You are on page 1of 10

Available online at www.sciencedirect.

com
Procedia
Engineering
Procedia
Procedia Engineering
Engineering 00 (2009)
2 (2010) 000–000
573–582
www.elsevier.com/locate/procedia
www.elsevier.com/locate/procedia

Fatigue 2010

Fatigue life evaluation through field measurements and laboratory


tests
Kiana Kashefia,b *, A. Peyman Zandia,b, Mostafa Zeinoddinia
a
Faculty of Civil Engineering, K. N. Toosi University of Technology, No. 1346, Vali-Asr St., Tehran, Postcode 1996715433, Iran
b
TAZAND Consulting Engineers Company, No. 19, Sarv Alley, Arash Blvd.,Farid Afshar St., Zafar, Tehran, Postcode 1916646664,Iran

Received 3 March 2010; revised 11 March 2010; accepted 15 March 2010

Abstract

The maintenance cost of old bridges is rapidly increasing, since many of them are reaching their design life. Hence requirements
for remaining life evaluation of existing bridges are overwhelmingly increasing for bridge authorities all over the world. Fatigue
phenomenon seems to be a great concern in this regard whereas no theoretical methods can evaluate the remaining life of existing
bridges accurately.
This paper presents the procedures used for assessment of remaining fatigue life of an aging orthotropic steel deck bridge in
Tehran including strain field measurements carried out on critical fatigue details as well as laboratory tests on limited samples,
taken from similar details of an identical disassembled bridge.
Field measurements have been carried out under normal traffic loading during an 8-day period of time. Laboratory tests are made
on samples taken from the most critical fatigue details on a region of high stress fluctuation (mid span) and the same details at a
stagnant region (supports location) and compared. Similar results were obtained from both approaches showing the investigated
bridges will not experience fatigue failure under the current traffic regime.
c 2010 Published by Elsevier Ltd. Open access under CC BY-NC-ND license.

Keywords: Remaining Fatigue Life; Orthotropic Steel Deck Bridge; Field measurement; Laboratory test; Structural health monitoring.

1. Introduction

The extent of deterioration of the existing infrastructures such as bridges all over the world has prompted many
authorities to investigate the existence of damages and make cost-effective decisions for their rehabilitation.
Structural Health Monitoring (SHM) is a new strategy that has been developed recently for non-destructive damage
evaluation of structures under various service loads. SHM can provide the authorities with useful information for
management of their structural inventories. For example in a bridge management system different sensors are used
to measure responses such as strain and acceleration and detect probable damages. [1]
One of the probable damages which bridges may experience, due to increasing traffic volume as well as
environmental degradations such as corrosion, is fatigue phenomenon. Specifically fatigue cracks have been
reported in several types of welded joints of orthotropic deck steel bridges [2]. However, orthotropic decks are

* Corresponding author. Tel.: +98-21-8877-9474; fax: +98-21-8877-9476.


E-mail address: kiana0k085@yahoo.com.

1877-7058 c 2010 Published by Elsevier Ltd. Open access under CC BY-NC-ND license.
doi:10.1016/j.proeng.2010.03.062
574 K. Kashefi et al. / Procedia Engineering 2 (2010) 573–582

