as non‐representative and excluded.
The comparison of PMT Pre‐2 and Post‐2a that were done in the almost same location
indicates that while the rock fill has been driven into the soft clay, its EM and PLM values have
also increased respectively by 118% and 132%. The average values of EM and PLM after
improvement were respectively 8.05 MPa and 1.40 MPa. The maximum PLM that was
499
recorded during the test exceeded 2.2 MPa. It can also be calculated that the harmonic mean
of EM in the rock fill after improvement is equal to 6.03 MPa.
Figure 3‐233: Trial zone and PMT locations
The Young modulus of the clay and rock fill can also be calculated using Equation 2‐214 with
α equal to 1/4 for rock fill and 1/2 for altered clay.
The shear strength parameters can also be estimated from the pressuremeter test using
Equation 2‐222; however, Equation 2‐224 or 2‐225 have been proposed for estimation of
500
drained friction angle of sands, and it is the experience of the author that their application
will underestimate the friction angle in rock fill. As shown in Figure 3‐234, during the project
Yee and Varaksin proposed to extend Menard’s line (see Figure 2‐168) to ϕ = 38o and 40o
(presumably the indicative ϕ angle for uncompacted crushed rock), suggested that the
relationship lies within this range, and proposed2 Equation 2‐224:
∗
4 2 3‐12
Based on these values presented, a finite element model can be constructed with the
parameters of Table 3‐25.
The offshore dynamic replacement trial has demonstrated that it is possible to perform this
technique and to verify the results by implementing the pressuremeter test at depths of 30
m. Test results can be used for constructing suitable models for required analyses.
Specific lessons and conclusions for this project can be summarised as:
1. The barge must be large enough to safely support the personnel and equipment. The
barge size will be influenced by the location of the project and the sea conditions.
2. Water resistance reduces the pounder’s dropping velocity, and consequently the
impact energy.
2
The author has previously published a paper (Hamidi, B., Yee, K., Varaksin, S., Nikraz, H. & Wong, L.
T. (2010) Ground Improvement in Deep Waters Using Dynamic Replacement. 20th International
Offshore and Polar Engineering Conference, Beijing, 20‐26 June, 848‐853) in which Equation
Equation 2‐224 has been mistyped, and appears in the corrected form in this thesis.
501
Figure 3‐234: Proposed method for estimation of rock fill friction angle from PMT
3. Offshore pounders must be designed to minimize water resistance. As water
resistance reduces impact energy, marine pounders are generally heavier than land
pounders.
4. The DC rig should be large enough to provide the required stability and winch
capacity.
5. Suitable material (rock fill) must be placed on the seabed prior to commencement of
works. In this project rock size ranged from 150 to 300 mm.
6. Grid size and number of blows appear not to be very different from the parameters
that would be used for land based DR; however, water resistance reduces impact
efficiency, and equations developed for estimation of depth of improvement in land
based DC are not applicable to marine DC.
7. Friction angle of rock fill can be estimated from PMT by using Equation 2‐224.
502
3.15 Palm Jumeira Trial
As described in Section 3.12.1 Palm Jumeira is a reclamation off the coast of Dubai in the
United Arab Emirates, and consists of a tree trunk, a crown with 17 fronds, three surrounding
crescent islands and two identical smaller islands on the sides of the trunk that are in the
shape of the logo of The Palm.
After it was established that the reclamation was in a loose state the project’s engineers
stipulated that ground improvement had to be undertaken to increase the soil strength.
Initially, the specifications required that relative density be at least 60%, and CPT qc be (Al
Hamoud and Wehr, 2006):
Depth < 4 m: qc ≥ 6 MPa
4 m > depth > 8 m: qc ≥ 8 MPa
Depth > 8 m: qc ≥ 10 MPa
Later, in consideration of the carbonate content of the soil (refer to Section 3.12.1) the
specification was revised to qc ≥ 6 MPa for all depths.
In order to demonstrate the ability of dynamic compaction to satisfy this requirement, a DC
trial was performed at Chainage +300 m of Frond N, which has since been renamed to Al
Naghal, at the location shown in Figure 3‐235.
Figure 3‐235: Location of dynamic compaction trial on Frond N (Al Naghal)
503
3.15.2 Ground Conditions
PMTs were carried out at the trial location to establish the existing ground conditions.
Pre‐treatment PMT results at intended print and in between print locations are shown in
Figure 3‐236. It can be observed that the approximately 3 to 4 m thick superficial crust of the
reclamation was very dense, but the sand then became loose to medium dense. PLM and EM
in the upper 4 m were respectively dominantly ranging from 1.4 to 2 MPa and from 4.5 to 20
MPa. At greater depths down to 11 m PLM and EM were respectively from 0.4 to 0.8 MPa and
from 3 to 4.5 MPa.
Figure 3‐236: Pre‐treatment PLM and EM values in the DC trial area
The dynamic compaction trail was carried out in a 30 x 30 m2 area that is shown in
Figure 3‐237using a 25 ton pounder that was dropped from 20 m. Three compaction phases
were performed based on grids of 11 x 11 m2. As a first step the locations of Phase 1A were
pre‐excavated by approximately 2 m, with material retained on site for backfilling the craters.
504
Then 30 blows were applied per print (additional blows were implemented during the HPTs),
and the site was levelled using the excavated sand. The same procedure was also followed
for Phases 1B and 2 (with the grid size reduced to 7.8 x 7.8 m2), also with 30 blows per print.
As a last step 3 blows were applied to each print (with the final grid size of 5.5 x 5.5 m2) to
compact the backfilled and levelled sand.
Figure 3‐237: Dynamic compaction trial area
As shown in Figure 3‐237 and as part of the trial, PMTs and CPTs were carried out inside and
in between the DC prints after dynamic compaction. Figure 3‐238 shows PLM and EM values
after dynamic compaction. It can be observed that the improvement profile of the average
PMT parameters have sickle shapes with the average maximum PLM and EM values at the
depth of about 4 m. Figure 3‐239 shows that in the absence of ironing and with consideration
of the original strength of the topsoil, the soil parameters of uppermost layer of sand has
remained almost unchanged, but both PLM and EM improvement ratios are above 1, up to 5,
and approximately 2 at the depth of 11 m. Average PLM and EM improvement ratios
throughout the 11 m of fill was respectively 2.7 and 3.1. This suggests that dynamic
compaction with blows inside pre‐excavated prints has significantly improved the PMT
parameters at even the deepest depths of the hydraulic fill.
505
Figure 3‐238: After dynamic compaction PLM and EM values in the DC trial area
Figure 3‐239: PLM and EM improvement ratios
506
(a) (b)
Figure 3‐240: (a) EM/PLM before and after dynamic compaction, (b) Ratio of after to before dynamic
compaction EM/PLM
EM/PLM values before and after dynamic compaction are shown in Figure 3‐240(a). As can be
observed except for the layer between the depths of approximately 3 to 4 m that initially had
a lower EM/PLM value, but whose value increased to an over consolidated status (see
Table 2‐22), average EM/PLM has remained unchanged. This can be better understood in
Figure 3‐240(b) that shows the ratio of after to before dynamic compaction EM/PLM. As can
be seen for most of the test points the ratio fluctuates around unity. Calculation shows that
the average ratio of all points is 1.14, but the average drops to 1.0 if the points at depths of
3 and 4 m are excluded.
CPT qc values inside and in between DC prints after dynamic compaction are shown in
Figure 3‐241. It can be seen that the qc profiles are also sickle shaped with maximum values
at depths of about 3 to 4 m. It can also be observed that while the minimum value of qc is
about 6 MPa at the bottom of the test profile, it was constantly much higher than the
acceptance value of 6 MPa throughout the depth and more than 20 MPa at its peak.
507
Figure 3‐241: After dynamic compaction CPT qc values
The result of these tests show that while the soil profile has improved several times more
than acceptance throughout its depth and calculations can readily demonstrate that bearing
capacity and settlement performance are better than what could have been achieved if the
soil strength was strictly as per the acceptance criterion of qc ≥ 6 MPa, in the event that cone
resistance had dropped to slightly less than 6 MPa at depth, then improvement would have
been non‐compliant. The results of this test is a good example of why it was emphasised in
Section 2.10.1 that acceptance should be based on design criteria rather than minimum test
results.
qc/PLM and EM/qc correlations for in between prints, in prints and for average values at
various testing depths are shown in Figure 3‐242. As can be seen, it does not appear that
qc/PLM is affected by the shallow and deep failure modes that were observed in Al Nakhilat
Ship Repair Yard (refer to Section 3.7.5.3). As in Al Nakhilat project, qc/PLM does not seem to
be influenced by depth.
508
Figure 3‐242: qc/PLM and EM/qc correlations
Average qc/PLM in between prints, in prints and for all tests, including tests carried out at the
uppermost levels, can be calculated to be respectively 4.50, 5.20 and 4.86. These figures are
either just below or just above the minimum q*c/P*LM value of the 5 to 12 range that has
been proposed by Baguelin et al. (1978), see Table 2‐26, but substantially less than qc/PLM =
1/(0.11) that has been suggested by Briaud et al. (1985), see Table 2‐27.
Baguelin et al. have related q*c/P*LM values to dilatancy (see Section 2.9.2.9.4) possibly non‐
dilant when q*c/P*LM is about 5 to 6 and dilant if it is 8 to 12, but as confirmed by the test
results the treated sand in the trial was well compacted and a higher ratio should have been
predicted. Noting that the location of the sands that were considered by Baguelin et al. are
in a region where sands are not calcareous (see Section 2.9.2.9.4), with the available data it
can be speculated that the low qc/PLM values originate from the soil mineralogy rather than
compaction and the potential of soil dilatancy in shear.
509
Standard deviations of these points were respectively 0.80, 1.75 and 1.08. Comparison of the
overall qc/PLM average and standard deviation with Al Nakhilat (respectively 4.54 and 1.48)
suggests that while the standard deviation in this project has been less and thus the results
can cautiously be deemed as having lesser fluctuation, the difference between the average
qc/PLM of the two studies is less than 8%.
qc versus PLM values of Palm Jumeira Trial and Al Nakhilat Ship Repair Yard have been plotted
in Figure 3‐243. Best fit linear, second and third degree polynomials and power curves were
compared within the data range. Although these different mathematical functions produced
non‐coinciding curves, they all appeared to be pseudo linear, which suggests that the best
curve correlation can be assumed to be a linear function. By forcing the function to pass
through the origin of the axes, the best curve can be expressed by:
4.82 3‐13
Figure 3‐243: qc versus PLM values of Palm Jumeira Trial and Al Nakhilat Ship Repair Yard
EM/qc correlations at depth for in between prints, in prints and average values are also shown
in Figure 3‐242. Here, EM/qc for the average of all points at the uppermost level seems to be
510
greater than deeper points, but the deviation seems to be equal in magnitude to some
deeper points of the in between prints and in print locations.
Average EM/qc of in between prints, in prints and all tests, including tests carried out at the
uppermost levels, can be calculated to be respectively 1.49, 1.60 and 1.52. Standard
deviations were respectively 0.54, 0.40 and 0.32. Comparison of the overall EM/qc average
and standard deviation with Al Nakhilat Ship Repair Yard (respectively 1.42 and 0.38 for 21
points) shows that the results of the two studies are compatible whereas there is less than
8% difference in the average EM/qc.
Similar to Al Nakhilat Ship Repair Yard, the average value of EM/qc in Palm Jumeira Trial is
somewhat higher than what Briaud et al. (1985) have proposed (1.15, see Table 2‐27).
EM versus qc values of Palm Jumeira Trial and Al Nakhilat Ship Repair Yard have been plotted
in Figure 3‐244. Best fit linear, second and third degree polynomials and power curves were
compared within the data range. While the power curve also appeared to be pseudo linear,
the polynomials slightly bent downwards towards the end of the range. In the studied range,
the linear curve still seems to be the best curve, and by forcing the function to pass through
the origin, the best curve can be expressed as:
3‐14
1.54
From Equations 2‐212 and 2‐214, the relationship between Young and oedometer moduli
are:
1
2‐212
1 1 2
511
Figure 3‐244: qc versus EM values of Palm Jumeira Trial and Al Nakhilat Ship Repair Yard
2‐214
From Equations 2‐212 and 2‐214 and 3‐14, for the calcareous sands of Palm Jumeira and Al
Nakhilat Ship Repair Yard:
1
3‐15
1.54 1 1 2
(see Section 2.9.2.5), for saturated calcareous sands:
3.12 3‐16
Lee and Salgado (2002) have cited from Schmertmann et al. (1978) and Robertson and
Campanella (1989):
For young normally consolidated silica sand:
512
2.5 3‐17
For aged normally consolidated silica sand:
3.5 3‐18
For over consolidated silica sand:
6 3‐19
The factor of 3.12 in Equation 3‐16 is in between the factors for young normally consolidated
and aged normally consolidated silica sands, and suggests that silica sand correlations are
not suitable for carbonate sands.
1. It is possible to improve the strength of loose saturated carbonate sand at depth
using dynamic compaction.
2. Depth of improvement can be improved by performing dynamic compaction in pre‐
excavated prints.
3. Ironing is required to improve the strength of the superficial strength.
4. Acceptance should be based on design criteria rather than minimum test results.
5. It appears that compared to before dynamic compaction, average EM/PLM after
dynamic compaction can be considered to remain unchanged if points with initially
low EM/PLM are excluded from the calculation.
6. It was observed that for the saturated carbonate sands of Palm Jumeira and Al
Nakhilat Ship Repair Yard the relationship between qc and PMT parameters was not
influenced by depth and could be formulated as:
513
a. qc = 4.82 PLM
b. qc = EM /1.54
7. It was observed that for the saturated carbonate sands of Palm Jumeira and Al
Nakhilat Ship Repair Yard the relationship between the CPT cone resistance and
Young modulus is Ey = 3.12 qc.
514
3.16 Dynamic Compaction Vibration Monitoring
The focus of this section of the thesis is not the actual ground improvement works
themselves, but rather the assessment and interpretation of the results of vibration
monitoring.
As shown in Figure 3‐245, Fujairah Power and Desalination Plants are located in Qidfa in the
United Arab Emirates, and have been subject to an expansion project to meet the increasing
power and potable water requirements of the UAE. The new desalination plant includes a
pumping station, reverse osmosis, lime re‐mineralisation, CO2, chemical, sludge, and
electrical buildings.
Figure 3‐245: Location of Fujairah Power and Desalination Plant
The geotechnical investigation reported that the ground was composed of multiple layers of
sandy soil. The top soil was composed of about 1.5 m of well compacted sand backfill that
was placed as the working platform. This man‐made layer was followed by about 9 m of loose
to very dense sand with SPT blow counts in the range of 11 to 30 and sometimes more. Fines
content was usually less than 10%, but occasionally higher and up to 15%. The ground then
became very dense sand with infrequent layers of very stiff silty soil down to the end of the
515
boreholes at the depth of about 30 m. SPT blow counts were constantly more than 50 and
fines content was generally less than 10% but as high as 55% in the very stiff deep silt layers.
The CPT log of one of the test locations is shown in Figure 3‐247.
Figure 3‐246: Before dynamic compaction CPT at Fujairah Desalination Plant
Groundwater level was approximately 6 m below working platform level.
Calculations by the engineering team suggested that the project requirements were not met,
and ground improvement with the below criteria were stipulated for an area of 28,000 m2:
Bearing capacity: 150 to 170 kPa for mat foundations and 250 kPa for single
footings
Total settlement: 25 mm
Differential settlement: 10 mm
Dynamic Compaction was proposed and implemented to increase the soil’s mechanical
properties and to satisfy the design criteria. The 15 ton pounder that was used for the works
had a base that was made of 2x2 m2 steel plates that were chamfered at the corners.
516
3.16.1.2 Vibration Monitoring
As shown in Figure 3‐247, dynamic compaction was sometimes carried out in the vicinity of
plant and pipelines; hence, there were concerns about the magnitude of induced vibrations
and the potential effects that they could have.
Figure 3‐247: Dynamic Compaction at Fujairah Desalination Plant
For informative purpose the vibrations were monitored and measured in three directions
(dir. 1= radial, dir. 2= vertical, dir. 3= transverse) during the project. The results of 22 readings
are presented in Table 3‐26. These measurements were carried out without the utilisation of
vibration reducing trenches. It can be observed that the registered vibration frequencies
associated with the peak particle velocities were most often in the range of 7 to 12 Hz and
occasionally up to 27 Hz. Vibration frequencies for all three directions are shown in
Figure 3‐248. It can be observed that the registered frequencies in the other two directions
were 11 to 21 Hz and are also within the same range order. The most occurring frequencies
appear to be in the range of 10 to 20 Hz.
Table 3‐26 is graphically presented in Figure 3‐249. For demonstrative purposes the multiple
equations proposed by Mayne et al. (1984), Mayne (1985), and Varaksin (1981) have also
been plotted onto the same figure. Mayne’s upper limit and conservative equation
(Equation 2‐150) has the largest overestimation and estimates the PPV values by 3 to 5 times
of what was measured on site. Although Mayne’s earlier proposed relationship (Equation 2‐
149) is less conservative, it is never‐the‐less still overestimating PPV by 3 to 4 times. It
appears that Equations 2‐149 and 2‐150 predict relatively similar particle velocities at greater
distances from pounder impact point, but the calculated values divert when PPV is predicted
closer to the source of vibration.
