J Engfailanal 2019 05 037

Engineering Failure Analysis 104 (2019) 154–176
Contents lists available at ScienceDirect
Engineering Failure Analysis

journal homepage: www.elsevier.com/locate/engfailanal
Review
Fatigue cracking of welded railway bridges: A review

T
Guilherme Alencara, , Abílio de Jesusa, José Guilherme S. da Silvab, Rui Calçadaa
⁎
a
Department of Civil Engineering, Faculty of Engineering of University of Porto, Dr. Roberto Frias St., 4200-465 Porto, Portugal
b
Faculty of Engineering, State University of Rio de Janeiro, São Francisco Xavier St, 524, Maracanã, 20550-900 Rio de Janeiro, Brazil
ARTICLE INFO ABSTRACT
Keywords: Steel railway bridges of all types comprise a large portion of the infrastructure inventory of
Railway bridges several countries. Due to the ageing of these structures and increasing traffic loadings, structural
Welded connections fatigue increasingly became an important concern, leading to bridges deterioration. In this
Fatigue cracking context, it is well-known that welded connections are among the weakest locations in steel
Distortion/vibration-induced fatigue
bridges since they are prone to stress concentrations leading to the initiation of fatigue cracks.
Although its importance, the study of the development history of welding technology applied to
steel bridges is often overlooked in introductory texts about fatigue of metallic bridges. However,
to understand the historical development of welding in steel bridges construction and its impact
on fatigue design is fundamental, mainly because fatigue evaluation depends on the age of
structures. The goal of this paper is to review the history of fatigue cracking of welded railway
bridges, discussing the structural behaviour of welded bridges under dynamic traffic loads and
the different causes of fatigue damage of welded connections. The presented case studies cover a
period from the beginning of application of welding to steel bridges in the 1930s, revisiting well-
known cases of cracks documented in the literature due to their historic relevance and presenting
also recent cases of the last decades due to new traffic conditions related to Heavy Traffic Loads
(HAL) and High-Speed Trains. For this purpose, an in-depth literature review covering more than
one-century history was carried out, providing a multidisciplinary analysis with insights into the
welding history and practice and fatigue of welded joints, especially applied to steel railway
bridges.
1. Introduction
Railway bridges of all types comprise a large portion of the infrastructure inventory of several countries. Due to their high
transportation capacity and effective energy usage, with lowest environmental harm, railways are one of the most important means of
transportation in the world, in which steel bridges have been widely used. The maintenance of railway bridges is costly and a major
concern for railway administrations. It was reported that > 60% of the railway bridges stock in Europe was over 50 years old
and > 30% was over 100 years old [1]. These bridges are subjected to higher loads and speeds than those for which they were
originally designed.
Due to the ageing of the materials combined with increasing traffic loadings, structural fatigue increasingly became an important
concern in bridges deterioration. Several authors mention fatigue as one of the main causes of structural failures in metallic bridges
[2–5], considering it in the front of the concerns in the update of the railway infrastructures to carry heavy loads. In this context, it is
well-known that welded connections are among the weakest points in steel bridges, since they are prone to stress concentrations
⁎
Corresponding author.
E-mail address: guilherme.alencar@fe.up.pt (G. Alencar).
https://doi.org/10.1016/j.engfailanal.2019.05.037
Received 27 September 2018; Received in revised form 22 May 2019; Accepted 30 May 2019
Available online 31 May 2019
1350-6307/ © 2019 Elsevier Ltd. All rights reserved.
G. Alencar, et al. Engineering Failure Analysis 104 (2019) 154–176
leading to initiation of fatigue cracks. Nowadays, many existing welded bridges are reaching their design working life, and a number
of fatigue damages, which cannot be ignored or neglected, are known. The inexperience and the lack of knowledge about fatigue
behaviour in the beginning of the application of welding to bridges is one of the main causes of existing fatigue damages [6]. The
application of welding to railway bridges started in the 1930s, with the majority of steel welded bridges built since 1950. However,
although its importance, the study of the development of welding technology to steel bridges is often overlooked in introductory texts
of fatigue of metallic bridges, with the history of mid-twentieth-century bridge technology being a relatively new field of research
[7].
Clearly, it is fundamental to understand the historical development of welding in steel bridge construction and its impact on the
fatigue design mainly because fatigue evaluation depends on the age of structures. Such study also can broaden the understanding of
old bridges, which today's engineers must maintain and inspect. In case of repairs of railway bridges due to fatigue damages, Lippi
et al. [8] highlighted the necessity of knowing the historical reconstruction or maintenance activities performed by railway autho-
rities which could permit to evaluate correctly the fatigue load cycles really occurred on bridge components. Initial key questions may
include: what were the factors that led some countries and their engineers to develop particular technologies and structural details?
What made one country's bridge division progressive and another's conservative in the application of the new technologies? What
were the significant works in advancing successful technologies, such as welded girders, orthotropic steel decks, composite (steel-
concrete) welded girders, etc.?
The present work is an attempt to answer these questions to some extent. The main goal is to review cases of fatigue cracking of
welded connections in railway bridges under traffic service loads over the years and make the reader aware of this important theme.
Herein, ‘welded bridge’ is understood as any bridge that have used welding to some extent, even combining it with other types of
joining techniques. Since welding is used for almost every new steel bridges, the applications can be very broad [6]. In some cases,
insights into how and why engineers chose to do what they did can even be of great importance. The work is organized in the
following manner: i) initially, a historical overview of welding application to steel railway bridges is presented; ii) secondly, several
cases of detected cracks are outlined; with the different causes of fatigue cracks in welded joints of steel bridges being discussed. The
reasons for the occurrence of cracks are examined, trying to extract some practical lessons. Employed repair and retrofit methods to
welded cracked joints are not covered in the present work, being only cited very briefly if necessary.
2. A historical overview of welded railway bridges construction
The adoption of welding was one of the most important refinements in structural steel design during the twentieth century, but it
took many decades to replace the long-customary method of bridge construction based on the use of rivets. Electrode welding, i.e.
using a fusion agent of flux which joins the parts to be welded was known in mechanical manufacturer since about 1917 [9].
However, railway bridges have probably been those steel structures where, due to very strong safety concerns, welding was admitted
very late [10]. From 1926 onwards, the railway industry used to apply arc welding in reinforcement of old metallic bridges while
keeping it in operation, mainly in U.S. [7], Germany and Belgium [11], United Kingdom and oversea territories [12], France [13] and
in Japan starting from 1931 [14], and other countries. Konishi [14] reported for example an impressive number of 1300 old railway
metallic bridges that were reinforced with welding from 1931 to 1940 in Japan. During the Second World War, thousands of welders
were trained in the repairing of bridges, ships, aircrafts, and other structures.
Already in the 1930s, electric arc welding offered a valuable solution to railway engineers facing the problem of increased traffic
loads and deterioration due to corrosion [12]. One of the main advantages claimed in the beginning for the application of arc welding
to bridge reinforcement was that it avoids or greatly reduces delays in the railway traffic during works. An emblematic case of success
was the reinforcement of the Austerlitz Bridge, carried out around 1936. The bridge is located in the Line 5 of the Parisian Me-
tropolitan Railway, and it is still operating today. It was constructed by 1906, with the design capacity to carry light trains of 121 t.
However, thirty years after its construction, it was carrying light trains of 420 t, requiring a reinforcement of its structure, which was
made with welding with no interruption of the rail traffic [15].
Thus, motivated by some relatively successful experiences with reinforcement of old metallic bridges with arc welding, the first
all-welded railway bridge began to be erected in 1928, in Chicopee-Falls, Massachusetts, U.S. [16]. The bridge with a span of 41 m
was built by The Westinghouse Electric Company, a leading promoter of welding technology in the early times, and it was used to
transport large generators from facilities to the rest of U.S. thru railways. One year later, the German National Railway Company
commissioned its first completely welded bridge to be installed on a track near Münster in Westphalia, with a span of 10 m and girder
height of 920 mm [17]. The 30 mm thick and 260 mm wide flange plates of the bridge were attached by strong non-rounded fillet
welds to the 15 mm thick web plates, due to the concern regarding its strength to cyclic loading, which was unknown at the time. The
width of the fillet weld was chosen equal to the thinner of the 2 plates being joined. The bridge was put in service in 1930 and after
six years of heavy traffic with the passage of 220,000 trains, their perfect integrity has been emphasized as an eloquent testimony to
the usability of the welding process to railway bridges [9].
Soon, engineers recognized the advantages of welding over riveting, and the technology had spread to several countries. Based on
those advantages, the following ones are highlighted:
i) Welded structures are usually lighter than riveted structures since when using welding, gussets or other connecting components
are not used.
ii) The welded joints provide high efficiency, which is not possible in the case of riveted joints.
iii) Modifications and additions can be easily performed in existing structures. Riveting requires enough clearance.
155
iv) Welding is less time consuming than riveting.

v) Welded joints provide better aesthetic finishing of joints.
vi) Welding can be easily adapted to a diversity of shapes and geometries like cylindrical pipes which are difficult to be riveted.
vii) Tension welded members are stringer than riveted counterparts.
viii) Less workforce is required to perform welded joints.
Nevertheless, welded joints also show some drawbacks such as:
i) Inspecting and maintaining welded structures takes a longer time than rivets.
ii) Welding requires a demanding preparation of joints. The welding process also requires a lot of equipment increasing the pro-
duction costs.
iii) Since welding involves an uneven heating and cooling, members may get distorted or additional stresses may develop. If no
provision is kept for expansion and contraction in the frame, there is a possibility of cracks development.
Besides weldments, bolts have also been applied as an alternative to the rivets in bridge construction. Bolts increases the joints
efficiency compared to riveted joints due the usage of high strength preloaded bolts. In general, welded details are more susceptible
to fatigue cracking in comparison to bolted or riveted ones. Weld defects and discontinuities such as weld undercut, entrapped
porosities and lack of fusion are stress raisers from which fatigue cracking may initiate. Unfavourable welding residual stresses and
stress concentrations due to weld geometry and welding induced micro-cracks are other factors that often accelerate the process of
fatigue damage in welded bridge details.
Table 1 lists some of the first welded bridges per some selected countries, with the indication of their location, year of con-
struction, maximum span length (in case of multiple spans), main traffic type, structural type, current condition and if some sort of
cracks have been reported. Bridges a) to e) represent several cases of all-welded trial railway bridges: a) and b) were the first welded
railway bridges in the world, constructed by Westinghouse Electric Company in the U.S., as already mentioned [16,18]. Bridge c) was
one of the first European welded bridge, operated by ELiN A.G. in Austria [19], with a similar plate girder with 5.20 m span
constructed in Leuk (Switzerland) in the same year by the Swiss Federal Railways [17,20]. Bridges d) and e) were the first fully-
welded railway bridges in Germany [20] and Finland [21], respectively. All these bridges were an initial experiment performed by
national state railway companies and some relevant private companies for the use of welding in bridges. They demonstrated the
capabilities and peculiarities of welding technique to bridge construction in the early times. However, most of these trial bridges are
not in use nowadays and seemed to have been in operation for only a few years, due to its experimental character.
Some notable cases of entirely welded railway bridges for passengers' traffic from this period which are still operating after more
than of 80 years of service are represented by cases f) and g). Bridge f) is a curved-frame steel bridge still in use in Zagreb (Croatia), in
very good state and preserved conditions regarding its first state presentation in the IABSE Berlin-Munich Congress in 1936 [22], and
bridge g) is an overpass in the Poznan-Szczecin Line in Poland built in 1935, which was subjected to damages caused by military
operations during the Second World War but was renovated and is still in service nowadays [23]. Many others old welded railway
bridges constructed during the period 1936–1975 in Poland are also reported in [23–26].
As can be seen from the presented cases, the introduction of welding to bridge construction was marked by short spans, as with
the introduction of every new material or technique in bridge construction [18]. After the relatively success of such trial bridges, the
steel bridge industry has set a clear course in the direction of welding technology in the early 1930s, mainly in Europe. Just a few
years later, spans beyond the 30 m mark had already been achieved [9], and in 1935 the first welded girder bridge (of road type)
beyond the 100 m mark were constructed [27], with railway bridges limited to spans up to 50 m due to safety concerns [17]. During
those times, welding was a key topic in the first two conferences of the newly formed International Association for Bridge and Structural
Engineering (IABSE), held in Paris at 1932 and Berlin at 1936 [28]. Those meetings congregated the state-of-the art of welding
technology applied to bridges at the time with experts highlighting the advantages of welding over riveting and describing it as more
cost-effective [11].
Thus, after achieving a relatively level of confidence in welding, entirely welded girder bridges with more demanding passengers
and freight traffic requirements began to be constructed [29]. Some of the early welded girder bridges per selected countries in
operational lines (not trial ones) are shown in Table 2, cases h) to k). Bridge h) was constructed around 1934, and it is still in
operation in the current longest railway line in Norway [30], being an example of one of the earliest railway bridges combining both
shop welding and field riveting. In this case, the girders were welded in workshops in Oslo and then transported to site, while the
lateral brackets were attached to the girder by means of rivets. Bridge i) is referred as the first welded bridge in France, located in
Saint-Denis, Paris, and it was designed for a high skew angle [13]. Bridge j) is a continuous girder bridge 592.6 m long with 32 spans
over the Miryang River, in South Korea and bridge k) is one of the oldest welded railway bridges in Sweden, constructed around 1950
and located over the Söderström River in Stockholm [31]. All these bridges, with the exception of the French one, are still in service
nowadays, some of them even being subjected to heavy traffic.
From the reviewed early technical literature on welded bridges, it was possible to observe that several methods of construction
using welding were first tested in road and highway bridges, due to the reason that the impact stress for a movable load being less
important than for railway bridges. Indeed, the earliest all-welded bridges in several countries were of road traffic type, e.g. Poland in
1928 [29], Romania in 1931 [32], Serbia in 1932 [22], Netherlands in 1934 [33], Sweden in 1934 [34] and Belgium in 1932 [35],
with the Belgian State Railways Department allowing the construction of railway bridges using only rivets during the 1930s, the same
occurring in Hungary [36]. The last Belgian riveted railway bridge was built in the 1960s [37]. Moreover, only after more or less a
156
Table 1
Some of the earliest entirely welded Railway Bridges per selected countries.
Bridge Location/year of Max span Traffic Bridge type Status Cracks
construction
Earliest welded a) Westinghouse Truss 41 m Freight Through Truss Out-of- X

bridges Bridge service
Chicopee-Falls, MA, U.S.
1928-?
b) Westinghouse Girder 16.8 m Freight Plate girder Out-of-