2 K. Kashefi et al. / Procedia Engineering 00 (2010) 000–000

widely used in construction of bridges. This is because they provide a considerable reduction in dead loads which
are a great concern in the bridge design. Hence, assessment of their remaining fatigue life is necessary for
rehabilitation, retrofitting or replacement plans.
Generally, there are two approaches for assessment of remaining fatigue life;
• A fracture mechanics approach, considering local stress intensity, crack initiation and propagation by further
loading cycles.
• A classical approach, using S-N curves/Palmgren–Miner rule for estimating cumulative damage using an
appropriate load spectra, to achieve an acceptable probability that fatigue failure will not occur before a given
number of cycles.
There is a requirement for accurate primary information on loading, stress ranges and loading cycles with both
approaches. In order to incorporate this information various methods have been adopted in relevant specifications.
Among them is an analytical evaluation based on design truck loads. It has been shown that this approach
underestimates the remaining fatigue life of a bridge, mainly because the real traffic load is usually less than the
truck load [3,4,5]. On the other hand, evaluations based on actual traffic load through field strain measurements can
help to obtain real stress time histories that a bridge may experience. The measured strains can reveal the actual load
distribution in the structure as well as the weight, volume, and traffic patterns of vehicle loads.
Today fatigue evaluation can be accomplished accurately by field measurements through SHM which is the
integration of a sensory system, a data acquisition system, a data processing system and an archiving system to
acquire real information regarding the in-service performance of structures [2].
There have been a number of researches focusing on fatigue damage analysis and fatigue life predictions through
SHM and field measurements. Mohammadi et.al. (1998) presented application of this method to condition
assessment of fifteen highway bridges in Illinois [3]. The field data, then, had been used to drive probability density
functions for the stress range data at fatigue critical locations and for fatigue life evaluation. Alampalli and Lund
(2006) [4] and Zhou (2006) [5] used this method for fatigue life evaluation of some bridges in the United States
based on fatigue strength (S-N) curves. Li et al. (2009) applied field strain measurement upon linear fatigue damage
law and performed fatigue analysis according to Continuous Damage Mechanics (CDM) on a bridge in Hong Kong.
The modeling was validated by comparing the analytical predictions with the field measurements. Life predictions
were also made for the bridge using SHM data [6].
In the following study methods used for fatigue life evaluation of Tehran orthotropic deck bridges have been
presented. These eight identical bridges were constructed in early 1970s. All of them were perceived as temporary
solutions for Tehran overwhelming traffic and were planned to be replaced after 5 years with permanent overpasses.
However, all remained under service at their primary conditions till 2007 when one of them was disassembled. The
other seven bridges stayed in place and are still under service. Tehran municipality intends to have them in place
even for more years. In order to investigate their current condition a thorough SHM has conducted on three of them.
This study aims to determine the remaining fatigue life for one of these typical bridges (Hafez-Taleghani bridge)
which is under quite lower traffic loads. To fulfill this duty, series of field strain measurements have been carried
out on the bridge and a classical fatigue life assessment approach has been adopted for interpreting the information.
The outcome has then been compared to laboratory tests on coupons taken from one girder of the disassembled
bridge. Fig 1.shows Hafez-Taleghani bridge and the disassembled bridge.

Fig. 1. (a) A bottom view of Hefez-Taleghani Bridge; (b) Saadi-Cheraghbargh girders


K. Kashefi et al. / Procedia Engineering 2 (2010) 573–582 575

K Kashefi et al. / Procedia Engineering 00 (2010) 000–000 3

2. Classical approach definition

In the current investigation, as mentioned before, the classical approach has been used to predict remaining
fatigue life of the bridges. This method is based on Palmgren–Miner rule [7]; the damage accumulation law which is
expressed as,
n
D= ¦ (n / N )
i =1
i i (1)

It says that a load cycle with amplitude Si adds to the cumulative damage D, a quantity ( 1 / N i ). Here, Ni
denotes the fatigue life under constant amplitude loading with amplitude Si and ni is the number of load cycles at
this amplitude .This rule is independent of sequence and fatigue failure is expected when these fractions sum to
unity (i.e. Dn = 1 ).
An S–N curve is a relation between the stress range under constant amplitude loading and the number of stress
cycles to failure. The standard S–N curve can be expressed in the form of:

m
A = NS (2)

where N is the number of stress cycles to failure at a constant amplitude stress range S, A and m are the material
parameters.
Specifications such as AASHTO [8,9] and AREMA [10] put forward a series of formula for S-N curves (see
Eq.3). They are in essence the best fits to S-N experimental data with a 2.3% probability of failure [11]

−3
N = AS (3)
r

where N is the number of stress cycles to fatigue failure; S r is the nominal stress range at a fatigue detail; and A is
detail category constant, varying for the types of details. (Fig.2)
. In AASHTO 1990 [8] and AASHTO 2003 [9] , which are basically developed for evaluation of existing high
way bridges, different bridge details are categorized as A, B, B', C, C', D, E, or E' considering their fatigue strengths.
Category A is for the base material steel and has the highest fatigue strength. Category E' represents the poorest
welded details that have the lowest fatigue strength.