517
PPV Frequency Drop
(mm/s) (Hz) Distance (m) Height (m)
4.57 7.5 46.6 20
5.33 7.6 46.6 20
5.84 7.4 46.6 20
5.97 7.4 46.6 20
6.22 7.3 46.6 20
5.97 7.2 46.6 20
5.97 10.6 36.6 20
6.60 12.1 36.6 20
7.11 12.4 36.6 20
7.62 26.9 36.6 20
7.49 12.8 36.6 20
7.62 20.4 36.6 20
10.80 11.3 26.6 20
11.94 11.6 26.6 20
12.32 12.1 26.6 15
11.05 12.1 26.6 15
12.95 12.8 26.6 15
11.18 12.1 26.6 15
10.67 11.9 16.6 10
11.81 12.1 16.6 10
12.70 11.6 16.6 10
Table 3‐26: Results of vibration monitoring
Mayne’s final relationship (Equation 2‐152) that is independent of the pounder weight, but
that takes into consideration the pounder’s drop height and radius appears to be closer to
the measured PPV values on site. This equation predicts PPV by an overestimation of 1.2 to
2.8 times. It appears that the larger variations are when prediction is made closer to the
vibration source.
Varaksin (Equation 2‐153) appears to be the closest to the site’s PPV measurements.
However, Equation 2‐153 sometimes underestimates PPV. This has also been noted by
518
Romana Ruiz and Jurado (1999) for dynamic compaction vibrations that were monitored for
an onshore fill project in the Canary Islands.
Figure 3‐248: PPV and vibration frequencies
Figure 3‐249: Comparison of PPV measurements with various methods for estimating PPV
Although Equation 2‐153 appears to predict PPV values quite reasonably, in some occasions
it may unsafely underestimate the actual field values. As underestimation of PPV can lead to
unanticipated problems during the execution of the works, the author proposes a
519
modification to this equation in the form of Equation 2‐153 that overestimates measured
PPV by 1.2 to 2.4 times in this case study.
.
560 mm/s 3‐20
Instead of solely a function of distance as presented in Equation 2‐153, PPV values of
Table 3‐26 can be formulated in the form of Equation 2‐148 to yield:
.
√
25 3‐21
The comparison of the field data with Equations 2‐152 and 2‐153 and 3‐21 are graphically
shown in Figure 3‐250. It appears that while providing a more reasonable estimate of PPV,
Equation 3‐21 can still sufficiently overestimate PPV during the planning stage of a project.
Figure 3‐250: Comparison of field measured PPV values with equations proposed by Mayne
(Equation 2‐152), Varaksin (Equation 2‐153) and the author (Equation 3‐21).
Al Medina A'Zarqa, translating to Blue City, is a multibillion dollar megacity project that has
been planned in Oman, and is to spread over an area of 32 km2 with 16 km of coastline
southeast of Al Sawadi and along the Gulf of Oman. The general location of the project is
520
shown in Figure 3‐251. The multiple phases of the project are to be built over a period of
several years. Phase 1 of this project is located within an area that measures 3 km along the
coastline and 2 km inland. Figure 3‐252 shows the plan of the various phases of Medina
A’Zarqa.
Figure 3‐251: Location of Medina A’Zarqa
Figure 3‐252: Plan of various phases of Medina A’Zarqa
The plan of Phase 1 of Medina A’Zarqa is shown in Figure 3‐253. Plots No. 1.1.2, 1.4.2, 1.3.1
and 1.3.2 with an approximated area of 225,000 m2 are to be the first construction areas.
Based on the topographical survey, as a whole, the original ground level of the site was from
521
+0.8 m to +2.7 m MSL (Mean Sea Level). Average ground level was approximately +1.5 m MSL
in Plots No. 1.1.2 and 1.4.2, and approximately +2.3 m MSL in Plot No. 1.3.1 and 1.3.2.
Minimum, maximum and average groundwater levels were reported to be respectively ‐0.4
m, +0.25 m and ±0.0 m MSL.
Excluding their partial basements, the buildings in the mentioned plots were anticipated to be
2 to 7 stories.
Figure 3‐253: Plan of Medina A'Zarqa Phase 1
Table 3‐27: Summary of ground conditions
522
A summary of the generalised ground conditions is tabulated in Table 3‐27.
Preliminary calculations indicated the presence of loose sand layers in the upper 8.5 m of
ground, and stipulated the application of ground improvement to allow the construction of
raft foundations. Ground improvement works were tendered, and the accepted proposal was
based on the implementation of dynamic compaction.
Dynamic compaction was carried out from elevation +1.8 m MSL in three deep phases using
a 23 t pounder that was dropped from 20 m, followed by ironing using a 15 t pounder that
was dropped from 15 m. Figure 3‐254 shows the application of dynamic compaction during
one of the deep phases.
Figure 3‐254: Application of dynamic compaction at Medina A’Zarqa using a 23 t pounder
As shown in Figure 3‐255 vibration monitoring by recording the peak particle velocities and
associated frequencies in the radial (Dir R), vertical (Dir V) and tangential (Dir T) directions
was performed using a seismograph. Measurements were carried out at different distances,
from 10 m to 100 m during the three deep and ironing phases for drops that was applied to
a specific print.
523
Figure 3‐255: Dynamic compaction vibration monitoring at Medina A’Zarqa
PPV versus the blow number for deep DC phases 1 and 3 are respectively shown in
Figure 3‐256 and Figure 3‐257 at various distances. As can be observed that regardless of the
blow number, distance to pounder impact point and phase of dynamic compaction, peak
particle velocity was always in the radial direction. Similarly, particle velocity in the tangential
direction was always the least value among the three directions. The comparison of the two
figures also indicates that at equal distances, PPV in phase 3 is higher than phase 1. The PPV
ratio of the two phases appears to be less at closer distances and more at further distances.
Figure 3‐256: PPV versus number of blow in DC phase 1
524
Figure 3‐257: PPV versus number of blow in DC phase 3
It can also be seen in Figure 3‐256 that in most instances of phase 1 monitoring, when
distance is kept constant, PPV initially increases with the number of blows, but then reduces
to values lower than what was measured in the first blow. This is most observable in all three
directions in the closest distances. As lesser blows were applied per impact point in phase 3
it cannot be said that the same has happened in the later phase, but Figure 3‐257 shows that
PPV of the first blow in this phase was also smaller than subsequent blows.
Figure 3‐258 shows the plot of peak particle velocity of each direction against frequency for
phases 1, 3 and ironing at various distances. It can be observed that almost all plotted points
are between the frequency range of 5 to 9 Hz.
Plotting PPV against distance for DC phases 1, 2 and 3 produces interesting results. It can be
seen that although there is PPV scatter for records made at the same distance during each
phase, the upper limit value of PPV for each phase appears to fit reasonably well with a line
drawn in a semi logarithm scale. It can also be understood that the slope of PPV attenuation
in the earlier phases is more than the later phases. This suggests that there is greater material
damping in loose soils than dense soils. As already noted for Figure 3‐256 and Figure 3‐257,
Figure 3‐259 also clearly indicates that the difference between PPV values of premier and
subsequent phases of dynamic compaction becomes greater as distance from impact point
525
increases. Also, it appears that at distances closer than a critical distance, seemingly about
19 m in this study, PPV becomes insensitive to the compaction phase.
Figure 3‐258: PPV versus frequency in phase 1, phase 3 and ironing
Figure 3‐259: PPV versus distance and comparison with prediction equations
Comparison of measured PPV in Figure 3‐259 with Equation 2‐151 indicates that this
empirical formula is able to predict PPV with sufficient accuracy in this project as well.
526
3.16.3 Um Quwain Marina
Umm Al Quwain (UAQ) Marina is a master‐planned community in the northern emirate of
Umm Al Quwain in the UAE (refer to Figure 3‐260). The plan of the project envisages 6,000
villas, 2,000 townhouses, 1,200 resort and hotel rooms, super markets, shopping centres,
schools and health clinics. UAQ Marina Phase 1 consists of 277 two floor (ground and first
floor) villas in Community 16 of which 127 villas are within the area (86,000 m2) that will be
reviewed in this thesis.
Figure 3‐260: General location of UAQ Marina
Groundwater was recorded to be 1 to 3 m below ground level. The preliminary geotechnical
investigation that was based on SPT boreholes suggested that the site was composed of 2 m
of very loose to medium dense silty sand with SPT blow counts ranging from 2 to 28 and with
fines content less than 10%. This stratum was followed by 4.5 m of very loose to dense silty
sand with blow counts ranging from 0 to 35 and fines content of less than 15%.The next 3.5
m of soil was silty sand with SPT blow counts ranging from 16 to more than 50 and with fines
content of less than 20%.
Based on the geotechnical investigation the project engineers designed strip footings on
improved ground. Maximum footing width was 1.5 m under a uniform load of 140 kPa.
Footing depth was defined as 1 m below ground level. Consequently, the project was
tendered and the design and construction ground improvement package was awarded to a
527
ground improvement specialist contractor who had proposed the implementation of
dynamic compaction.
Once the contractor was on site, further geotechnical testing revealed a different soil profile
and the presence of a previously unidentified 0.3 m thick layer of very soft very silty sand to
sandy silt (fines content in the range of 40 to 60%) at an upper depth of approximately 1.7 m
to 2.1 m. PLM in this layer was 200 kPa.
Therefore, the ground improvement technique was modified to pre‐excavated dynamic
replacement. The saturated soft material was excavated from below groundwater level, the
excavated pits were backfilled with sand to groundwater level, and mixed soil was placed
above groundwater level.
As dynamic replacement works were to be carried out as close as 20 m from existing and
under construction structures concerns were raised that vibrations generated by the ground
improvement works could damage the buildings. Hence, a vibration monitoring programme
was developed to study the vibration parameters with and without the installation of
vibration isolators.
In this programme initially a 14.5 ton pounder was dropped from 20 m, and radial, transversal
and vertical particle velocities and associated vibration frequencies were measured at
distances of 10 to 40 m, see Figure 3‐261(a).
(a) (b)
Figure 3‐261: (a) Vibration monitoring without a trench, (b) vibration monitoring with a trench
Next, as shown in Figure 3‐261 (b), a trench that was 25 m long, 2.5 m wide and 2.5 deep (to
groundwater level) was excavated 13.3 m away from the pounder’s drop point and the same
528
parameters were measured again. In this phase the pounder used was 13 tons and the
pounder drop height varied from 5 m to 18 m.
Vibration Monitoring without an Isolation Trench
The measured PPVs and their corresponding frequencies for the case with no trench are
tabulated in Table 3‐28, which also shows PPV values that have been estimated by using
Equation 2‐151 and the ratio of estimated PPV to measured PPV. It can be observed that the
frequencies corresponding to peak particle velocity are within the range of 12 to 24 Hz. It can
also be seen that in this monitoring programme the frequency of the peak particle velocity
increases with distance.
While Equation 2‐151 has underestimated PPV at close distances (by 31% at 10 m and by
14% at 15 m), it has been able to conservatively estimate PPV at other distances. The
underestimations at 10 m and 15 m are not of major concern as they are still above what
would be statistically deemed as non‐damaging (Siskind et al., 1980). Underestimation at 20
m is approximately 9%, but the predicted values become an overestimation of 63% at 40 m.
Measured and estimated PPVs plotted against distance are shown in Figure 3‐262 for the
case with no vibration isolation trench. Noting that PPV is dependent on a number of
parameters other than those appearing in Equation 2‐151, such as number of pounder drops,
the reliability of Equation 2‐151 can be deemed as satisfactory as a starting point.
529
Figure 3‐262: Comparison of measured and estimated PPV
Vibration Monitoring with an Isolation Trench
The measured PPVs and their corresponding frequencies for the case with a trench are
tabulated in Table 3‐29, which also shows the estimated PPVs had there been no trench.
The ratios of estimated PPV (without a trench) to measured PPV for the two cases of
monitoring with and without an isolation trench have also been included in Table 3‐29. It can
be seen that the ratio of estimated to measured PPV is considerably higher when a trench
has been excavated. For example at the distance of approximately 35 m the ratio of
estimated to measured PPV is 1.57 without a trench but 2.91 and 5.46 when a trench has
been excavated. The ratios of estimated to measured PPV for the trenched case are not
demonstrative of the true efficiency of the trench as these ratios were also not unity when
there was no trench. To obtain a more accurate estimation of the efficiency of the trench it
is more proper to calibrate the results by dividing the ratios of estimated to measured PPV
(last column in Table 3‐29).
The division of the ratios indicates that the trench has been efficient with PPV possibly having
been 1.09 to 3.47 times more had there not been any trench. Graphically presenting the
trench efficiency, Figure 3‐263 shows that while the data scatter does not allow a conclusive
interpretation of any specific trend of the PPV reduction factor, it can still be observed that
the best linear fit is almost a horizontal line with an average value of 2.05.
530
Distance Drop Frequency PPV (mm/s) Ratio
(m) height (Hz) Measured Estimated Estimated to trenched
(m) (trench at without measured to no
13.3 m) trench trenched no trench
trench
10.0 5 5.2 35.6 19.7
10.0 8 10.0 47.2 25.5
16.3 8 18.2 7.4 14.9 2.02 0.86 2.35
25.8 12 16.5 5.7 11.3 1.97 1.17 1.69
25.8 15 18.9 7.9 12.7 1.61 1.17 1.38
30.8 12 24.3 3.2 9.3 2.89 1.25 2.31
30.8 15 24.3 3.7 10.5 2.83 1.25 2.26
35.8 10 28.4 1.3 7.1 5.46 1.57 3.47
35.8 12 23.2 2.7 7.9 2.91 1.57 1.85
35.8 18 22.2 5.7 9.8 1.72 1.57 1.09
40.8 5 10.2 1.4 4.2 3.00 1.63 1.84
40.8 10 12.1 1.7 6.2 3.62 1.63 2.22
Table 3‐29: Vibration monitoring summary for the case with the isolation trench
Figure 3‐263: PPV reduction factor when using a vibration reduction trench
531
3.16.4 Comparison of Vibrations Generated by Dynamic Compaction and
Vibro Compaction
Vibro compaction, also known as vibroflotation, is a deep ground compaction technique that
was developed almost 80 years ago (Mitchell, 1981) with the invention of the first vibroprobe
by Degen and Steuermann (Better Ground) in Germany.
This technique is best suitable for the treatment of soils with limited amounts of fines.
Mitchell (1981) proposes that the best desirable soils for vibro compaction are when the
soil’s fines content is limited to 18%. Woodward (2005) suggests that best results can be
achieved when fines content is less than 10%.
The vibroflot, also referred to as vibroprobe or vibrating poker is a hollow steel tube
containing an eccentric weight mounted on a vertical axis in the lower part so as to give a
horizontal vibration. The vibroflot itself is connected to extension tubes that are supported
by a rig, which is usually a crane.
The vibroflot is either flushed down to the required depth in the soil using water jets or
vibrated dry with air jets. When the vibroprobe reaches the required depth, material is added
from the ground surface during withdrawal, and the vibroflot is moved in an up and down
motion at certain intervals. The horizontal vibrations form a compacted cylinder of soil with
a depression at the surface due to the reduction of void ratio in the ground. Depending on
the vibroflot power, the zone of improved soil extends from 1.5 m to more than 4 m from
the vibrator.
Similar to dynamic compaction, application of vibro compaction also generates vibrations.
The intent of this section of this thesis is to develop a method using a case study to estimate
vibro compaction generated vibrations and to compare the peak particle velocity with
estimated dynamic compaction PPV (using Equation 2‐151.
Unfortunately not much research is available on particle velocity generated by vibro
compaction. Without providing details about the measurements and the scatter of data,
Woods and Jedele (1985) have presented the graph of peak vertical velocity (which may have
been PPV) for vibroflots with motor powers of 22 kW (30 hp) and 75 (100 hp). Similarly,
Dowding (2000) has presented the graph for a vibroflot with a motor power of 120 kW, 18
532
mm vibration amplitude and 30 Hz frequency. Neither of the publications has referred to the
depth of the vibroflot during particle velocity measurements. Figure 3‐264 shows the results
of Woods and Jedele’s and Dowding’s publications.
Figure 3‐264: : PPV generated by vibro compaction as per previous research
As shown in Figure 3‐265, vibration monitoring was carried out on points located on Plot 106
of Frond D, now renamed to Al Barhi, of Palm Jumeira whose general description and ground
conditions were presented in Section 3.12.1.
The power of the vibroflot’s motor that was used in this study was 96 kW. The amplitude at
the tip of the vibro probe was rated at 6 mm. Centrifugal force and eccentric moment were
respectively 193 kN and 17 Nm. Vibration frequencies were up to 53 Hz.
Vibration monitoring was carried out using an accelerometer at distances of 5, 10, 15, 20, 30,
40 and 70 m along 4 lines for two treatment points. PPV was recorded at depths ranging from
seabed at approximately 12.5 m to 2 m below the ground surface.
Measured PPV at various depths and distances from the vibroflot are shown in Figure 3‐266.
Although the scatter of data does not allow the realisation of curve fitting processes with
high amounts of reliability, it can be observed that PPV reduces not only with distance, but
also with depth. The lines that are shown in Figure 3‐266 are not the best fitted curves for
533
PPV versus vibroflot depth, but are drawn to visualise the trend of changes. What is
noticeable is that the rate of PPV reduction with depth increases as distance reduces.
Figure 3‐265: Location of vibro compaction vibration monitoring on Frond D of Palm Jumeira
Figure 3‐266: PPV at various depths and distances from the vibroflot
As PPV monitoring has not included data above the depth of 2 m, extrapolation must be
carried out to predict PPV at shallower depths. Considering the amount of data scatter, this
534
approach is justifiable. Using this concept, it is possible to develop a curve for PPV when the
vibroflot nose is at ground level. This is shown as a dotted curve in Figure 3‐267. It can be
seen that even though the plot of the extrapolated PPV values on ground surface is not
strictly parallel to the other curves, the curve that best fits these points (the solid line in
Figure 3‐267) is parallel with the other lines and falls in between the 75 and 120 kW vibroflots.
