Proceedings of 3rd International Conference on Advances in Civil Engineering, 21-23 December 2016, CUET, Chittagong, Bangladesh

Islam, Imam, Ali, Hoque, Rahman and Haque (eds.)

A DETAILED INVESTIGATION OF REINFORCED CONCRETE

COLUMN JACKETING: EXPERIMENT, THEORY AND NUMERICAL
ANALYSIS

K. S. Ahmed & M. Riaz*

Department of Civil Engineering, Military Institute of Science & Technology, Dhaka, Bangladesh
*
Corresponding Author: riazeity7239@gmail.com

ABSTRACT
In recent days, Reinforced Concrete (RC) column jacketing is been increasingly used in structural
strengthening in Bangladesh. This research work investigates the structural capacity enhancement of
column by RC jacketing. Twelve jacketed short column samples composed of 25 mm and 31.5 mm
jacket thickness were experimented for axial capacity. Samples vary in terms of use of surface
preparation, welded ties, and change of clear cover in jacketed part. Analytical equations in terms of
Interaction Diagram are formulated. As an outcome, a software is developed to analyse and compare the
capacity under combined compression with uniaxial bending. Tested sections are modelled in ETABS
2015 and SAP 2000 for Finite Element analysis. Experimental result shows that new concrete collapses
earlier at the interface than that of the old concrete. Welding jacket ties contribute to axial capacity by
resisting rebar buckling. Proposed analysis accounts the effect of interface bonding thus differs with
Japanese code, ETABS, and SAP. However, test results validated the analytic axial capacity at an
accuracy of 89 to 96%. Hence a bondage coefficient of 0.85-0.95 is proposed in determining axial
capacity.

Keywords: Column; retrofitting; jacketing; interaction diagram; Finite element

INTRODUCTION
In order to avoid potential earthquake hazard, latest Bangladesh National Building Code (BNBC)-2015
guideline demands more structural resistance that suggests to strengthen many existing building
structures of Bangladesh. Recent earthquakes, Rana Plaza incident and some other structural hazards
raised the attentions towards the structural strengthening. Previously, many buildings were designed
neither following the guideline nor considering lateral load. In addition, changes in live loads and user
facilities, deterioration of the load carrying elements, design errors, poor construction quality during
erection, and aging of structure, addition and alteration of existing structure force the users to strengthen
the structural elements. Column being the most important structural element requires the utmost
priority to be retrofitted. In recent years, column jacketing is commonly used to enhance the strength
and stiffness of existing RC structure. Applications of RC column jacketing has already been executed
in some garments buildings of Bangladesh. Still there is a large number of building structures that
requires strengthening work immediately.
Effect of surface preparation, failure criteria and capacity of concrete jacketing has been experimentally
and analytically investigated in the past by Bett et al. (1988), Saatcioglu and Ozcebe (1989), Alcocer
and Jirsa,(1993), Erosy et al.(1993), Park and Rodriguez(1994), Stoppenhagen et al. (1995), Abu-Tair
et al.(1996), Austin et al.(1999), Climaco and Regan (2001), Julio et al.(2003), Eduardo et al. (2005),
Beushausen and Alexander (2008), Yuce et al. (2007), Roberto et al. (2008), H. Sezen and Eric A.
Miller (2009), Stephanos E. Dritsos et al. (2010), D. W. Zhang et al.(2013), Veena M and Mini Soman
(2014) and M. G. Marques et al. (2015). Indian code [3] and Japanese code [4] has separate
methodology and equations for designing jacketed section. In Bangladesh, analysis and constructional
methodology are recently published as guideline [6] that is modified and based on Japanese guidelines.
Due to increasing use of RC jacketing, engineers need a simplified, time saving design and analysis
approach. However, no significant research work has been reported yet and also the analysis and design
guideline for RC column jacketing is yet to be established authentically.

591
Proceedings of 3rd International Conference on Advances in Civil Engineering, 21-23 December 2016, CUET, Chittagong, Bangladesh
Islam, Imam, Ali, Hoque, Rahman and Haque (eds.)