Bridge service
East Pittsburgh, PA, U.S.
1928-?
c) ELiN A.G. Bridge 9.7 m Freight Plate girder Out-of-

Weiz, Austria service
1929-?
d) Reichsbahn Trial Bridge 10 m Mixed Half-through plate Out-of-

Münster, Westphalia, girder service
Germany
1930–1945
e) Finnish State Railways 12 m Mixed Plate girder Out-of-

Bridge service
Helsinki, Finland
1933-?
Commuters bridges f) Tratinska Street Bridge 18 m Commuter Curved Framed In use

Croatia, Zagreb
1934
g) Zdroje Szczecin Bridge 13.5 m Commuter Curved frame In use X

Poznan-Szczecin Line,
Poland
1935
decade of a very consistent set of successful applications to highway bridge construction, a new construction method making use of
welding was extensively adopted in railway bridges, e.g.: i) Composite (steel-concrete) construction, with the first composite roadway
bridge built in 1936 in Switzerland and the first railway composite bridge built in 1950 in Germany [38], herein adopting the concept
of composite beams those that have welded shear connectors between the steel beam and the reinforced concrete (RC) slab; ii)
orthotropic steel decks (OSD), firstly adopted for a roadway bridge in 1950′, and later in a railway bridge in 1959, both in Germany,
see Table 3, bridge l) [39,40].
As a consequence of this precaution in railway industry, there are much more cases of fatigue cracks reported in the literature for
highway bridges [41,42], although the more severe railway traffic loads. One exception to this rule was the adoption of high strength
bolts, since it was first used in metallic railway bridges by 1947 in the United States to replace loosening rivets, still in experimental
character [43]. Bridges m) and n) (Table 3) are some examples of the early railway bridges which used high-strength bolts to join
elements on site since the design: bridge m), a rail-road bridge built in 1957 in Australia [44], and bridge n), the first bolted-welded
bridge in China, built in 1963 [3]. There was, of course, nothing new about the use of bolts in steel construction in the 1950s.
157
Table 2
Some of the earliest welded girder bridges in operational lines by selected countries.
Bridge Location/year of Max span Traffic Bridge type Status Cracks
construction
Some of the earliest welded h) Bridge over River 17 m Mixed Plate girder In use
bridges in operational Namsen
lines Nordslandbanen,
Norway
~ 1934
i) Bridge at Saint-Denis 35 m Mixed Half-through Out-of-

La plaine Saint-Denis, plate girder service
Paris
1935-?
j) Miryang River Bridge 13.5 m Mixed Plate girder In use X

Miryang, South Korea
1945
k) Söderström Bridge 33.7 m Mixed Half-through In use X

Stockholm, Sweden plate girder
~ 1950
However, until laboratory tests were performed in order to assess the behaviour of high strength bolts [45], their use was limited to
applications where the applied stresses were simpler and lower than those found in main bridges components. Another important
advancement, which was exclusively related with railway engineering was the development of bridges designed to high-speed trains,
firstly in Japan, by using plate girders only, conventional composite beams or through truss bridges in some parts of the Tokaido
Shinkansen, represented by cases o) and p) (see Table 3) [46], and then in France for the TGV Northern line, with the first bridges
built in 1991, adopting modern composite steel-concrete girders, case q) (see Table 3) [47]. However, it is important to highlight that
although composite construction was very popular in highway bridges since the 1940s, their use in railway bridges were more restrict
in the same period.
A representation of the hitherto discussed historical development of the use of welding in railway bridge construction is shown in
Fig. 1. The figure was made based on the presented literature review, and the remarked dates delimiting each method of construction
may vary depending on the technological development level, disseminated knowledge, and mastery of technique present in each
country in each period.
This information can be useful to estimate the maximum age of existing steel railway bridges per construction method with the
goal of evaluating its remaining service life. Furthermore, the adopted classification in the diagram is not strictly closed. Of course,
there might exist bridges that combine two or more methods. The graph also shows that due to lack in the understanding of metal
weldability, poor design and execution at early times, the use of welding for the purpose of bridge reinforcement lasted more or less
only up to the end of Second World War because of the risk of fatigue and brittle cracking. Another important development in bridge
construction, although not directly linked to welding, was the usage of hot-rolled profiles in railway bridges superstructure with short
spans starting in 1910s [48], because of the provided economy in construction. Sometimes it was combined with concrete, but
without taking advantage yet of the composite action between steel and concrete by means of shear connectors, which occurred for
railway bridges only in the 1950s.
Despite available welding related technologies, the graph highlights that riveting technique was the predominant method for
bridge construction in the beginning of the twentieth century. One of the influencing aspects for a persistent preference of riveted
construction for many years, with the last riveted bridges being constructed until in the end of 1960s, was related with engineer's
decisions regarding safety concerns with welding technique, which can vary for each country for the covered period. For example, in
some states of the U.S. that did not early adopt welding, engineers either held on conservatively to tried-and-trusted riveting or
expressed concerns that welding did not have enough track record to prove its safety in the long-term [7]. By 1955, only 22 states of
U.S. among 48 reported that they had used welded beams [7], while 300 welded railway bridges have been constructed in Germany
by 1939 [35], from which 100 are still in use today [17]. The viability of welded connections under high stresses and the risk of crack
initiation and propagation were under some apprehension for bridge engineers. These concerns were not unjustified, the welding
158
Table 3
Some bridges with important advancements in railway bridge construction making use of welding.
Bridge Location/year of Max Traffic Bridge type Status Cracks
construction Span
OSD l) Bridge over 36 m Mixed Ortho- tropic Steel Deck Out-of-

Autobahn A1 service
Lest Wuppertal,
Germany
1959
Weld & High m) Burdekin Rail-Road 76 m Freight Bolt-rivet-welded In use

strength bolts Bridge through truss
Queensland,
Australia
1957
n) Langjiang Bridge 62 m Mixed Bolt-welded through In use X

Changsha-Liuzhou truss
Line, China
1964
High-speed traffic o) Shinkansen Girder 35 m High-speed Plate or box girders only In use X
Bridges traffic and Composite beams
Tokaido
Shinkansen, Japan
1964
p) Shinkansen Through 60 m High-speed Through truss In use X

Truss traffic
Tokaido
Shinkansen, Japan
1964
q) TGV Northern HSR 40 m High-speed Composite deck In use

Bridge traffic
HSR Paris to Lille
line, France
1991
technique, as well as its strength to static and dynamic loads, was new and unproven, and there was a lack of convenient methods for
inspecting and repairing defective welds. Moreover, the knowledge of why cracks do occur was something that was not common
knowledge among the bridge engineers in the 1930s, as riveted bridges had not been experiencing this kind of problems to such
extent before [42,49].
An important book summarizing the state-of-the-art in the beginning of twentieth century in hot-driven rivet construction is due
to Frémont [50], which was published in 1906. In this work, it is possible to observe the presence of a general caution with the
material brittleness of rivets and metallic plates. However, after almost 60 years of experience with the use of rivets in metallic
bridges and large structures, such as the Britannia riveted bridge erected in 1848 [51] and the Eiffel Tower constructed in 1889,
fatigue – in the sense of crack propagation under service loads – was not described or included among the several types of damages
that could evolve in riveted structures in Frémont's work. It was only verified that after riveted railway bridges being in operation for
many years that cracks started to occur, mainly due to increased stress levels with the introduction of new and heavier rolling stocks
with increased frequency [52]. One of the earliest confirmed fatigue failures in a riveted railway bridge was discovered in 1930 in the
shear connection of the floor beam of a through truss bridge built in 1898 in the United States [53], taking place at the riveted web
angles due to cyclic end rotation of the connected member. The bridge was subjected to 182,000 stress cycles in a year, lasting >
30 years in service. A few number of railway bridges developed fatigue cracks in similar details approximately after 30 years of
service in the first half of the twentieth century [53]. Cavaco [52] mentions the case of a riveted bridge having almost 100 years of
service with no indications of crack development by 2008, after intensive use.
On the other hand, fully-welded bridges and bridges strengthened with welding have developed fatigue cracks after only few
159
- Steels with improved

Entirely riveted bridge construction chemical composition
- Higher understanding
of brittle fracture
Use of hot-rolled sections in superstructure steel bridge construction
Welding reinforcement of Resumption of welding of old metallic

old metallic bridges structures under restricted conditions
State of development
First brittle failures

Entirely welded bridge construction
Chicopee-Falls Modern composite
Bridge, U.S. (1928) bridges for high-speed
Hybrid riveted / welded bridges traffic: TGV Northern
HSR Paris to Lille (1991)
Bridve over the
River Namsen, Composite (steel-concrete) welded bridges
Overpass in Göttingen,
Norway (~1934)
Germany (1950)
Hybrid bolted / welded bridges
Burdekin Bridge,
Australia (1957)
Welded orthotropic steel deck (OSD)
Bridge over A1,
Germany (1959)
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010 2020
Fig. 1. Penetration of welding in railway bridge construction.
years of service. Regarding the first all-welded bridges presented in Table 1, with the exception of the through truss bridge over the
River Chicopee built in 1928, all first entirely-welded bridges are of plate girder type. Reported cracks on butt welds of the lower
flanges of the world's first railway bridge only after a few days encouraged national railway companies to keep constructing truss
bridges only with rivets [17]. Surveys of other industries demonstrated that fatigue and brittle cracking was much more common for
welded than for riveted structures in the period of 1900–1950 [54]. In the next section, the main causes of fatigue cracking of welded
railway bridges are investigated.
3. Fatigue cracking of welded railway bridges
Several authors distinguish between the different causes of fatigue cracking in metallic bridges into different categories
[6,8,41,55,56]. Regarding the type of effect or design characteristic that induces the fatigue damage, usually these authors agree with
the following category subdivision: a) Fatigue induced by welding crack-like defects: fatigue cracks originating from weld defects that
were introduced at the time of fabrication; b) Load-induced fatigue: fatigue cracking due to an inappropriate structural component of
low fatigue strength in terms of nominal stress; c) Distortion-induced fatigue: fatigue cracks originating and/or fatigue crack growth
due to out-of-plane stresses and deformations that were unforeseen in the design; d) Vibration-induced fatigue: Fatigue cracks ori-
ginating and/or fatigue crack growth due to unexpected structural behaviour such as vibrations induced by wind or traffic. One fifth
category usually considered is fatigue assisted by corrosion, which is however out-of-the-scope of the present work.
Based on the chronological review of reported fatigue cracks occurring in railway bridges, it was constructed the qualitative
diagram shown in Fig. 2, where the horizontal axis represents the most predominant cause of fatigue cracks for bridges constructed in
the corresponding covered period (1900 to 2020), and the vertical axis represents the recognition order over time by the community
of bridge engineers to the related phenomenon giving rise to fatigue cracks. For example, bridges constructed in the period between
Low fatigue strength category in terms of nominal stresses