This suggests that earlier research may have also measured PPV at ground surface.
Figure 3‐267 also suggests that while the magnitude of vibro compaction generated
vibrations can be expected to be less than dynamic compaction vibrations at the same
distance, implementation of vibro compaction can also lead to exceeding proposed vibration
limits (Siskind et al., 1980) at distances that are less than 10 m.
Figure 3‐267: vibro compaction generated PPV versus distance
In Equation 2‐148 Wiss (1981) expressed PPV as a direct function of distance, d, and
reciprocal function of energy, E:
2‐148
√
K is the intercept with the ordinate and n is the slope or attenuation rate.
535
As power, P, is energy per unit of time, it is possible to express vibro compaction generated
PPV in terms of vibroflot power and distance. By applying Equation 2‐148 to the PPV
measurements by Woods and Jedele (1985), Dowding (2000) and this study, it can be
calculated that n is from 1.52 to 1.64 with an average value of 1.59. Calculation shows that K
ranges from 2.08 to 8.16 with an average value of 6.2 for the four set of curves. Noting that
the lowest K value is for the 24 kW vibroflot, the average K for the remaining 3 vibroflots will
be 7.5 when the value of the 24 kW vibroflot is excluded. As the 24 kW vibroflot, if ever at
all, is less commonly used than the other types of vibro probes, and since using a larger K
value is more conservative (as it over estimates); thus, when P is in kW and d is in metres:
.
7.5 mm/s 3‐22
√
In the projects that were studied in this Section, vibrations were monitored during dynamic
compaction works. Measurements included vibration frequencies and peak particle
velocities for various pounder weights, drop heights and at various distances. From the
results it can be understood that:
1. Vibration frequencies were generally in the range of 5 to 20 Hz.
2. Measured PPVs were 3 to 5 times less than Equation 2‐150. Equation 2‐152 also
appeared to be conservative, but was able to estimate more accurate results.
3. Equation 2‐153 was able to predict reasonably close values to field‐measured PPVs;
however, the equation sometimes underestimated the results.
4. Equation 2‐153 suggests that without application of specific vibration reduction
techniques, it may be possible to carry out dynamic compaction as close as 25 to 35
536
m from structures without exceeding the maximum PPV values that have been
recommended by USBM (Siskind et al., 1980).
5. Equation 2‐151 is able to predict PPV reasonably well and with sufficient accuracy;
however, it may underestimate PPV at distances less than 20 m. As both measured
and predicted vibrations within the underestimation zone are probably above
recommended tolerances and specific vibration reduction measures have to be
adapted, Equation 2‐151 can still be used for distances that are further away than 20
m and with caution and as a first assessment at closer distances.
6. Vibration parameters have been monitored for different number of blows and
distances during several phases of dynamic compaction. It has been observed that
peak particle velocities are greater during later phases of compaction as compared
to the premier phases. The differences are greatest at farther distances, and it
appears that when the distance is closer than a critical value, compaction phase
influence becomes unnoticeable.
7. Vibration frequency associated with PPV increases with distance.
8. Excavation of a trench that was 25 m long, 2.5 m deep (to groundwater level), 2.5 m
wide, and that was approximately 13 m away from a DC print was able to reduce PPV
on average to half of its value when there was no trench. This experiment suggests
that isolation trenches can be an effective means for reducing dynamic compaction
vibrations.
9. The result of this study indicates that maximum PPV can be expected to be highest
when the vibroflot is closest to the ground surface. In other words, the most critical
time that vibroflot vibrations can damage a nearby structure is when it is just
penetrating the ground or being pulled out of the treatment point.
10. Vibro compaction generated PPV can be estimated as a function of vibroflot power
and distance from Equation 3‐22.
11. It appears that while the magnitude of vibro compaction generated vibrations can
be expected to be less than dynamic compaction vibrations at the same distance,
vibro compaction generated vibrations can also exceed vibration limits (Siskind et al.,
1980) at distances that are less than 10 m.
537
12. The attenuation rate of dynamic compaction is less than vibro compaction. In other
words, dynamic compaction vibrations will reduce over a longer distance.
538
3.17 Predicting PLM and EM from Dynamic Compaction Induced
Subsidence
3.17.1 Introduction
Common practice in dynamic compaction to carry out a calibration programme before
production and execution of actual ground improvement works to optimise the design
parameters. The ground is tested before soil treatment to provide an understanding of its
condition and geotechnical parameters. Then, dynamic compaction is carried out on a
predefined grid size with a number of blows that are thought to provide the best post
improvement data. A number of heave and penetration tests may be carried out during the
calibration to provide an understanding of the trend and amount of compaction and heave
during the pounder impacts.
Upon completion of dynamic compaction the ground will be tested again to ensure that the
desired parameters have been achieved. Occasionally, a number of patterns may be tried in
the calibration to provide the engineer with more design options and sometimes poor test
results force the repetition of the calibration with alternative patterns. Testing consumes
time, and it would be very displeasing to realise that the results were not acceptable after
the completion of the tests and interpretation of the data. Thus, it would be beneficial to be
able to predict the amount of post treatment soil condition and to take possible corrective
measures if the prediction indicates that the tests will not meet the design criteria.
In this section of this thesis a new method is proposed for the estimation of post dynamic
compaction PLM and EM based on the amount of induced ground subsidence by dynamic
compaction.
Ground subsidence due to treatment, per se, is not a criterion that should be used for
assessing or verifying soil improvement, but can be used as an indication of the amount of
improvement. With the same amount of energy, loose soil will subside more than dense soil.
Likewise, the greater the applied energy is, the larger the induced subsidence will be.
However, large or small amounts of settlements do not necessarily mean that design criteria
have or have not been satisfied.
539
If it was possible to relate subsidence to the amount of improvement of the measurable soil
parameters, then subsidence could be used in conjunction with actual pre‐treatment test
results to predict and to estimate the ground condition after dynamic compaction.
3.17.2 The Relation between Induced Strain and Subsidence with PLM and
EM
Varaksin et al. (2005) have developed a relationship between dynamic compaction induced
strain and PLM improvement for the dune sands of Al Quo’a (refer to Section 3.3). The
hypothesis used was that PLM will double for every 3% of strain. Thus:
3‐23
2 3‐24
ε= strain
(PLM)i= limit pressure before soil improvement
(PLM)j= limit pressure after soil improvement
n= number of times the limit pressure has doubled
a= percentage of strain induced for doubling PLM (3%)
Solving Equation 3‐24 for n, and replacing its result in Equation 3‐23 will yield:
3‐25
2
Further expanding this notion will result in a relationship between ground subsidence and
the increase in limit pressure:
3‐26
2
, ,
540
m= number of pressuremeter tests in the borehole within the improvement zone (i.e., the
depth where PLM has increased), and hk is the testing spacing. If (PLM)j/( PLM)i is denoted by r
(r ≥ 1), then the subsidence can be calculated to be:
3‐27
2
,
Replacing the values of a =0.03 and log 2 in Equation 3‐27 will yield:
0.1 3‐28
,
As illustrated in Figure 2‐63, Lukas (1986) studied lateral ground movement due to dynamic
compaction, and measured lateral movements in three types of soils at distances of 3 and 6
m from the pounder. Figure 2‐63 suggests that the magnitude of ground movement initially
increases to some depth, but then reduces and eventually becomes negligible. Noting that
we have already established a relationship between strain and PLM in Equation 3‐25, it would
be rational to assume that the soil’s post dynamic compaction PLM profile will also follow a
similar trend. This has been observed by Lukas as well (refer to Figure 2‐62).
Similar results have also been achieved by Hajialilue‐Bonab and Rezaei (2009) (refer to
Figure 2‐64) who carried out laboratory scale dynamic compaction using particle image
velocimetry techniques. A sickle shape curve characterises the soil displacement profile along
vertical lines of soil, and the curves become flatter as they go farther away from the impact
centre. It was observed that the displacement of soil in any given horizontal plane has a bell
shaped curve.
Berry et al. (2004) proposed that for simplicity a Rayleigh distribution be used for void ratio
reduction. The probability density function Rayleigh distribution can be mathematically
written in the form of Equation 3‐29:
541
3‐29
z= depth from ground surface
= depth of maximum strain. Lukas (1986) assumes maximum improvement to be a depth
between one third to one half the depth of improvement.
The procedure that is proposed in this Section is subject to the below conditions and
assumptions:
1. A pre‐treatment pressuremeter test has been carried out in the calibration area.
2. The material grading in the ground is relatively uniform throughout the treatment
depth.
3. The soil parameters are fairly uniform before treatment; i.e., there are not any very
loose or very dense layers.
4. Average ground settlement is measured; either by levelling the ground after dynamic
compaction or by using the average measured settlement of the heave and
penetration test for the pounder’s cell.
5. As geometric mean values are used in bearing capacity calculations (see
Section 2.9.2.6), even if prediction of PLM at individual points varies from reality, the
difference in bearing capacity calculations would remain within acceptable
derivations.
Ground subsidence due to dynamic compaction is the accumulation of the vertical
deformation of the layers within the depth of improvement; thus:
_ 3‐30
,
εDC_k= dynamic compaction induced strain in layer (test spacing in borehole) k
If we assume that test spacing is the same throughout the borehole, then Equation 3‐30 can
be re‐written as:
542
_ 3‐31
,
h= test spacing in borehole
If all layers had the same initial PLM value, then it could have been said that dynamic
compaction induced strain could have been defined in a simplistic approach to follow the
Rayleigh distribution; however, PLM values will most probably vary in reality. Thus, it is
necessary to introduce a pre‐strain, εo_k, for each layer that demonstrates the strain
difference of that layer compared to the lowest PLM value. In other words, considering the
pre‐treatment loosest soil level as the local origin of computations, it can be assumed that
all other layers have undergone a strain to reach their initial pre‐treatment state. As shown
in Figure 3‐268, the summation of εo_k and εDC_k will form the Rayleigh distribution strain, εR:
_ _ 3‐32
Figure 3‐268: Rayleigh distribution strain as a function of pre‐treatment and DC induced strain
εR follows and is proportional to the Rayleigh distribution, thus:
_ _ 3‐33
= scale parameter of the distribution
543
cR= proportion coefficient
εo_k is not known, but can be back calculated by determining the amount that each layer had
strained to reach its pre‐improvement limit pressure, PLM_k, as compared to the minimum
pre‐treatment limit pressure value, PLM_min. This calculation can be done using Equation3‐25,
re‐written below:
_
_ 3‐34
_
2
Once εo_k and εDC_k for each layer has been determined, it is then possible to add them all up,
and determine the coefficient:
_ 3‐35
_
or:
_
_ 3‐36
0.03
2
_
_
∑ 0.03 3‐37
2
∑
can be assumed to be an arbitrary value between one half to one third of the depth of
influence. However, experimenting with this parameter in this research suggested that
predicted post improvement limit pressures are very sensitive to , and lesser values tend
to result in unrealistically high limit pressure values at depths that are in the vicinity of . It
appears that a value of 0.45 to 0.5 times the depth of influence can predict more rational PLM
peak values. It is recommended that peak values be controlled and the value of be
reviewed accordingly.
544
Once cR has been determined, Equation 3‐35 can be used to calculate εDC_k for each layer:
_
_
∑ 0.03
2 3‐
∑ 38
_
_
0.03
2
As a final step, using Equations 3‐23 and 3‐24 in conjunction with Equation 3‐38, the post
improvement PLM value of each layer can be computed to be equal to:
_
_ _ 2 . 3‐39
If calculations yield a negative value for εDC_k, then the calculation procedure is predicting a
negative strain or expansion of soil, which signifies a reduction in PLM. In such a case the
calculation should be repeated without considering the rows associated with the negative
strain values.
Menard (Centre D'Etudes Menard, 1975) has identified a correlation between PLM and EM for
different soil types (refer to Table 2‐22); hence, it seems rational to be able to predict post
dynamic compaction EM values using a similar approach.
EM will double in value every time the soil is strained by a certain percentage, b, which does
not necessarily have to be 3%. Equations 3‐34, 3‐37, 3‐38, and 3‐39 can be re‐written in terms
of b and EM:
_
_ 3‐40
_
2
545
_
_
∑
2
_ 3‐41
∑
_
_ 2 .
_
_
∑
2
_
∑ 3‐42
_
_
2
_
_ _
2 3‐43
3.17.5 Verification
Al’Quo’a New Township (Section 3.3) and Marjan Island Road Corridor (Section 3.6) projects
have been chosen for verification of the proposed method for predicting PLM and EM from
dynamic compaction induced ground subsidence.
In Al Quo’a New Township PMTs were used for quality control and verification purposes.
Before and after dynamic compaction PLM values for four locations are shown in Figure 3‐30
to Figure 3‐33. Measured site settlements after dynamic compaction were in the range of
0.60 to 0.80 m. For the purpose of verification calculations and as summarised in Table 3‐30,
settlements have been calculated according to Equation 3‐28, and depths of improvement
have been assumed to be testing termination depth. Also shown in Table 3‐30 are the
values and the ratio of actual to predicted geometric mean PLM within the depth of influence.
Figure 3‐269 shows the ratios of actual and predicted limit pressure improvements for the
PMTs presented in Figure 3‐30 to Figure 3‐33. As can be observed that in all diagrams the
predicted ratio of improvement is initially less than the actual ratio of improvement, and
becomes greater after the depth of maximum improvement ratio has been reached, and
546
ultimately reverses back to being less at the deepest treatment zones. In all tests the
geometric mean of the limit pressure within the improvement depth was quite close to
reality with a maximum over estimation of about 5%.
Ground subsidence due to dynamic compaction at Marjan Island Road Corridor was 0.29 m,
and the PLM and EM values before and after dynamic compaction are shown in Figure 3‐104.
While it could have been expected that the improvement profiles would look like sickles,
which could be defined with a Rayleigh distribution, that was not exactly the case, and
improvements seem to have been more uniform in depth with occasional points of higher
strength. Due to the non‐classical shape of the improvement profiles it is interesting to study
the applicability of the proposed calculation procedure for the prediction of PLM and EM.
Figure 3‐269: Ratios of actual and predicted limit pressure improvements for PMTs of Figure 3‐30 to
Figure 3‐33
547
For calculation purposes it is assumed that depth of improvement was 10 m (any strain in the
deeper denser layers has been assumed to be negligible), = 4.5 m, and a= b= 3%.
While subsidence calculated PLM values was from 0.27 to 0.34 m, which is close to the actual
average subsidence, EM based settlement calculations were larger and from 0.33 to 0.42 m.
This is not necessarily inaccurate as all tests were performed in print locations. The greatest
difference was observed in PMT‐102 with the most fluctuating profile.
Actual and predicted PLM and EM improvement ratios are shown in Figure 3‐270. For both PLM
and EM, the general shape of the ratios of predicted post improvement to initial values are
rather similar to the ratios of actual post improvement to initial values; however, the actual
harmonic mean PLM is 94% to 117% of predicted harmonic mean. In line with the predicted
settlements based on EM, the actual geometric mean EM is 112% to 144% of the predicted
geometric mean. It appears that in general the largest variation in the ratios is in the level
where improvement is maximal. This may be due to the fact that all tests were carried out in
the print itself; however, more research is required to confirm this explanation. The biggest
difference is once again observed in PMT‐102. Also, the review of EM to PLM ratios in the post
improvement tests shows that the range of values in PMT‐102 has had the most amount of
variation.
(a) (b)
Figure 3‐270: Actual and predicted improvement ratios for (a) PLM, (b) EM
548
3.17.6 Conclusions
Although it is the belief of the author that actual field tests are the best means for verification
of ground improvement works, a calculation procedure has been proposed that can be used
as a tool for a first estimation of PLM and EM values after dynamic compaction. It is noted that:
1. In the proposed procedure it is assumed that PLM and EM double for every 3% of strain. It
is also assumed that the final induced strain in the ground has a Rayleigh distribution.
2. Eliminating the errors of correctly predicting the depth of improvement and ground
settlement, it was observed that the Rayleigh distribution is quite sensitive to the scale
parameter of the distribution, which is the depth at which strain is maximal. It has been
observed that the shallower this depth is the more strain is predicted. It appears that the
depth of maximum strain should be assumed to be about 0.45 to 0.5 times the depth of
improvement.
3. It has also been observed that the Rayleigh distribution initially underestimates the ratio
of PLM improvement; however, beyond the point of maximum ratio of improvement the
ratio is overestimated and at the lowest depths the ratio is once again underestimated.
Furthermore, in the PMTs that have been studied the predicted geometric limit pressures
after dynamic compaction were very close to reality.
4. Application of the proposed procedure was also implemented to PMTs inside DC prints
that did not have a sickle shaped improvement profile. Predicted harmonic means PLM in
the treatment depth were reliably close to measured harmonic mean PLM values. Back
calculated settlement based on measured EM was more than reality and predicted
geometric mean EM values were less than measured geometric mean values; however,
the profile shapes were still reasonably similar to reality. The biggest differences
between predicted and measured parameters occurred in PMTs that had the least
resemblance to the shape of the sickle.
549
4 Conclusion
Dynamic compaction and dynamic replacement were respectively patented by Louis Menard
in 1970 and 1975. Menard died in 1978, but his legacy was continued by his first generation
of engineers. It has been the author’s privilege to have been exposed to dynamic compaction
and dynamic replacement through the mentoring and guidance of Mr Serge Varaksin who is
the first engineer that was employed by Louis Menard and who is the world’s leading expert
on the subject.