This paper investigates the structural capacity enhancement of column by RC Jacketing. Hence,
experimental investigation on jacketed column were carried out to determine load carrying capacity
under pure compression. Simplified analytical equations are proposed to estimate jacketed column
capacity in terms of Interaction Diagram which was compared with that of derived from Japanese
retrofitting codes. Tested samples are modelled in ETABS 2015 & SAP 2000 v17 for Finite Element
(FE) analysis. Finally, a computer program is developed which is able to analyse and compare the
capacity of jacketed sections under combined compression with uniaxial bending.

EXPERIMENTAL INVESTIGATION
Twelve short column samples of 610 mm height with a cross section of 102 mm x 102 mm were used as
Reference Sample (RS). Four 8 mm bar were used as longitudinal reinforcement. Tie (2.5 mm) with 6.5
mm clear cover (CC) were used at 150 mm spacing. To depict old column section comparatively
weaker concrete was used in RS. Eleven samples were retrofitted using RC jacketing with 25 mm and
31.5 mm jacket thickness. Eight longitudinal bars were used with the same diameter as RS. Ties were
spaced at 100 mm. Other than two samples as mentioned in Table 1, typical CC was kept as 12.7 mm.
Brick and stone chips of 19 mm and 12.7 mm downgraded respectively were used. Local and Sylhet
sand of Fineness Modulus (FM) 1.1 and 2.2 respectively were used in test. Mix ratio was 1:2:4 and
1:1.5:3 by percent of volume for RS and jacketed samples respectively. Before jacketing surface were
roughened using hand chisel. Afterwards, sand blasting was carried out using coarse sand of FM 2.2.
Samples are named according to their thickness. Suffix letter used to describe the jacketing process;
‘N’-no bonding agent, ‘B’-surface prepared and bonding agent, ‘M’- monolithic casting, ‘W’-welded
ties, and ‘C’ change of clear cover. CC was changed from 12.7 to 8.5 mm and 15 mm for 25 mm and
31.5 mm jacket thickness respectively.
Compressive capacity was tested in Universal Testing Machine (UTM) as in Fig. 1. Steel base plate of
205 mm x 205 mm cross section and 15 mm thick was used for uniform distribution of axial load. Load
was applied at a rate of 1 mm/ min rate. Machine was programmed to stop at 40 % strain after reaching
to the maximum axial load. Peak axial capacity was displayed in both the dial meter and in computer
monitor system generating required graph of load, stress and strain.

(a) (b) (c) (d) (e) (f)

Fig. 1: (a) Surface roughening (b) Sand blasting (c) application of epoxy bonding agent (d) Jacketing of sample
(e) UTM machine (f) Test Setup

ANALYTIC ASSESSMENT
Both weighted average concrete strength ( fc avg ) and old concrete strength ( fc old ) are used for analysis
of jacketed section as in [6] and [7]. However, lower elastic modulus need to be considered in design as
mentioned in [2]. Compressive strength is increased for the active confinement determined by various
equationsby Scott et al. (1982), Uzumeri (1982), Mander et al. (1988) and Yong et al. (1988). In this
regard, a concrete model is proposed to account confining stress generated by the jacket thickness.
Effect of longitudinal reinforcement is negligible in pure compression according to P. Christou et al.
(2013). Thus, thickness of jacket concrete in between longitudinal bar and old column face tjacketinner is
only used to determine confining stress. Volumetric ratio of the concrete to old column sections
 concrete  (4  tjacketinner ) / b0 [Modified to concrete according to FRP formula of R. Benzaid and H.A.

592
Proceedings of 3rd International Conference on Advances in Civil Engineering, 21-23 December 2016, CUET, Chittagong, Bangladesh
Islam, Imam, Ali, Hoque, Rahman and Haque (eds.)

Mesbah, (2013)]. Confinement coefficient ke  1/ (1   concrete) . Tensile strength of concrete frc and
volume of confine concrete to the volume of column section ratio  cc is determined to find out
confining stress fl  0.5  ke   cc  frc . Finally jacketed compressive stress fc ' Jacket is determined using
Mander concrete model [5].