Recognition order of the cause
by the bridge engineers
Fatigue cracks originating from weld defects
Secondary and out-of-plane stresses
Vibration-induced fatigue
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010 2020
Fig. 2. Prevailing periods for the main causes of fatigue cracking of welded railway bridges.
160
Table 4
– Different railway bridges with reported cracks due to fatigue of welded connections.
Bridge Location/year of construction Max Span Traffic Bridge type Status Cracks
r) Zoo Train Station Viaduct 50 m Commuter Plate girder Out-of- X

Hardenbergstraße, Berlin service
1936–1938
s) Welded bridge at FAST 19.8 m Freight Plate girder Out-of- X

Pueblo, Colorado, U.S. service
1956–2014
t) Bridge over the Panaro River 75.6 m Mixed Full through truss bridge In use X
Bologna-Verona Line, Italy
~ 1970
u) Potomac River Metrorail Bridge 36 m Commuter Box girder bridge In use X

Yellow Line, Washington DC, U.S.
1976
v) Wei River Railway Bridge 26.2 m Freight Bolted-welded girder bridge In use X
West-central China
1982
w) Bridge over Córrego do Ouro 35 m Freight Curved Box girder bridge In use X
Vitória-Minas HAL Line, Brazil
1987
1930 – beginning of welding technique – and 1970 – when non-destructive inspection methods for welding became further devel-
oped– are most prone and well-known to have developed cracks due to weld defects. Of course, any type of damage might be found at
any time in any country during the referred period, and the diagram is an overall evaluation of the reported cases in literature.
Naturally, a combination of two causes is very frequent too, e.g. cracks that originated from a weld defect but grew due to high
secondary stresses.
More recently, this classification was adopted by Yokoyama & Miki [57] in the construction of a large database of previously
repair cases containing a comprehensive information of welded structures that have been damaged by fatigue, accounting for 204
cases from different countries. According to this database, today's most noticeable cases of fatigue cracks are due to secondary stresses
and complex and unexpected behaviour in the service phase, such as vibrations caused by traffic, representing together 63% of the
total cases of reported fatigue cracks [57]. Fatigue cracks have also been a problem with some more recent welded bridges over the
years. Table 4 presents some recent case studies in different countries that presented fatigue cracks, which were subjected to the
different causes of fatigue cracking. In the following subsections, the following bridge case studies, together with some of those
presented in Tables 1-3 will be related to specific fatigue cracking causes in welded connections.
The database constructed by Yokoyama & Miki [57] allowed to make interesting deductions, such as the relation between the
mentioned causes of fatigue crack growth and years of service at the date of crack detection, as illustrated in Fig. 3. From this graph, it
is possible to observe that fatigue cracks caused by welding defects were discovered during the first years of service for most bridges,
a remark that is usually attributed to the state of development of welding technique and the structural detailing practice of welding
connections in early times. Based on these results, it is also possible to confirm that vibration is another cause of fatigue cracking
which is most likely to occur early in-service periods. Furthermore, Fig. 3 shows that damage detection from causes related to low-
161
Fig. 3. Relation between years of service and fatigue causes [adapted from [57]].
fatigue strength and secondary stresses tend to peak between 15 and 25 years of service, and that all causes of fatigue damage are
usually detected within 45 years of service. In bridges life cycle, a time interval of 25 years corresponds in equivalence to their
‘infancy’, drawing attention to the fact that when new and unproven designs are employed in bridges, their integrity must be more
frequently inspected during this period. In the following subsections, the different causes of fatigue cracking in railway bridges will be
outlined, with the subsequent discussion of selected case studies.
3.1. Fatigue induced by weld crack-like defects
3.1.1. Examples of fatigue cracking induced by welding defects

Probably the first case of fatigue crack occurring in a metallic bridge with welding took place in an old lattice railway bridge by
1930, which was strengthened with welding in order to carry heavy train loads [11], see Fig. 4. Reinforcing cover plates ending near
existing rivets were welded to the webs of the diagonals of the truss bridge (represented by A, Fig. 4a), and the C-sections were
connected to the gusset plate by fillet welds. Adjacent L-sections previously riveted to the flanges of the C-profiles were also welded to
the gusset.
After only nine months of service, cracks starting in the welds in front of the cover plates caused the failure of one diagonal. Then,
the other reinforced diagonals were corrected by increasing the redundancy through additional plates welded to the gussets (re-
presented by B, see Fig. 4b). However, in 1936, after the bridge had a further period of service of four years and five months, cracks
appeared again more or less simultaneously on the web of five C-channel diagonals, stopping at the nearest rivet hole. In both cases,
the reasons for crack initiation and growth were attributed to the high constraint created by excessive welding near the existing rivet
holes, leading to high stress concentration, and also due to shrinkage stresses, as the joint components were not heated before
welding.
In spite of this, one of the most illustrative cases of fatigue cracks (leading up to local brittle failure) in early all-welded railway
(a) (b)
Fatigue B B
Fatigue
cracks
cracks
Fig. 4. Illustration of one of the first reported fatigue cracks in welded structures which occurred in 1930 and 1936 in the reinforcement of an old
metallic bridge: (a) first and (b) second reinforcement option [11].
162
(a) (b) Cracked

25 mm
area
Groove weld
between flange Drilled holes
and web (Ø20 mm)
20 mm
60 mm
90 mm
500 mm
Fig. 5. (a) As-built groove weld of the girder railway welded bridge in the Train Zoo Station (Berlin), (b) Cracked area and drilled crack arrest holes,
Bridge case study r), Table 4 [58].
bridges were still to occur, two years before the well-known fracture of the road bridge near Hasselt [49]. These cracks took place in
the overpass of the Zoological Gardens Station in Berlin (bridge r), Table 4), which was erected in 1936 as two double frames girder
bridges spanning 50 m – one, single track, the other, double track. Girder sections were made of welded plates, with a web 3 m in
depth. On the single-track bridge, the flange was made of 60 mm plates and 500 mm wide. On the double-track bridge, the flange was
620 mm by 65 mm in cross section.
In the end of 1936, after the single-track bridge had been in use for half a year, transverse cracks visible to the naked eye were
noted in the welds between the lower flanges and the web, extending around 20 mm into the parent metal [58,59]. As soon as the
cracks were discovered, stop-holes were drilled both in the flange and in the web (see Fig. 5b), and temporary supports were
provided. The suggested fracture scenario was that a fatigue crack initiating from a weld defect, in the transition area between the
flange and the web in the lower part of the girder, gradually would have propagated to a certain critical length into the parent metal
(with respect to brittle fracture) and thus caused the fracture, which was restricted to a small zone probably because of load shedding
effects (load redistribution during crack propagation). A roadway welded girder bridge that suffered from fractures originating from
the same detail in the same period in Germany did not have the same “luck”, since the crack extended 1.50 m long in the web [10].
In this case, several factors have contributed to create the necessary conditions to the crack growth and fracture. For example,
groove welds between the flange and the web in the form of that presented in Fig. 5a became common practice in German welded
girders in those times, mainly for a good intention, i.e. to shift the longitudinal flange-to-web weld – seen as an introducer of the so-
called notching effect and as weakest than the parent metal made of rolled plates – slightly above the point of maximum stress in the
lower flange to web intersection. However, such welds are prone to lack of fusion and it is difficult to achieve a good weld penetration
in these kind of details than for the now customary full-penetration butt welds. Even flange-to-web fillet welds might result in less
defective welds if the root is carefully welded. Indeed, groove welds in the web-to-flange weld was then considered a poor detailing
practice [41,49]. Besides that, investigations carried out in 1938 by Kommerell in web-to-flange welds of actual I-girders bridges
revealed the appearance of weld crack-like defects as soon as the weldment was completed [60], which were later attributed to the
high content of hydrogen and phosphorus of the electrodes [49,54], thus leading to the creation of brittle welds. The fatigue cracks
and the later fracture of the I-girders of the Berlin Zoo Station bridge has almost certainly started from a weld crack-like defect, as
later noted by Kommerell. Moreover, Wichtowski [26] reported several cracks and weld defects detected by means of periodically
radiographic inspection tests performed in 155 welded bridges constructed between 1930 and 1960 in Poland. Such cracks were
mainly attributed to the low quality of the welded joints performed immediately after and before the Second World War, and some of
them even gave rise to fatigue crack growth and brittle fracture.
Another factor with implications in the fatigue strength of early welded structures was the advent of high-strength steels in the
1960's. The increase in strength helped to save material in the new bridges that were built, however, because of the high carbon
content (in combination with the high allowable stresses) the material was not suitable for welded bridges, in particular since steels
with higher carbon content can experience hardening through welding. Later investigations performed by German bridge regulators
on brittle cracking due to shrinkage stresses caused by welding led to the establishment of some requirements, such as: i) mandatory
pre-heating before welding of steel plates with thickness > 20 mm (to avoid high weld shrinkage stresses); ii) prescribed limits for the
chemical composition of high-strength steels in order to control the carbon content [49]; and iii) the obligation of use only steel plates
with thickness lower than 50 mm, which then led to development of a particular detail with sound fatigue strength constituted by
several layers of welded plates, also called lamellae joint, in order to achieve higher thicknesses [61]. Some of these rules spread
worldwide, and the construction regulatory agencies of some countries even followed it until recently, e.g. in South Korea where
plates with thickness > 50 mm were not allowed in steel bridge construction until 1999 [62].
3.1.2. Designing fatigue resistant welded connections to crack-like welding defects

Of course, the high number of reported cracks in welding structures between 1920 and 1945 may also be attributed to the level of
the knowledge regarding fatigue of metals in this period [63]. Thus, in order to achieve a higher knowledge on the fatigue of welded
connections, several fatigue tests mainly on small-scale specimens were carried out in different countries, the most important being:
a) Gottwalt Schaper and Otto Kommerell (German State Railways) and Otto Graf (University of Struttgart) in Germany [64–66]; b)
163
Mirko Roš (Empa) in Switzerland [67]; c) Henri Dustin (University of Brussels) in Belgium, mainly for welded nodes of Vierendeel
bridges [35] and d) Wilbur Wilson and Frank Thomas (University of Illinois) in United States for bridge riveted connections [68].
In general, these early tests underlined relevant aspects of the welding technique on the fatigue strength of bridge details, both
from the metallurgical and structural points of view, such as: a) the fractures in the fatigue tests occurred mostly in the parent metal
and started from notches in the surface at the transition between weld metal and plate, an effect that was then called the “notch
effect”; b) as consequence of the “notch effect” on the geometry, it was noted that the use of weld electrodes of superior strength did
not result in higher fatigue lives. Tests performed later in the 1970s by the Lehigh University Laboratory also confirmed that the steel
material strength was not a significant factor to fatigue strength [69], also as a consequence of the “notch effect” created by the
welding process; c) proper executed butt welds can achieve even higher fatigue strengths than riveted connections; d) the fact that
thin welds with smooth transition increase the fatigue strength, unlike what was initially thought before 1935 (e.g. bridge d, Table 1,
where thick fillet welds were adopted) and e) the necessity of constant welding work supervision both in the shop and on site in order
to achieve the expected fatigue strengths.
These findings obtained from early cyclic load tests using pulsator-machines represented a significant contribution to welding
technique, improving its understanding and solving its first technical issues. Later definitive improvements in the quality of steel and
in detailing provisions allowed engineers to design new welded structures with confidence [7]. On the other hand, the fact that
welding could introduce high residual stresses which could be critical to fatigue crack initiation had set a direction in industry to
avoid welding to reinforcement of old metallic structures. This trend remained until the further development of weldability tests and
non-destructive methods from 1970s onwards. Later, the use of welding to strengthening of steel bridges was resumed to very
restricted conditions, although according to Kühn et al. [6] common recommendation is to avoid it on existing structures if possible
and use it only to members subjected to compression.
3.2. Load-induced fatigue
Load-induced fatigue is a result of the fluctuation of the nominal primary stresses, i.e. the stress range induced by the applied
loads using standard first-order design calculations. Load-induced cracking occurs primarily at poor details when these details are
subjected to significant stress ranges exceeding the fatigue limit [70]. Before the advent of welding, engineers already knew that some
structural details with inappropriate geometry often result in fatigue cracking when subjected to uniaxial loads. Details with low
fatigue strength due to the presence of sharp notches were known to be problematic, mainly from railway axles failures. Therefore,
with the development of fatigue tests of welded connections from 1930 onwards and also learning from failures, bridge engineers
gained knowledge about some widespread details which result in very low fatigue strength and should be avoided in construction of
new bridges. In the meantime, however, many bridges were constructed and are still in operation with load-induced fatigue prone
details. In the following subsection, some of these details are reminded and discussed.
3.2.1. Examples of load-induced fatigue cracking