It has come to the author’s attention that while dynamic compaction and dynamic
replacement have become popular ground improvement techniques due to cost, execution
speed, reliability and ability to meet specifications, information about these methods are
scattered, and to the knowledge of the author, no attempt has been made since the
comprehensive publications of Lukas (1986, 1995) to gather, compile and update previous
research and experience on the subject. Hence, Chapter 2 of this thesis was allocated to a
thorough and comprehensive literature review.
Chapter 3 focused on distinguished and sometimes globally renowned projects in which the
author was directly involved in or indirectly associated with. This provided the author with
the opportunity not only to record the approach, construction and quality control
methodology in these projects as reference for future studies, but also to compare the results
with previous research, and to develop new formulations that were previously not available.
In conclusion it can be noted that:
1) It is more likely that non‐engineered backfilling will be loose and potentially subject to
low bearing and excessive settlements. Experience in the projects that have been studied
in this thesis indicates that regardless of the filling process being on land or at sea, all
dumping methods, either tipped or hydraulically placed, result in a range of densities of
loose sand fills. It is recommended that ground improvement be perceived during
planning stage to avoid any surprises and disappointments when construction
commences.
2) Proper determination of design criteria is very important, and failure to adopt a suitable
specification can lead to unnecessary treatment, additional costs and delay.
550
3) The most suitable acceptance criteria are based on design criteria and actual project
requirements. It is not necessary to specify minimum test results as verification of
satisfying design criteria can be achieved without this limitation. One criterion does not
have to govern throughout the depth of treatment. In Al Quo’a the upper layers were
treated for bearing capacity and settlement under structural loads while the deeper soils
were treated for self‐bearing.
4) It was observed in Abu Dhabi New Corniche Road and Al Nakhilat Ship Repair Yard
projects that relative density is not a reliable acceptance criterion. It has been seen that
in practice fines content can exceed the acceptable limits defined for applicability of
relative density, which will result in the project becoming criteria‐less in those conditions.
Also, it was observed that it is possible to improve the ground more and provide better
results (higher bearing capacity and lesser settlements) without necessarily satisfying
relative density requirements. The ground works as a mass, not in individual layers, and
not satisfying the requirement at one level does not necessarily lead to inferior results.
5) Dynamic compaction is an affordable and effective ground improvement technique for
treating thick dry desert dune sands or saturated sands.
6) If the project size justifies, in an optimal dynamic compaction design the treatment area
should be broken down to sub areas based on design and acceptance criteria, loading,
and ground conditions. Pounder weight and drop height, compaction intensity, grid size,
and the other design parameters can consequently be determined. In large projects, it is
preferable to mobilise dynamic compaction rigs with different capacities to optimise the
relationship between equipment capacity and required impact energy.
7) Classical approaches where activities are clearly separated into preliminary works, design
and construction may not yield desirable results in special projects in which the site is
unusually large, preliminary field tests are most probably insufficient, and design and
construction period is relatively short. In these projects construction may have to
commence simultaneously with the preparation of architectural and structural drawings.
It is possible to successfully proceed in such projects if design and construction phases
are merged smartly and efficiently, a combination of ground improvement techniques
are used for treating different ground conditions rather than attempting to implement
one technique with its limitations to a variety of different conditions, possible production
rates are well understood and sufficient amounts of plant and equipment are mobilised,
551
and site observations are given value and changes in actual ground conditions are taken
into account during the process of the works.
8) The mass mechanical properties of sand layers can be further improved by introducing
dynamically compacted crushed stone inclusions.
9) Early planning will allow mobilisation of sufficient numbers of plant and equipment to
perform grand size projects in relatively short periods.
10) Energy intensity can be optimised based on the calibration programme results.
11) It is not necessary to compact the ground until reaching the asymptote of ground
settlement (or compaction) versus number of drops (or accumulative energy). What is
relevant is to demonstrate that design criteria (settlements, bearing capacity, etc) have
been satisfied.
12) It is possible to meet the project specifications in a shorter time by using heavy pounders
with high drops rather than using lighter pounders with the same energy intensity.
13) The combination of treatment methods; e.g. dynamic replacement and surcharging or
dynamic surcharging can expand the applicability limits of the ground improvement
techniques.
14) Pre‐excavation of soft saturated soils can be used to accelerate pore water pressure
dissipation and to allow more repetition of pounder dropping during each sub‐phase of
dynamic replacement before creating the state of liquefaction. At critical locations pre‐
excavation can be further adopted in the form of DR trenches.
15) By pre‐excavation of soft saturated soils and allowing accelerated dissipation of pore
water pressure it is possible to use sand for backfill material; otherwise pounder impact
will liquefy the sandy backfill material.
16) Dynamic surcharging can be used to accelerate the consolidation of saturated silty sands
and sandy silts. It is the author’s experience that dynamic surcharging is applicable to
silts, but not to clays. Additionally the degree of consolidation must be roughly in the
range of 50 to70%. This technique was used successfully in King Abdulla University of
Science and Technology when building loads were considerably increased during the
period at which ground improvement works were proceeding. In Palm Jumeira Sewage
552
Treatment Plant dynamic surcharging was used to increase induced foundation
settlement under static surcharge by 1.3 to 5 times depending on the distance from the
pounder impact point. Although the settlement magnitude of dynamic compaction was
much more than dynamic surcharging, the latter induced critical settlement at depths
that were treated less effectively with the allocated pounder.
17) It is possible to perform dynamic compaction at close vicinity to the shoreline. Dynamic
compaction was literally carried out to the edge of the reclamation in Abu Dhabi New
Corniche.
18) It is possible to efficiently improve the depth of influence in dynamic compaction by
implementation of the free falling and automatic grabbing MARS technology. This
technology was used in Al Quo’a New Township and Al Nakhilat Ship Repair Yard to drop
a 35 ton pounder.
19) To the knowledge of the author, the world record for the most number of DC and DR rigs
was achieved in King Abdulla University of Science and Technology with 13 rigs working
in double shifts.
20) To the knowledge of the author, at the time Al Falah Community Development was under
construction, it held the world records for largest (in size) single ground improvement
contract at 4.84 million m2 and highest production rate of 966,000 m2 of soil
improvement in one month.
21) Should the grid pattern be close enough, compaction of first phase prints will also
somewhat densify the ground influenced by the second phase prints. Consequently, it
may be possible to apply lesser amounts of energy in the second phase of compaction.
22) The efficiency of dynamic compaction rapidly deteriorates with the increase of fines
content to a point where dynamic compaction may no longer be able to satisfy
acceptance criteria, and the technique must be switched to dynamic replacement. It was
observed in King Abdulla University of Science and Technology that increase of fines
resulted in lesser improvement with the same amount of impact energy intensity, and it
appeared that the limit of efficient dynamic compaction is 30% fines content. In Abu
Dhabi Ritz‐Carlton Hotel, while PMT parameters of the soft layer increased after DR by
up to 100%, the improvement was not sufficient to satisfy acceptance criteria without
the introduction of compacted granular material in the form of DR columns. PLM
553
improvement inside the DR column was more than 900% the value of the original ground,
and 5.15 times more than the improvement in between the columns. A similar behaviour
was also observed at Al Jazira Steel Pipe Factory.
23) Pounder impacts create heave around the crater when the ground contains saturated
cohesive soils, but the periphery of the crater also settles (negative heave) when the soil
is granular.
24) Depth of improvement can be increased by applying DC blows to pre‐excavated prints.
25) Application of ironing phase of DC will improve the superficial ground parameters.
26) For marine DC, the barge must be large enough to safely support the personnel and
equipment. The barge size will be influenced by the location of the project and the sea
conditions.
27) Water resistance reduces the pounder’s dropping velocity, and consequently the impact
energy.
28) Offshore pounders must be designed to minimise water resistance. As water resistance
reduces impact energy, marine pounders are generally heavier than land pounders.
29) Suitable material (rock fill) must be placed on the seabed prior to commencement of
dynamic compaction works. In the Quay Expansion in Southeast Asia rock size ranged
from 150 to 300 mm.
30) Grid size and number of blows for marine application of DC and DR does not appear to
be very different from the parameters that would be used for land based DC and DR;
however, water resistance reduces impact efficiency, and equations developed for
estimation of depth of improvement in land based DC are not applicable to marine DC.
31) In Abu Dhabi New Corniche average PLM improvement ratios after dynamic compaction
were from 1.5 to 7. In Al Quo’a New Township it was observed that it is possible to
improve maximum PLM values of dune sands on average by up to 570% and at least in
one instance by up to 1950%. Improvement beyond 400% was observed over a thickness
of 20% of treatment depth. In Al Falah Community Development improvement ratio in
the calibration programme was in the order that has been suggested by Lukas (1986).
However, it could be cautiously expected that higher energy intensities will result in
554
better results. In Marjan Island Road Corridor PMT parameters’ improvement ratios were
in the order suggested by Lukas (1986). In Reem Island Causeway PLM improvement ratio
in the upper several metres of soil was 500 to 700%. The parameters’ improvements
were still substantial and in the range of 80 to 130% at depth. It is possible to achieve PLM
values greater than 2.4 MPa after dynamic compaction in saturated sands; however, the
peak value decreases with depth. Noting that the amount of improvement is a function
of impact energy, it is possible to improve PLM of saturated sands by 500%, but the peak
amount of improvement decreases with depth. The maximum amount of improvement
appears to be in the upper half of the depth of improvement. In HFO Tanks the maximum
percentage of improvement was much more than what is suggested by earlier literature
and observed values for other projects that have been studied in this thesis. PMT
parameters’ Improvement ratios were in the order of 40 and maximum PLM was in the
order of 5,000 MPa. In Palm Jumeira STP project, excluding the highest values, average
PLM and EM respectively ranged from approximately 20 to 40 MPa and from 23 to 30 MPa
at depths of about 4 m to 8 m. These values are significantly higher than what has been
suggested by Lukas (1986). Maximum improvement ratios were in the range of 10 to 18,
which are significantly higher than the range that has been suggested by Lukas. In
summary, it appears that while post improvement upper bound values of 1900 to 2400
kPa is a reasonable range, it is possible to exceed these values if sufficient energy is
applied. Likewise, while Lukas proposes that properties of soils will improve in the order
of 100 to 400% after dynamic compaction, it has been observed in the research
conducted in this thesis that it is possible to substantially exceed these figures.
32) SPT can result in misleading interpretations when the ground is composed of large size
particles.
33) It is possible to perform dynamic compaction in populated areas. DC works were carried
out in an urban area that was highly populated with high rise buildings throughout the
length of Abu Dhabi New Corniche, and a 35 ton pounder was dropped in free fall not far
away from town houses in Al Quo’a New Township.
34) Buildings that are at very close to dynamic compaction works may sustain damage not
only due to vibrations, but also from DC induced ground movements.
35) Dynamic compaction generated vibration frequencies were measured to be generally in
the range of 5 to 20 Hz.
555
36) Measured PPVs in the projects that were studied in this thesis were 3 to 5 times less than
what is estimated from Equation 2‐150 (Mayne, 1985). Equation 2‐152 also appeared to
be conservative, but was able to estimate more accurate results.
37) Equation 2‐153 (Varaksin, 1981) was able to predict reasonably close values to field‐
measured PPV; however, the equation sometimes underestimated the results.
38) Peak particle velocity can be estimated for more commonly used impact energies by
.
560 mm/s 2‐1533‐20
Or more generally by
.
√
25 2‐1513‐21
39) Equation 2‐1533‐20 suggests that without application of specific vibration reduction
techniques, it may be possible to carry out dynamic compaction as close as 25 to 35 m
from structures without exceeding the maximum PPV values that have been
recommended by USBM (Siskind et al., 1980).
40) Equation 2‐1513‐21 is able to predict PPV reasonably well and with sufficient accuracy;
however, it may underestimate PPV at distances less than 20 m. As both measured and
predicted vibrations within the underestimation zone are probably above recommended
tolerances and specific vibration reduction measures will have to be adapted, Equation
3‐21 can still be used for distances that are further away than 20 m and with caution and
as a first assessment at closer distances.
41) Vibration parameters have been monitored for different number of blows and distances
during several phases of dynamic compaction. It has been observed that, with the same
impact energy, peak particle velocities are greater during later phases of compaction as
compared to the premier phases. The differences are greatest at farther distances, and
it appears that when the distance is closer than a critical value, compaction phase
influence becomes unnoticeable.
42) Vibration frequency associated with PPV increases with distance.
556
43) Excavation of a trench that was 25 m long, 2.5 m deep (to groundwater level), 2.5 m wide,
and that was approximately 13 m away from a DC print was able to reduce PPV on
average to half of its value compared to when there was no trench. This experiment
suggests that isolation trenches can be an effective means for reducing dynamic
compaction vibrations.
44) A study carried out for comparative purposes on vibro compaction indicates that
maximum PPV can be expected to be highest when the vibroflot is closest to the ground
surface. In other words, the most critical time that vibroflot vibrations can damage a
nearby structure is when it is just penetrating the ground or being pulled out of the
treatment point.
45) Vibro compaction generated PPV can be estimated as a function of vibroflot power and
distance from Equation 3‐22:
.
7.5 mm/s 3‐22
√
46) While the magnitude of vibro compaction generated vibrations can be expected to be
less than dynamic compaction vibrations at the same distance, implementation of vibro
compaction can also exceed the proposed vibration limits (Siskind et al., 1980) at
distances that are less than 10 m.
47) The attenuation rate of dynamic compaction is less than vibro compaction. In other
words, dynamic compaction vibrations will reduce over longer distances.
48) At very close distances, both vibrations and ground deformation can result in cracking
and damage to buildings.
49) The analyses of the data from the projects that were studied in this thesis suggest that
there is a relationship between the square root of compaction volume and crater depth
after the first few pounder blows. In Marjan Island corridor this ratio was about 2.15 to
2.4 (in SI units). In Palm Jumeira STP the ratio was 2.26. The upper and lower bound
values of 2.15 and 2.75 in Equation 3‐11 resulted in 15% underestimation and 40%
overestimation in that case. A ratio of 2.5 resulted in 16% overestimation. In Al Nakhilat
Ship Repair Yard the ratio was about 2.11 (when only the crater volume is considered) to
557
2.32 (when the net compaction volume, being the resultant of the crater and periphery
volume changes is considered). In Reem Island Causeway the ratio was approximately
2.6 m after the first several blows. In HFO Tanks the ratio was measured to be
approximately 2.75. This figure was higher than the other projects that have been
studied in this thesis. In summary, it was observed that:
√
2.15 2.75 3‐11
50) In Al Nakhilat Ship Repair Yard it was observed that 90% and 10% of the compaction
volume respectively originated from the crater and the ground around the crater. In
Reem Island Causeway measurements suggest that 95% of the net compaction volume
originated from the crater, and in HFO Tanks project the crater volume was about 60 to
65% of net compaction volume.
51) There appears to be a linear relationship between the logarithm of accumulative impact
energy and (DT‐DB)/Dc. More research is required to determine if the relationship is
unique for a given soil type or whether it is also a function of the drop energy.
52) It was observed from Al Nakhilat Ship Repair Yard in Ras Laffan and Palm Jumeira Trial in
Dubai that the relationship between CPT cone resistance and PMT parameters was not
influenced by depth in saturated carbonate sands, and could be formulated by:
4.82 3‐13
3‐14
1.54
53) Silica sand correlations for Young modulus – cone resistance are not suitable for
carbonate sands. It was observed that for the saturated carbonate sands of Palm Jumeira
and Al Nakhilat Ship Repair Yard the relationship between CPT cone resistance and Young
modulus is:
3.12 3‐163‐16
558
54) Friction angle of rock fill can be estimated from PMT by:
∗
4 2 2‐2243‐12
55) A calculation procedure has been proposed that can be used as a tool for a first
estimation of PLM and EM values after dynamic compaction. It is noted that:
a) In the proposed procedure it is assumed that PLM and EM double for every 3% of
strain. It is also assumed that the final induced strain in the ground has a Rayleigh
distribution.
b) Eliminating the errors of correctly predicting the depth of improvement and ground
settlement (by adopting the actual depth of improvement), it was observed that the
Rayleigh distribution is quite sensitive to the scale parameter of the distribution,
which is the depth at which strain is maximal. It has been observed that the shallower
this depth is the more strain is predicted. It appears that the depth of maximum
strain should be assumed to be about 0.45 to 0.5 times the depth of improvement.
c) It has also been observed that the Rayleigh distribution initially underestimates the
ratio of PLM improvement; however, beyond the point of maximum ratio of
improvement the ratio is overestimated and at the lowest depths the ratio is once
again underestimated. Furthermore, in the PMTs that have been studied the
predicted geometric limit pressures after dynamic compaction were very close to
reality.
d) Application of the proposed procedure was also applied to PMTs inside DC prints that
did not have a sickle shaped improvement profile. Predicted harmonic means PLM in
the treatment depth were reliably close to measured harmonic mean PLM values.
Back calculated settlement based on measured EM was more than reality and
predicted geometric mean EM values were less than measured geometric mean
values. However, the profile shapes were still reasonably similar to reality. The
biggest differences between predicted and measured parameters occurred in PMTs
that had the least resemblance to the shape of the sickle.
559
References
Aboshi, H., Ichimoto, E. & Harada, K. (1979) The Compozer: A Method to Improve
Characteristics of Soft Clays by Inclusion of Large Diameter Sand Columns.
International Conference on Soil Reinforcement; Reinforced Earth and Other
Techniques, 2, Paris, 20‐22 March, 211–216.
Aboshi, H., Mizuno, Y. & Kuwabara, M. (1991) Present State of Sand Compaction Pile in Japan.
Deep Foundation Improvements: Design, Construction and Testing, ASTM STP 1089‐
EB, Philadelphia, 32‐46.