Fig. 2. Fatigue strength S–N curves as AASHTO (2004) [12]


576 K. Kashefi et al. / Procedia Engineering 2 (2010) 573–582

4 K. Kashefi et al. / Procedia Engineering 00 (2010) 000–000

Constant Amplitude Fatigue Limit (CAFL) in Fig 1. is a stress range threshold which no fatigue crack would
occur below that. The values of constant A and CAFL for all fatigue categories are summarized in Table 1.
(AASHTO 2004) [12]

Table 1. Constant A and CAFL of fatigue categories [12]

Category Constant A CAFL


MPa3 MPa
11
A 82.0 × 10 165.0
B 39.3 × 1011 110.0
B' 20.0 × 1011 82.7
C 14.4 × 1011 69.0
C' 14.4 × 1011 82.7
D 7.21 × 1011 48.3
E 3.61 × 1011 31.0
E' 1.28 × 1011 17.9

3. Fatigue life evaluation

Fatigue evaluation/design of steel bridges, can be accomplished using several specifications such as AASHTO
for highway bridges and [8,9] AREMA (2002) [10] for railway bridges.
The Guide Specifications for Fatigue Evaluation of Existing Steel Bridges (AASHTO, 1990) [8] and the Guide
Manual for Condition Evaluation and Load and Resistance Factor Rating (LRFR) of Highway Bridges (AASHTO,
2003) [9] have been developed specifically for evaluation of existing highway bridges. Both specifications suggest
different reliable methods to obtain stress ranges under variable cyclic loads. According to the AASHTO Guide
Specifications for Fatigue Evaluation of Existing Steel Bridges, if a primary live load analysis with Fatigue Truck
Load performed on susceptible tension /reversal members shows that the fatigue life based on the specification
criteria is insufficient for bridge service life, decisions on further studies should be made. This is because analytical
evaluations using truck loads usually underestimate fatigue life. More accurate studies can be accomplished then by
applying SHM and field strain measurements.
For fatigue monitoring, it is necessary to first find fatigue sensitive components. Fatigue cracks typically initiate
from discontinuities within weld or base metal [13]. The weld geometry itself can induce stress concentrations
higher than those associated with the aforementioned weld discontinuities. For instance one of the most fatigue
sensitive weld details is a fillet weld termination oriented perpendicular to the applied stress field.
In a next step suitable strain gauges should be installed on critical components. While choosing strain gauges it
should be noted that their fatigue life is adequate for test duration. The actual strain histories are then measured by
strain gauges at the areas of concern. Other information such as effects of varying vehicle weights, their random
combinations and also load distributions in bridge members can be investigated using the measured strains. The
duration of strain measurement should be at least 1 week so that it can represent stress histories that bridge
experiences for extended time periods [4, 5, 6].
Prior to the measurement under real traffic, performing calibration tests are recommended. These tests should be
carried out using a vehicle of known axle weights, crossing each traffic lane at the time that the bridge is closed to
all regular traffic [4,5]. This test shows the relationship between strain responses and the vehicle weight and strain
gauges' locations. It also helps to assess the weight and load distribution of regular vehicles from measured strain
histograms.
After processing the result, stress range histograms can be obtained from strain time histories, using an
appropriate cycle counting algorithms such as the rainflow proposed by Downing and Socie, 1982 [14]. The stress
K. Kashefi et al. / Procedia Engineering 2 (2010) 573–582 577

K Kashefi et al. / Procedia Engineering 00 (2010) 000–000 5

range histogram presents the occurrence of stress cycles in terms of the number of cycles for each stress range
magnitude captured during the measurement period.
For interpreting the information, the fatigue category of each tested part should be defined using the fatigue
strength S-N curves. The highest stress range ( S r , max ) in the measured histogram is then compared to the fatigue
threshold (CAFL) for the corresponding detail category. If ( S r , max ) is less than fatigue threshold, fatigue life is
considered infinite. Otherwise, the finite fatigue life of the detail should be calculated.
Since the fatigue strength S–N curves were developed through tests under constant-amplitude cyclic loading an
effective stress range should be obtained as a representative of the actual variable-amplitude cyclic loading on the
bridge. The effective stress which represents the stress spectrum of a certain detail is estimated using Miner’s rule
(AASHTO, 1990) [8] and can be obtained from:
1/ 3
S re = §¨ ¦ γ i S ri 3 ·¸ (4)
© ¹

where S re is the effective stress range for a variable amplitude stress range histogram; S ri is the ith stress range in the
stress range histogram; γ i = ni / N is the fraction of occurrence of the stress range S ri in the histogram; ni is number
of occurrences of the stress range Sri; and N is total number of occurrences of all stress cycles in the histogram.
The number of effective stress cycles (Ne) can be computed from the corresponding S-N curve. The remaining
fatigue life is finally obtained by subtracting the current age of the bridge from the total fatigue life with the
consideration of loading history and predicted growth rates of traffic volume and weight.