One of the best examples of load-induced fatigue prone welded details are cover plates, for which the characteristic fatigue
strength ranges between 36 MPa and 56 MPa [24,71]. Furthermore, although cover plates are usually avoided in new bridges, it can
be still used in the form of sole plates, a plate attached to the bottom flange of a beam that distributes the reaction of a spherical
bearing to the beam. Recently, fatigue cracks were found in the sole plates of composite steel-concrete twin girder bridges (bridge o),
see Table 3) located in the Shinkansen high-speed line in Japan [72,73].
In this case, fatigue cracking occurred from the sole plate weld, progressing to the girder flange and web. Stress measurements
showed that a drop in support function can greatly increase stresses on fillet welds [75]. For sole plates, when rotating function is lost,
very high stresses occur at the front surface of the sole plate, and these local stresses are made even higher by the deformation of the
flange, see Fig. 6. Three types of cracks occur: in the web to flange weld extending into the web (type L1), in the web to flange weld
propagating in the longitudinal direction (type L2) and in the welds between the sole plate and the flange (type L3).
Fig. 6. Spherical bearing support with (a) good mobility and (b) degraded mobility [adapted from [74]] (Bridge case study o), Table 3).
164
Fig. 7. Illustration scheme of the fatigue cracks in flange lateral attachments in welded railway bridges: (a) bridge cross-section and (b) top view of
the lower flange attachment detail (Bridge case study r), Table 4).
Besides cover plates and similar details, many fatigue cracks were reported in other details due to load-induced fatigue, e.g. lateral
attachment plates welded to the bottom flange of railway girder bridges (bridges j) and k), see Table 2, and bridge v), see Table 4),
[76–79], which serve as connection between the horizontal bottom bracing system and the main bridge members. Horizontal and
vertical bracings are intended to distribute traffic loading and to offer resistance to wind, but most important, it functions as bracing
against girders lateral buckling during construction, before the concrete deck is poured. Usually, for constructional reasons the
position of these lateral attachments is matched with the position of transverse attachments welded to the web. The gusset is welded
over the lower flange by means of fillet welds or in the flange thickness, when it is called a flange tip attachment, creating a very
prone fatigue detail, with characteristic fatigue resistances in the order of 50 MPa at 2 million constant amplitude cycles [80]. The
mechanism of crack formation is very simple and illustrated in Fig. 7. The crack usually initiates from the fillet weld toes extending
into the parent metal normal to principal direction in a zone of high primary tension stresses.
Alternatively, lateral attachments could be welded to the girder web instead of the tensioned flanges. However, some cautions are
required in order to position the gusset in a zone of lower tension stresses. An example of fatigue cracking in this kind of connection
which was placed very near to the tension flange was reported for a simply-supported completely welded railway girder bridge
spanning 19.80 m built in 1956 in the U.S. and replaced by a new bridge in 1997 [81,82]. With the bridge replacement, the old bridge
was moved to a test line at the Facility for Accelerated Service Testing (FAST) in Pueblo, Colorado (U.S.) to investigate the effects of
Heavy Axle Loads (HAL) traffic – composed by trains with axle weights usually higher than 22.5 t – on the fatigue life of welded
details. After being subjected to 1900 million gross tons of HAL traffic between 1997 and 2014 (equivalent to 12 million load cycles),
several fatigue cracks were reported in different locations, and the bridge is now out-of-service (bridge s), see Table 4). The cracks
that developed in the gusset-to-web are shown in Fig. 8, where it is possible to observe that the gusset was placed very near to the
lower flange, in a location with high bending stress concentration.
Another type of fatigue cracking due to load-induced effects which was referred by China Academy of Railway Sciences [83]
occurred in the tube-gusset joints of longitudinal diaphragms of cable-stayed bridges. Although such connection being not very
common in old steel bridges, it is presenting an increasing adoption in the construction and rehabilitation of steel and composite
steel-concrete bridges in the last years [84–87], mainly due to the following advantages: i) facilitation of girder installation and to
Fig. 8. Fatigue cracking in the gusset plate of the welded railway bridge at FAST [adapted from [82]] (Bridge case study s), Table 4).
165
Fig. 9. Typical fatigue failures of tube-gusset joint of the longitudinal diaphragm of cable stayed-bridge with orthotropic deck, China Academy of
Railway Sciences [adapted from [83]].
obtain desirable deck performance, being very efficient in compression; ii) easy internal access; iii) weight saving and iv) elimination
of a considerable amount of weld by replacing plate diaphragms. However, important differences are found between fatigue
structural codes concerning the characteristic fatigue strength of tube-gusset joints on the basis of nominal SeN curves, being FAT 71
or 63 according to EN 1993-1-9 [88], FAT 63 or 45 according to IIW [89], FAT 68 or 43 according BS 7608 [90] and FAT 90 or 71
according to DNV [91].
The cracks reported in the tube-gusset connections of the longitudinal diaphragms appeared only after few years of service and
initiated from the weld toe near the slot end and propagated into the tube along the circumference with crack lengths ranging from
several to hundreds of millimetres, and some of them eventually resulted in a circumferential failure, as shown in Fig. 9. The fatigue
strength of the connection could have been increased by adopting a chamfer angle in the end of the tube and full-penetration welds,
allowing a smooth stress flow and reducing the stress concentration. However, in the reported case, a tube without chamfer and
welded to the gusset by means of fillet welds was adopted. The cracks can also be attributed to the overly optimistic fatigue nominal
strength used in the design phase and based on current standards. For example, according to Baptista et al. [84] Eurocode curves
(FAT 71 or FAT 63 for a chamfer angle smaller or higher than 45°, respectively) are clearly unrepresentative and nonconservative
when compared to old and recent fatigue tests results of tube-gusset joints.
3.2.2. Designing to load-induced fatigue

The way to design against load-induced cracking is by performing extensive laboratory fatigue tests of large and small-scale
specimens, which is an attempt to simulate typified welded details following the so-called principle of similitude [92]. Most design
rules of current fatigue standard bridge codes were created bearing in mind this type of fatigue cause by performing uniaxial fatigue
tests [88,93]. Regarding cover plates, nowadays this type of detail is usually used only for the strengthening of existing metallic
bridges, but with precautions [94,95]. Recent experimental and analytical research have proposed the upgrading of the fatigue
strength of cover plates welded to existing steel girder bridges up to FAT 80 by extending their ends to zones of low stress variation
[96]. Concerning lateral connection plates, later fatigue tests performed with large-scale specimens indicated that the fatigue re-
sistance of flange tip attachments can be improved by adopting a smooth transition radius between the gusset plate and the flange,
however just up to FAT 90, because after a certain radius the fatigue life began to be limited by cracks initiating from the weld roots
[97]. Following this recommendation, railway bridges built after 1980s usually adopted a transition radius in the flange-tip at-
tachment in order to improve its fatigue performance. This was the case of the 24 steel bridges constructed for the Korea High-speed
Railway line, connecting Seoul to Busan, see Lee et al. [62].
Another complementary way to avoid load-induced fatigue is by designing welded connections with good detailing practices,
mainly based on past steel bridge construction experience regarding fatigue behaviour. Most railway bridges that presented load-
induced fatigue cracks in the last two to three decades were built before 1975, when the procedures to design welded connections to
be fatigue resistant with good detailing practices were not well developed in structural codes. Currently, detailing rules are seen by
some authors as perhaps the most important part of the fatigue and fracture design and assessment procedures [98]. The detailing
rules are intended to avoid notches and other stress concentrations. Because of the detailing rules, modern steel bridges are detailed
in a way that appears much cleaner than those built before the 1970s [98]. There are fewer connections and attachments in most
recent bridges and the connections use more fatigue-resistant details. For example, in the case of lateral attachments welded to the
web, a minimum distance from the flange is required in order to reduce the stress concentration in the gusset-to-web weld to a
tolerable magnitude and improve its fatigue strength. It also provides adequate electrode access and places the connection plate
closer to the neutral axis of the girder, thus reducing the impact of the weld termination on fatigue strength. In this manner, the
Supplement S6–14 of the Canadian code for bridge construction [99], specifically in the clause 8.9.5.1 specifies a minimum vertical
166
distance of 125 to 150 mm between web lateral attachments and the girder flange, depending on the attachment length.
Furthermore, welded bridges built in the 1950s until the late 1960s used the fatigue design specifications developed on the basis
of fatigue tests largely performed with small-scale specimens in the 1930s (referred in the last section). However, fatigue cracking due
to primary stresses was then often found in many highway and railway bridges. New fatigue tests performed in the 1970s with girder
and full-scale specimens indicated that the tests performed with small-scale specimens overestimated the fatigue resistance of
structural details [100], a fact that is well-known today [101]. As consequence, some fatigue design provisions in use since the 1930s
were inadequate and overly optimistic, particularly at longer lives, because the assumption of a fatigue limit of 2 million cycles
proved to be incorrect (a number arbitrarily defined by Wöhler in 1860). On the other hand, another cause of fatigue cracking, due to
secondary stresses instead of primary stresses revealed to be even more critical to railway bridges in the same period. In the following
subsection, this cause will be outlined.
3.3. Distortion-induced fatigue
Distortion-induced fatigue cracking results from second-order stresses, typically due to out-of-plane deformations and in-
compatible deformations at intersecting structural elements [70]. Nishikawa et al. [102] mentioned that almost all fatigue cracks
reported in Japan are caused by distortion-induced stresses. Connor & Fisher [103] have estimated that nearly 90% of all fatigue
cracking is the result of out-of-plane distortion or other unanticipated secondary stresses at fatigue-sensitive details. Bowman et al.
[104] have conducted a survey of US state transportation officials in order to evaluate current fatigue inspection and evaluation
procedures. The results of the survey revealed that distortion-induced fatigue cracking is the most frequently encountered type of
fatigue distress observed by various state transportation agencies is United States. The AASHTO LRFD [93] specification, as well as
the Manual for Bridge Evaluation (MBE) [105] defines it as “fatigue effects due to secondary stresses not normally quantified in the typical
analysis and design of a bridge”. Early reported cracks due to distortion-induced fatigue in welded joints were observed in the mid-
1970s at the end of vertical stiffeners or connection plates for floor beams and transverse diaphragms of railway bridges [71,106].
The cracks appeared in a region that was previously designated as “web gap”, i.e. a space let between the transverse stiffeners and the
tension flanges during girder construction, a practice that had its origin from the German codes for railway bridge construction
developed in the 1930s due to the concern of fatigue cracking related with welding an element normal to primary stresses [64].
This concern, which is frequently mentioned in introductory texts to distortion-induced fatigue, is herein recovered from a
conference paper translated to English by the secretariat of the second IABSE meeting held in Berlin in 1936:
“While in early days stiffeners were welded unhesitatingly to both flanges, later on endurance fatigue tests with fillet welds placed vertically
to the direction of force and with the zones in which the lateral fillets start and end led to the recognition that, on account of the considerable
reduction of fatigue resistance, it would be necessary to prohibit welding of stiffeners and girder connections in the tension flange for bridge
construction.” [64].
Although distortion-induced fatigue is a problem more frequent in steel girder bridges, it also has developed in many other types
of bridges, such as trusses, suspension bridges, tied arch bridges and box girder bridges and have been known to crack frequently and
relatively early in the service life of many bridges [82]. Indeed, initiation of fatigue cracks in the web-gap regions for structures not
designed to current code specifications is anticipated to occur within 1/5 of the design service life of a bridge, which is equivalent to
more or less 20 years for most bridges, as can be testified by the data obtained by Yokoyama & Miki [57], see Fig. 3. In the following
subsection, most representative distortion-induced fatigue prone details are outlined and discussed.
3.3.1. Examples of distortion-induced fatigue cracking