Abramson, L. W., Lee, T. S., Sharma, S. & Boyce, G. M. (2002) Slope Stability and Stabilization
Methods, 2nd Edition, John Wiley & Sons
Adalier, K. & Elgamal, A. (2004) Mitigation of Liquefaction and Associated Ground
Deformations by Stone Columns. Engineering Geology, 72, 3‐4, 275‐291.
Adam, D. & Brandl, H. (2009) Innovative Dynamic Compaction Techniques & Integrated
Compaction Control Methods. 17th International Conference on Soil Mechanics and
Geotechnical Engineering, Alexandria, Egypt, 5‐9 October, 2216‐2219.
Adam, D., Brandl, H., Kopf, F. & Paulmichl, I. (2007) Heavy Tamping Integrated Dynamic
Compaction Control. Ground Improvement, 11, 4, 237‐243.
Ahmad, S. & Al Hussaini, T. M. (1991) Simplified Design for Vibration Screening by Open and
in‐Filled Trenches. Journal of Geotechnical Engineering, ASCE, 117, 1, 67‐88.
Airey, D., Nazhat, Y., Moyle, R. A. & Avalle, D. L. (2012) Dynamic Compaction ‐ Insights from
Laboratory Tests. International Conference on Ground Improvement and Ground
Control ‐ Transport Infrastructure Development and Natural Hazards Mitigation
(ICGI2012), Wollongong, NSW, Australia, 30 October ‐ 2 November, 987‐996.
Akiyoshi, T., Fuchida, K., Matsumoto, H., Hyodo, T. & Fang, H. L. (1993) Liquefaction Analysis
of Sandy Ground Improved by Sand Compaction Piles. Soil Dynamics and Earthquake
Engineering, 12, 299‐307.
Al Hamoud, A. S. & Wehr, W. (2006) Experience of Vibrocompaction in Calcareous Sand of
Uae. Journal of Geotechnical and Geological Engineering, 24, 3, 757‐774.
Almeida, M. S. S., Jamiolkowski, M. & Peterson, R. W. (1992 ) Preliminary Result of CPT Tests
in Calcareous Quiou Sand. Calibration Chamber Testing: First International
Symposium on Calibration Chamber Testing (ISCCT1), Potsdam, NY, 28‐29 June 1991,
41‐53.
Amar, S., Clarke, B. G. F., Gambin, M. P. & Orr, T. L. L. (1991) The Application of Pressuremeter
Test Results to Foundation Design in Europe ‐ Part 1: Predrilled Pressuremeters/ Self‐
Boring Pressuremeters. International Society of Soil Mechanics and Foundation
Engineering, European Regional Technical Committee No. 4 ‐ Pressuremeter, 1‐23.
560
Amar, S. & Jezequel, J. F. (1972) Essais En Place Et En Laboratoire Sur Sols Coherents,
Comparaison Des Resultats. Bulletin de Liaison des Laboratoires des Ponts et
Chaussées, 58, March.
Anderegg, R. & Kaufmann, K. (2004) Intelligent Compaction with Vibratory Rollers ‐ Feedback
Control Systems in Automatic Compaction and Compaction Control. Transportation
Research Record: Journal of the Transportation Research Board, 1868, 124‐134.
Arslan, H., Baykal, G. & Ertas, O. (2007) Influence of Tamper Weight Shape on Dynamic
Compaction. Ground Improvement, 11, 2, 61‐66.
Asaoka, A. (1978) Observational Procedure of Settlement Prediction. Soils and Foundations,
Japanese Society of Soil Mechanics and Foundation Engineering, 18, 4 (December),
87‐101.
ASTM (1969) Test Method for Relative Density of Cohesionless Soils. D 2049‐69.
ASTM (2006a) Standard Test Methods for Maximum Index Density and Unit Weight of Soils
Using a Vibratory Table. D4253–00 (Reapproved 2006). 15.
ASTM (2006b) Standard Test Methods for Minimum Index Density and Unit Weight of Soils
and Calculation of Relative Density. D 4254–00 (Reapproved 2006). 9.
ASTM (2007) Standard Methods for Prebored Pressuremeter Testing in Soils. D 4719‐07.
Atherton, G. H., Polensek, A. & Corder, S. E. (1976) Human Response to Walking and Impact
Vibration of Wood Floors. Wood Product Journal, 26, 10 (October), 40‐47.
Avalle, D. L. (2004) Ground Improvement Using the Square Impact Roller ‐ Case Studies. 5th
International Conference on Ground Improvement Techniques, Kuala Lumpur, March
2004.
Baez, J. & Martin, G. R. (1993) Advances in the Design of Vibro Systems for the Improvement
of Liquefaction Resistance. 7th Annual Symposium of Ground Improvement,
Vancouver, 27‐28 May, 1‐16.
Baguelin, F., Jezequel, J. F. & Shields, D. H. (1978) The Pressuremeter and Foundation
Engineering, Aedermannsdorf, Trans Tech Publications,617.
Balaam, N. P. & Booker, J. R. (1981) Analysis of Rigid Rafts Supported by Granular Piles.
International Journal for Numerical and Analytical Methods in Geomechanics, 5, 4,
379‐403.
Baldi, G., Bellotti, V. N., Ghionna, N., Jamiolkowski, M. & Pasqualini, E. (1986) Interpretation
of Cpt's and Cptu's ‐ 2nd Part: Drained Penetration of Sands. 4th International
Geotechnical Seminar Field Instrumentation and In‐Situ Measurements, Nanyang
Technological Institute, Singapore, 25‐27 November 1986, 143‐ 156.
Banerjee, P. K. (1994) The Boundary Element Methods in Engineering, London and New York,
McGraw Hill Book Co
Barkan, D. D. (1962) Dynamics of Bases and Foundations (Translated from Russian by L
Drashevska and Edited by G P Tschbotarioff), New York, McGraw Hill Book Co,434.
561
Barksdale, R. D. & Bachus, R. C. (1983) Design and Construction of Stone Columns, Volume 1,
FHWA/Rd‐83/026. 194.
Bazaraa, A. R. (1967) Use of Standard Penetration Tests for Estimating Settlements of Shallow
Foundations on Sand. University of Illinois at Urbana,
Bely, L. D., Doudler, I. V., Mosiakov, E. F., Potapov, A. D. & Julin, A. N. (1975) Research
Methods and Evaluation of Various Genesis Sand Grain Morphology Role in
Formation of Their Geological‐Engineering Properties. Bulletin of International
Association of Engineering Geology, 11, 27‐31.
Bergado, D. & Long, P. V. (1994) Numerical Analysis of Embankment on Subsiding Ground
Improved by Vertical Drains and Granular Piles. 13th International Conference on Soil
Mechanics and Foundation Engineering, New Delhi, 1361‐1366.
Berry, A., Visser, A. & Rust, E. (2004) A Simple Method to Predict the Profile of Improvement
after Compaction Using Surface Settlement. International Symposium on Ground
Improvement, Paris, 9‐10 September 2004, 371‐386.
Berthier, D., Debats, J. M., Shcarff, G. & Vincent, P. (2009) Marine and Land Based
Compaction Works at the Port Botany Project, Sydney. Coasts and Ports 2009
Conference, Wellington, 16‐18 September.
Biot, M. A. (1956) Theory of Propagation of Elastic Waves in Fluid‐Saturated Porous Solids.
Journal of Acoustical Society of America, 28, March, 168‐191.
Bishop, R. F., Hill, R. & Mott, N. F. (1945) The Theory of Indentation and Hardness Tests. The
Proceedings of the Physical Society, 57, 321, Part 3, 147‐159.
Bo, M. W., Na, Y. M., Arulrajah, A. & Chang, M. F. (2009) Densification of Granular Soil by
Dynamic Compaction. Ground Improvement, 162, August, 121‐132.
Bolton, M. & Lau, C. K. (1993) Vertical Bearing Capacity Factors for Circular and Strip Footings
on Mohr‐Coulomb Soil. Canadian Geotechnical Journal, 30, 1024‐1033.
Bolton, M. D. & Gui, M. W. (1993) The Study of Relative Density and Boundary Effects for
Cone Penetration Tests in Centrifuge. Technical Report CUED/D‐Soils TR 256,
Cambridge University, 31.
Bornitz, G. (1931) Uber Die Ausbreitung Der Von Groszkolbernmaschinen Erzeugten
Bodenschwingungen in Die Tiefe, Berlin, J. Springer
Boulard, J. (1974) La Forme De Radoub Prefabriquee No. 10 Du Port Militaire De Brest.
Travaux, October, 17‐29.
Bowles, J. E. (1996) Foundation Analysis and Design, 5th Ed., New York, McGraw Hill,1175.
562
Bozbey, I. & Togrul, E. (2010) Correlation of Standard Penetration Test and Pressuremeter
Data: A Case Study from Istanbul, Turkey. Bulletin of Engineering Geology and the
Environment, 69, 4, 505‐515.
Briaud, J. L. (1992) The Pressuremeter, Rotterdam, A A Balkema,322.
Briaud, J. L. (2011) Email to Varaksin, S. Board Decision on "Louis Menard" Lecture.
Briaud, J. L., Noubani, A., Kilgore, J. & Tucker, L. M. (1985) Correlation between
Pressuremeter Data and Other Paramaters, Research Report. Texas A&M University,
Briaud, J. L. & Saez, D. (2012) Soil Compaction: Recent Developments. International
Conference on Ground Improvement and Ground Control ‐ Transport Infrastructure
Development and Natural Hazards Mitigation (ICGI2012), 1, Wollongong, NSW,
Australia, 30 October ‐ 2 November, 3‐30.
British Standards Institution (1990) Evaluation and Measurement for Vibration in Buildings.
Part 1. Guide for Measurement of Vibrations and Evaluation of Their Effects on
Buildings. BS 7385‐1:1990. London, British Standards Institution,
British Standards Institution (1992) Evaluation of Human Exposure to Vibration in Buildings
(1 Hz to 80 Hz). BS 6472: 1992. London, British Standards Institution,
British Standards Institution (1993) Evaluation and Measurement for Vibration in Buildings.
Part 2. Guide to Damage Levels from Groundborne Vibration. in INSTITUTION, B. S.
(Ed.) BS 7385‐2:1993. London, British Standards Institution,
British Standards Institution (1995) Strengthened‐Reinforced Soils and Other Fills. BS
8006:1995. London, British Standards Institute,
British Standards Institution (2009) Code of Practice for Noise and Vibration Control on
Construction and Open Sites. Part 2. Vibration. BS 5228‐2:2009. British Standards
Institution,
British Standards Institution (2010) Mechanical Vibration and Shock. Vibration of Fixed
Structures. Guidelines for the Measurement of Vibrations and Evaluation of Their
Effects on Structures BS ISO 4866:2010. London, British Standards Institution,
British Steel Piling (Bsp) Rapid Impact Compaction, Viewed 7 Feb 2009, http://www.bsp‐
if.com/RICS.html.
Bureau of Reclamation (1960) Earth Manual, Denver, US Department of the Interior
Bureau of Reclamation (1998) Earth Manual, Part 1, 3rd Edition, Denver, US Department of
the Interior,329.
Burmister, D. M. (1948) The Importance and Practical Uses of Relative Density in Soil
Mechanics. American Society for Testing and Materials, 48, Philadelphia, PA, USA,
1249‐1268.
563
Campanella, R. G., Berzins, W. E. & Shields, D. H. (1979) A Preliminary Evaluation of Menard
Pressuremeter, Cone Penetrometer and Standard Penetration Tests in the Lower
Mainland, British Colombia, Vancouver, University of British Colombia,49.
Cascante, G. & Santamarina, J. C. (1996) Interparticle Contact Behavior and Wave
Propogation. Journal of Geotechnical Engineering, ASCE, 122, 10, 831‐839.
Cassan, M. (1968) Les Essais in Situ En Mécanique Des Sols. Construction, 10 (October), 337‐
347.
Cassan, M. (1969) Les Essais in Situ En Mécanique Des Sols. Construction, 7‐8 (July ‐ August),
244‐256.
Castro, J. & Sagaseta, C. (2009) Consolidation around Stone Columns. Influence of Column
Deformation. International Journal for Numerical and Analytical Methods in
Geomechanics, 33, 7, 851–877.
Cattaneo, C. (1938) Sul Contatto Di Due Corpi Elastici. Accademia die Lincei Rendiconti, 27, 6,
342‐348, 434‐436, 474‐478.
Centre D'etudes Menard (1975) The Menard Pressuremeter, D60. Sols Soils, 26, 5‐43.
Chang, F. (1973) Human Response to Motion in Tall Buildings. Proceedings, ASCE, 99, 2, 1259‐
1272.
Chapot, P. & Et Al. (1981) Revue Francaise De Geotechnique (from Queyroi D, Chaput D, and
Pilot G: Amelioriation Des Sols De Fondation. Note Technique de Labratoire Central
des Ponts et Chaussées, 55.
Chaumeny, J. L., Hecht, T., Kirstein, J., Krings, M. & Lutz, B. (2008) Dynamic Consolidation for
the Intersection of an Active Sinkhole Area in the Course of the Federal Highway Bab
a 71. 4th Hans Lorenz Symposium, Berlin.
Chen, J. W. & Lin, C. Y. (2002) Effects of the Fines Content in Soil on Stress Attenuation During
Dynamic Compaction. 12th International Offshore and Polar Engineering Conference,
Kitakyushi, Japan, 26‐31 May 2002, 616‐621.
Chen, W. F. (1969) Soil Mechanics and Theorems of Limit Analysis. Journal of Soil Mechanics
and Foundations Division, ASCE, 95, 2, 493‐518.
Chen, W. F. (1975) Limit Analysis and Soil Plasticity, New York, Elsevier
Choa, V. (1980) Geotechnical Aspects of a Hydraulic Fill Reclamation Project. 6th Southeast
Asian Conference on Soil Engineering, Taiwan, 469‐484.
Choa, V., Karunartne, G. P., Ramaswamy, S. D., Vijaratnam, A. & Lee, S. L. (1979) Compaction
of Sand Fill at Changi Airport. 6th Asian Regional Conference on Soil Mechanics and
Foundation Engineering, Singapore, 137‐140.
Chow, Y. K., Yong, D. M., Yong, K. M. & Lee, S. L. (1994) Dynamic Compaction of Loose
Granular Soils: Effect of Print Spacing. Journal of Geotechnical Engineering, ASCE,
120, 7 (July), 1115‐1133.
564
Chu, J., Varaksin, S., Klotz, U. & Mengé, P. (2009) State of the Art Report: Construction
Processes. 17th International Conference on Soil Mechanics & Geotechnical
Engineering: TC17 meeting ground improvement, Alexandria, Egypt, 7 October 2009,
130.
Communication Department of Menard (2007) Menard ‐ Half a Century of History ‐ Un Demi
Siècle D'histoire, Mame à Tours,176.
Cooke, H. G. & Mitchell, J. A. (1999) Guide to Remedial Measures for Liqufaction Mitigation
at Existing Highway Bridge Sites, Mceer Technical Report No 99‐0015,216.
Cubrinovski, M. & Ishihara, K. (1999) Empirical Correlation between SPT N‐Value and Relative
Density for Sandy Soils. Soils and Foundations, Japanese Society of Soil Mechanics
and Foundation Engineering, 39, 5, 61‐71.
D'appolonia, D., Poulos, H. & Ladd, C. (1971) Initial Settlement of Structures on Clay. Journal
of Soil Mechanics and Foundations Division, ASCE, 97, SM10, 1359‐1377.
Dan, G., Sultan, N. & Savoye, B. (2007) The 1979 Nice Harbour Catastrophe Revisited: Trigger
Mechanism Inferred from Geotechnical Measurements and Numerical Modelling
Marine Geology, 245, 1‐4, 40‐64.
Dauvisis, J. P. & Menard, L. (1964) Etude Expérimentale Du Tassement Et De La Force
Portante De Fondations Superficielles. Sols Soils, 10, September, 11‐32.
Dayte, K. R. (1982) Note on Design Approach and Calculations for Stone Columns. Indian
Geotechnical Journal.
Denisov, N. Y. & Reltov, B. F. (1961) The Influence of Certain Processes on the Strength of
Soils. 5th International Conference on Soil Mechanics and Foundation Engineering, 1,
Paris, 17‐22 July, 75‐78.
Deresiewicz, H. (Ed.) (1974) In R D Mindlin and Applied Mechanics, New York.
Dggt (2004) Ebgeo: Bewehrte Erdkörper Auf Punkt‐ Oder Linienförmigen Traggliedern,
Chapter 6.9.,
Dimaggio, J. A. (1987) Stone Columns for Highway Construction. Demonstration Project No.
46. Report No. FHWA‐Dp‐46‐1, Federal Highway Administration,
Dimaggio, J. A. (2009) Technical Question Regarding Calculation of Equivalent Tangent of Soil
‐ Column. in HAMIDI, B. (Ed.).
Dobson, T. & Slocombe, B. (1982) Deep Densification of Granular Fills. 2nd Geotechnical
Conference, ASCE National Convention, Las Vegas, April 1982.
Dolan, J. (2011) Request for Report Dacw 38‐76‐M 6646 by Schmertmann. in HAMIDI, B. (Ed.).
ERDC Library,
Dowdall Stapleton, M. (2008) Helping to Create a New Map. Gulf News. Dubai,
Dowding, C. H. (2000) Construction Vibrations,
565
Duvall, W. I. & Fogelson, D. E. (1962) USBM Report of Investigation 5968 ‐ Review of Criteria
of Estimating Damage to Residences from Blasting Vibrations, US Bureau of Mines
Edwards, A. T. & Northwood, T. D. (1960) Experimental Studies of the Effects of Blasting on
Structures. The Engineer, 260, 30 September, 538‐546.