4. Case study

4.1. Bridge structure

Hafez-Taleghani Fly-over is a one-way 31-simple-span orthotropic deck steel bridge in Tehran (one of 8 identical
bridges) which carries three traffic lanes. The bridge deck consists of six I-shape girders made out of EN
10027,S355 steel ( Fy=355 MPa) . The bottom flange with a segment of the web forms an inverted T which is
attached to the orthotropic deck with two rows of bolts (Fig 3. (a)). Rib profiles are HEA 200, with 200 mm section
depth and 200 mm flange width attached to deck plate by fillet welds. The girders are all 24 meters long and have
splices at the mid span. A typical deck plan with a girder numbering is shown in Fig 3. (b). The girders were
assumed to act independently which means that a girder does not theoretically receive contributions from its
adjacent girders when it is subjected to a vertical load due to low stiffness of transverse tie beams. This assumption
has been later argued by extensive field measurements.
Analytical evaluations based on Iranian Standard No.139 for bridges (2000) [15] demonstrated that the bridges’
fatigue life was insufficient, for their planned service life.

Traffic direction
N

Fig. 3. (a) A representative section of a girder with transverse beam, longitudinal ribs and deck ; (b) Typical bridge deck plan and girder
numbering
578 K. Kashefi et al. / Procedia Engineering 2 (2010) 573–582

6 K. Kashefi et al. / Procedia Engineering 00 (2010) 000–000

4.2. Field measurements

Calculations showed that high values of tensile stress would develop in bottom flange of girders at mid spans
while two side by side 40-ton trucks were passing over the bridge (the critical loading condition according to Iranian
specification). Since welded connections were used to join the bottom flange to the web, these high stresses may
cause horizontal fatigue cracks in the web or in the web to flange welds.
Since all spans are identical, the forth span from the northern abutment was instrumented in order to accomplish
field strain measurements. Strain gauges were installed on bottom flange, at mid span of each girder (Fig 4). The
gauges were placed at least 500 mm away from the splices to avoid stress concentrations. Strain histories were
collected every 0.125 second over an eight-day period under normal traffic.
A 275 KN truck was used for the field calibration of the measuring devices. While the bridge was closed to all
regular traffic, the truck was passed, at a very low speed, separately over each traffic lane. The strain time history
were recorded and analyzed to verify the calibration factors. Subsequently, stress time histories under regular traffic
were recorded using all gauges over an eight-day period. Stress cycles/ranges were then extracted using a program
which was developed in MATLAB based on a rain flow counting algorithm [14,16,17]. Since presence of large
number of very low strains will decrease the effective stress range obtained from strain histogram, there should be a
criterion for eliminating strain cycles below a defined threshold. After a research by Fisher (1977) [18] for the
American Institute of Steel Construction, a truck with a gross vehicle weight less than 20 kips (89 KN) has a very
low probability of causing fatigue damages in a steel bridge. Alampalli and Lund (2006) conducted a test using a
165 KN truck crossing the bridge and linearly scaled strain ranges to find the fatigue threshold [4]. In the current
study the results of the linearly scaled calibration test using a 275 KN truck (Fisher’s method) have been compared
to .5CAFL which was proposed in a previous research [5]. Considering those methods and the observations made
while strain measurements, a value of 10 MPa has been obtained for all girders to define their fatigue threshold.
As mentioned before in order to obtain the remaining fatigue life through S-N curves, it is necessary to calculate
the effective stress according to Eq.4. The longitudinal fillet welds between the flange and web of a plate girder can
be reasonably considered to adhere to a category B fatigue detail. According to AASHTO (1990) the CAFL value
for this class is 110 MPa, the remaining safe life of this detail in the bridge may be considered infinite and no further
fatigue calculations are required if [8]:

RS S r e < CAFL (5)

where RS is reliability factor; Being 1.35 for redundant members and 1.75 for nonredundant members.
Considering RS as 1.75 for al girders, the effective stress ranges calculated after introducing the reliability factor
still remain below CAFL.

