ÉCOLE DE TECHNOLOGIE SUPÉRIEURE
DÉPARTEMENT DE GÉNIE DE LA CONSTRUCTION
DRSR : DEVELOPMENT & RESEARCH FOR STRUCTURES AND REHABILITATION

FATIGUE TESTING OF TECHSTAR SINGLE SUPPORT

BAR MODULAR BRIDGE EXPANSION JOINT WITH
WELDED STIRRUPS

 
 

 
 
April 2012
 
 

FATIGUE TESTING OF TECHSTAR SINGLE SUPPORT BAR

MODULAR BRIDGE EXPANSION JOINT WITH WELDED
STIRRUPS

By

Omar CHAALLAL, P.E., Ph.D., Full Professor and DRSR Director

Lotfi GUIZANI, P.E., Ph.D., Associate Professor
William BONNELL, Graduate Student

DRSR: Development & Research for Structures and Rehabilitation

Department of Construction Engineering
Université du Québec
École de technologie supérieure
No. 1100, Notre-Dame West Street
Montréal (Québec)
Canada H3C 1K3
Tel.: (514) 396-8852
Fax : (514) 396-8584
E-mail : omar.chaallal@etsmtl.ca

Submitted to

TechStar Inc.
1219 West Main Cross Street
Findlay, Ohio
USA 45840

April 2012
 ACKNOWLEDGEMENTS
 
This work was sponsored by TechStar Inc. The research agency is DRSR (Development and
Research for Structures and Rehabilitation), Department of Construction Engineering, École de
Technologie Supérieure, Université du Québec, Montreal, Canada.

Professor R. J. Connor, Purdue University, co-author of the NCHRP-402 Report (NCHRP, 1997)
on which this work was based, visited our facilities. He verified the test setup during the fatigue
testing. He provided valuable guidance in the course of this research.

The authors are especially grateful to John Lescelleur, senior technician at DRSR, who worked
laboriously in the setup, instrumentation and tuning of the experimental investigation. The
contribution of Juan Mauricio Rios, structural technician at DRSR and Fernando Avendano,
application engineer, to the test setup and instrumentation is gratefully acknowledged.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
   

ii.
  
SUMMARY
 
This study, which was sponsored by TechStar Inc., reports on the fatigue testing of the TechStar
Single Support Bar (LG Type) Modular Bridge Expansion Joint (MBEJ) with Welded
Stirrups carried out at the Structures Laboratory at École de Technologie Supérieure, University
of Quebec in Montreal, Canada. The main objective of the study is to evaluate the fatigue
resistance of critical details of this modular joint system, with particular emphasis on the welded
stirrups. Specific objectives are as follows: (a) establish an experimental fatigue curve, with at
least 10 points, for the details of the stirrups welding to the center-beam, and (b) verify that the
stirrup details of the MBEJ identified above meet fatigue Category C of the AASHTO LRFD
2007 Bridge Design Code requirements.
 
All the tests were carried out in conformity with the requirements and guidelines of the procedure
of AASHTO LRFD 2007 Bridge Design Code, which is based on the National Cooperative
Highway Research Program, NCHRP-402 Report entitled: Fatigue Design Criteria for Modular
Bridge Expansion Joints (NCHRP, 1997).

Three identical subassemblies of the modular joint system were tested in fatigue. Each specimen
is made of a center-beam supported by four equally spaced 1081 mm long support bars and
forming three spans of 910 mm each. Support bars are maintained in place through stirrups which
are welded to the center-beam.
 
The fatigue testing was undertaken in the finite life range. Both vertical and horizontal load
ranges were applied to the test specimen simultaneously in the following proportions: (a) Vertical
Load Range = Pv; and (b) Horizontal Load Range = 0.2 Pv. To that end, the specimen was
seated on the testing frame at an inclination of 11.3 from the horizontal plane and the resultant
load was applied through actuators oriented in the vertical plane. The loads were applied using
two synchronized actuators under force control conditions (distance between the two actuators =
1946 mm). Different load ranges, varying between 130kN/actuator and 175kN/actuator, were
applied for each specimen in order to obtain a sufficient number of points defining the

iii.
experimental fatigue (S-N) curve. The number of cycles varied between 1 200 550 and 3 163 276
cycles.

Static calibration tests were performed prior to the fatigue testing for each specimen. These
calibration tests confirmed the validity of the structural models and also established the
repeatability of experimental data, in conformity with the requirements of the AASHTO LRFD
2007 Bridge Design Code and the NCHRP-402 Report (NCHRP, 1997). 
 
Structural models (1-D, 2-D) were used for the theoretical analysis of joints. The 1-D model was
found to provide reasonably accurate results for the stress ranges on the center-beam. However,
the experimental strain readings at the stirrups locations where concentrated stresses (near welds)
are present were captured only grossly and could not be captured with good accuracy. This is
attributed to the fact that the developed model could not capture satisfactorily all the complexities
and detailing of the welds and the eccentricities inherent to the manufacturing process of these
sensitive elements.
 
Characteristic fatigue crack types affecting the stirrups were observed, defined and reported. The
experimental S-N fatigue curves of the MBEJ with welded stirrups clearly showed that the
studied stirrup detail is compatible with and meets category C requirements as defined by
the AASHTO LRFD 2007 Bridge Design Code. Furthermore, the present study revealed that
the contribution of the secondary bending moment within the stirrup into the stress range should
be included in the calculation of the stress range. This report defines how to calculate the stress
ranges at different critical sections of the stirrup.
 

iv.
 TABLE OF CONTENTS 

ACKNOWLEDGEMENTS ................................................................................................ ii
SUMMARY ....................................................................................................................... iii
LIST OF TABLES ............................................................................................................. vi
LIST OF FIGURES… ...................................................................................................... viii

1.  INTRODUCTION .................................................................................. 1 

1.1.  State of the problem............................................................................................. 1 

1.2.  Objectives ............................................................................................................ 2 
1.3.  Layout of Report .................................................................................................. 2 

2.  EXPERIMENTAL PROGRAM ................................................................. 3 

2.1.  Description of the Specimen and Instrumentation .............................................. 3 

2.2.  Experimental Program and Procedure ............................................................... 8 
2.3.  Failure Criteria and Stress Range Calculations ............................................... 10 

3.  MODELLING AND ANALYSIS OF THE JOINT SPECIMENS ..................... 11 

3.1.  Objectives .......................................................................................................... 11 

3.2.  Presentation of the Structural Models and Loading ......................................... 11 
3.3.  AASHTO LRFD Fatigue Curves ....................................................................... 17 
3.4.  Evaluation of Stress Range ............................................................................... 18 
3.5.  Evaluation of Number of Cycles........................................................................ 22 
3.6.  Predicted Strains at Gage Locations ................................................................ 23 

4.  RESULTS AND OBSERVATIONS .......................................................... 25 

4.1.  Summary of Results ........................................................................................... 25 

4.2.  Comparison with AASHTO Fatigue Categories ............................................... 26 
4.3.  Repeatability and Calibration Tests .................................................................. 29 
4.4.  Stirrups Detailed Testing Results ...................................................................... 32 

5.  CONCLUSIONS .................................................................................. 41 

REFERENCES .................................................................................................................42

v.
 LIST OF TABLES

Table 2.1 – Location of Strain Gages on Stirrups ............................................................................ 6 