Proposed Interaction Diagram Equations

Simplified five-point interaction diagram is formulated by deriving equations which are based on
column interaction diagram with ACI code 318-08. It is considered that, old rebar is corroded hence
cross sectional area is considered negligible in analysis to contribute in compression which is agreed
with V. C. Marlapalle et al. (2014). Thus, pure compression for jacket sections can be written as:
P  CB  [0.85 fc ' jacket  ( A - As old - As jacket )]  ( Ast jacket  Asc jacket )  fy jacket (1)
CB = Coefficient of bondage; to account the reduction factor due to bonding effect at interface of
different concrete. A = cross section after jacketing, As old and As jacket =area of longitudinal steel in
existing column and jacket respectively, Ast jacket and Asc jacket = area of tension and compression steel
respectively in jacket, fy jacket = yield strength of jacket longitudinal steel.
Bending capacity of old rebar is considered to contribute along with jacketed bar. To account for the
remaining contribution of old longitudinal bar, a partial value of their original bending capacity is
assumed. It is denoted as Coefficient of moment (CM). It is considered that, these bar had reached to
yield strength and their remaining ductility is added in the bending capacity of jacketed section.
Thus CM  1  ( fyold / fuold ) in which ( fyold / fuold ) is ratio of yield and ultimate strength of old column
rebar. Generally CM gives a value ranging from 0.20-0.35. Thus pure bending point is denoted as:
a
M  [ Ast jacket  fy jacket  CM  Ast old  fy old ]  (dn - )
n
(2)
2
Ast old = area of tensile steel in old column, dn = distance of centroid of tensile jacketed steel from top
fibre, an = dimension of equivalent stress block in jacketed section.
Jacketed concrete strain  ujacket is taken as 0.003 in analysis. However, old concrete strain  uold is
considered up to 0.005. Basing on strain compatibility a single concrete strain is considered for
simplification. Basing on strain and force, weighted average concrete strain  uavg is calculated. Finally
following equations are proposed for balanced, compression and bending control points:
P  0.85 fc jacket  an  bnew  Asc jacket  fs ' jacket  Ast jacket  fy jacket (3)
hnew an hnew
M  0.85 fc jacket  an  bnew  (  )  Asc jacket  fs ' jacket  (  d ' new) 
2 2 2
hnew hold
Ast jacket  fy jacket  (dnew  )  CM  Ast old  fy old  (dold  ) (4)
2 2

Interaction Diagram using Japanese equations

Following [4] and [6] maximum theoretical axial capacity P max and bending equations are formulated.
P max  Ast jacket  fy jacket  Ast old  fy old  bnew  hnew  fc avg (5)
[According to 3.3.4-2 a of Japanese guidelines]
M  ( Ast jacket  fy jacket  gjacket  Ast old  fy old  gold  0.12  bnew  hnew  f )
2
c  avg

P max  P
( ) (6)
P max  0.4  bnew  hnew  f c  avg

gold and gjacket are distance between tensile and compressive longitudinal steel in jacket portion and
existing column respectively

FINITE ELEMENT MODELLING USING ETABS 2015 AND SAP 2000

Section designer is used to model jacketed column section in ETABS 2015 and SAP 2000 v17. Material
properties were defined according to the test samples. Rectangle and box section was used to model

593
Proceedings of 3rd International Conference on Advances in Civil Engineering, 21-23 December 2016, CUET, Chittagong, Bangladesh
Islam, Imam, Ali, Hoque, Rahman and Haque (eds.)

samples. Longitudinal and tie bar were placed in the model following the actual dimensions of samples.
After modelling, interaction diagram and moment vs. curvature diagram can be extracted as output.
SAP 2000 produces advance features for analysis stress and strain in different conditions as well as
moment vs. concrete and steel strain and compression data. Interaction diagram are formulated using
these data. Fig. 2 displays the model and stress distributions.