Recent examples of welded girder railway bridges with distortion-induced fatigue cracking in web gaps can be found in Åkesson
et al. [107], Grondin e Kulak [108], Aygül et al. [109] and Dick et al. [81] the latter two shown in Table 2 and Table 4, bridges k) and
s), respectively. Grondin e Kulak [108] reports an incredible number of 300 distortion-induced fatigue cracks in the transverse
stiffeners web gaps of a railway composite steel-concrete multi-girder bridge in Canada, removed from service in 1998. The overall
driving mechanism for distortion-induced fatigue cracking of welded railway bridges is similar throughout the different existing
types, and it is well-illustrated in Fig. 10 [110]. Out-of-plane forces are developed in the diaphragm members when one girder
deflects a different amount than an adjacent girder. Such deflections give rise to diaphragm rotation, which in turn imposes a
deformation on the transverse stiffeners. The diaphragm members that are typically attached to the connection plates introduce a
pumping on the corresponding web region and tend to develop elevated out-of-plane stresses. Because transverse stiffeners are rigidly
welded to relatively flexible girder webs, the latter take up most of the out-of-plane displacement and rotation, leading to web
cracking, usually in the form of a ‘horse shoe’ crack (Fig. 10c). The effect of the out-of-plane stresses can be summed to the effect of
the primary stresses generated by the overall flexural behaviour of the steel girders, leading to a combination of both load-induced
and distortion-induced fatigue phenomena. It was observed that even though the out-of-plane deformation being relatively small in
comparison to the deformation of the entire structure, it can induce very high stresses. For example, some authors put the localized
minimum out-of-plane deformations that can produce this type of cracking in the order of 0.10 mm, which are quite frequent in
actual bridges under normal rail traffic [82,111]. Experimental measurements on actual bridges showed that secondary stresses in
web gaps can even be much larger (as much as 20 times larger) than primary stresses in flanges [112], and it can occur either in the
top or in the bottom flange web gaps.
Despite distortion in web gaps of girder bridges, another type of distortion occurring in the top flange has often caused fatigue
cracks in the welds between transverse stiffener and top flanges in many bridges (Fig. 11). In railway bridges, the rail track is either
lying on cross ties which may be directly on the girders or lying on sleepers which rest on ballast. One aspect that was often
167
Fig. 10. (a) Half cross-section of a typical steel girder bridge; (b) Out-of-plane distortion and (c) Crack location [adapted from [110]] (Bridge case
study k), Table 2).
Fig. 11. Cracks due to top flange distortion in a welded railway bridge [113] (Bridge case study j), Table 2).
overlooked in early welded railway bridges with pure steel, wood or RC slab decks was the existing eccentricity created between the
main longitudinal bending axis of the girders and the rail (see Fig. 12a). This eccentricity combined with horizontal loads induced by
the wheel-rail contact (Fig. 12b) can induce additional out-of-plane restraint bending in the flanges, leading to fatigue cracking of
flange welded attachments. Recent examples of welded girder railway bridges with this type of crack can be found in Kyung et al.
[113], Lee & Lee [114] and Lippi et al. [115], the latter two shown in Table 2 and Table 4, bridges j) and t), respectively.
Several cracks were found in brackets of the walkway of steel welded railway bridges in recent years, due to displacement of the
floor-beams in the longitudinal direction. According to Fisher [41], this is one of the most oldest cases of cracks due to out-of-plane
distortion, firstly detected in highway bridges in U.S. in the beginning of the 1970s. The cracks usually occur in two main locations,
i.e. the connection of the walkway floor beams to the brackets (as shown in Figs. 13–14) and between the brackets and the web of the
e
(a) (b)
M1 M1
M2 M2
Fig. 12. Mechanism of distortion in the flanges due to (a) vertical loads and (b) horizontal loads [adapted from [56]].
168
Fig. 13. Fatigue crack due to out-of-plane displacement of the walkway under the crossing of HAL traffic in a bridge located in Vitória-Minas Line
(Brazil) [119] (Bridge case study w), Table 4).
Fig. 14. (a-c) Mechanism of creation of distortion in the walkways of continuous bridges [adapted from [41]].
steel girders. Recent examples of bridges that suffered from this effect can be found in [116–118]. It seems to be common in brackets
of bridges with continuous spans. The overall mechanism of crack formation is illustrated in Fig. 14a-c.
In general, the rotation angle due to train crossing and girder bending (θ) led to an additional longitudinal displacement (Δ),
which, due to the rigid welded connection between floor beams and brackets creates secondary stresses at the level of the bracket to
the girder weld and at the weld between the brackets and the floor beams.
A famous case of distortion-induced fatigue cracking took place at the Dan Ryan welded railway viaducts located in Chicago
(Illinois). Following the discovery of cracking in the bents of the Dan Ryan Rapid transit, the cracks were temporary retrofitted with
two stop holes at both crack tips of the through thickness crack [82]. Later, the fractured bents were replaced by new bents. I-girders
are currently resting on the upper flange of the bents (Fig. 15a–b). At the time of the discovery of the cracked bents, the structure was
analysed using the finite element method for computing stresses. The web of the box bent was modelled with shell elements. The
analysis of the deformations in the area intersection between the east web and bottom flange indicated stress concentrations due to
out-of-plane bending of the web, see Fig. 16 [120]. Another factor contributing for the stress concentration was attributed to the lack
of fusion in the welds of the intersections, which were connected by partial penetration groove welds that resulted in a large unfused
region in the box girder web [41].
3.3.2. Designing to distortion-induced-fatigue

First attempts to characterize the fatigue strength of distortion-induced fatigue sensitive details were performed by Fisher et al.
[121] in NCHRP no. 206 (National Cooperative Highway Research Program Report) for the web gaps of welded transverse stiffeners.
The goal of this NCHRP study was to develop recommended criteria for designing steel girders so that distortion-induced fatigue
169
Fig. 15. (a) Old [82] and (b) current Dan Ryan Expressway viaduct state [Google Street View].
top flange
girder web
Lack of fusion of the bottom-

flange to web bent weld
bottom flange
Out-of-plane bending of the

Box bent
web due to live and dead load
without reverse cycle
Fig. 16. Curved cross girder structure with fatigue crack in the bottom flange to box bent web weld [adapted from [82]].
cracking problems are minimized. The fatigue testing demonstrated that for cyclic out-of-plane displacement of web gap regions, the
fatigue strength increased as the web gap length increased. However, a relationship between web gap length and the initiation of
fatigue cracking revealed not to be directly proportional. Small web gaps, less than five times the web thickness, were found to have
very unpredictable behaviour. Later, a comprehensive experimental study was conducted by Fisher et al. [122] in NCHRP no. 336
with the goal of recommend criteria for designing steel girders in order to minimize distortion-induced fatigue cracking problems.
When the data was plotted on an SeN curve it was verified that all the data exceeded FAT 80 category, according to the nominal
stress method. From this experimental study, some recommendations were made. Rigid connections of the plate to the top and
bottom flanges by bolted or welded connections are needed to prevent fatigue cracks from out-of-plane deformation. Successful arrest
of cracking at details with distortion-induced cyclic stresses higher than 105 MPa requires a positive attachment between the con-
nection plate and the girder flange in order to bridge the web gap and reduce the magnitude of the out-of-plane web bending stress.
These early tests were an initial attempt to characterize the fatigue strength based on the nominal applied stress and detail category.
More recently, some fatigue tests to characterize the characteristic fatigue strength of distortion-induced fatigue bridge sensitive
details in terms of local applied stresses (such as the stress at the weld hot-spot) have been performed [123]. In this context, the hot-
spot stress method revealed to be suitable to analyse the fatigue performance of distortion-induced fatigue sensitive details of
composite steel-concrete bridges through the use of the finite element sub modelling technique [124].
According to Okelo [125], although current bridge design specifications provide rules to estimate the fatigue life for load-induced
cases, fatigue cracking due to out-of-plane distortion is not adequately addressed in standard structural codes, despite its severity. In
general, most bridge design specifications codes intend to prevent distortion-induced fatigue issues by recommending detailing
guidelines for common welded joints. The AASHTO LRFD [93] specification, Article 6.6.1.3, as well as the MBE [105] provide a
separate section on distortion-induced fatigue. They contain general statements that stress the importance of the proper connection of
transverse connection stiffeners to longitudinal (i.e. main) components, however the development of web gap stresses is limited
through improved detailing requirements and it does not explicitly classify detail categories for distortion-induced fatigue. Pre-
scriptive provisions on how distortion-induced fatigue cracking can be evaluated or treated are not provided. AASHTO LRFD [93]
also requires that stiffeners used as connection plates for diaphragms or cross-frames shall be attached to both the compression and
tension flanges of the longitudinal girders. In Europe, although the Eurocode standard for fatigue assessment of steel structures [88]
170
(a) hangar (b)

crack
flange
Fig. 17. (a), (b) Cracks induced by vibration, e.g. by wind or traffic in welded hangars of railway bridges [131].
does not deal directly with ‘distortion-induced fatigue-sensitive details’, design rules to avoid fatigue web gap cracking are given by
the British Steel Construction Institute in Hendy & Iles [126]. It is interesting to note that these rules for intermediate transverse
stiffener design are quite similar in terms of broad concepts to the approach taken in AASHTO [93].
3.4. Vibration-induced fatigue
Vibration-induced fatigue cracking in steel bridges can result from either primary or secondary stresses, and can originate from
different type of actions, e.g. wind loading, railway traffic, earthquakes and others. Due to low local stiffnesses in correspondence of
critical details, vibration phenomena can occur simultaneously or not with distortion-induced fatigue, and can in fact strongly affect
steel bridges safety [55,127]. Vibration caused by wind was already recognized as being very critical since the well-known failure of
Tacoma Narrows Bridge, which took place in 1940. In this case, the wind caused an excessive number of stress reversals of the bridge
deck (made of steel riveted girders), which resulted in resonance with a 0.2 Hz torsional vibration mode, leading to failure due to low-
cycle bending fatigue [49]. The sources of railway traffic-induced vibration in welded components can be separated as follows: i) due
to the complex dynamic interaction phenomena between the train and the bridge; ii) due to dynamic effects caused by the contact
forces between the wheel and the train considering the rail irregularities and the wheel defects and iii) due to the global and local
resonance that can be created by the crossing of repeated and equally spaced axle loads. In the following subsection some examples of
vibration-induced fatigue regarding the type of phenomenon are cited and discussed.
3.4.1. Examples of vibration-induced fatigue cracking

More recently, fatigue cracks on tensioned welded hanger connections of railway bowstring bridges, due to wind and traffic
vibrations, were reported by Haghani et al. [56] and Klinger et al. [128], with an illustration example shown in Fig. 17. Hangers
usually have very low bending stiffness, which makes them very sensitive to resonance with many stress cycles in the free vibration
phase [129,130]. Since the goal of current work is to review fatigue due to traffic loads, herein a focus will be given for cracks caused
by traffic loads.
One of the early reported cases of cracks due to vibration induced by traffic loads in railway bridges occurred in the Tokaido
Shinkansen high-speed line [55]. Although high-speed passenger trains spacing are not considered very relevant in terms of ultimate
limit state condition (ULS) due to their relatively low weight compared to freight trains [132], they are able to originate high
dynamic effects on bridges. The advent of high-speed traffic led to an increase in reported fatigue cracks of welded bridges due to
secondary stresses created by excessive vibrations observed at certain circulating high-speeds. Since the start of operation in 1964,
speeds have been increased progressively and from about the time that 200 km/h was exceeded, vibrations of bottom flanges of
stringers in plate girder bridges and truss girder bridges in directions perpendicular to bridge axes began to appear prominently when
crossed by trains, after only 8 years of service [133]. Vibrations in out-of-plane directions have also been reported in diaphragms of
box section girder bridges in the Tokaido Shinkansen, Fig. 18 [134]. The presence of local vibration modes of the web characterized
by high-frequencies in the range 20–45 Hz, typically excited during train transit, see Fig. 18b could cause high local “modal” stresses
leading to fatigue damage and reducing the remaining fatigue life [134]. The natural frequencies of these local modes were identified
by means of the FFT from both numerical simulations and experimental measurements in the actual bridge due to the crossing of
high-speed trains with speeds of about 270 km/h, considering a coupled dynamic analysis of the train-bridge system for the former.
Accordingly, peak frequencies at about 30 Hz were distinctive with relatively large magnitude, which indicated the possibility of
resonance of this high-frequency mode with the train moving loads [135].
Another structure typology often used in long-span high-speed railway bridges which is very sensible to vibration-induced da-
mages according to Lippi et al. [8] are orthotropic steel decks (OSDs). It is well-known that fatigue phenomenon is a worldwide
problem to orthotropic bridge decks [137]. Recently, several cases of fatigue cracks in railway bridges with orthotropic decks were
reported and investigated by the China Academy Railways [83,138]. In China, OSDs were adopted in the Chengdu-Kunming Railway
for the first time during 1960s–1970s and have been developing at an unprecedented speed since 1990s. Cracks in OSDs are caused by
171
Fig. 18. (a) fatigue cracking in a steel box girder of the Tokaido Shinkansen and (b) out-of-plane vibration modes of the web leading to increase
“modal” stresses in the web gap [134–136] (Bridge case study o), Table 3).
the secondary bending stress concentration generated by the out-of-plane distortion at the welding restraint provided by the so-called
rib-to-deck connection. OSDs have been widely adopted in high-speed railway bridges in China, and other numerous bridges are
planned to be built in near future [139]. The study of the vibration phenomena caused by high-speed trains in orthotropic decks and
hence the impact on the fatigue life has increasingly attracted research interest in recent years [140,141].
On the other hand, fatigue cracks due to vibration were also reported for train speeds lower than 200 km/h, specifically for HAL
traffic for the welded bridge at FAST (bridge s), see Table 4). In particular, it was observed a strong correlation between rail
irregularities and increased rate of fatigue crack initiation and growth. In railway bridges, the irregularities of the rail have a strong
influence in the dynamic behaviour of the structure, and it can be separated between two parts: (a) static track irregularity, due to the
inherent defects of the rails and the train wheels and (b) dynamic track irregularity, due to rail elastic deformation under train loads
[142]. In the welded bridge at FAST, the static irregularities, i.e. without the effect of the live load were measured to have an
amplitude of 2.30 mm on the north rail and 2.80 mm on the south rail, which were considered of moderate level. However, they were
enough to cause strains about 10% higher than with smoother rails. The higher strain ranges observed for non-smoothed rails resulted
in about a 50% increase in fatigue accumulation per train crossing [81].
Other example which occurred for speeds under 200 km/h was at the steel structure of one of the elevated parts of the Washington
D.C. Metro, specifically in the crossing over the Potomac River (bridge u), Table 4). The Washington D.C. Metrorail system was built
in 1970's and inaugurated in 1976. Each train is 22.86 m long with a unique track gauge of 1429 mm and being able to operate at a
maximum speed of 121 km/h. The crossing within this portion is composed of multiple simple spans across concrete piers, varying
between 21 and 36 m approximately, with two steel box girders section, which support a RC slab. Within each box girder, K-type
diaphragms are weld connected to transverse connection plates which are fillet weld connected to the box girder webs. The referred
fatigue cracks occurred in the transverse connection plate weld ends [98]. Herein, vibration with frequencies with magnitude in the
order of 20 Hz were observed (Fig. 19), in a very similar manner with the steel box girders fatigue cracking found in the Tokaido
Shinkansen bridges.
3.4.2. Designing to vibration-induced fatigue