Enoki, E., Yagi, N., Yatabe, R. & Ichimoto, E. (1991) Shearing Characteristic of Composite
Ground and Its Application to Stability Analysis. Deep Foundation Improvements:
Design, Construction and Testing, ASTM STP 1089‐EB, Philadelphia, 19‐31.
Etude Pressiometrique Louis Menard (1970) Détermination De La Pousée Exercée Par Un Sol
Sur Une Pario De Soutènement, D38/70,
European Standard (2000) Dd Env 1997‐3:2000, Eurocode 7: Geotechnical Design ‐ Part 3:
Design Assisted by Field Testing 149.
European Standard (2004) En 1997‐1: 2004, Eurocode 7: Geotechnical Design ‐ Part 1:
General Rules 171.
Ewing, W. M., Jardetzky, W. S. & Press, F. (1957) Elastic Waves in Layered Media, New York,
McGraw Hill Book Co
Fei, X. Z., Wang, Z. & Zhou, Z. B. (2002) Model Test of Improvement Depth of Dynamic
Compaction. Chinese Journal Sichuan University, 34, 4, 56‐59.
Feng, T. W., Chen, K. H., Su, Y. T. & Shi, Y. C. (2000) Laboratory Investigation of Efficiency of
Conical‐Based Pounders for Dynamic Compaction. Geotechnique, 50, 6, 667‐674.
Feng, T. W. & Ke, C. C. (2005) A Study on Using Conical Bottom Tamper for Dynamic
Compaction in Platy Sand. 15th International Offshore and Polar Engineering
Conference, Seoul, 19‐24 June 2005, 674‐678.
Feng, T. W. & Yuan, C. (2009) Discussion: Influence of Tamper Weight Shape on Dynamic
Compaction by Arslan Et Al. Ground Improvement, 162, GI3, 153‐154.
Forssblad, L. (1980) Compaction Meter on Vibrating Rollers for Improved Compaction
Control. International Conference on Compaction, 2, Paris, 541‐546.
Gambin, M. (1979) Utilisation Du Module Pressiométrique Et De La Pression Limite Pour Le
Calcul Des Fondations. Sols Soils, 28, 14‐25.
Gambin, M. & Frank, R. (2009) Direct Design Rules for Piles Using Menard Pressuremeter
Test. International Foundation Congress and Equipment Expo: Contemporary Topics
in In Situ Testing, Analysis, and Reliability of Foundations, ASCE Geotechnical Special
Publication No 186, Orlando, Florida, 15‐19 March.
566
Gambin, M. P. (1982) Menard Dynamic Compaction, a New Method for Improving
Foundation Beds Off‐Shore. International Symposium, Brugge, 5‐7 May, 3.91‐93.95.
Gambin, M. P. (1983) The Menard Dynamic Consolidation at Nice Airport. 8th European
Conference on Soil Mechanics and Foundation Engineering, Helsinki, 231‐239.
Gambin, M. P. (1997) Le Compactage Profond Des Sables, Idées Des Base. 3rd International
Conference on Ground Improvement Geosystems: Densification and Reinforcement,
London, 3‐5 June 1997, 1‐27.
Gazetas, G. (1982) Vibrational Characteristics of Soil Deposits with Variable Wave Velocity.
International Journal for Numerical and Analytical Methods in Geomechanics, 6, 1‐
20.
Geoquip Geoquip Note: I‐01 ‐ What Defines an Impact Roller and What Is It's Purpose, Viewed
8 Feb 2009, http://www.geoquip.com.au/pub‐impact‐roller‐purpose.html.
Geoquip Geoquip Note: I‐02 ‐ Comparing Impact Rollers Viewed 8 Feb 2009,
http://www.geoquip.com.au/pub‐comparing‐impact‐rollers.html.
Ghassemi, A., Pak, A. & Shahir, H. (2009a) A Numerical Tool for Design of Dynamic
Compaction Treatment in Dry and Moist Sands. Iranian Journal of Science &
Technology, 33, B4, 313‐326.
Ghassemi, A., Pak, A. & Shahir, H. (2009b) Validity of Menard Relation in Dynamic
Compaction Operations. Ground Improvement, 162, 1 (February), 37‐45.
Gibbs, K. J. & Holtz, W. G. (1957) Research on Determining the Density of Sands by Spoon
Penetration Testing. 4th International Conference on Soil Mechanics and Foundation
Engineering, 1, London, 35‐39.
Gibson, R. E. & Anderson, W. F. (1961) In Situ Measurements of Soil Properties with the
Pressuremeter. Civil Engineering and Public Works Review, 56, 658.
Goh, S. H., Lee, F. H. & Tan, T. S. (1998) Effects of Lateral Constraints and Inertia on Stress
Wave Propagation in Dry Soil Columns. Geotechnique, 48, 4, 449‐463.
Goldman, D. E. (1948) A Review of Subjective Responses of Vibration Motions of the Human
Body in the Frequency Range, 1 to 70 Cycles Per Second, Naval Medical Research
Institute Report No. 1 Project Nm 004001,
Goughnour, R. R. & Bayuk, A. A. (1979) Analysis of Stone Column‐Soil Matrix Interaction
under Vertical Load. International Conference on Soil Reinforcement; Reinforced
Earth and Other Techniques (Colloque International sur le Renforcement des Sols), 2,
Paris, 20‐22 March, 271 – 278.
Goughnour, R. R. & Pestana, J. M. (1998) Mechanical Behavior of Stone Columns under
Seismic Loading. 2nd International Conference on Ground Improvement Techniques,
Singapore, October, 157‐162.
Greenwood, D. A. (1991) Load Tests on Stone Columns. Deep Foundation Improvements:
Design, Construction and Testing, ASTM STP 1089‐EB, Philedelphia, 148‐171.
567
Gu, Q. & Lee, F. H. (2002) Ground Response to Dynamic Compaction of Dry Sand.
Geotechnique, 52, 7, 481‐493.
Gupta, R. C. & Mckeown, J. D. (1973) Effect of Variations in Minimum Density on Relative
Density. Evaluation of Relative Density and its Role in Geotechnical Projects Involving
Cohesionless Soils: ASTM STP523‐EB.7744‐1, Los Angeles, 25‐30 June 1972, 85‐97.
Hajialilue‐Bonab, M. & Rezaei, A. H. (2009) Physical Modelling of Low‐Energy Dynamic
Compaction. International Journal of Physical Modelling in Geotechnics, 3, 21‐32.
Hajialilue‐Bonab, M. & Zare, F. S. (2014) Investigation on Tamping Spacing in Dynamic
Compaction Using Model Tests. Ground Improvement, 167, GI3, 219–231.
Haldar, A. & Tang, W. H. (1979) Uncertainty Analysis of Relative Density. Journal of
Geotechnical Engineering, ASCE, 107, 7 (July), 899‐904.
Hamidi, B., Varaksin, S. & Nikraz, H. (2010) Correlations between CPT and PMT at a Dynamic
Compaction Project. 2nd International Symposium on Cone Penetration Testing
(CPT10), Huntington Beach, California, 9‐11 May, paper 2‐04.
Harada, M. & Suzuki, M. (1984) Improvement of Sandy Soil and Measurement of Socking
Earth Pressure by Dynamic Consolidation (in Japanese). 16th Soil Engineering
Research Seminar, Tokyo, 1689‐1692.
Hardin, B. (1985) Crushing of Soil Particles. Journal of Geotechnical Engineering, ASCE, 111,
10 (October), 1177‐1192.
Hatanaka, M. & Feng, L. (2006) Estimating Relative Density of Sandy Soils. Soils and
Foundations, Japanese Society of Soil Mechanics and Foundation Engineering, 46, 3,
299‐313.
Hausler, E. A. (2002) Influence of Ground Improvement on Settlement and Liquefaction: A
Study Based on Field Case History Evidence and Dynamic Geotechnical Centrifuge
Tests. University of California, Berkley, 364.
Heh, K. S. (1990) Dynamic Compaction of Sand. Department of Civile Environmental
Engineering. Brooklyn, NY, Polytechnic University,
Hendy, M. S. & Muir, I. C. (1997) Experience of Dynamic Replacement on a 40 M Deep
Reclamation in Hong Kong. Third International Conference on Ground Improvement
Geosystems: Ground Improvement Geosystems ‐ Densification and Reinforcement,
London, 3‐5 June 1997, 76‐80.
Hertz, H. (1881) Über Die Berührung Fester Elastischer Körper. Journal Reine U. Angew Math.,
92, 156‐171.
Hewlett, W. J. & Randolph, M. (1988) Analysis of Piled Embankments. Ground Engineering,
21, 3, 12‐18.
Hill, R. (1950) The Mathematical Theory of Plasticity, Oxford, University Press
Hiller, D. M. & Hope, V. S. (1998) Ground Borne Vibration Generated by Mechanized
Construction Activities. Geotechnical Engineering, 131, 223‐232.
568
Hirona, T. (1948) Mathematical Theory on Shallow Earthquake. The Geophysical Magazine,
18, 1‐4, October.
Hobbs, N. B. & Dixon, J. C. (1969) In‐Situ Testing for Bridge Foundations in the Devonian Marl.
Conference on In Situ Investigations in Soils and Rocks, London, 13‐15 May, 31‐38.
Holtz, R. D. (1973a) Discussion on Determination of Relative Density of Sand Below
Groundwater Table. Evaluation of Relative Density and its Role in Geotechnical
Projects Involving Cohesionless Soils: ASTM STP523‐EB.7744‐1, Los Angeles, 25‐30
June 1972, 376‐377.
Holtz, W. G. (1973b) The Relative Density Approach ‐ Uses, Testing Requirements, Reliability,
and Shortcomings. Evaluation of Relative Density and its Role in Geotechnical Projects
Involving Cohesionless Soils: ASTM STP523‐EB.7744‐1, Los Angeles, 25‐30 June 1972,
5‐17.
Holubec, I. & D'appolonia, E. (1973) Effect of Particle Shape on the Engineering Properties of
Granular Soils. Evaluation of Relative Density and its Role in Geotechnical Projects
Involving Cohesionless Soils: ASTM STP523‐EB.7744‐1, Los Angeles, 25‐30 June 1972,
304‐318.
Hughes, J. M. & Withers, N. J. (1974) Reinforcing of Soft Cohesive Soils with Stone Columns.
Ground Engineering, 47, 3 (May), 42‐49.
Hwang, J. H. & Tu, T. Y. (2003) Ground Vibration Due to Dynamic Compaction. 13th
International Offshore and Polar Engineering Conference, Honolulu, 25‐30 May 2003,
490‐497.
Hwang, J. H. & Tu, T. Y. (2006) Ground Vibration Due to Dynamic Compaction. Soil Dynamics
and Earthquake Engineering, 26, 337‐346.
Iai, S., Koizumi, K. & Kurata, E. (1987) Basic Consideration for Designing the Area of the
Ground as a Remedial Measure against Liquefaction (in Japanese). 590, PHRI, 2‐67.
Ichese, Y., Yamakoda, A. & Takano, S. (1971) High Pressure Jet‐Grouting Method. in UNITED
STATES PATENT AND TRADEMARK OFFICE (Ed.). United States, 7.
Ishihara, K. (1985) Stability of Natural Deposits During Earthquakes. 11th International
Conference on Soil Mechanics and Foundation Engineering, 1, San Francisco, 321‐
376.
Ishihara, K. & Yamazaki, F. (1980) Cyclic Simple Shear Tests on Saturated Sand in Multi‐
Directional Loading. Soils and Foundations, Japanese Society of Soil Mechanics and
Foundation Engineering, 20, 1, 49‐59.
Jahangiri, G., Pak, A. & Ghassemi, A. (2011) Numerical Modelling of Dynamic Compaction in
Dry Sandy Soils for Determination of Effective Print Spacing. Journal of Structural
Engineering and Geo‐techniques, 1, 1‐9.
Jamiolkowski, M., Lo Presti, D. C. F. & Manassero, M. (2001) Evaluation of Relative Density
and Shear Strength of Sands from CPT and Dmt. Soil Behavior and Soft Ground
Construction: Geotechnical Special Publication No. 119, Boston, Massachusetts, 5‐6
October, 201‐238.
569
Jamiolkowski, M. & Pasqualini, E. (1992) Compaction of Granular Soils ‐ Remarks on Quality
Control. Grouting, Soil Improvement and Geosynthetics: ASCE Geotechnical Special
Publication No. 30, 2, New Orleans, 25‐28 February, 902‐914.
Japanese Geotechnical Society (1998) Remedial Measures against Soil Liquefaction, from
Investigation and Design to Implementation, Rotterdam, A A Balkema,443.
Jezequel, J. F., Lemasson, H. & Touzé, J. (1968) Le Pressiomètre Louis Ménard: Quelques
Problèmes De Mise En Oeuvre Et Leur Influence Sur Les Valeurs Pressiométriques.
Bulletin de Liaison des Laboratoires des Ponts et Chaussées, 32 (June ‐ July), 97‐120.
Jullienne, D. (2008) Ras Laffan Port Expansion ‐ Dc Presentation. Ras Laffan,
Kelly, D. & Gil, J. (2012) Monitoring Heic Using Landpac Cir and Cis Technologies. International
Symposium on Ground Improvement (IS‐GI) Brussels 2012, 1, Brussels, 31 May ‐ 1
June, 273‐285.
Kempfert, H. G. & Gebreselassie, B. (2006) Chapter 7 Soil Stabilisation with Column Like
Elements. Excavations and Foundations in Soft Soils. Berlin, Springer.
Kempfert, H. G., Gobel, C., Alexiew, D. & Heitz, C. (2004) German Recommendations for
Reinforced Embankments on Pile Similar Elements. EuroGeo3: 3rd European
Geosynthetics Conference, Geotechnical Engineering with Geosynthetics, Munich,
279‐284.
Kopf, F. & Paulmichl, I. (2004) Deep Dynamic Compaction; Compaction Control Using
Dynamic Measurements. Institute for Soil Mechanics and Geotechnical Engineering,
Research Report (in German). Vienna University of Technology,
Kopf, F., Paulmichl, I. & Adam, D. (2010) Modelling and Simulation of Heavy Tamping Dynamic
Response of the Ground. 14th Danube‐European Conference on Geotechnical
Engineering: From Research to Design in European Practice, Bratislava, Slovak, June
2‐4.
Kröber, W., Floss, E. & Wallrath, W. (2001) Dynamic Soil Stiffness as Quality Criterion for Soil
Compaction. in GOMES CORREIRA, A. & BRANDL, H. (Eds.) Geotechnics for Roads, Rail
Tracks and Earth Structures. A.A. Balkema.
Lacroix, Y. & Horns, H. M. (1973) Direct Determination and Indirect Evaluation of Relative
Density and Its Use on Earthwork Construction Projects. Evaluation of Relative
Density and its Role in Geotechnical Projects Involving Cohesionless Soils: ASTM
STP523‐EB.7744‐1, Los Angeles, 25‐30 June 1972, 251‐280.
Lamb, H. (1904) On the Propagation of Tremors over the Surface of an Elastic Solid.
Philosophical Transactions of the Royal Society, London, 203, Ser. A, 1‐42.
Lambe, T. W. & Whitman, R. V. (1979) Soil Mechanics, Si Version, New York, Wiley
Langefors, U., Kilhstrom, B. & Westerberg, H. (1958) Ground Vibrations in Blasting. Water
Power, February, 335‐338, 390‐395, 421‐424.
570
Leblanc, J. (1982) Menard Pressuremter Testing. Symposium on the Pressuremter and its
Marine Applications, Paris, 19‐20 April 1982, 23‐37.
Lee, F. H. & Gu, Q. (2004) Method for Estimating Dynamic Compaction Effect on Sand. Journal
of Geotechnical and Geoenvironmental Engineering, ASCE, 113, 2 (February), 139‐
152.
Lee, J. & Salgado, R. (2002) Estimation of Footing Settlement in Sand. International Journal
of Geomechanics, ASCE, 2, 1, 1‐28.
Lee, K. L. & Albeisa, A. (1974) Earthquake Induced Settlement in Saturated Sands. Journal of
Soil Mechanics and Foundations Division, ASCE, 100, 4, 387‐406.
Lee, K. M. (2001) Influence of Placement Method on the Cone Penetration Resistance of
Hydraulically Placed Sand Fills. Canadian Geotechnical Journal, 38, 9, 592‐607.
Lee, K. M., Shen, C. K., Leung, D. H. K. & Mitchell, J. K. (1999) Effects of Placement Method on
Geotechnical Behavior of Hydraulic Fill Sands. Journal of Geotechnical and
Geoenvironmental Engineering, ASCE, 125, 10, 832‐846.
Lee, K. M., Shen, C. K., Leung, D. H. K. & Mitchell, J. K. (2000) Closure: Effects of Placement
Method on Geotechnical Behavior of Hydraulic Fill Sands. Journal of Geotechnical
and Geoenvironmental Engineering, ASCE, 126, 10, 943‐944.
Leonards, G. A., Cutter, W. A. & Holtz, R. D. (1980) Dynamic Compaction of Granular Soils.
Journal of Geotechnical Engineering, ASCE, 106, 1 (January), 35‐44.
Li, D. (2005) Transition of Railroad Bridge Approaches. Journal of Geotechnical and
Geoenvironmental Engineering, ASCE, 131, 11 (November), 1392‐1398.
Liao, S. & Sangrey, D. A. (1978) Use of Piles as Isolation Barriers. Journal of Geotechnical
Engineering, ASCE, 104, 9, 1139‐1152.