(a) (b) (c)

Fig. 2 (a) Old column section (b) Jacketed column (c) Stress distributions

RESULTS AND DISCUSSIONS

Results of the test and analytical assessment are shown in Table 1. Bonding does not follow a liner
behaviour and vary greatly according to construction, material property and types of surface
preparation, moisture content of substrate [8]. Effect of creep and any direct tension stress caused by
shrinkage are ignored in analysis. All these factors contribute in the deviations of test with analytic
results. Monolithic samples 25-M and 31.5-M had the less deviation due to absence of different
concrete interface. Whereas 25-N and 31.5-N had the larger deviation. Since there is no provision of
reduction factors due to bonding interface in FE analysis and Japanese code therefore proposed pure
compression point differs as much as 26% as in Fig. 3. Other points agree with the Japanese code by a
variation of 8-12% while pure bending resulted deviation by only 1-2.5%. On the other hand, for
maximum bending capacity, proposed analysis agrees well with an accuracy of 93.5-98 % with the
Japanese code. However, both differ with ETABS and SAP with a deviation of 16.5-25% due to liner
addition and composite action account which is shown in Fig. 7.
Table 1. Experiment vs. Analytical Results

Sample fc'old fc'jacket Analytic Test % Area Capacity ratio

Name (MPa) (MPa) capacity capacity variation Ratio Pn/Pold
(KN) (KN) A/Aold Analytic Test
25-N 11.12 19.35 243.29 226.21 7.02 1.522 1.521
25-B 21.63 289.64 258.51 10.74 1.948 1.739
25-B-W 21.98 289.75 264.38 8.78 2.220 1.813 1.778
25-M 11.13 21.80 289.69 302.5 4.36 1.812 2.030
25-B-C 21.89 329.83 298.83 9.46 2.064 2.010
31.5-N 11.12 24.98 336.9 305.67 9.26 2.108 2.056
31.5-B 29.42 441.00 416.78 5.51 2.759 2.803
31.5-M 12.20 29.87 441.72 452.86 2.69 2.617 2.764 3.046
31.5-B 2 11.12 20.85 403.70 379.45 6.01 2.526 2.552
31.5-B-W 20.21 402.43 382.78 4.88 2.518 2.575
31.5-B-C 11.12 21.32 372.08 342.68 7.74 2.328 2.305
RS - 159.80 148.65 6.97 - - -

Failure Pattern
Samples failed with generation of full and partial depth longitudinal crack as in Fig. 5. Few lateral
cracks were generated also. Local failure occurred due to stripping out of concrete at the new and old
concrete interface which agrees with [7]. This happens as the jacketed column is unable to maintain
strain compatibility at the interface of new and old concrete. Crushing of jacket concrete along with
failure of ties were seen. Concrete failed at the corner due to the high concentration of stress which
agrees with [9]. Significant increase of axial capacity was found up to 178% depending on compressive
strength of jacket. From Table 1 it is seen that, for same area ratio bonding and surface prepared sample
had increased capacity ratio from 18-22% than the nonbonding sample.

594
Proceedings of 3rd International Conference on Advances in Civil Engineering, 21-23 December 2016, CUET, Chittagong, Bangladesh
Islam, Imam, Ali, Hoque, Rahman and Haque (eds.)

Fig. 3: Comparison of Interaction Diagram. Fig. 4: Developed software output interface

Development of Analysis Software

A software is developed to analyse jacketed column using Eq. (1)-(6) in the programming language of
‘Microsoft C#’. Interface takes the input of material properties and dimensions to generate interaction
curve both with and without phi. A comparison is also generated with Japanese code in the output
interface as shown in Fig. 4. User facility like print, save as image, and zoom options are incorporated.
Jacket design verification can be performed using the dynamically generated point.

Stress-strain Behaviour
Maximum loading occurred in an average strain of 0.012-0.013 mm/mm for all the sample as in Fig. 6.
However, 25-M, 31.5-M and RS samples had lesser rate of 0.011 and 0.01 respectively. Thus ultimate
strain of confined concrete increases due to jacketing as tensile reinforcement undergoes strain
hardening well agreed with S. Chun and H.C. Park (2012). 0.3% ductility is achieved in axial loading
corresponding to 1.61% increase of size. This approves the seismic effectiveness of jacketing.