Due to their lightness and low damping capacities, it is likely that steel bridges could experience resonance-induced vibrations
produced by trains running slower than the maximum design speed usually set to 350 km/h for high-speed lines. If a frequent
operating speed of a HS train passing through high-speed railway bridge corresponds to the resonant speed, excessive span deflections
and vibration may be caused and, therefore, the dynamic behaviour of railway bridges subjected to high-speed train loading should
Outter face:
Compression (–)
Inner face:
Tension (+)
θ
Out-of-plane
movement of
Web from Box Deflection of bottom flange
corner rotation
Fig. 19. View of the Metrorail double-box steel girder corner of the bridge over the Potomac River with out-of-plane vibration of the girder web
[adapted from [98]] (Bridge case study u), Table 4).
172
be taken into consideration in the fatigue assessment [143]. Most of the first bridges built for the Tokaido Shinkansen in the 1960s,
where most fatigue cracks due to vibration-induced fatigue were early reported, adopted pure steel solutions (ballastless track half-
through girder, through truss, box girder or girder bridges), which are well-known to have very low mass and damping ratios. In the
case of modern high-speed railway bridges, very strict design criteria in terms of deformability (vertical, horizontal and torsional) and
dynamic behaviour are now recommended by standards and technical-scientific literature [143–145]. These criteria can be fulfilled
by designing a main steel or composite section, providing the required stiffness, incorporating a concrete slab underneath the track
and adopting ballasted tracks. The ballast and the concrete introduce additional mass and damping under the track, thus reducing the
dynamic vibrations induced by the live load. This design philosophy has been employed throughout the French HS lines since the
TGV Nord, opened to traffic in 1990. The efficient use of materials, alongside with the design of simple details, lead to a good fatigue
behaviour of these structures and, subsequently, to a longer lifecycle and lower maintenance [146].
Phenomena like vibration-induced fatigue are still partially uncovered by actual design codes and represent a critical aspect for
the assessment of existing bridge remaining life and for the design of new bridges. One relevant attempt to change this scenario was
performed by the European project FADLESS [8], whose main objective was to clarify the uncertainties concerning vibration/dis-
tortion induced fatigue into railway steel bridges. In particular, the most important aspects outlined include the dynamic behaviour,
the influence of train-bridge interaction phenomena, the local vibration/distortion effects, the actual stress patterns into critical
details, the real composition of traffic spectra and the influence of secondary stresses on the fatigue resistance of components. Several
bridge case studies in Europe were selected in order to apply the methodology developed under FADLESS project, which included a
variety set of bridges with different construction methods (riveted-only, welded, composite steel-concrete, etc.). Regarding vibration-
induced fatigue on welded joints, the FADLESS project highlighted that the structural damping has an important effect on the
dynamical behaviour of a railway bridge because it is one of the main factors limiting the vibration amplitudes. Its knowledge is
therefore essential for the assessment of existing bridges subjected to high-speed trains. Moreover, the computation of the local stress
effects by vibration phenomena in order to perform the fatigue assessment of critical details requires the consideration of the whole
dynamic system for the global model. Because the task of numerical simulation is to check the local vibration of the steel bridge, one
sufficiently precise numerical model is necessary, both from a global and a local point of view.
4. Conclusions
In the present work an in-depth literature review covering more than one century was carried out, providing a multidisciplinary
analysis with insights into the welding history and practice, especially applied to steel railway bridges construction. The combined
results of the historical, technical and structural investigations emphasize the importance of confronting the original design of a
welded joint with its historical context, characterized by joining typologies, construction techniques and calculation methods peculiar
to a certain time period. Advances in welding technology and shop practice played a significant role in fatigue design of steel railway
bridges, but few studies have emphasized yet how those practices determined, for example, the changes that made possible the
construction of welded bridges with high fatigue performance. This is an important area of study that needs to comprehend a
worldwide context.
The first technical issues related to welding and the generated distrusted climate followed by many reported cases of fatigue
cracking and brittle failures of dynamic loaded structures after few years of service in the early times explain why riveted connections
were still in use until 1970s. Both the mentioned cases of early reported damages, the cracks of a lattice bridge reinforced with
welding and the fracture of the girders of the Berlin Zoological Garden station bridge were caused by insufficient experience with
welding technology applied to steel bridges. Such problems were initially perceived as setbacks by the engineering community;
however, they contributed later to the further development of welding technology in the correct direction. Thus, slowly welding
technology became the dominant technology for the fabrication of built-up bridge components, surely overtaking riveting as the
preferred joining method.
The following individual cases provided valuable insight into the causes of cracking, the importance of details, and the sig-
nificance of defects on the performance of railway bridges. Throughout the years, further advice regarding caution with fatigue
cracking has been given, certainly with good intentions. Some, however, such as the warning to not weld transverse stiffeners to
tensioned flanges or the recommendation to use groove welds instead of butt welds in the flange-to-web weld, were unrealistic. In
most cases, fatigue cracking in bridges resulted from an inadequate experimental base and overly optimistic specification provision
developed from early experimental data with small-scale specimens. For the design of new steel railway bridges, some of the causes of
cracking discussed in this work can be prevented with the engineering knowledge acquired today. The lessons learned from the past
assist with an understanding of the behaviour of railway bridges and the importance of detail and execution.
As outlined in this work, distortion/vibration phenomena represent nowadays a cause for concern for most railway bridge
managers, and one that continues to attract research. Therefore, some aspects of the distortion and vibration-induced fatigue were
addressed, as were referred some important works related with investigation of the structural behaviour of connections with out-of-
plane displacements in steel railway bridges. Identification of vulnerable structures and assessment of distortional fatigue stresses are
very important tasks to evaluate the remaining service life of railway infrastructure inventory. However, the lack of standard
guidelines for the evaluation of distortion-fatigue crack initiation and propagation is one of the challenges that structural engineers
could face in practice when dealing of the ageing and deterioration of old steel bridges and in the design of non-standardized
structural details for new bridges. In this case, the application of numerical modelling techniques with the Finite Element Method,
considering updated models to experimental measurements together with advanced methods for fatigue assessment based on local
approaches may be useful to evaluate fatigue-prone details of existing bridges. For design purposes, FE modelling may offer a unique
173
perspective to analyse stress concentrations, distortion/vibration effects, and fatigue crack propagation caused by cyclical loadings.
The stresses acting in the mentioned details are difficult to determine without a high-level finite element analysis of the bridge and,
hence, are difficult to design for.
Declarations of interest
None.
Acknowledgements
The first author gratefully acknowledges the financial support for this research work provided by the Brazilian Science
Foundation'sCNPq- National Council for Scientific and Technological Development (doctoral scholarship process 203662/2014-8).
The authors also gratefully acknowledge the financial support for this research work provided by CAPES, FAPERJ and the Portuguese
Foundation for Science and Technology, particularly through the Projects FiberBridge - Fatigue strengthening and assessment of
railway metallic bridges using fiber-reinforced polymers, POCI-01-0145-FEDER-030103, and CONSTRUCT - Institute of R&D in
Structures and Construction, POCI-01-0145-FEDER-007457, both funded by FEDER funds through COMPETE2020 – Operational
Thematic Program for Competitiveness and Internationalization (POCI).
References
[1] Sustainable Bridges, D1.2 European railway bridge, Demography vol. 1, (2004) 1–15.
[2] P. Oehme, Schäden an stahltragwerken: eine analysis, in German, IABSE Proc, vol. 13, 1989, pp. 121–140.
[3] S. Yongji, Y. Yenman, C. Zegan, Fatigue failures of steel railway bridges in China, IABSE Colloq. International Association for Bridge and Structural Engineering,
Lausanne, Switzerland, 1982, pp. 516–524.
[4] A. Lédeczi, T. Hay, P. Völgyesi, D.R. Hay, A. Nádas, S. Jayaraman, Wireless acoustic emission sensor network for structural monitoring, IEEE Sensors J. 9 (2009)
1370–1377, https://doi.org/10.1109/JSEN.2009.2019315.
[5] T. Yan, B.E. Jones, Acoustic emission testing on large structures, Meas Control 37 (2004) 168–173.
[6] B. Kühn, M. Lukić, A. Nussbaumer, H.P. Günther, R. Helmerich, S. Herion, et al., Assessment of existing steel structures: recommendations for estimation of
remaining Fatigue Fife (JRC scientific and technical reports), Jt Rep Prep under JRC-ECCS co-operation Agreem Evol Eurocode 3 (2008).
[7] P. Harshbarger, I. Wuebber, Robert John Prowse, A Monograph, (2009).
[8] F. Lippi, W. Salvatore, A. Braconi, M. Finetto, H. Wenzel, Roeck G. De, et al., FADLESS: Fatigue Damage Control and Assessment for Railway Bridges, (2012).
[9] M. Braun, Beginn der Brückenschweißung vor 80 Jahren (in German), Stahlbau 81 (2012) 803–812, https://doi.org/10.1002/stab. 201201604.
[10] M. Braun, Die Autobahn-Talbrücke bei Rüdersdorf, Germanica 1904 (2015).
[11] O. Kommerell, Experience obtained with structures executed in Germany, IABSE Congr. Rep. International Association for Bridge and Structural Engineering,
Berlin, 1936.
[12] J. Caldwell, Calculation, design and construction of welded steel structures, First Congr. Int. Assoc. Bridg. Struct. Eng., Paris, 1932.
[13] A. Goelzer, Experience obtainet with structures executed in France, Second Congr. Int. Assoc. Bridg. Struct. Eng., Berlin-Munich, 1936.