Table 2.2 – Calibration Tests under Static Loading ......................................................................... 8 
Table 2.3 – Experimental Program – Fatigue Test........................................................................... 8
Table 3.1 – Sections and Properties of the Component ................................................................. 13
Table 3.2 – Material Properties of the Members ............................................................................ 13
Table 3.3 – Member Lengths for the 3–D Model ......................................................................... 13
Table 3.4 – Detail Category Constant A ........................................................................................ 17
Table 3.5 – Section Properties and Analysis Results Used for Calculation of the Nominal Stress
Range for Type A, B, and B’ Cracking. ..................................................................... 19
Table 3.6 – Nominal Stress Range for Identified Cracking Types ................................................ 21
Table 3.7 – Stress Range after Scaling ........................................................................................... 22
Table 3.8 – Life Expectancies to Qualify Stirrup Details for AASHTO Category C .................... 23
Table 4.1 – Test Results for Stirrups of the Three Specimens ...................................................... 26
Table 4.2 – Experimental Strains for Three Static Calibration Test Runs for MJ–3 ..................... 29
Table 4.3 – Comparison of Calibration Test Results with 1–D Theoretical Values ...................... 30
Table 4.4 – Comparison of Calibration Test Results with 3–D Theoretical Values ...................... 31
Table 4.5 – Loading Program during Testing of MJ1 .................................................................... 32
Table 4.6 – Summary of Test Results for MJ1 .............................................................................. 33 
Table 4.7 – Stress Range during Testing of MJ1 ........................................................................... 33 
Table 4.8 – Loading Program during Testing of MJ2 .................................................................... 36 
Table 4.9 – Summary of Test Results for MJ2 .............................................................................. 36 
Table 4.10 – Stress Range during Testing of MJ2 ......................................................................... 36 
Table 4.11 – Loading Program during Testing of MJ3 .................................................................. 38 
Table 4.12 – Summary of Test Results for MJ3 ............................................................................ 38 
Table 4.13 – Stress Range during Testing of MJ3 ......................................................................... 39 
 
 

vi.
 LIST OF FIGURES

Figure 2.1 – Details of the Specimen ............................................................................................... 3 

Figure 2.2 – Details of the Specimen and Connections ................................................................... 4 
Figure 2.3 – Gages Locations on the Top of the Center-beam ........................................................ 5 
Figure 2.4 – Gage Locations on Bottom of Support Bars ................................................................ 5 
Figure 2.5 – Strain Gage Locations on Stirrups ............................................................................... 6 
Figure 2.6 – View of the Bolted Stop System .................................................................................. 7 
Figure 2.7 – Specimen Installed on Inclined Seats of the Test Bench ............................................. 7 
Figure 2.8 – View of Specimen under Actuator Loads .................................................................... 9 
Figure 2.9 – Types A, B and B’ Cracking and Failure Criteria ..................................................... 10
Figure 3.1 – 1–D Structural Model of Center–Beam ..................................................................... 12 
Figure 3.2 – 2–D Structural Model of Stirrup ................................................................................ 12 
Figure 3.3 – 3–D Structural Model of MBEJ Test Specimen ........................................................ 12 
Figure 3.4 – Generic Loading Components ................................................................................... 14 
Figure 3.5 – Generic Loading Applied on the 3–D Model of Joint ............................................... 14 
Figure 3.6 – MVST Stirrup Results from 2–D Analysis under a reaction of 50kN/stirrup ............... 16 
Figure 3.7 – MV and MH Results from 3–D Analysis .................................................................... 16 
Figure 3.8 – S–N Curves from AASHTO LRFD (2007) Bridge Design Code ............................. 17 
Figure 4.1 – Twelve S-N Points Obtained for Stirrup Details from Fatigue Tests. ....................... 28 
Figure 4.2 – Type B Cracking in Connection S2 at 2 300 000 and 2 927 965 cycles ................... 32 
Figure 4.3 – Comparison of Results with S–N Curves of AASHTO for Stirrups of MJ1 ............. 34 
Figure 4.4 –Type A Cracking at Connections S2 (left) and S3 (right). ......................................... 35 
Figure 4.5 – Type B’ Cracking on S1 at 2 299 521 Cycles............................................................ 35 
Figure 4.6 – Comparison of Results with S–N Curves of AASHTO for Stirrups of MJ2 ............. 37 
Figure 4.7 – Type A Cracking at Connection S3. Test Stopped at 1 200 500 Cycles ................... 38 
Figure 4.8 – Comparison of Results with S–N Curves of AASHTO for Stirrups of MJ3 ............. 40 

vii.
1. INTRODUCTION 

1.1 State of the problem

TechStar Inc., hereafter called TechStar, is a manufacturer of bearings and expansion joints for
road and railway bridges. Over the last two decades, TechStar has expanded its fabrication base
considerably. Today TechStar supports a network of partners around the world producing
TechStar Engineered Products in North America, China, South Korea, India, Taiwan, Mexico,
and soon South America. TechStar is now one of the leading manufacturers of modular joints
with more than 5,000 installations in over 20 countries.

Increasingly, the tendency among bridge design engineers is to eliminate the number of smaller
capacity expansion joints in a structure, which have short life expectancies and need greater
maintenance, in favour of multiple-cell or Modular Bridge Expansion Joints (MBEJ). These latter
units are considerably more expensive to purchase and to replace, so manufacturers as well as
bridge owners and engineers are constantly seeking ways to enhance their performance including
their long-term service life.

One area of concern has been the performance of the various MBEJ designs with respect to
fatigue. Several studies have been carried out in this regard and most notably that conducted by
NCHRP for AASHTO. The NCHRP-402 Report (NCHRP, 1997) is becoming a worldwide
reference standard for the fatigue design of MBEJ.

Recently, TechStar brought an improvement to their MBEJ with welded stirrups. Mindful of the
critical importance of the long term fatigue resistance of its MBEJ, Z-TECH wanted to validate
this experimentally. The company wanted to ensure that the MBEJ and in particular the stirrups
welded to the center-beam and supporting the transverse support bars would respectively meet
the fatigue category C as defined in the AASHTO Bridge Design Code.
 
 
 
 

1.
1.2 Objectives
 
The main objective of this study is to evaluate the fatigue resistance of the modular single
support bar (LG) with welded stirrup joints with particular emphasis into the critical details of the
welded stirrups.

Specific objectives are as follows:

(a) Establish an experimental fatigue curve, with at least 10 points, for the stirrups
welding to the center-beam of the MBEJ with welded stirrups in full accordance with the
test procedure recommended by the NCHRP-402 Report (NCHRP, 1997); and
(b) Evaluate the fatigue rating of the stirrup details of the MBEJ identified above and
verify that they meet at least the fatigue category C of the AASHTO LRFD 2007 Bridge
Design Code.
 

1.3 Layout of Report

 
Chapter 2 covers the experimental program. The description of specimens, details of the testing
programs, including the load cases and conditions, as well as the test setup and test procedure are
outlined in this chapter. The structural models and results, including strains and stresses, from
the theoretical and analytical predictions are presented in Chapter 3. The experimental results
are reported in Chapter 4. For each specimen, two types of tests are presented: (a) calibration
tests where experimental strains and stresses under static loading are compared to those predicted
by the structural models presented in Chapter 3; (b) fatigue testing results including the number
of cycles to failure, the type of failure, and the fatigue rating category according to AASHTO
LRFD Bridge Design Code requirements. Finally, Chapter 5 presents the conclusions of the
investigation.
 