25-N 25-B 25-M 31.5-N 31.5-B 31.5-M

Fig. 5: Typical failure pattern of samples
Effect of Surface Preparation, Confinement Stress and Welded Ties
Increase capacity can be achieved without using any bonding agent and surface preparation. However,
use of this increased the capacity to 24-30%. A detailed trial method was carried out with different
value of CB ranging from 0.85-1 for bonded and 0.60-0.70 for non-bonded. Hence CB in the range
0.85-0.95 is proposed. For perfect bonding the value is 1. Tabulated capacity is calculated with the
minimum value of CB as 0.85 and 0.65 for surface prepared and non-prepared sample respectively. Test
result verified the analysis with an accuracy of 89 to 96%. For same jacket thickness, clear cover of bars
in jacket determines the effective thickness of confinement concrete. Thus the reduction of clear cover
will increase the capacity with a ratio varying from 0.45-0.50 and vice versa. For constant clear cover,
confinement stress increased due to increase in the thickness and compressive strength of jacket
concrete. Welding of ties prevent buckling of main and tie bar in failure with an increase of axial

595
Proceedings of 3rd International Conference on Advances in Civil Engineering, 21-23 December 2016, CUET, Chittagong, Bangladesh
Islam, Imam, Ali, Hoque, Rahman and Haque (eds.)

capacity at a range of 1-2.5% only. Thus for construction convenient welded ties may not be applicable
except joint.

Fig. 6 Stress vs. Strain Fig. 7 Maximum bending capacity comparison

CONCLUSIONS
Based on the results and findings presented in the paper, following conclusions can be drawn:
 Capacity enhanced significantly (up to 178% in this research) by RC jacketing depending on
compressive strength, clear cover, use of surface treatment and bonding agent. Outer concrete and
interface bonding governs the failure pattern.
 Developed compressive stress of RC jacketing can be used instead of weighted average or existing
value. Proposed equations can be used for simplified analysis of RC column jacketing.
 Idea of using commercial software (such as ETABS, SAP) in analysing and designing jacketed
column is controversial. Engineers should be careful in case of retrofitting column design.
 The developed program can be a useful tool in jacketed column capacity prediction. It can be
effectively used by the engineers to analyse and design for simplicity and enhanced time efficiency.
 Proposed CB and bending capacity analysis can further be evaluated for precision by axial and
bending test of actual size column.
 This research work may contribute to develop a design guideline for column strengthening using
RC jacketing.

REFERENCES
Aysha H, Ramsundar, K. R. Arun M and Velrajkumar. G, “An Overview of Interface Behavior between
Concrete to Concrete”, International Journal of Advanced Structures and Geotechnical Engineering,
Vol. 03, No. 02, April 2014.
CEB bulletin no. 162, “Assessment of Concrete Structures and Design Procedures for Upgrading
(Redesign)”. Bulletin Information, Euro international concrete committee, Paris, 1983.
Indian Standard, “Seismic Evaluation and Strengthening of Existing Reinforced Concrete
Buildings-Guidelines”, Bureau of Indian Standards, February 2013, chapter 8.
Japanese Standard, “Guidelines for Seismic Retrofit of Existing Reinforced Concrete Buildings”, Japan
Building Disaster Prevention Association, 2001, chapter 3, pp. 2-23.
Mander, J.B., Priestley, M.J.N, and Park, R. "Theoretical Stress-Strain Model for Confined Concrete"
Journal of Structural Engineering, American Society of Civil Engineers, vol.114, No. 8, 1988, pp.
1804-1825.
“Manual for Seismic Retrofit Design of Existing Reinforced Concrete Buildings”, Public Works
department, Bangladesh, 2015, chapter 3.
Stephanos E. Dritsos, “Seismic Strengthening of Columns by Adding New Concrete”, bulletin of the
New Zealand Society for Earthquake Engineering, Vol. 40, No. 2, June 2007, pp. 65-66
S. Mohsen S. Asaei, Tze Liang Lau and Norazura Muhammad Bunnori, “Modelling `FRP-confined RC
Columns using SAP2000”, World Applied Sciences Journal 27 (12): pp.1717-1720, 2013.

596