[14] J. Konishi, Railways and bridges, Japan Railw Transp Rev (2011) 48–55.
[15] M. Fauconnier, Renforcement du viaduct d'Austerlitz par soudure à l'arc électrique (in French), Second Congr. Int. Assoc. Bridg. Struct. Eng., Berlin-Munich,
1936.
[16] G.D. Fish, Arc-Welded Steel Frame Structures: Designing, Estimating, and Construction Data for Engineers, Architects, and Contractors, McGraw-Hill Book
Company, inc, 1933.
[17] F. Voormann, The use of welding in civil engineering – conditions of a technological innovation in the 1920s, in: K.E. Kurrer, W. Lorenz, V. Wetzk (Eds.), Proc.
Third Int. Congr. Constr. Hist. Neunplus1, Berlin, 2009, pp. 1471–1478.
[18] M.V. Biezma, F. Schanack, Collapse of steel bridges, J. Perform. Constr. Facil. 21 (2007) 398–405, https://doi.org/10.1061/(ASCE)0887-3828(2007)21:5(398).
[19] F. Zelisko, Experience obtained with structures executed in Austria, Second Congr. Int. Assoc. Bridg. Struct. Eng., Berlin-Munich, 1936.
[20] F. Voormann, M. Pfeifer, M. Trautz, Die ersten geschweißten Stahlbrücken in Deutschland - Über die wechselvollen anfänge der schweißtechnik (in German),
Stahlbau 75 (2006) 287–297, https://doi.org/10.1002/stab. 200610029.
[21] F.L. Lehtinen, Experience obtained with structures executed in Finland, Second Congr. Int. Assoc. Bridg. Struct. Eng., Berlin-Munich, 1936.
[22] N. Lancos, Experience obtained with structures executed in Yugoslavia, Second Congr. Int. Assoc. Bridg. Struct. Eng., Berlin-Munich, 1936.
[23] B. Wichtowski, R. Pakos, Diagnostic inspections of butt joints in the oldest steel welded bridges in Poland, Weld. Int. 26 (2012) 493–503, https://doi.org/10.
1080/09507116.2011.581346.
[24] B. Wichtowski, J. Hołowaty, Assessment of fatigue limits in historical welded railway bridges in Poland, Procedia Struct Integr 5 (2017) 1035–1042, https://
doi.org/10.1016/j.prostr.2017.07.066.
[25] B. Wichtowski, An analysis of butt joints in a railway viaduct, Weld. Int. 15 (2001) 180–184, https://doi.org/10.1080/09507110109549339.
[26] B. Wichtowski, Assessment of fatigue and selection of steel on constructions of steel bridges welded according to Eurocode 3, Weld. Int. 27 (2013) 345–352,
https://doi.org/10.1080/09507116.2011.600042.
[27] S. Ewert, Die groBten Stahi-Vollwandtrager-Briicken (in German), Stahlbau 68 (1999).
[28] T.F. Peters, IABSE-the First 80 Years. IABSE, (2011).
[29] S. Bryla, Experience obtained with structures executed in Poland, Second Congr. Int. Assoc. Bridg. Struct. Eng., Berlin-Munich, 1936.
[30] A. Ledang, Experience obtained with structures executed in Norway, Second Congr. Int. Assoc. Bridg. Struct. Eng., Berlin-Munich, 1936.
[31] J. Leander, A. Andersson, R. Karoumi, Monitoring and enhanced fatigue evaluation of a steel railway bridge, Eng. Struct. 32 (2010) 854–863, https://doi.org/
10.1016/j.engstruct.2009.12.011.
[32] C. Miklósi, Experience obtained with structures executed in Roumanie, Second Congr. Int. Assoc. Bridg. Struct. Eng., Berlin-Munich, 1936.
[33] P. Joosting, Experience obtained with structures executed in Holland, Second Congr. Int. Assoc. Bridg. Struct. Eng., Berlin-Munich, 1936.
[34] M.E.J. Nilsson, Experience obtained with structures executed in Sweden, Second Congr. Int. Assoc. Bridg. Struct. Eng., Berlin-Munich, 1936.
[35] B. Espion, The Vierendeel bridges over the Albert Canal, Belgium - their significance in the story of brittle failures, Steel Constr 5 (2012) 238–243, https://doi.
org/10.1002/stco.201210029.
[36] P. Algyay-Hubert, Experience obtained with structures executed in Hungary, Second Congr. Int. Assoc. Bridg. Struct. Eng., Berlin-Munich, 1936.
[37] Q. Collette, Riveted Connections in Historical Metal Structures (1840–1940), Unpubl Dr thesis Brussels Vrije Univ Brussel (2014).
[38] E. Pelke, K.E. Kurrer, des Zur Entwicklungsgeschichte, Stahlverbundbaus (in German), Stahlbau 85 (2016) 764–780, https://doi.org/10.1002/stab.
201610429.
[39] A.R. Mangus, Sun S. “Orthotropic Deck Bridges”, Chapter 14, Bridge Engineering Handbook, 1ed ed., CRC Press, Boca Raton, 2000.
[40] R. Wolchuk, Design Manual for Orthotropic Steel Plate Deck Bridges, (1963).
[41] J.W. Fisher, Fatigue and Fracture in Steel Bridges. Case Studies, (1984).
[42] B. Åkesson, Fatigue Life of Riveted Steel Bridges, CRC Press, 2010.
174
[43] AREA, Use of high strength structural bolts in steel railway bridges, Proc. Am. Railw. Eng. Assoc. 1950, pp. 506–540.
[44] D.T. Wright, E.M. Lewis, Fundamental concepts, and the development of specifications for high-tensile bolted joints, IABSE Congr. Report1, 1960.
[45] J.W. Fisher, L.S. Beedle, Bibliography on Bolted and Riveted Structural Joints, (1964).
[46] K. Sakamoto, C. Miki, A. Ichikawa, M. Abe, Vibration fatigue of steel bridges of the bullet train system, IABSE Work. - Remain. Fatigue Life Steel Struct. vol. 59,
1990, pp. 157–166 Lausanne, Switzerland.
[47] W. Hoorpah, Steel bridges for high speed railways – design regarding fatigue and durability, Bridg. Maintenance, Safety, Manag. Heal. Monit. Informatics -
IABMAS ‘08 Proc. Fourth Int. IABMAS Conf. Ed. Koh Frangopol. Taylor & Francis Group, 2008, pp. 646–654.
[48] E. Pelke, K.-E. Kurrer, On the Evolution of Steel-Concrete Composite Construction, (2016).
[49] B. Åkesson, Understanding Bridge Collapses, Taylor & Francis Group, 2008.
[50] C. Frémont, Étude expérimentale du rivetage (in French), Siège de la Société (1906).
[51] Q. Collette, I. Wouters, C. de Favereau, Peters A. Development of riveting technology through an analysis of Belgian patents (1830-1940), in: J. Jasienko (Ed.),
Proc. Int. Conf. Struct. Anal. Hist. Constr. (SAHC 2012), vol. 2, 2012, pp. 1071–1079 Wroclaw.
[52] J.A. Cavaco, Managing fatigue susceptible details on critical railway bridges at CN, Bridg. Maintenance, Safety, Manag. Heal. Monit. Informatics - IABMAS ‘08
Proc. Fourth Int. IABMAS Conf. 2008, pp. 1262–1269.
[53] AREA, Proceedings of the Annual Conference of American Railway Engineering, Vol. 41 (1940), pp. 1–1045.
[54] M.E. Shank, A Critical Survey of Brittle Failure in Carbon Plate Steel Structures Other than Ships, National Research Council's Committe on Ship Structural
Design, Washington, D.C., 1953.
[55] C. Miki, Fatigue and fracture issues of welds in civil structures, Sci. Technol. Weld. Join. 5 (2000) 347–355, https://doi.org/10.1179/136217100101538416.
[56] R. Haghani, M. Al-Emrani, M. Heshmati, Fatigue-prone details in steel bridges, Buildings 2 (2012) 456–476, https://doi.org/10.3390/buildings2040456.
[57] K. Yokoyama, C. Miki, Participatory database of repair cases on fatigue damaged welded structures, Int. J. Fatigue 101 (Part 2) (2017) 385–396, https://doi.
org/10.1016/j.ijfatigue.2017.01.010.
[58] N. Stranghöner, Der Aufschweißbiegeversuch oder: Nichts ist beständiger als ein Provisorium? (in German), Stahlbau 78 (2009) 815–821, https://doi.org/10.
1002/stab. 200910097.
[59] G. Schaper, Der hochwertige Baustahl St 52 im Bauwsen (in German), Die Bautechnik 16 (1938) 649–655.
[60] O. Kommerell, Die neuen Lieferbedingungen für St 52 als Folge neuerer Versuche und Erfahrungen (in German), Stahlbau 11 (1938) 49–54.
[61] K. Kudla, U. Kuhlmann, Fatigue strength of lamellae joints in steel and composite bridges, 37th IABSE Symp, vol. 102, 2014, pp. 668–675, , https://doi.org/10.
2749/222137814814066780 Madrid.
[62] J. Lee, T. Yoon, S. Chang, The Fatigue performance of plate girder in Korea high-speed railway bridge, IABSE Symp. Rep. vol. 87, International Association for
Bridge and Structural Engineering, Antwerp, Belgium, 2003, pp. 129–135, , https://doi.org/10.2749/222137803796328944.
[63] W. Schütz, A history of fatigue, Eng. Fract. Mech. 54 (1996) 263–300.
[64] O. Kommerell, The influence of frequently alternating loading on welded structures, Second Congr. Int. Assoc. Bridg. Struct. Eng., Berlin-Munich, 1936.
[65] O. Graf, Influence of the form of welded connections to strength and resitance, Second Congr. Int. Assoc. Bridg. Struct. Eng., Berlin-Munich, 1936.
[66] O. Graf, Fatigue strength of riveted connections, Second Congr. Int. Assoc. Bridg. Struct. Eng., Berlin-Munich, 1936, https://doi.org/10.1061/(ASCE)0733-
9445(1998)124:7(792).
[67] M. Ros, Fatigue strength and safety of welded structures (bridges, structural steel work and pressure pipes), Second Congr. Int. Assoc. Bridg. Struct. Eng., Berlin-
Munich, 1936, https://doi.org/10.5169/seals-3168.
[68] W.M. Wilson, F.P. Thomas, Fatigue Tests of Riveted Joints, Univ Illinois Bull, 1938, pp. 1–116.
[69] J.W. Fisher, Effect of Weldments on the Fatigue Strength of Steel Beams, TRB, 1970.
[70] R.J. Dexter, J.W. Fisher, Fatigue cracking of orthotropic steel decks, IABSE REPORTS, 1997, pp. 203–212.
[71] J.W. Fisher, S. Roy, Fatigue damage in steel bridges and extending their life, Adv. Steel Constr. 11 (2015) 250–268.
[72] M. Seki, Large-scale renovations of the civil engineering structures along the Tokaido Shinkansen, Japan Railw Transp Rev (2013) 16–31.
[73] Y. Niwa, K. Matsumoto, S. Yajima, Y. Kobayashi, Reinforcement for fatigue damage of welded joints of sole plate in steel railway composite girder, in Japanese,
J. Struct. Eng. 58 (2012) 611–621.
[74] Y. Matsuda, Maintenance of railway structures with aging deterioration and amarube, Bridge 1999 (2013) 32–39.
[75] C. Miki, T. Konishi, Retrofit engineering for steel bridge structures in Japan, IABSE Symp. Rep. vol. 93, International Association for Bridge and Structural
Engineering, 2007, pp. 49–79.
[76] C. Wang, L. Duan, M. Zhai, Y. Zhang, S. Wang, Steel bridge long-term performance research technology framework and research progress, Adv. Struct. Eng. 20
(2017) 51–68, https://doi.org/10.1177/1369433216646005.
[77] A. Andersson, J. Leander, R. Karoumi, Extending the fatigue service life of a railway bridge by local approaches, IABSE Symp. Rep. vol. 99, International
Association for Bridge and Structural Engineering, 2013, pp. 1369–1376.
[78] C. Wang, B.T. Yen, H.T. Li, L. Duan, Fatigue life evaluation of in-service steel bridges by using bi-linear S-N curves, Adv. Steel Constr. 11 (2015) 269–282,
https://doi.org/10.18057/IJASC.2015.11.3.2.
[79] H.H. Lee, J.C. Jeon, K.S. Kyung, Determination of a reasonable impact factor for fatigue investigation of simple steel plate girder railway bridges, Eng. Struct. 36
(2012) 316–324, https://doi.org/10.1016/j.engstruct.2011.12.021.
[80] C. Baptista, Multiaxial and variable amplitude fatigue in steel bridges, Ecole Polytechnique Fédérale de Lausanne, 2016, https://doi.org/10.5075/EPFL-
THESIS-7044.
[81] S.M. Dick, R.J. Connor, D. Otter, J. Lloyd, Cracks in welded steel girders on the steel bridge at FAST, AREMA Annu. Conf. 2013, pp. 711–728.
[82] R.J. Connor, J.B. Lloyd, Maintenance Actions to Address Fatigue Cracking in Steel Bridge Structures, (2017).
[83] CARS. Reflections on Fatigue Cracks in Steel Bridges in China (in Chinese), (2009).
[84] C. Baptista, S. Kannuna, J.O. Pedro, A. Nussbaumer, Fatigue behaviour of CHS tubular bracings in steel bridges, Int. J. Fatigue 96 (2017) 127–141.
[85] T. Guo, Z. Liu, J. Liu, D. Han, Diagnosis and mitigation of fatigue damage in longitudinal diaphragms of cable-stayed bridges, J. Bridg. Eng. 21 (2016) 5016007,