2.
2. EXPERIMENTAL PROGRAM 

2.1 Description of the Specimen and Instrumentation

Specimen
The test specimen is a full-scale center-beam/support bar assembly representative of that to be
used in field applications. Each specimen consists of three continuous center-beam spans over
four equally spaced support bars. Spans between support bars are equal to 0.91 m (3’-00”).
Support bars are made of the same section as the center-beam which is 3020 and their spans
(center to center of supports) is 0.95 m. The center-beam and the support bars are connected by
welded stirrups made of three steel welded plates (2 leg plates and 1 bottom thicker plate) which
are made of ASTM A572 grade steel. Stirrups are welded on center-beam and support bars are
clamped (in stirrups) by two steel discs and a Nylatron bearing. Details of specimens and fixtures
are provided in figure 2.1 and figure 2.2.
 

P1 P2 P3

S1 S2 S3 S4

where: X= 225mm for MJ1

185 mm for MJ2 and MJ3  
Figure 2.1 – Details of the Specimen
(Adapted from drawings provided by ALG Groupe-Conseil Inc.)
 

3.
  

 
 
 
 
 

Figure 2.2 – Details of the Specimen and Connections

(Adapted from drawings provided by ALG Groupe-Conseil Inc.)
 
Note: Prior to testing, specimens were visually inspected to double check/confirm the measurement and dimensions
and to detect any defects that can affect the fatigue testing.

4.
Instrumentation
The test specimens were instrumented in order to measure nominal strain ranges for the planned
load range. As pointed out in the NCHRP-402 Report (NCHRP, 1997), best results are usually
obtained when the load range interval does not pass through zero loads. To that end, a preload is
applied on specimens for each test: (1) 1kN per actuator for static loading, and (2) 30kN per
actuator for dynamic test (fatigue test).
The test specimens were all instrumented as indicated in figure 2.3 and figure 2.4. Additionally,
it was required that some strain gages be added on stirrups. Their locations are listed in table 2.1
for each test specimen and presented in figure 2.5.
 

P1 P2 P3

 
S1   S2 S3 S4
Figure 2.3 – Gages Locations on the Top of the Center beam
(Source: Drawing provided by ALG Groupe-Conseil Inc.) 
 
 

 
S1 S2   S3 S4
Figure 2.4 – Gage Locations on Bottom of Support Bars
(Source: Drawing provided by ALG Groupe-Conseil Inc.)
 

5.
 Table 2.1 - Location of Strain Gages on Stirrups
Stirrup S1 Stirrup S2 Stirrup S3 Stirrup S4
Specimen
(Gages – 36) (Gages – 37) (Gages – 38) (Gages – 39)
Bo1
Bo2
MJ1 Bo Bo Bo
To
To To To
Bi
Ti
ToL1 Bo
ToR1 To
MJ2 To To
ToR2 Bi
ToR3 Ti
Bo Bo
ToR1 To
MJ3 Bo To
ToR2 Bi
ToR3 Ti
Note: Bi: Bottom weld on the inside face of the stirrup leg
Bo: Bottom weld on the outside face of the stirrup leg (1 for the left side; 2 for the right side; if not
specified gage is on left side)
Ti: Top weld on the inside face of the stirrup leg
To: Top weld on the outside face of the stirrup leg
ToL1: Top outside face left side position 1
ToR1: Top outside face right side position 1 – middle place of aligned position (see on figure 2.5 (b))
ToR2: Top outside face right side position 2 – left place of aligned position (see on figure 2.5 (b))
ToR3: Top outside face right side position 3 – right place of aligned position (see on figure 2.5 (b))
 
 
 
 
(a) (b)

2    1    3 

Figure 2.5 – Strain Gage Locations on Stirrups

(Adapted from drawing provided by ALG Groupe-Conseil Inc.)

6.
Fixtures
The fixtures were designed as precisely as possible to avoid additional stresses being generated in
the specimen as a result of fixture misalignment. A bolted stop system as shown in figure 2.6 was
installed on each support bar to prevent the specimen to translate. Links between the fixtures and
the specimen are provided by steel bearings. Note that MBEJ fatigue tests were conducted at a
greater load ranges than those encountered at service loads in order to reduce the number of cycle
and hence the testing duration. The specially designed fixtures, as well as the experimental setup,
are shown in figure 2.7.

 
Figure 2.6 – View of the Bolted Stop System

         
Figure 2.7 – Specimen Installed on Inclined Seats of the Test Bench

7.
2.2 Experimental Program and Procedure

Experimental Program
Three specimens were tested. Two series of tests were performed on each specimen. The first
series were static calibration tests. These were aimed at: (i) validating the structural models
outlined in Chapter 3; and (ii) ensuring repeatability of data; and ensuring that no additional
stress is induced on the specimen by the setup. This series of tests was performed under static
load conditions (see table 2.2) prior to fatigue testing, as well as when change of loading or
support conditions occur. The second series involved dynamic fatigue testing. The loads were
applied in the form of sinus at a frequency of 2 Hz for the first two specimens and 1.5 Hz for the
third specimen. Details of the testing program are presented in table 2.3.

Table 2.2 – Calibration Tests under Static Loading

Static load applied per actuator Frequency
Specimen
(kN) (Hz)
MJ1 4 runs: 30 – 45 – 90 – 160 NA

MJ2 3 runs: 30 – 45 – 90 NA

MJ3 4 runs: 30 – 45 – 90 – 205 NA

 
Table 2.3 – Experimental Program – Fatigue Test
Applied static load Load range per actuator Frequency
Specimen
(kN) (kN) (Hz)
MJ1 95  65 2.0

MJ2 105  75 2.0

MJ3 117.5  87.5 1.5(a)

(a)
The frequency was reduced with increasing stress range to maintain an acceptable level of vibration. 

8.
Loading
Both vertical and horizontal load ranges were applied to the test specimen simultaneously in a
ratio of 5:1. This was achieved by inclining the specimen at 11.3 from the horizontal plane and
by applying the resultant load in the vertical plane, as illustrated in figure 2.7. It follows
therefore that the components of a resultant load range ∆P are:
- Vertical load range = PV = 0.98P
- Horizontal load range = PH = 0.20PV = 0.196P

Loads were applied under force control condition using hydraulic actuators through two 250mm-
long steel plates with hard rubber bearing pads, placed in contact with the bottom flange of the
center-beam. The actuators were not centered on outer spans, as the theoretical models showed
that a better load sharing between support bars and stirrups is achieved when loads are applied
slightly off-center, as shown in figure 2.8. Load ranges were monitored continuously throughout
the test to ensure that the desired load range was maintained.

P1 P2 P3

S1 S2 S3 S4  
Figure 2.8 – View of Specimen under Actuator Loads
 
Note: As pointed out in the NCHRP-402 Report (NCHRP, 1997), best results are usually obtained when the load
range interval does not pass through zero loads. Therefore, in this study a preload of 30kN was maintained
throughout the fatigue tests.
 