Liu, L. & Dorby, R. (1997) Seismic Response of Shallow Foundation on Liquefiable Sand.
Journal of Geotechnical and Geoenvironmental Engineering, ASCE, 123, 6, 557‐567.
Liu, Q. B. & Lehane, B. M. (2012) The Influence of Particle Shape on the (Centrifuge) Cone
Penetration Test (CPT). Geotechnique, 62, 11, 973‐984.
Lukas, R. G. (1980) Densification of Loose Deposits by Pounding. Journal of Geotechnical
Engineering, ASCE, 106, 4 (April), 435‐446.
Lukas, R. G. (1986) Dynamic Compaction for Highway Construction, Volume 1: Design and
Construction Guidelines, FHWA Report Rd‐86/133. Federal Highway Administration,
Lukas, R. G. (1995) Geotechnical Engineering Circular No. 1: Dynamic Compaction,
Publication No. FHWA‐Sa‐95‐037. Federal Highway Administration,
Luongo, V. (1992) Dynamic Compaction: Predicting Depth of Improvement. Grouting, Soil
Improvement and Geosynthetics: ASCE Geotechnical Special Publication No. 30, 2,
New Orleans, 25‐28 February, 927‐939.
Lysmer, J. (1966) Personal Communication with F E Richart.
571
Madhav, M. R. & Vitkar, P. P. (1978) Strip Footings on Weak Clay Stabilized with a Granular
Trench or Pile. Canadian Geotechnical Journal, 15, 4, 605‐609.
Marcuson Iii, W. F. (1978) Determination of in Situ Density of Sands. Dynamic Geotechnical
Testing, ASTM STP 654, Denver, 28 June 1977, 318‐340.
Marcuson Iii, W. F. & Bieganousky, W. A. (1977) Laboratory Standard Penetration Tests on
Fine Sands. Journal of Geotechnical Engineering, ASCE, 103, 6 (June), 565‐589.
Marston, A. & Anderson, A. O. (1913) The Theory of Loads on Pipes in Ditches and Tests of
Cement and Clay Drain Tile and Sewer Pipe. Iowa Engineering Experiment Station
Bulletin, 31.
Massarsch, K. R. & Fellenius, B. H. (2012) Early Swedish Contributions to Geotechnical
Engineering. Full‐Scale Testing and Foundation Design: Honoring Bengt H. Fellenius,
ASCE Geotechnical Special Publication No. 227, Reston, VA, USA, 25‐29 March, 239‐
256.
Mayne, P. W. (1985) Ground Vibrations During Dynamic Compaction. Symposium on
Vibration Problems in Geotechnical Engineering: ASCE Special Publication, Detroit, 22
October 1985, 247‐265.
Mayne, P. W. (2006) The Second James K Mitchell Lecture: Undisturbed Sand Strength from
Seismic Cone Tests. Geomechanics and Geoengineering: An International Journal, 1,
4, 239‐257.
Mayne, P. W., Jones, J. S. & Dumas, J. C. (1984) Ground Response to Dynamic Compaction.
Journal of Geotechnical Engineering, ASCE, 110, 6 (June), 757‐774.
Mayne, P. W. & Kulhawy, F. H. (1982) Ko‐Ocr Relationship in Soil. Journal of Geotechnical
Engineering, ASCE, 108, 6, 851‐872.
Mccamy, K., Meyer, R. P. & Smith, T. J. (1962) Generally Applicable Solutions of Zoeppritz'
Amplitude Equations. Bulletin of the Seismological Society of America, 52, 4
(October), 923‐955.
Menard, L. (1957) Msc Thesis, an Apparatus for Measuring the Strength of Soils in Place. Civil
Engineering. Urbana, Il, USA, University of Illinois, 50.
Menard, L. (1963a) Calcul De La Force Portante Des Fondation Sur La Base Des Resultats Des
Essais Pressiometriques. Sols Soils, 5, 9‐32.
Menard, L. (1963b) Tendances Nouvelles En Mécanique Des Sols. Bulletin de l'A.I.A., 1, 1‐12.
Menard, L. (1965) Rules to Obtain Bearing Capacity and Foundations Settlements from
Pressuremeter Results (in French). 6th International Conference on Soil Mechanics
and Foundation Engineering, Montreal.
Menard, L. (1971) A Low Cost Method of Consolidating Fills Dumped into the Sea. Sols Soils,
24, 9‐20.
Menard, L. (1972) La Consolidation Dynamique Des Remblais Recents Et Sols Compressibles.
Travaux, November, 56‐60.
572
Menard, L. (1974) Fondation D'une Cale De Radoub À Brest. 6th International Harbour
Conference, Antwerp, 12‐18 May.
Menard, L. (1978) La Consolidation Dynamique Comme Solution Aux Problemes De
Fondation: Pour La Construction De Quais, Terminaux, Reservoirs De Stockage Et Iles
Artificielles Sur Sols Compressibles. 7th International Harbour Conference, Antwerp.
Menard, L. (1981) L'utilisation De La Consolidation Dynamique Pour La Réalisation Du
Nouveau Port De Pêche De Sfax En Tunisie. Navires, Ports et Chantiers, April.
Menard, L., Bourdon, G. & Houy, A. (1964) Etude Expérimentale De L'encastrement D1un
Rideau En Fonction Des Caractéristiques Pressiométriques Du Sol De Fondation. Sols
Soils, 3, 9, 11‐41.
Menard, L. & Broise, Y. (1975) Theoretical and Practical Aspects of Dynamic Compaction.
Geotechnique, 25, 1, 3‐18.
Menard, L. & Lambert, P. (1966) Etude Expérimentale D'un Massif De Fondation Soumis À
Des Vibrations. Sols Soils, 17, 9‐30.
Menard, L. & Rousseau, J. (1962) L'évaluation Des Tassements ‐ Tendances Nouvelles. Sols
Soils, 1, 1, 13‐29.
Menard Soltraitement Soil Improvement Specialists ‐ Design, Construct ... Perform,
Mesri, G., Feng, T. W. & Benak, J. M. (1990) Postdensification Penetration Resistance of Clean
Sands. Journal of Geotechnical Engineering, ASCE, 116, 7, 1095‐1115.
Meyerhof, G. G. (1957) Discussion on Research on Determining the Density of Sands by Spoon
Penetration Testing. 4th International Conference on Soil Mechanics and Foundation
Engineering, 1, London, 110.
Meyerhof, G. G. (1959) Compaction under Impact. Journal of Soil Mechanics and Foundations
Division, ASCE, 85, 4, 1292‐1323.
Meyerhof, G. G. (1976) Bearing Capacity and Settlement of Pile Foundations. Journal of
Geotechnical Engineering, ASCE, 102, 1, 197‐259.
Michalowski, R. L. & Nadukuru, S. S. (2011a) Delayed Increase in Cone Penetration Resistance
of Sand after Dynamic Compaction. 2nd International Symposium on Computational
Geomechanics: Com Geo II, Dubrovnik, Croatia, 251‐261.
Michalowski, R. L. & Nadukuru, S. S. (2011b) Stress Corrosion Cracking and Delayed Increase
in Penetration Resistance after Dynamic Compaction of Sand. Geo‐Frontiers 2011:
Advances in Geotechnical Engineering, ASCE Geotechnical Special Publication No 211,
Dallas, 13‐16 March, 519‐528.
573
Michalowski, R. L. & Nadukuru, S. S. (2012) Static Fatigue, Time Effects, and Delayed Increase
in Penetration Resistance after Dynamic Compaction of Sands. Journal of
Geotechnical and Geoenvironmental Engineering, ASCE, 138, 5, 564‐574.
Miller, G. F. & Pursey, H. (1954) The Field and Radiation Impedance of Mechanical Radiators
on the Free Surface of a Semi‐Infinite Isotropic Solid. Royal Society, 223, London,
521‐541.
Miller, G. F. & Pursey, H. (1955) On the Partition of Energy between Elastic Waves in a Semi‐
Infinite Solid. Royal Society, 233, London, 55‐69.
Mindlin, R. D. (1949) Compliance of Elastic Bodies in Contact. Journal of Applied Mechanics,
Transactions ASME, 71, 259‐268.
Mindlin, R. D. & Deresiewicz, H. (1953) Elastic Spheres in Contact under Varying Oblique
Forces. Journal of Applied Mechanics, Transactions ASME, September, 327‐344.
Mitchell, J. K. (1981) Soil Improvement State‐of‐the‐Art Report. 10th International
Conference on Soil Mechanics and Foundation Engineering, 4, Stockholm, 509‐565.
Mitchell, J. K., Baxter, C. & Munson, T. (1995) Performance of Improved Ground During
Earthquake. Soil Improvement for Earthquake Hazard Mitigation: ASCE Geotechnical
Special Publication No. 49, 1‐36.
Mitchell, J. K. & Gardner, W. S. (1975) State of the Art: In Situ Measurement of Volume
Change Characteristics. ASCE Conference on In situ Measurements of Soil Properties,
2, Raleigh, North Carolina, 279‐345.
Mitchell, J. K. & Huber, T. R. (1985) Performance of a Stone Column Foundation. Journal of
Geotechnical Engineering, ASCE, 111, 2 (Feb), 205‐223.
Mitchell, J. K. & Solymar, Z. V. (1984) Time Dependent Strength Gain in Freshly Deposited or
Densified Soil. Journal of Geotechnical Engineering, ASCE, 110, 11 (November), 1559‐
1576.
Mobil (1990) Chapter 6 ‐ Foundation Design. Mobil Engineering Guide: Egs 262‐90. Mobil.
Modoni, G., Croce, P. & Mongiovi, L. (2006) Theoretical Modelling of Jet Grouting.
Geotechnique, 56, 5, 335‐347.
Moretto, O. (1954) Subsoil Investigation for a Bridge over the Parana River, Argentina.
Geotechnique, 4, 4, 137‐142.
Mostafa, K. (2010) Phd Thesis, Numerical Modeling of Dynamic Compaction in Cohesive Soils.
Akron, University of Akron, 182.
Müller, H. (1970) Baugrunduntersuchung Mit Dem Pressiometerverfahren Nach Menard. Die
Bautechnik, 9, 289‐295.
Murali Krishna, A. & Madhav, M. R. (2009) Engineering of Ground for Liquefaction Mitigation
Using Granular Column Inclusions: Recent Developments. American Journal of
Engineering and Applied Sciences, 2, 3, 526‐536.
574
Murayama, S. (1962) Vibro‐Compozer Method for Clayey Ground (in Japanese).
Mechanization of Construction Work, 150, 10‐15.
Na, Y. M., Choa, V., Teh, C. I. & Chang, M. F. (2005) Geotechnical Parameters of Reclaimed
Sandfill from Cone Penetration Test. Canadian Geotechnical Journal, 42, 91‐109.
Nagase, H. & Ishihara, K. (1988) Liquefaction Induced Compaction and Settlement of Sand
During Earthquakes. Soils and Foundations, Japanese Society of Soil Mechanics and
Foundation Engineering, 28, 1, 66‐76.
Nashed, R. (2005) Liquefaction Mitigation of Silty Soils Using Dynamic Compaction.
Department of Civil, Structural & Environmental Engineering. Buffalo, State
University of New York, 282.
Naval Facilities Engineering Command (1997) Soil Dynamics and Spectral Design Aspects,
Norfolk, Department of Defense
Navdocs (1962) Design Manual Dm‐7, Soil Mechanics, Foundations and Earth Structures.
Nazaret, M. (1972) Influence Du Mode De Mise En Oeuvre De La Sond Pressiométrique,
Rapport De Recherche Du Laboratoire Régional Des Ponts Et Chaussées D'angers,
F.A.E.R 1.05.11.1.
Nicholls, H. R., Johnson, C. F. & Duvall, W. I. (1971) USBM Bulletin 656 ‐ Blasting Vibrations
and Their Effects on Structures, US Bureau of Mines
Office of Surface Mining Reclamation Enforcement (1983) Federal Register. in OFFICE OF
SURFACE MINING RECLAMATION AND ENFORCEMENT (Ed.). Department of the
Interior, 9805‐9811.
Oshima, A. & Takada, N. (1994) Effect of Ram Momentum on Compaction by Heavy Tamping.
13th International Conference on Soil Mechanics and Foundation Engineering, New
Delhi, 5‐10 January, 1141‐1144.
Oshima, A. & Takada, N. (1997) Relation between Compacted Area and Ram Momentum by
Heavy Tamping. 14th International Conference on Soil Mechanics and Foundation
Engineering, 1641‐1644.
Oshima, A. & Takada, N. (1998) Evaluation of Compacted Area of Heavy Tamping by Cone
Point Resistance. Centrifuge 98, Tokyo, 813‐818.
Osterberg, J. O. & Varaksin, S. (1973) Determination of Relative Density of Sand Below
Groundwater Table. Evaluation of Relative Density and its Role in Geotechnical
Projects Involving Cohesionless Soils: ASTM STP523‐EB.7744‐1, Los Angeles, 25‐30
June 1972, 364‐378.
Parkin, A. K. (1977) The Friction Cone Penetrometer: Laboratory Calibration for the Prediction
of Sand Properties. Internal Report No. 52108‐5, Norwegian Geotechnical Institute,
Parkin, A. K. & Lunne, T. (1982) Boundary Effects in the Laboratory Calibration of a Cone
Penetrometer for Sand. 2nd European Symposium on Penetration Testing, 2,
Amsterdam, 761‐767.
575
Peck, R. B. & Bazaraa, A. R. (1969) Discussion of Settlement of Spread Footings on Sand.
Journal of Soil Mechanics and Foundations Division, ASCE, 95, 3, 905‐909.
Perucho, A. & Olalla, C. (2006) Dynamic Consolidation of a Saturated Plastic Clayey Fill.
Ground Improvement, 10, 2, 55‐68.
Pettijohn, F. J., Potter, P. E. & Sievin, R. (1972) Sand and Sandstone, New York, Springer‐
Verlag
Poran, C. J., Heh, K. S. & Rodriguez, J. A. (1991) Impact Response to Granular Soils. 2nd
International Conference on Recent Advances in Geotechnical Earthquake
Engineering and Soil Dynamics, 2, St Louis, MO, USA, 1387‐1398.
Poran, C. J. & Rodriguez, J. A. (1992) Design of Dynamic Compaction. Canadian Geotechnical
Journal, 29, 5, 796‐802.
Port and Harbour Research Institute (1997) Handbook of Liquefaction Remediation of
Reclaimed Land, Rotterdam, A A Balkema
Power, D. V. (1966) A Study of Complaints of Seismic Related Damage to Surface Structures
Following the Salmon Underground Nuclear Detonation. Bulletin of the Seismological
Society of America, 56, 6 (December), 1413‐1428.
Prandtl, L. (1920) Uber Die Härte Plastischer Körper Nachrichten Von Der Königlichen
Gesellschaft Der Wissenschaften, Gottingen, Math.‐ Phys. Klasse,74‐85.
Prange, B. (1977) Parameters Affecting Damping Properties. Dynamical Methods in Soil and
Rock Mechanics, 1 Dynamic Response and Wave Propagation in Soils, Karlsruhe, 5‐
16 September 1977, 6‐78.
Raithel, M., Kirchner, A. & Kempfert, H. G. (2008) German Recommendations for Reinforced
Embankments on Pile‐Similar Elements. 4th Asian Regional Conference on
Geosynthetics, Shanghai, 17‐20 June, 697‐702.
Rayleigh (1885) On Waves Propagated Along the Plane Surface of an Elastic Solid. London
Mathematical Society, 17, 4‐11.
Reiher, H. & Meister, F. J. (1931) The Effect of Vibration on People. Forschung auf dem
Gebeite des Ingenieurwesens, 2, 11, 381.
Renault, J. & Tourneur, P. (1974) La Forme De Radoub No. 10 À Brest. 6th International
Harbor Conference, Antwerp, 12‐18 May.
Richart, F. E. (1962) Foundation Vibrations. Transactions, ASCE, 127, Part 1, 863‐898.
Richart, F. E., Hall, J. R. & Woods, R. D. (1970) Vibrations of Soils and Foundations, Prentice‐
Hall
Robertson, P. K. & Campanella, R. (1985) Liquefaction Potential of Sands Using the CPT.
Journal of Geotechnical Engineering, ASCE, 111, 3, 384–403.
576
Robertson, P. K. & Campanella, R. G. (1989) Guidelines for Geotechnical Design Using the
Cone Penetrometer Test and CPT with Pore Pressure Measurement, 4th Ed.,
Columbia, MD, Hogentogler & Co.
Rogers, J. D. (2004) Notes for Standard Penetration Test, Viewed 23/11/2010,
http://web.mst.edu/~rogersda/umrcourses/ge441/NOTES%20for%20STANDARD%
20PENETRATION%20TEST.pdf.
Rollins, K. & Kim, J. (2010) Dynamic Compaction of Collapsible Soils Based on Us Case
Histories. Journal of Geotechnical and Geoenvironmental Engineering, ASCE, 136, 9,
1178‐1186.
Rollins, K. M. & Kim, J. H. (1994) Us Experience with Dynamic Compaction of Collapsible Soils.
In‐Situ Deep Soil Improvement: ASCE Geotechnical Special Publication No. 45, Atlanta,
9‐13 October 1994, 26‐43.
Romana Ruiz, M. & Jurado, C. (1999) Vibration Monitoring of a Dynamic Compaction on a Fill
at Santa Cruz De La Palma Harbour (Canary Island). 12th European Conference on Soil
Mechanics and Geotechnical Engineering: Geotechnical Engineering for
Transportation Infrastructure, 591‐600.