Strain Gauges

Fig. 4. Typical arrangement of strain gauges at mid span of a girder


K. Kashefi et al. / Procedia Engineering 2 (2010) 573–582 579

K Kashefi et al. / Procedia Engineering 00 (2010) 000–000 7

Table 2. shows the threshold strains, maximum stress, effective stress and the remaining fatigue life for each
girder. According to Table 2., based on the existing traffic regime the fatigue life of the structural details examined
in the girders is estimated to be infinite.

Table 2. Result summary for one week fatigue monitoring of the girders

Girder Strain threshold (mV) Maximum Effective Effective Remaining


no. stress (MPa) stress (MPa) stress × Rs safe life
Fisher 0.5 Adopted by
<?> CAFL
method CAFL author
1 48.6 262 48 140 17.38 30.4 < 110 Infinite
2 35.7 262 48 157.5 13.27 23.2 < 110 Infinite
3 58.4 262 48 132 15.13 26.5 < 110 Infinite
4 53.5 262 48 209 22.73 39.8 < 110 Infinite
5 31 262 48 228 17.91 31.3 < 110 Infinite
6 41.3 262 48 145 19.42 34.0 < 110 Infinite

5. Laboratory tests

For further investigation of the field measurement results a novel experiment methodology has been applied in
this study in which a series of fatigue tests on coupons have been carried out to determine the remaining fatigue life
of the bridge girder. In the sections below a summary of the experimental program has been given.

5.1. Preparation of specimens

Specimens were designed according to ASTM-Standard Practice for Conducting Force Controlled Constant
Amplitude Axial Fatigue Tests of Metallic Materials E 466-96 and ASTM-Standard Practice for Strain-Controlled
Fatigue Testing E 606-92. The T shaped specimens were carefully cut from bottom flange/web connection a
disassembled main girder. Axial force fatigue tests are commonly performed on flat plates with rectangular cross
sections. In the current study T shaped specimen were considered to demonstrate possible defects in the welds and
the heat affected zones (HAZ). An axial cyclic loading was considered for the tests so that the resulting stresses
could be in agreement with those real stresses on a web/flange connection in the main girder. A reduced section was
also produced by wire cutting in the specimen to facilitate the failure in a predefined and limited zone. Fig 5 shows a
specimen while testing along with geometrical details of the specimens.

Fig. 5. (a) Geometrical details of the specimens (b) a specimen while testing
580 K. Kashefi et al. / Procedia Engineering 2 (2010) 573–582

8 K. Kashefi et al. / Procedia Engineering 00 (2010) 000–000

5.2. Test set up , procedure and results

All tests were made under laboratory air conditions by a 25-ton hydraulic testing machine under pure axial
loading. The loading frequency was 3.5 Hz, the load amplitude was kept constant and from zero to maximum
tension. The cyclic loading was maintained until complete disjoining of the specimen. This was considered at the
fatigue failure limit. Initiation of a crack in the reduced section and complete failure of a specimen are shown in Fig
6.
Since all girders are simply supported, a web to flange connection in the main girder, at a support location has
experienced no bending stresses during its past history. So this specific detail may be regarded as a fatigue damage
free connection. This is while a similar connection at the mid span of the girder has experienced maximum fatigue
damage on account of maximum live load fluctuations. Therefore, specimens chosen from the bottom flange at
support locations (S coupons) could demonstrate the total fatigue life of a newly constructed girder and the ones
chosen from the bottom flange of the mid span would indicate the remaining fatigue life of the girder (M coupons).
Axial force controlled fatigue tests based on ASTM standard; E 466-96, were conducted on three specimens from
mid span (S coupons) and three from the support location (M coupons). All tests were carried out, at constant
amplitude of 0.6 Fy=213 MPa. A summary of test results is shown in Table 3.

Fig. 6. (a) Initiation of a crack in the reduced section (b) complete failure of a specimen

Table 3. Summary of the coupon test results

Coupon Number of Average number


type cycles to failure
P1 105000
P2 92700 120421
P3 163564
S1 160500
S2 88200 119758
S3 110575

It was observed that the failure of all six specimens initiated at the root of the weld and happened at nearly the
same number of cycles which validated the results of field strain measurements.
K. Kashefi et al. / Procedia Engineering 2 (2010) 573–582 581