9.
 2.3 Failure Criteria and Stress Range Calculations
 
Definition of Cracking (Failure) Types
The NCHRP-402 Report defines modes of failure (cracking types) for standard MBEJ. However,
the report does not particularly cover the special case of MBEJ with welded stirrups. Therefore,
in this study, the NCHRP-402 failure criteria were adapted to the MBEJ with welded stirrups as
follows:
Type A Cracking
Type A cracking originates at the center-beam
weld toe and propagates straight into the center-
beam. A specimen shall be considered as failed
due to Type A cracking when a crack has grown
on any vertical face to a length of d/2 from the
point of origin, where d = depth of the center-
beam, as illustrated. 

 
 
WELDED LIPS 
Type B (Top) and B’ (bottom) Cracking
Type B (or B’) cracking originates at the top (or
bottom for B’) of the stirrup leg near the weld
head and grows in a plane parallel to the
transversal axis of the stirrup leg. A specimen
shall be considered as failed due to Type B (or
B’) cracking when a crack has reached a length
of b from the point of origin through any stirrup
leg, where b = depth of the stirrup leg. 
 
   

Figure 2.9 – Types A, B and B’ Cracking and Failure Criteria

(Inspired from NCHRP-402 Report (NCHRP, 1997))
 
Note: If a welded stirrup fails, it is replaced by a fixed support (i.e., rigid) and the fatigue test is resumed until the
targeted number of cycles is achieved.

10.
3. MODELLING AND ANALYSIS OF THE JOINT SPECIMENS 

3.1 Objectives

The objective of this chapter is twofold: (i) Present the structural modelling (1-D, 2-D and 3-D)
of the test specimen at different stages of their testing; and (ii) Present the theoretical equations
including numerical examples of the stresses and strains at locations of interest for the calibration
tests as well as for analysis and interpretation of experimental results. Note that stresses at
locations of the welded connections are calculated according to the guidelines of the NCHRP-402
Report (NCHRP, 1997), adapted herein to test conditions, whereas stresses and strains at the
locations of the strain gages are calculated according to the classical beam theory.

3.2 Presentation of the Structural Models and Loading

The structural models were created using SAP2000 software. The latter allows the analysis and
the design of multi-dimensional structures. First, a 1-D model was used for the analysis of the
center-beam which is the main element of the MBEJ specimen. This was followed by a 2-D
modelling of the stirrup to compute the stresses and the strains in the stirrups. Finally a 3-D
model was considered to predict the overall behaviour of the MBEJ test specimens including the
stirrups under static loading. The complexity of this 3-D model given the multiple components
of the specimen is noteworthy.

The 1-D, 2-D and 3-D models are respectively illustrated in figure 3.1, figure 3.2 and figure 3.3.
Table 3.1 and table 3.2 present the section and materials properties of the joint structural
elements used as input parameters for the theoretical models, whereas table 3.3 provides the
lengths of the members forming the joint specimen structural model.

11.
 Figure 3.1 – 1-D Structural Model of Center-Beam

Figure 3.2 – 2-D Structural Model of Stirrup

Figure 3.3 – 3-D Structural Model of MBEJ Test Specimen

12.
 Table 3.1 – Section Properties of the Components
(b)
Stirrup Center-beam
(a) Nylatron
Properties Bottom &
Leg bearing
plate Support bar
Material Steel Steel Steel Nylatron
Area (mm2) 991 2476 4500 5026
6 4
Moment of inertia about x-x axis : Ix (10 mm ) 0.013 0.208 10.7 -
6 4
Moment of inertia about y-y axis,: Iy (10 mm ) 0.502 1.256 1.82 -
Section modulus about x-x axis : Sx (mm3) 2 100 13 100 165 000 -
3
Section modulus about y-y axis : Sy (mm ) 12 880 32 200 45 150 -
Notes:
(a) For center-beam x-x- is the strong axis, whereas it is the weak axis for stirrup elements.
(b) Welded lips were added to the center-beam as shown in figure 2.9. This will change the location of the center of
gravity, which result in center-beam properties to be used for calculations, as follows:
- Moment of inertia about strong axis, x-x: Ixx = 10642453 mm4
- Section modulus of center-beam for bottom fiber: Sxx.bot = SxCBbot = 158606 mm3
- Section modulus of center-beam for top fiber: Sxx.top = SxCBtop = 170279 mm3
- Section modulus about weak axis, y-y: Syy = 46542 mm3

Table 3.2 – Material Properties of the Members

Properties Support bars & center-beam Stirrups Nylatron
Grade ASTM A572, 50ksi ASTM A572, 50ksi NSM
Modulus of elasticity : E (GPa) 200 200 2.764
Shear modulus : G (MPa) 76923 76923 1 107
Poisson’s ratio : ν 0.3 0.3 0.4
Density (kN/m3) 76.98 76.98 11.5
Yield stress : Fy (MPa) ± 350 350
Ultimate stress : Fu (MPa) ± 450 450

Table 3.3 – Member Lengths for the 3-D Model

Member Designation Member length (mm)
Center-beam 3110.00
Support bar 950.00
Steel disk 25.40
Stirrup leg 190.30
Stirrup bottom plate 140.00
Nylatron bearing 19.05

13.
Results for a generic loading of 102kN per actuator are presented herein. The loading is a 100kN
vertical load (= 102kN  cos(11.3°)) combined to a horizontal component of 20kN (= 0.2∆Pv) as
illustrated in figure 3.4. Loads are applied on a 0.25m plate by each of the two actuators,
resulting thereby in two distributed loadings of 400kN/m (=100kN/0.25m) and 80kN/m applied
as shown in figure 3.5. Loads were applied to the downside face of the bottom center-beam
flange, therefore the horizontal component of the load induced an additional bending moment of
5.18kNm relative to the centroid of the center-beam (= PH  dcb/2 = 50kN  0.1296/2 m). These
results are then to be linearly scaled as a function of the load actually applied on the specimen.

PV=0,98∆P 

PH=0,2∆P
Figure 3.4 – Generic Loading Components

Figure 3.5 – Generic Loading Applied on the 3-D Model of Joint

14.
Degrees of Freedom for the MBEJ Modeling
 1-D Model – Center-beam. The specimen is modeled as a continuous beam over four
elastic supports with a vertical stiffness of 83.3kN/mm. This value takes into account the
flexibilities of : (i) the support bar (117.7kN/mm), (ii) the stirrup legs (466.7kN/mm) and (iii) the
Nylatron bearing (731.2kN/mm). Other displacements are blocked along the longitudinal and
transversal directions. Finally, the rotation about the longitudinal axis is blocked to ensure the
stability of the model.
 2-D Model – Stirrup. The stirrup is modeled by 1-D beam elements assembled in a 2-D
structure. The modeling integrates components such as stirrup leg, bottom plate and a T-section is
used to represent the attachment plate welded to the stirrup left leg. The stirrup legs are fixed
(i.e., blocking all degrees of freedom) at the level of their connection to the center-beam.
 3-D Model – Global joint. The MBEJ is modeled by 1-D beam elements assembled in a 3-
D structure. This model integrates most components of the specimen. The stirrup 2-D model is
assembled to the center-beam 1-D model. The four stirrups are linked to the support bars by
Nylatron bearings and steel disks. Each support bar seats on two rigid supports and the
displacements are blocked in the three directions but the rotations are allowed around the three
axes. The device used to avoid the transversal displacement is also modeled.