Salencon, J. (1966) Expansion Quasi ‐Statique D'une Cavite a Symetric Spherique Ou
Cylindrique Dans Un Milieu Elasto‐Plastique. Annales des Ponts et Chaussées, 3, May‐
June, 175‐187.
Sandström, A. (1994) Numerical Simulation of a Vibratory Roller on Cohesionless Soil, Internal
Report. Geodynamik,
Satibi, S. (2009) Numerical Analysis and Design Criteria of Embankments on Floating Piles,
Stuttgart, DCC Siegmar Kästl e. K., Ostfildern,166.
Scherocman, J., Rakowski, S. & Uchiyama, K. (2007) Intelligent Compaction, Does It Exist?
2007 Canadian Technical Asphalt Association (CTAA) Conference, Victoria, BC,
Canada, July.
Schmertmann, J. H. (1970) Static Cone to Compute Static Settlement over Sand. Journal of
Geotechnical Engineering, ASCE, 96, SM3 (May) Interpretation of Cone Penetration
Tests. Part I: Sand, 1101‐1143.
Schmertmann, J. H. (1975) State of the Art Paper: Measure of in Situ Strength. ASCE
Conference on In situ Measurements of Soil Properties, 2, Raleigh, North Carolina,
57‐138.
Schmertmann, J. H. (1976) An Updated Correlation between Relative Density and Fugro‐Type
Electric Cone Bearing Qc. Contract Report, Dacw 38‐76‐M 6646. Waterways
Experiment Station, 145.
Schmertmann, J. H. (1978) Guidelines for Cone Penetration Test, Performance and Design,
Report FHWA‐Ts‐78‐209. Federal Highway Administration, 145.
Schmertmann, J. H. (1987) Discussion on "Time Dependent Strength Gain in Freshly
Deposited or Densified Soil" by Mitchell and Solymar. Journal of Geotechnical
Engineering, ASCE, 113, 2, 173‐175.
577
Schmertmann, J. H., Hartman, J. P. & Brown, P. R. (1978) Improved Strain Influence Factor
Diagrams. Journal of Geotechnical Engineering, ASCE, 104, GT8 (August), 1131‐1135.
Seed, H. B. & Booker, J. R. (1977) Stabilization of Potentially Liquefiable Sand Deposits Using
Gravel Drains. Journal of Geotechnical Engineering, ASCE, 103, 7, 757‐ 768.
Seed, H. B. & Idriss, I. M. (1967) Analysis of Soil Liquefaction, Niigata Earthquake. Journal of
Soil Mechanics and Foundations Division, ASCE, 93, SM 3, 83‐108.
Seed, H. B. & Idriss, I. M. (1971) Simplified Procedure for Evaluating Soil Liquefaction
Potential. Journal of Soil Mechanics and Foundations Division, ASCE, 97, SM 9, 1249‐
1273.
Seed, H. B., Martin, P. P. & Lysmer, T. (1976) Pore Water Pressure Changes During Soil
Liquefaction. Journal of Geotechnical Engineering, ASCE, 102, 4, 323‐346.
Selig, E. T. & Ladd, R. S. (1973) Evaluation of Relative Density Measurements and Applications.
Evaluation of Relative Density and its Role in Geotechnical Projects Involving
Cohesionless Soils: ASTM STP523‐EB.7744‐1., Los Angeles, 25‐30 June 1972, 487‐504.
Serridge, C. J. (2002) Dynamic Compaction of Loose Sabkha Deposits for Airport Runway and
Taxiways. 4th International Conference on Ground Improvement Techniques, Kuala
Lumpur, 26‐28 March, 649‐656.
Shibazaki, M. & Yoshida, H. (1995) Method for Controlling a Final Pile Diameter in a Cast‐in‐
Place of Solidification of Pile by a Jet Process. in UNITED STATES PATENT AND
TRADEMARK OFFICE (Ed.). United States, Chemical Grouting Company, 4.
Silver, M. L. & Seed, H. B. (1971) Deformation Characteristics of Sands under Cyclic Loading.
Journal of the Soil Mechanics and Foundations Division, ASCE, 9, SM8, 1081‐1098.
Siskind, D. E., Stagg, M. S., Kopp, J. W. & Dowding, C. H. (1980) USBM Report of Investigations
8507 ‐ Structure Response and Damage Produced by Ground Vibration from Surface
Mine Blasting, US Bureau of Mines
Skempton, A. W. (1986) Standard Penetration Test Procedure and the Effects in Sands of
Overburden Pressure, Relative Density, Particle Size, Ageing and Overconsolidation.
Geotechnique, 36, 3, 425‐447.
Skipp, B. & Buckley, J. (1977) Ground Vibration from Impact. 9th International Conference on
Soil Mechanics and Foundation Engineering, 2, Tokyo, 397‐400.
Sladen, J. A. & Hewitt, K. J. (1989) Influence of Placement Method on the in‐Situ Density of
Hydraulic Sand Fills. Canadian Geotechnical Journal, 26, 3, 453‐466.
Slocombe, B. (2004) Dynamic Compaction. in MOSELEY, M. P. & KIRSCH, K. (Eds.) Ground
Improvement. 2nd ed. New York, Spon Press.
578
Spaulding, C. & Zanier, L. (1997) Apron Densification at Macau International Airport Using
Dynamic Consolidation and Replacement Methods. International Conference on
Ground Improvement, Macau, 525‐530.
Standards Australia (2006) Explosives ‐ Storage and Use, Part 2: Use of Explosive. in
STANDARDS AUSTRALIA (Ed.) AS 2187.2 ‐ 2006. Sydney, NSW, Australia, Standards
Australia,
Studer, J. & Suesstrunk, A. (1981) Swiss Standard for Vibrational Damage to Buildings. 10th
International Conference on Soil Mechanics and Foundation Engineering, 3,
Stockholm, 307‐312.
Tan, S. A., Tjahyono, S. & Oo, K. K. (2008) Simplified Plane‐Strain Modeling of Stone‐Column
Reinforced Ground. Journal of Geotechnical and Geoenvironmental Engineering,
ASCE, 134, 2 (February), 185‐194.
Tatsuoka, F., Tokida, K., Yasuda, S., Hirose, M., Imai, T. & Konno, M. (1978) A Method for
Estimating Undrained Cyclic Strength of Sandy Soils Using Standard Penetration
Resistance. Soils and Foundations, Japanese Society of Soil Mechanics and
Foundation Engineering, 18, 3, 43‐58.
Tavenas, F. A. & La Rochelle, P. (1970) Problems Related to the Use of the Relative Density,
Report S‐21, Laval University.
Tavenas, F. A., Ladd, R. S. & La Rochelle, P. (1973) Accuracy of Relative Density
Measurements: Results of a Comparative Test Program. Evaluation of Relative
Density and its Role in Geotechnical Projects Involving Cohesionless Soils: ASTM
STP523‐EB.7744‐1, Los Angeles, 25‐30 June 1972, 18‐60.
Terra Systems Rapid Impact Compaction Another Form of Dynamic Compaction, Viewed 7
February 2009, http://www.terrasystemsonline.com/terranotes/ric.pdf.
Terzaghi, K. (1925) Erdbaumechanik Auf Bodenphysikalisher Grundlage, Vienna, Deuticke
Terzaghi, K. (1943) Theoretical Soil Mechanics, New York, John Wiley & Sons,510.
Terzaghi, K. & Frohlich, O. K. (1936) Theorie Der Setzung Von Tonschichten, Leipzig, Franz
Deuticke
Terzaghi, K. & Peck, R. B. (1948) Soil Mechanics in Engineering Practice, New York, John Wiley
and Sons
Thau, S. A. & Pao, Y. (1966) Diffraction of Horizontal Shear Waves by a Parabolic Cylinder and
Dynamic Stress Concentration. Journal of Applied Mechanics, Transactions ASME,
December, 785‐792.
Thevanayagam, S., Martin, G. R., Nashed, R., Shenthan, T., Kanagalingam, T. & Ecemis, N.
(2006) Liquefaction Remediation in Silty Soils Using Dynamic Compaction and Stone
Columns: Technical Report Mceer‐06‐0009, Buffalo, NY, MCEER,101.
Thoenen, J. R. & Windes, S. L. (1942) USBM Bulletin 442 ‐ Seismic Effects of Quarry Blasting,
US Bureau of Mines
579
Thurner, H. & Sandström, A. (1980) A New Device for Instant Compaction Control.
International Conference on Compaction, Paris, 611‐614.
Thurner, H. & Sandström, A. (2000) Continuous Compaction Control, Ccc. Workshop on
Compaction of Soils and Granular Materials, Modeling of Compacted Materials,
Compaction Management and Continuous Control, ISSMGE (European Technical
Committee), Paris, 237‐246.
Tiedemann, D. A. (1973) Variability of Laboratory Relative Density Test Results. Evaluation of
Relative Density and its Role in Geotechnical Projects Involving Cohesionless Soils:
ASTM STP523‐EB.7744‐1, Los Angeles, 25‐30 June 1972, 61‐73.
Timoshenko, S. P. & Goodier, J. N. (1951) Theory of Elasticity, New York, McGraw Hill Book
Co.
Tokimatsu, K. (1979) Generation and Dissipation of Pore Water Pressures in Sand Deposits
During Earthquakes. Tokyo, Tokyo Institute of Technology,
Tokimatsu, K. & Yoshimi, Y. (1983) Empirical Correlation of Soil Liquefaction Based on SPT‐N
Value and Fines Content. Soils and Foundations, Japanese Society of Soil Mechanics
and Foundation Engineering, 23, 4, 56‐74.
Tokoro, T., Kashima, S. & Murata, M. (1984) Grout Injection Method and Apparatus. in
UNITED STATES PATENT AND TRADEMARK OFFICE (Ed.). United Source, Nihon Soil
Engineering Co., Nihon Sobo Bosui Co., Yamaguchi Kikai Kogyo Co., 12.
Tsai, P. H. & Chang, T. S. (2009) Effects of Open Trench Siding on Vibration ‐ Screening
Effectiveness Using the Two‐Dimensional Boundary Element Method. Soil Dynamics
and Earthquake Engineering, 29, 5 (May), 865‐873.
Us Army Corps of Engineers (1999) Guidelines on Ground Improvement for Structures and
Facilities, Engineer Technical Letter Elt‐1110‐1‐185, Washington, DC, USACE,115.
Van Eekelen, S. & Bezuijen, A. (2008) Design of Piled Embankments, Considering the Basic
Starting Points of the British Standard Bs8006. EuroGeo4, Edinburgh, September.
Van Impe, W. F. (1989) Soil Improvement Techniques and Their Evolution, Rotterdam,
Balkema
Van Impe, W. F. & Bouazza, A. (1996) Densification of Domestic Waste Fills by Dynamic
Compaction. Canadian Geotechnical Journal, 33, 6, 879‐887.
Van Wambeke, A. (1962) Méthodes D'investigation Des Sols En Place ‐ Etude D'une
Campagne D'essais Comparatifs. Sols Soils, 2, September, 9‐18.
Van Wieringen, J. B. M. (1982) Relating Cone Resistance and Pressuremeter Test Results. 2nd
European Symposium on Penetrating Testing, Amsterdam, 951‐955.
Varaksin, S. (1970) Determination of Sand Density Below Groundwater Table. Northwestern
University at Evanston, Ill,
Varaksin, S. (1981) Recent Development in Soil Improvement Techniques and Their Practical
Applications. Sols Soils, 38‐39, 7‐32.
580
Varaksin, S. (2014) Discussions on Dynamic Compaction and Dynamic Replacement. in
HAMIDI, B. (Ed.).
Varaksin, S., Hamidi, B. & D'hiver, E. (2005) Pressuremeter Techniques to Determine Self
Bearing Level and Surface Strain for Granular Fills after Dynamic Compaction.
International Symposium 50 Years of Pressuremeters (ISP5‐ Pressio 2005), Paris, 22‐
24 August, 687‐.
Varaksin, S. & Racinais, J. (2009) Etude Des Paramètres D’application De La Consolidation
Dynamique Et De Ses Techniques Dérivées. 17th International Conference on Soil
Mechanics and Geotechnical Engineering, Alexandria, Egypt, 5‐9 October, 2407‐
2410.
Vesic, A. S. (1975) Shallow Foundations. in WINTERKORN, H. F. & FANG, H. Y. (Eds.)
Foundation Engineering. New York, Van Nostrand Company.
Villet, W. C. B. & Mitchell, J. K. (1981) Cone Resistance, Relative Density and Friction Angle.
Symposium on Cone Penetration Testing and Experience, ASCE National Convention,
St Louis, October, 178‐208.
Vuola, P. & Hartikainen, J. (1999) Aspects for Dynamic Compaction of Saturated Sand. 12th
European Conference on Soil Mechanics and Geotechnical Engineering: Geotechnical
Engineering for Transportation Infrastructure, 3, 1603‐1606.
Waschkowski, E. (1976) Comparisons Entre Des Resultats Des Essais Pressiometriques Et Le
SPT. Rapport De Recherche Du Laboratoire Regional Des Ponts Et Chaussées De Blois,
Faer 1.05.23.5. not published,
Watts, G. R. (1990) Traffic Induced Vibrations in Buildings. Transport and Road Research
Laboratory Research Report 246. Transport Research Laboratory,
White, D. & Vennapusa, P. (2010) A Review of Roller‐Integrated Compaction Monitoring
Technologies for Earthworks, Final Report Er10‐04. Earthworks Engineering Research
Center (EERC), Department of Civil Construction and Environmental Engineering,
Iowa State University, 36.
Windes, S. L. (1942) Damage from Air Blast. Progress Report 1, USBM Report of Investigation
3622, US Bureau of Mines
Windes, S. L. (1943) Damage from Air Blast. Progress Report 2, USBM Report of Investigation
3708, US Bureau of Mines
Wiss, J. F. (1981) Construction Vibrations: State of the Art. Journal of Geotechnical
Engineering, ASCE, 107, 2 (February), 167‐182.
Wiss, J. F. & Parmelee, R. A. (1974) Human Perception of Transient Vibrations. Journal of the
Structural Division, 100, ST4 (April), 773‐787.
Wolf, J. P. (1994) Foundation Vibration Analysis Using Simple Physical Models, New Jersey,
Prentice Hall
581
Woods, R. D. (1968) Screening of Elastic Surface Waves by Trenches. Journal of Soil Mechanics
and Foundations Division, ASCE, 94, 4 (July), 951‐979.
Woods, R. D. & Jedele, L. P. (1985) Energy Attenuation Relationships from Construction
Vibration. Symposium on Vibration Problems in Geotechnical Engineering, Detroit,
MI, 229‐246.
Woods, R. D. & Richart, F. E. (1967) Screening of Elastic Surface Waves by Trenches.
International Symposium on Wave Propagation and Dynamic Properties of Earth
Materials, Albuquerque, New Mexico, August 1967.
Woodward, J. (2005) An Introduction to Geotechnical Processes, Taylor & Francis,123.
Wu, T. H. (1957) Relative Density and Shear Strength of Sands. Journal of the Soil Mechanics
and Foundations Division, ASCE, 83, 1 (January), 1161‐1161 ‐ 1161‐1123.
Yahiro, T. & Yoshida, H. (1977) High Velocity Jet Digging Method. in UNITED STATES PATENT
AND TRADEMARK OFFICE (Ed.). United States, Chemical Grout Company & Kajima
Corporation, 7.
Yee, K. (1999) Upgrading of Existing Landfills by Dynamic Consolidation: A Geotechnical
Aspect. Master Builders Journal, September.
Yoshimi, Y. & Tohno, I. (1973) Statistical Significance of the Relative Density. Evaluation of
Relative Density and its Role in Geotechnical Projects Involving Cohesionless Soils:
ASTM STP523‐EB.7744‐1, Los Angeles, 25‐30 June 1972, 74‐84.
Youd, T. L. (1973) Factors Controlling Maximum and Minimum Densities of Sands. Evaluation
of Relative Density and its Role in Geotechnical Projects Involving Cohesionless Soils:
ASTM STP523‐EB.7744‐1, Los Angeles, 25‐30 June 1972, 98‐112.
Youd, T. L., Idriss, I. M., Andrus, R. D., Arango, I., Castro, G., Christian, J. T., Dobry, R., Finn, W.
D. L., Harder, L. F., Hynes, M. E., Ishihara, K., Koester, J. P., Liao, S. S. C., Marcuson Iii,
W. F., Martin, G. R., Mitchell, J. A., Moriwaki, Y., Power, M. S., Robertson, P. K., Seed,
R. B. & Stokoe, K. H. (2001) Liquefaction Resistance of Soils: Summary Report from
the 1996 Nceer and 1998 Nceer/Nsf Workshops on Evaluation of Liquefaction
Resistance of Soils. Journal of Geotechnical and Geoenvironmental Engineering,
ASCE, 127, 10 (October), 817‐833.
Zaeske, D. (2001) Zur Wirkungweise Von Unbewehrten Und Bewerhrten Mineralischen
Tragschichten Über Pfahlartigen Gründungselementen, Phd Thesis. Universität Gh
Kassel,
Zoeppritz, C. & Kosten, C. W. (1919) Nach D. Konigl. Gessell D. Wissen, Z. Gottingen. math‐
Phys., 66‐94.
Zou, W. L., Wang, Z. & Yao, Z. F. (2005) Effect of Dynamic Compaction on Placement of High
Road Embankment. Journal of Performance of Constructed Facilities, ASCE, 19, 4
(November), 316‐323.
582
Every reasonable effort has been made to acknowledge the owners of copyright material. I
would be pleased to hear from any copyright owner who has been omitted or incorrectly
acknowledged.
583
Appendix: First Page of Publications
584
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
This page is intentionally blank
1