K Kashefi et al. / Procedia Engineering 00 (2010) 000–000 9

6. Conclusions

This paper discussed the results of field measurements as well as laboratory tests carried out for obtaining the
remaining fatigue life of an orthotropic deck bridge in Tehran.
Analytical evaluations based on Iranian Standard No. 139 (2000) for Bridges showed that remaining fatigue life
of the bridge girders were insufficient for future services. A set of field strain measurements was then considered to
evaluate the remaining fatigue life accurately. For this, the girders were instrumented at bottom flange and relevant
strain data were collected during an eight-day monitoring period. Strain data results revealed that as per AASHTO
(1990) and under the existing traffic regime the fatigue life of the main girders can be regarded as infinite and an
appropriate safety margin regarding the fatigue loadings is available for their future service. This is while analytical
calculations indicated otherwise, most probably because contributions from secondary structural elements such as
transverse tie beams were neglected in the analytical models. In addition, the analytical approach was based on
design truck loads; while the real traffic load is considerably bellow that. Further more the field results were verified
by conducting laboratory tests on coupons taken from the support location and mid span of the same girders.

Acknowledgements

This study was a part of contract for Tehran bridges repair and rehabilitation which was awarded to TAZAND
Consulting Engineers by Tehran Municipality. The authors would like to express their sincere thanks to Tehran
Municipality (Civil Engineering Organization) and K.N Toosi University of Technology, for their supports and
invaluable assistance throughout this research.

References

[1] Sikorsky C. A strategy to implement structural health monitoring bridges. In: Ansari F editor. Sensing Issues in Civil Structural Health
Monitoring, Netherland: Springer; 2005, p. 43–53.
[2] Battistaa RC, Pfeila MS, Carvalhob EML. Fatigue life estimates for a slender orthotropic steel deck. J. Constructional Steel Research
2007:134–43.
[3] Mohammadi J, Guralnick AS, Polepeddi R. Bridge Fatigue Life Estimation from Field Data. Practice Periodical on Structural Design and
Construction 1998;3:128-33.
[4] Alampalli S, Lund R. Estimating Fatigue Life of Bridge Components Using Measured Strains. J. Bridge Engineering 2006;11:725–36.
[5] Zhou Y E. Assessment of Bridge Remaining Fatigue Life through Field Strain Measurement. J. Bridge Engineering 2006;11:737-44.
[6] Li ZX, Chan THT, Ko JM. Fatigue damage model for bridge under traffic loading application made to the Tsing Ma Bridge, Theoretical
and Applied Fracture Mechanics 2001;35:81–91.
[7] Miner M A. Cumulative damage in fatigue. J. Applied Mechanics 1945;12:159–64.
[8] AASHTO, 1990. Guide specifications for fatigue evaluation of existing steel bridges. Interim specifications in 1993 and 1995,
Washington, D. C.
[9] AASHTO, 2003. Guide manual for condition evaluation and load and resistance factor rating (LRFR) of highway bridges. Washington,
D.C.
[10] AREMA, 2002. American Railway Engineering and Maintenance-of-Way Association, Steel Structures. Washington D.C.
[11] Keating PB, Fisher JW. Evaluation of fatigue tests and design criteria on welded details. NCHRP Report 286, Transportation Research
Board, Washington, D.C., 1986.
[12]AASHTO, 2004. AASHTO LRFD bridge design specifications, 3rd Ed., Washington, D.C.
[13] Barsom JM, Rolfe ST. Fracture and Fatigue Control in Structures. Applications of Fracture Mechanics, 3rd Edition, ASTM manual
series: MNL41, 1999.
[14] Downing SD, Socie DF. Simple rain flow counting algorithms. Int. J. Fatigue 1982;4:31–40.
[15] Iranian Standards No. 139. Standard Loads for Bridges. Office of the Deputy for Technical Affairs, Bureau of Technical Affairs and
Standards, Management and Planning Organization, Tehran, Iran, 2000, 105 p.
[16] ASTM E 1049-85, Re-approved 1997. Standard practices for cycle counting in fatigue analysis, in: Annual Book of ASTM Standards,
Vol. 03. 01, Philadelphia 1999, p. 710-718.
582 K. Kashefi et al. / Procedia Engineering 2 (2010) 573–582

10 K. Kashefi et al. / Procedia Engineering 00 (2010) 000–000

[17] ASTM, 2004. Standard practices for cycle counting in fatigue analysis. E1049-85, ASTM, West Conshohocken, Pa.
[18] Fisher J. Bridge fatigue guide: Design and detail. American Institute of Steel Construction, New York, 19770.

You might also like