Response to loadings
Internal forces in the center-beam and support bars obtained from the 3-D structural analysis of
the specimen during initial phase (before any failure) and under the generic loading of 102kN per
actuator are presented in figures 3.7. For the stirrup, the results obtained from 2-D modeling are
presented in figure 3.6. These results are used for the calculation sample of the stress range and
the number of cycles to failure in the welded connection detail, as described in the section 3.4
and section 3.5 of this report.

15.
 MvST = 1.360kN.m
50,4 -0.0778kN.m 0.0710kN.m

0.0289kN.m -0.0357kN.m

Figure 3.6 – MvST Stirrup Results from 2-D Analysis under a reaction of 50kN/stirrup

MV CB min= -6.18kN.m MH CB min= -1.03kN.m

MV SB max= 12.47kN.m MH SB max= 0.73kN.m

MV CB max= 16.27kN.m MH CB max= 3.02kN.m

Figure 3.7 – MV and MH Results from 3-D Analysis

16.
3.3 AASHTO LRFD Fatigue Curves

The different AASHTO LRFD (2007) fatigue categories are defined in terms of S-N (Stress
range versus Number of cycles) curves, as presented in figure 3.8. The relation between Stress
range and Number of cycles in the finite life range is given by:

N  A
3
(3.1)
S reff

Where: N = Number of cycles; A = Detail constant, as presented in Table 3.4; Sreff = Nominal
effective stress range at the detail under consideration (in ksi).

Table 3.4 – Detail Category Constant A

Source: NCHRP-402 Report (NCHRP, 1997)
Constant A Constant A CAFL
Detail Category
(imperial values) (metric values) (MPa)
B 120.0  108 3.93  1012 110
B’ 61.0  108 2.00  1012 83
C 44.0  108 1.44  1012 69
C’ 44.0  108 1.44  1012 83
D 22.0  108 7.21  1011 48

For the welded connection detail, the effective stress range used in Eq. (3.1) will be the largest
value from the three possible cracking types: A, B and B’. These cracking types were defined in
section 2.3 and illustrated on figure 2.9.

Figure 3.8 – S-N Curves from AASHTO LRFD (2007) Bridge Design Code

17.
3.4 Evaluation of Stress Range

Stirrup Cracking Types

According to the NCHRP-402 Report (NCHRP, 1997), the nominal stress range (Δf) for each
cracking type and for each connection can be determined using Eq. (3.2). These stress ranges are
calculated differently for Type A, B and B’ cracking, as presented below.

∆ (3.2)

Where: Δf = Nominal stress range at weld toe,

SRB = Horizontal bending stress range in the center-beam or support bar,
SRZ = Vertical stress range in the weld or the base metal of the stirrup.

Table 3.5 presents a summary of internal forces and bending moments obtained from the 1-D
structural model and used to calculate the stress ranges for the cracking patterns Type A, B, and
B’.

18.
 Table 3.5 – Section Properties and Analysis Results Used for Calculation of the Nominal
Stress Range for Type A, B and B’ Cracking
Parameters Type A Type B & B’

Awcb (mm2) 3432 -

SXcb bot (mm3) 158 606 -
SXcb top (mm3) - -
SYcb (mm3) 46 542 -
AST (mm2) - 990.6
SXST (mm3) - 2 090
SYST (mm3) - -
llp (mm) - 190.3

A B (top) B’ (bottom)
Outer Inner Outer Inner Outer Inner
RV (kN) 50.22 49.78 50.22 49.78 50.22 49.78
RH (kN) - - - - - -
MVCB (kNm) 4.13 6.10 0.0359 0.0355 0.0780 0.0774
MHCB (kNm) 0.82 1.24 - - - -

Notations used in table 3.5:

RV = Vertical reaction in the connection,
RH = Horizontal reaction in the connection,
MVCB = Vertical bending moment in center beam due to the applied vertical force,
MHCB = Horizontal bending moment in center beam due to the applied horizontal force,
AWcb = Area of weld at the top of the stirrup leg = (lwelding Lwelding  Nwelding/leg  Nleg),
SXcb and SYcb = Section modulus of the center-beam with welded lips,
SXlp and SYlp = Section modulus of one stirrup leg about weak (x-x) and strong (y-y) axes,
dcb = Depth of the center beam,
dlp = Length of the lateral plate,
llp = Length of the lateral plate,
AST = Area of one stirrup leg.

The following section presents detailed calculation examples for the inner connections at the
initial phase of testing (i.e. before any crack formation). It should be noted that the stress ranges
at the outer connections are calculated using the same equations.

19.
Type A Cracking (Center-beam Weld Toe Cracking)
The generic load used for structural analysis of 1-D joint model is ∆ 102 .

6 097 481 . 1 244 300 .

65.2
158 606 46 541.7

49 777.0
14.5
3 432.0

∆ 66.8

Type B Cracking (Stirrup Top Leg Cracking)

The generic load used for structural analysis of stirrup model is ∆ 50 / . Note that
the comparison between data of 1-D and 2-D models shows that reactions are very close but not
quite identical since:
Rvouter = 50.22kN and Rvinner = 49.78kN

49 777.0 35 511.9 .
42.1
2 990.6 2 096.8

∆ 42.1

Where: MvST = maximum bending moment in the stirrup at the welded connection of the
stirrup plates (see figure 3.7).

20.
Type B’ Cracking (Stirrup Bottom Leg Cracking)
The generic load used for structural analysis of stirrup model is ∆ 50 / . Note that
the comparison between data of 1-D and 2-D models shows that reactions are very close but not
quite identical since:
Rvouter = 50.22kN and Rvinner = 49.78kN

49 777.0 77 430.1 .
62.1
2 990.6 2 096.8

∆ 62.1 (see note below)

Note: The distortions in the stirrup legs due to the welding may induce a secondary type moment in the stirrup,
resulting thereby in an increase of SRB. In this study, the distortions of the stirrups of specimen MJ3 have been
carefully measured and a maximum distortion of 1.5mm was obtained. From simulation, this yielded an
increase of SRB from 62.1MPa to 75MPa, which is an increase of 21%.

Stress Ranges Recap

The nominal stress range (∆f) corresponding to the cracking patterns are presented in table 3.6.
Similarly, the stress ranges in outer connections were calculated using the same equations as
inner connections, and results are also presented in table 3.6.

Table 3.6 – Nominal Stress Range for Identified Cracking Types

Stirrup Cracking Types

A B B’
Inner stirrup/
Stress 66.8 42.1 62.1
Inner span
range
Outer stirrup/
∆f (MPa) 46.1 42.4 62.6
Outer span

21.
From the table, it is seen that the maximum stress range for an applied load of 102kN per actuator
is 66.8MPa for inner stirrup and corresponds to type A cracking. For outer stirrup, a maximum
stress range of 62.6MPa is observed and corresponds to type B’ cracking.

Linear scaling is used to obtain the stress values corresponding to the fatigue loading program of
the present study, as presented in table 3.7.

Table 3.7 – Stress Range after Scaling

∆  65kN/actu. Stirrup Cracking Types
MJ1 95kN/actu. A B B’
Inner Stirrup/
Stress 85.1 54.3 79.1
Inner span
range
Outer Stirrup/
∆f (MPa) 58.8 54.0 79.8
Outer span

3.5 Evaluation of Number of Cycles

Sample calculation of the minimum required number of cycles

The calculation is based on 1-D joint model as it is the simplest model and yet sufficiently
accurate. As shown in table 3.6 the calculated stress range in the inner connections for the
generic resultant load range of 102kN/actuator is 66.8MPa. Note that no crack is considered for
the calculation.
For the specimen MJ1, the load range applied during the initial phase of fatigue testing is
130kN/actuator. So the stress range in the inner connections for this load range (ftest) is:
130
∆ 66.8 85.1 12.3
102
The following equation gives the minimum required number of cycles considering AASHTO
Categories C coefficient:
1.44 10 44.0 10
2.335 10
∆ 85.1 12.3

The minimum life expectancies were evaluated for the stirrup details of specimens to meet
Category C requirements, as presented in table 3.8.

22.
 Table 3.8 – Life Expectancies for Stirrup Details to Meet AASHTO Category C
Number of cycles required for the detail to meet Category C requirements
Outer Inner
Gov. stress range Ncycles Gov. stress range Ncycles
MJ1 79.8 2 833 000 85.1 2 335 000
MJ2 92.1 1 844 000 98.2 1 520 000
MJ3 107.4 1 161 000 114.6 957 000

3.6 Predicted Strains at Gage Locations

Theoretical strain values at strain gage locations are calculated from corresponding stresses using
the classical beam theory as follows:

Step 1: Determine vertical (V) and horizontal (H) stresses at strain gage locations, where V
and H are stresses induced by the perpendicular and the tangential load component, respectively.
As a reminder: Vertical Load Range = PV = 0.981P; and (ii) Horizontal Load Range = PH =
0.196P.

- For center-beam (1-D and 3-D model) and support bars (3-D model) the moment values are
given by theoretical model and stresses are determined as follows:

. .

- For stirrups (2-D and 3-D model) Mbot and Mtop (given by theoretical models) are interpolated
to obtain bending moment values at gage locations.

23.
Step 2: Determine strains using the modulus of elasticity, E.

V   V H   H
E E

Step 3: Compute the resultant strain: ε = εV + εH

The theoretical strain values obtained at gage locations are presented in Chapter 4, and
compared with experimental values.

24.
4. RESULTS AND OBSERVATIONS    
A summary of the testing results and comparison with AASHTO fatigue categories are presented
in sections 4.1 and 4.2. The repeatability and calibration test results are presented in section 4.3,
whereas the detailed results are presented in section 4.4 for each specimen.

4.1 Summary of Results

Repeatability and Calibration Tests

The stability of the test setup, the repeatability of the readings and the validity of the models were
established for the center-beam and the support bars. However, the readings of the stirrup gages
were found to be sensitive and their repeatability was less regular. Manufacturing tolerances,
misalignment and residual deformations induced in the stirrup legs during the welding process
(heat) are pointed out as having an important effect on the strain and stress ranges at the critical
section of the stirrups. It follows that applying the calibration criterion of 25% as a maximum
difference between the measured and the predicted theoretical strains for the stirrups were found
to be difficult to systematically achieve. In this study a 2-D model was used for the analysis of
the stirrups as it was found to capture the global stress patterns and to yield a good prediction of
the maximum stress range at critical stirrup sections.

Fatigue testing Results

The test results for stirrups of the three specimens are presented in table 4.1. The three cracking
types, inspired from the NCHRP-402 Report (NCHRP, 1997) and defined in Figure 2.9, occurred
during the testing as presented below:

Type A cracking occurred three times. Two of the cracks progressed in center-beam at inner
stirrups (S2 and S3) left side (where attachment plates are welded) during the second test. The
third crack occurred during the third test at the right inner stirrup (S3) and the same location as
the first two cracks.

Type B cracking occurred one time during the first test. The crack progressed in the left inner
stirrup (S2) left leg and reached failure criterion.

Type B’ cracking occurred one time during the second test. The crack progressed in the first
stirrup (S1) right leg and reached failure criterion.

25.
 Table 4.1 – Test Results for Stirrups of the Three Specimens

Nb. of cycles Crack Nb. of cycles Crack

Comments
at first crack type at failure Propagation
MJ1 S1 - - - - No crack

Test stopped at
3 163 276
S2 In 2 300 000 B 2 927 965 Top Stirrup Failure
S3 - - - - No crack
S4 - - - - No crack
MJ2 S1 In 2 125 000 B’ 2 299 521 Stirrup leg Failure

Test stopped at
2 877 500
S2 In 1 334 000 A - Center-beam No failure
S3 Out 1 334 000 A - Center-beam No failure
S4 - - - - No crack

MJ3 S1 - - - - No crack

Test stopped at
1 200 550
S2 - - - - No crack
S3 Out 710 000 A - Center-beam No failure
S4 - - - - No crack
Note: In = Inner side = cracks occurs between stirrups and loading point
Out = Outer side = cracks occurs between the two central stirrups

4.2 Comparison with AASHTO Fatigue Categories

When a connection fails, the stress ranges in the remaining connections change. Although
adjustments were made to keep the stress range quasi-constant in some locations, stress range
variations were observed elsewhere. The linear Miner damage rule was therefore used to take
into account such variations by computing the equivalent stress range for each connection as
presented hereafter. Thus according to Miner’s rule, the equivalent stress range Seq, which is the
stress range that produces the same damage fraction as the real test, is determined as follows:

n S i i
3

Seq  3 i 1
m (4.3)
n i 1
i

26.
Where: Si = stress range during the ith loading phase, ni is the number of cycle of the ith loading phase,
Σni = Neq = Ntest

The Seq – Ntest points obtained from tests on the three specimens are plotted in figure 4.1 along
with the S-N lower limit curves specified by AASHTO LRFD 2007 Bridge Design Specification.

As can be seen from this figure, the fatigue resistance of the welded MBEJ is compatible and
meet category C (for stirrup details) requirements as defined by the AASHTO LRFD 2007
Bridge Design Code.

Results for each of the specimen are detailed in the following sections.

27.
Figure 4.1 – Twelve S-N Points Obtained for Stirrup Details from Fatigue Tests in Conformity to AASHTO and NCHRP-402
4.3 Repeatability and Calibration Tests

Repeatability and calibration tests were performed for all the specimens. However, for
convenience only representative results are shown here. Table 4.2 presents experimental strains
obtained at different locations for the three static calibration test runs (corresponding to applied
load of 30kN, 45kN and 90kN) that were performed prior the fatigue testing of the specimen
MJ3.

Table 4.2 – Experimental Strains for Three Static Calibration Test Runs for MJ3
Measurements
Maximum variation %
average
Load 30 45 90 30 45 90
J1 72 105 208 0 1 0
J2 31 46 98 1 1 1
J3 139 208 419 0 1 0
J4 74 112 235 0 1 0
J5 72 103 199 0 2 0
J6 32 48 100 0 2 1
J7 -2 -6 -22 - 5 3
J8 -13 -18 -32 3 2 0
J9 -71 -107 -217 0 2 0
J10 -49 -72 -141 1 1 1
Center-beam

J11 -73 -108 -215 0 2 0

J12 -45 -65 -128 1 1 1
J13 -81 -117 -224 0 1 0
J14 -46 -66 -128 1 2 0
J15 -87 -123 -233 0 2 0
J16 -51 -72 -139 0 2 0
J17 -11 -13 -16 0 5 0
J18 -17 -21 -32 0 3 2
J19 59 92 195 0 1 0
J20 28 45 99 2 1 0
J21 134 201 407 0 1 0
J22 77 116 238 0 1 0
J23 66 98 198 1 1 1
J24 27 42 90 0 1 0
J25 138 205 411 0 1 0
J26 -60 -89 -172 0 2 0
J27 134 204 419 0 1 0
J28 -56 -84 -173 1 2 1
Support bars

J29 -69 -103 -209 0 1 0

J30 -56 -85 -171 1 2 0
J31 -69 -102 -204 0 2 0
J32 -57 -83 -167 0 2 0
J33 -72 -106 -210 0 2 0
J34 -62 -91 -183 1 2 0
J35 -72 -107 -216 1 1 0

29.
These experimental values were then compared with results from SAP2000 analysis to validate
the models. Table 4.3 and table 4.4 provide results for 1-D and 3-D models of MJEB.
Note that the values recorded by the gages near support bars are not shown in the tables as they
were deemed meaningless. These gauges, being located at the inflection zone of the beam,
experienced very small values of strain and often generate variations (expressed in percent)
greater than 25%. The comparison of results clearly shows that the 1-D model is accurate enough
for the purpose of this study.

Table 4.3 – Comparison of Calibration Test Results with 1-D Theoretical Values
         30kN        45kN        90kN    
   εth  Gap  εth  Gap  εth  Gap 
Gage  Aver.  Aver.  Aver. 
   (μe)  (%)  (μe)  (%)  (μe)  (%) 
J3  150  139  8%  225  208  8%  450  419  7% 
J4  85  74  15%  127  112  13%  254  235  8% 
J5  78  72  9%  118  103  14%  235  199  18% 
J6  45  32  39%  67  48  39%  134  100  33% 
J11  ‐70  ‐73  4%  ‐106  ‐108  2%  ‐211  ‐215  2% 
Center-beam 

J12  ‐39  ‐45  13%  ‐59  ‐65  9%  ‐118  ‐128  8% 
J13  ‐70  ‐81  13%  ‐106  ‐117  10%  ‐211  ‐224  6% 
J14  ‐39  ‐46  15%  ‐59  ‐66  11%  ‐118  ‐128  8% 
J19  78  59  33%  118  92  28%  235  195  21% 
J20  45  28  57%  67  45  50%  134  99  35% 
J21  150  134  12%  225  201  12%  450  407  11% 
J22  85  77  11%  127  116  10%  254  238  7% 
J25  142  138  3%  213  205  4%  426  411  4% 
J26  ‐55  ‐60  9%  ‐82  ‐89  7%  ‐164  ‐172  4% 
J27  142  134  6%  213  204  4%  426  419  2% 
                                
J28  ‐81  ‐56  43%  ‐121  ‐84  43%  ‐242  ‐173  40% 
Support bars 

J29  ‐75  ‐69  8%  ‐113  ‐103  9%  ‐225  ‐209  8% 
J30  ‐80  ‐56  42%  ‐120  ‐85  41%  ‐240  ‐171  40% 
J31  ‐74  ‐69  8%  ‐112  ‐102  10%  ‐223  ‐204  10% 
J32  ‐80  ‐57  40%  ‐120  ‐83  44%  ‐240  ‐167  43% 
J33  ‐74  ‐72  3%  ‐112  ‐106  6%  ‐223  ‐210  7% 
J34  ‐81  ‐62  31%  ‐121  ‐91  33%  ‐242  ‐183  32% 
J35  ‐75  ‐72  5%  ‐113  ‐107  5%  ‐225  ‐216  4% 

30.
 Table 4.4 – Comparison of Calibration Test Results with 3-D Theoretical Values
        30kN        45kN        90kN    
   εth  Gap  εth  Gap  εth  Gap 
Gage  Aver.  Aver.  Aver. 
   (μe)  (%)  (μe)  (%)  (μe)  (%) 
J3  144  139  3%  216  208  4%  432  419  3% 
J4  85  74  15%  128  112  14%  255  235  9% 
J5  74  72  3%  111  103  7%  222  199  11% 
J6  44  32  37%  66  48  37%  132  100  31% 
J11  ‐69  ‐73  6%  ‐103  ‐108  5%  ‐206  ‐215  4% 
Center-beam 

J12  ‐41  ‐45  9%  ‐62  ‐65  5%  ‐123  ‐128  3% 
J13  ‐69  ‐81  15%  ‐103  ‐117  12%  ‐206  ‐224  8% 
J14  ‐41  ‐46  11%  ‐62  ‐66  7%  ‐123  ‐128  4% 
J19  74  59  26%  111  92  20%  222  195  14% 
J20  44  28  55%  66  45  48%  132  99  33% 
J21  144  134  7%  216  201  8%  432  407  6% 
J22  85  77  11%  128  116  10%  255  238  7% 
J25  139  138  1%  209  205  2%  418  411  2% 
J26  ‐55  ‐60  9%  ‐82  ‐89  7%  ‐165  ‐172  4% 
J27  139  134  4%  209  204  2%  418  419  0% 
                                
J28  ‐59  ‐56  5%  ‐89  ‐84  5%  ‐178  ‐173  3% 
J29  ‐45  ‐69  36%  35%  36% 
Support bars 

‐67  ‐103  ‐134  ‐209 

J30  ‐64  ‐56  14%  ‐97  ‐85  14%  ‐193  ‐171  13% 
J31  ‐57  ‐69  18%  ‐85  ‐102  17%  ‐170  ‐204  17% 
J32  ‐64  ‐57  13%  ‐97  ‐83  16%  ‐193  ‐167  15% 
J33  ‐57  ‐72  22%  ‐85  ‐106  20%  ‐170  ‐210  19% 
J34  ‐59  ‐62  4%  ‐89  ‐91  2%  ‐178  ‐183  3% 
J35  ‐45  ‐72  38%  ‐67  ‐107  38%  ‐134  ‐216  38% 

31.
4.4 Stirrups Detailed Testing Results

Specimen MJ1

A cracking initiated in S2 connection at 2 300 000 cycles. It progressed in the top left stirrup leg,
indicating a type B cracking. It reached the rupture criterion at 2 927 965 cycles (see figure 4.2).
After its failure, connection S2 was blocked using a rigid support and the test program resumed to
test the other connections.

Figure 4.2 – Type B Cracking in Connection S2 at 2 300 000 and 2 927 965 Cycles

No other cracks occurred in stirrups during the fatigue test. Table 4.5 summarizes the different
loading steps for each connection, whereas a summary of results is provided in table 4.6. The
equivalent stress ranges for the four connections were computed and are provided in table 4.7.

Table 4.5 – Loading Program during Testing of MJ1

Connections Cycles Actuators Load
S1 S2 S3 S4 From 0 to 3 163 276. A1 and A2 95 kN +/- 65 kN

32.
 Table 4.6 – Summary of Test Results for MJ1

Nb. of cycles at Crack Nb. of cycles at Crack

Comments
first crack type failure Propagation
S1 - - - - No crack

Test stopped at
3 163 276
S2 2 300 000 B 2 927 965 Top Stirrup Failure
S3 - - - - No crack
S4 - - - - No crack

Table 4.7 – Stress Range during Testing of MJ1

1st Step 2nd Step Seq
Actuators A1 and A2 A1 and A2
Resultant load 130kN 130kN
range
Structural model

Nb. of cycles: 2 927 965 235 311

S1 79.8 72.6 79.4
Stress Range

S2 85.1 Failure (rigid support installed) 85.1

(MPa)

S3 85.1 72.3 84.4

S4 79.8 83.0 80.0

The Seq-Ntest (Equivalent Stress range versus Number of cycles from test) failure points obtained
experimentally are plotted in figure 4.3. The four points obtained for MJ1 are above the S-N
lower limit curve specified by AASHTO Bridge Design Specifications for category C.

33.
Figure 4.3 – Comparison of Results with S-N Curves of AASHTO for Stirrups of MJ1

34.
Specimen MJ2

Two cracks initiated in S2 and S3 connection at 1 340 000 cycles. They progressed in the center-
beam from stirrups left legs weld toe, indicating a type A cracking (see figure 4.4). The cracks,
which progressed in S3 connection, nearly reach the rupture criterion at 2 877 500 cycles (end of
test) but never did.

Figure 4.4 –Type A Cracking at Connections S2 (left) and S3 (right).

The third crack in S1 connection is a type B’ cracking (see figure 4.5). The crack progressed in
the bottom of stirrup left leg and reached rupture at 2 229 521 cycles. Connection S1 was blocked
after failure using a rigid support and the test program resumed to test the other connections.

Figure 4.5 – Type B’ Cracking on S1 at 2 299 521 Cycles

No other cracks occurred in stirrups during the fatigue test. Table 4.8 summarize the different
loading steps for each connection, whereas a summary of results is provided in table 4.9. The
equivalent stress ranges for the four connections were computed and are provided in table 4.10.

35.
 Table 4.8 – Loading Program during Testing of MJ2
Connections Cycles Actuators Load
S1 S2 S3 S4 From 0 to 2 877 500 A1 and A2 105kN +/- 75kN

Table 4.9 – Summary of Test Results for MJ2

Nb. of cycles at Crack Nb. of cycles at Crack

Comments
first crack type failure Propagation
S1 In 2 125 000 B’ 2 299 521 Stirrup leg Failure

Test stopped at
2 877 500
S2 In 1 334 000 A - Center-beam No failure
S3 Out 1 334 000 A - Center-beam No failure
S4 - - - - No crack

Table 4.10 – Stress Range during Testing of MJ2

1st Step 2nd Step Seq
Actuators A1 and A2 A1 and A2
Resultant load 150kN 150kN
range
Structural model

Nb. of cycles: 2 299 521 577 979

S1 92.1 Failure (rigid support installed) 92.1
Stress Range

S2 98.2 83.9 95.7

(MPa)

S3 98.2 97.1 98.0

S4 92.1 92.0 92.1

The Seq-Ntest failure points obtained experimentally are plotted in figure 4.6. Again, the four
points obtained for MJ2 are above the S-N lower limit curve specified by AASHTO Bridge
Design Specifications for category C and C’.

36.
Figure 4.6 – Comparison of Results with S-N Curves of AASHTO for Stirrups of MJ2

37.
Specimen MJ3

The first crack initiated in connection S3 at 710 000 cycles. It progressed in the center-beam from
stirrups left legs weld toe, indicating a type A cracking (see figure 4.7). The crack progressed in
connection S3 and nearly reached rupture at the 1 200 550 cycles (end of test) but never did.

Figure 4.7 – Type A Cracking at connection S3. Test stopped at 1 200 500 cycles

No other cracks occurred in stirrups during the fatigue test. Table 4.11 summarize the different
loading steps for each connection, whereas a summary of results is provided in table 4.12. The
equivalent stress ranges for the four connections were computed and are provided in table 4.13.

Table 4.11 – Loading Program during Testing of MJ3

Connections Cycles Actuators Load
S1 S2 S3 S4 From 0 to 1 200 550 A1 and A2 117.5kN +/- 87.5kN

Table 4.12 – Summary of Test Results for MJ3

Nb. of cycles at Crack Nb. of cycles at Crack

Comments
first crack type failure Propagation
S1 - - - - No crack
Test stopped at
1 200 550

S2 - - - - No crack
S3 out 710 000 A - Center-beam No failure
S4 - - - - No crack

38.
 Table 4.13 – Stress Range during Testing of MJ3
Unique Step Seq
Actuators A1 and A2
Resultant load range 150kN

Structural model

Number of cycles: 1 200 550

S1 107.4 107.4
Stress Range

S2 114.6 114.6
(MPa)

S3 114.6 114.6
S4 107.4 107.4

The Seq-Ntest failure points obtained experimentally are plotted in figure 4.8. Likewise MJ1 and
MJ2, the four points obtained for MJ3 are above the S-N lower limit curve specified by
AASHTO Bridge Design Specifications for category C and C’.

39.
Figure 4.8 – Comparison of Results with S-N Curves of AASHTO for Stirrups of MJ3

40.
5. CONCLUSIONS 

Three specimens of Single Support Bar Modular Bridge Expansion Joint (MBEJ) with Welded
Stirrups were tested in fatigue according to the requirements of AASHTO LRFD 2007 Bridge
Design Code and the NCHRP-402 Report (NCHRP, 1997). The main objective of the study was
to evaluate the fatigue resistance of critical details of this modular joint system, with particular
emphasis on the welded stirrups. Specific objectives are as follows: (a) establish an experimental
fatigue curve, with at least 10 points, for the stirrups welding to the center-beam detail, and (b)
verify that the stirrup details of the MBEJ identified above meet fatigue Category C of the
AASHTO LRFD 2007 Bridge Design Code requirements.

The following conclusions can be drawn:

Fatigue cracks and failures have been observed at the locations of the welded stirrup connections.
Test results were used to produce an experimental fatigue curve of the stirrup details with all the
twelve (12) points located above the AASHTO Fatigue curve for category C. Hence, it can be
concluded that the fatigue resistance of the welded stirrup connection details of the TechStar
Single Support Bar MBEJ is compatible with and meets the fatigue Category C
requirements of the AASHTO LRFD 2007 Bridge Design Code.

Furthermore, this study confirms that this type of MBEJ can be modelled by simple 1-D beam
element models with reasonable accuracy. However, bending moment within the stirrups was
found to be important and was therefore considered in this study. To this end, a 2-D model for the
stirrups was successfully used. Clearly, design of stirrups for fatigue should take into account the
bending moment to achieve an enhanced accuracy of the calculations.

41.
 REFERENCES

American Association of State Highways and Transportation Officials (AASHTO), AASHTO

LRFD BRIDGE DESIGN SPECIFICATIONS, 4nd Edition, SI units, Washington, D.C., 2007.

Dexter, R.J., Connor, R.J. and Kaczinski, M.R. (1997) National Cooperative Highway Research
Program (NCHRP) Report 402 - Fatigue Design Of Modular Bridge Expansion Joints,
Transportation Research Board, Washington, D.C.

42.