, https://doi.org/10.1061/(ASCE)BE.1943-5592.0000946.
[86] A. Reis, C. Baptista, Rehabilitation of the suspension bridge over Zambezi River in Mozambique, Struct. Eng. Int. (1) (2013) 89–93.
[87] N. Gimsing, East Bridge, Storebælt Technical Publications, Copenhagen, 1998.
[88] EN 1993-1-9. Eurocode 3, Design of Steel Structures - Part 1–9: Fatigue, (2005).
[89] IIW, Recommendations for fatigue design of welded joints and components. IIW docume, International Institute of Welding (IIW), 2016.
[90] BS 7608, Guide to Fatigue Design and Assessment of Steel Products, (2014).
[91] D.N.V. DNVGL-RP-C203, Fatigue Design of Offshore Steel Structures, Recommended Practice (2016), p. 176.
[92] S.V. Petinov, H.S. Reemsnyder, A.K. Thayamballi, The similitude of fatigue damage principle: application in S-N curves-based fatigue design, Eur Struct Integr
Soc 23 (1999) 219–228, https://doi.org/10.1016/S1566-1369(99)80044-2.
[93] AASHTO, AASHTO LRFD Bridge Design Specifications. 7th Ed, American Association of State Highway and Transportation Officials, Washington, DC, 2014.
[94] A. Pipinato, Extending the lifetime of steel truss bridges by cost-efficient strengthening interventions, Struct. Infrastruct. Eng. (2018) 1–17.
[95] C. Cremona, B. Eichler, B. Johansson, T. Larsson, Improved assessment methods for static and fatigue resistance of old metallic railway bridges, J. Bridg. Eng.
18 (2013) 1164–1173, https://doi.org/10.1061/(ASCE)BE.1943-5592.0000466.
[96] V. Grigoriou, A. Nussbaumer, D.G. Lignos, Fatigue strength upgrading of cover plate ends by welded extensions in existing steel bridge girders, J. Bridg. Eng. 23
(2018) 4018037.
[97] J.W. Fisher, B.M. Barthelemy, D.R. Mertz, J.A. Edinger, Fatigue Behavior of Full-scale Welded Bridge Attachments, Final Rep Lehigh Univ, Bethlehem, PA,
1980.
[98] R.J. Dexter, J.W. Fisher, S. Roy, Fatigue and fracture, Chapter 19, Bridge Engineering Handbook, 2nd ed, CRC Press, Taylor & Francis Group, LLC, Boca Raton,
2014.
[99] Supplement to CHBDC S6-14, Bridge Standards and Procedures Manual, Can Highw Bridg Des Code (2016), pp. 1–216.
[100] J.W. Fisher, Evolution of fatigue-resistant steel bridges, Transp. Res. Rec. 18 (1997), https://doi.org/10.3141/1594-01.
[101] C. Miki, M. Tai, Fatigue strength improvement of out-of-plane welded joints of steel girder under variable amplitude loading, Weld World 57 (2013) 823–840.
175
[102] K. Nishikawa, J. Murakoshi, T. Matsuki, Study on the fatigue of steel highway bridges in Japan, Constr. Build. Mater. 12 (1998) 133–141, https://doi.org/10.
1016/S0950-0618(97)00015-9.
[103] R.J. Connor, J.W. Fisher, Identifying effective and ineffective retrofits for distortion Fatigue cracking in steel Bridges using field instrumentation, J. Bridg. Eng.
11 (2006) 745–752.
[104] M. Bowman, G. Fu, Y. Zhou, R. Connor, A. Godbole, NCHRP: Fatigue Evaluation of Steel Bridges (Report 721), (2012), https://doi.org/10.17226/22774.
[105] AASHTO MBE, The Manual for Bridge Evaluatin, Second edition, (2011).
[106] J.W. Fisher, T.A. Fisher, C.N. Kostem, Displacement induced fatigue cracks, Eng. Struct. 1 (1979) 252–257.
[107] B. Åkesson, B. Edlund, D. Shen, Fatigue cracking in a steel railway bridge, Struct Eng Int J Int Assoc Bridg Struct Eng 7 (1997) 118–120.
[108] G.Y. Grondin, G.L. Kulak, Distortion-induced fatigue cracking of bridge girders – design issues, Struct. Congr. Structural Engineering Institute of the ASCE,
Orlando, Florida, 2010, pp. 484–495, , https://doi.org/10.1061/41130(369)45.
[109] M. Aygül, M. Al-Emrani, Z. Barsoum, J. Leander, Investigation of distortion-induced fatigue cracked welded details using 3D crack propagation analysis, Int. J.
Fatigue 64 (2014) 54–66, https://doi.org/10.1016/j.ijfatigue.2014.02.014.
[110] M. Aygül, M. Al-Emrani, Z. Barsoum, J. Leander, An investigation of distortion-induced fatigue cracking under variable amplitude loading using 3D crack
propagation analysis, Eng. Fail. Anal. 45 (2014) 151–163, https://doi.org/10.1016/j.engfailanal.2014.05.015.
[111] J.W. Fisher, P.B. Keating, Distortion-induced fatigue cracking of bridge details with web gaps, J. Constr. Steel Res. 12 (1989) 215–228, https://doi.org/10.
1016/0143-974X(89)90056-4.
[112] D. Jajich, A.E. Schultz, Measurement and analysis of distortion-induced fatigue in multigirder steel bridges, J. Bridg. Eng. 8 (2003) 84–91.
[113] K.S. Kyung, J.E. Park, S.S. Jun, J.G. Kim, A study of the evaluation of fatigue crack at welded joint for steel plate girder, Procedia Eng 14 (2011) 1543–1548,
https://doi.org/10.1016/j.proeng.2011.07.194.
[114] S.J. Lee, H.H. Lee, The analysis on the characteristic of fatigue crack in railway plate girder bridge and its retrofit method, Bridg. Maintenance, Safety, Manag.
Resil. Sustain. Proc. Sixth Int. IABMAS Conf. Ed. Fabio Biondini Dan Frangopol., Stresa, Lake Maggiore, CRC Press, Italy, 2012, pp. 1302–1307.
[115] F. Lippi, M. Orlando, W. Salvatore, Assessment of the dynamic and fatigue behaviour of the Panaro railway steel bridge, Struct. Infrastruct. Eng. 9 (2013)
834–848, https://doi.org/10.1080/15732479.2011.625955.
[116] L.R.T. Melo, R.M. Teixeira, A.P. da Conceição Neto, T.N. Bittencourt, Evaluation of fatigue crack formation in cantilever brackets of a multi-span railway steel
box girder bridge, Multi-Span Large Bridg - Proc Int Conf Multi-Span Large Bridg, 2015 2015.
[117] C. Chao-hui, H. Qi, Reason analysis and countermeasures of the cracks at the juncture of longitudinal beam and horizontal beam of a railway steel truss girder
(in Chinese), Railw Stand Des 58 (2014) 42–46, https://doi.org/10.13238/j.issn.1004-2954.2014.02.011.
[118] Takagi et al., Countermeasure for fatigue cracks at connections between bracket and main girder, in Japanese, Prc. Annu. Conf. JSCE, Japan Society of Civil
Engineers, 2009.
[119] T.N. Bittencourt, L.R. Ticona Melo, R.M. Teixeira, A.P. Conceição Neto, M. Choquepuma, T. Yanagihara, et al., Monitoramento da ponte metálica sobre o
córrego do Ouro, in Portuguese (2013).
[120] M. Suloway, L. Koncza, J. Hanson, J. Fisher, Final Report on Causes of Fractures in Bent Nos. 24, 25 and 26 of Dan Ryan Rapid Transit Structure, (1979).
[121] J.W. Fisher, H. Hausammann, M.D. Sullivan, A.W. Pense, NCHRP Report 206: Detection and Repair of Fatigue Damage in Welded Highway Bridges,
Washington, D.C. (1979).
[122] J.W. Fisher, J. Jin, D.C. Wagner, B.T. Yen, NCHRP Report 336: Distortion-Induced Fatigue Cracking in Steel Bridges, Washington, D.C (1990).
[123] K. Ghahremani, A. Sadhu, S. Walbridge, S. Narasimhan, Fatigue testing and structural health monitoring of retrofitted web stiffeners on steel highway bridges,
Transp Res Rec J Transp Res Board, 2013, pp. 27–35.
[124] G. Alencar, A.M.P. de Jesus, R.A.B. Calçada, J.G.S. da Silva, Fatigue life evaluation of a composite steel-concrete roadway bridge through the hot-spot stress
method considering progressive pavement deterioration, Eng. Struct. 166 (2018) 46–61.
[125] R. Okelo, Fatigue life prediction for distortion-induced cracking of steel bridges, Int J Steel Struct 17 (2017) 801–820, https://doi.org/10.1007/s13296-017-
6033-5.
[126] C. Hendy, D. Iles, Steel Bridge Group: Guidance Notes on Best Practice in Steel Bridge Construction (5th Issue).(P185), (2010).
[127] E. Pantoli, L. Vincenzi, M. Savoia, R. Testa, The effect of local vibrations on fatigue in old steel riveted bridges. A case study: the Manhattan Bridge, in: G. De
Roeck, G. Degrande, G. Lombaert, G. Muller (Eds.), Int. Conf. Struct. Dyn. EURODYN 2011, vol. 1, 2011, pp. 1–8.
[128] C. Klinger, T. Michael, D. Bettge, Fatigue cracks in railway bridge hangers due to wind induced vibrations–failure analysis, measures and remaining service life
estimation, Eng. Fail. Anal. 43 (2014) 232–252.
[129] D. Ribeiro, R. Calçada, R. Delgado, M. Brehm, V. Zabel, Finite element model updating of a bowstring-arch railway bridge based on experimental modal
parameters, Eng. Struct. 40 (2012) 413–435.
[130] A. Andersson, A. O'Connor, R. Karoumi, Passive and adaptive damping systems for vibration mitigation and increased fatigue service life of a tied arch railway
bridge, Comput Civ Infrastruct Eng 30 (2015) 748–757, https://doi.org/10.1111/mice.12116.
[131] Sustainable Bridges. D6.2 Repair and Strengthening of Railway Bridges, (2007).
[132] Sustainable Bridges. D4.3 Loads and Dynamic Effects, (2007).
[133] M. Iványi, Fatigue damage in steel retrofitting works: Case studies from Japan, Refurbishment of Buildings and Bridges, Chapter 2, “Refurbishment of Steel
Bridges”, CISM International Centre for Mechanical Sciences, 2002.
[134] I. Sugimoto, A. Ichikawa, A study to extend the Fatigue life of welded railway Bridges, Q Rep RTRI 40 (1999) 74–79.
[135] D. Su, Y. Fujino, T. Nagayama, T. Miyashita, Local train-induced vibration in high-speed train steel bridge, IABSE Symp. Rep. vol. 96, International Association
for Bridge and Structural Engineering, 2009, pp. 44–53.
[136] T. Miyashita, H. Ishii, Y. Fujino, A. Shoji, M. Seki, Clarification of the effect of high-speed train induced vibrations on a railway steel box girder bridge using
laser Doppler vibrometer, Proc. Int. Conf. Exp. Vib. Anal. Civ. Eng. Struct. Fr. 2005, pp. 349–357.
[137] Z. Qinghua, B. Yizhi, L. Qiao, Review on Fatigue problems of orthotropic steel bridge deck, China J Highw Transp 30 (2017) 14–39.
[138] B. Cheng, X. Ye, X. Cao, D.D. Mbako, Y. Cao, Experimental study on fatigue failure of rib-to-deck welded connections in orthotropic steel bridge decks, Int. J.
Fatigue 103 (2017) 157–167.
[139] S. Qin, Z. Gao, Developments and prospects of long-span high-speed railway bridge technologies in China, Engineering 3 (2017) 787–794, https://doi.org/10.
1016/j.eng.2017.11.001.
[140] Y.S. Song, Y.L. Ding, G.X. Wang, A.Q. Li, Fatigue-life evaluation of a high-speed railway bridge with an orthotropic steel deck integrating multiple factors, J.
Perform. Constr. Facil. 30 (2016) 1–10, https://doi.org/10.1061/(ASCE)CF.1943-5509.0000887.
[141] B. Cao, Y. Ding, H. Zhao, Y. Song, Damage identification for high-speed railway truss arch bridge using fuzzy clustering analysis, Struct Monit Maint 3 (2016)
315–333.
[142] H. Xia, Y. Deng, Y. Zou, G. De Roeck, G. Degrande, Dynamic analysis of rail transit elevated bridge with ladder track, Front Archit Civ Eng China 3 (2009) 2–8.
[143] G. Chellini, L. Nardini, W. Salvatore, DETAILS: a research project for improvement of analysis, design and durability of HS railway bridges, Bridg. Maintenance,
Safety, Manag. Heal. Monit. Informatics - IABMAS ‘08 Proc. Fourth Int. IABMAS Conf. Ed. Hyun-Moo Koh Dan Frangopol. Taylor & Francis Group, 2008, pp.
682–690.
[144] J. Hseih, F.-H. Wu, High-speed railway bridges, Chapter 8, Bridge Engineering Handbook, 2nd ed, CRC Press, Taylor & Francis Group, LLC, Boca Raton, 2014.
[145] ERRI D, Rail bridges for speeds > 200 km/h, Final Report, 1999.
[146] H. Figueiredo, R. Calçada, R. Delgado, Fatigue assessment of composite bridges for high speed railway traffic, in: Hyun-Moo Koh, Dan Frangopol (Eds.), Bridg.
Maintenance, Safety, Manag. Heal. Monit. Informatics - IABMAS ‘08 Proc. Fourth Int. IABMAS Conf. Taylor & Francis Group, Seoul, Korea, 2008, pp. 708–715.
176

J Engfailanal 2019 05 037

Uploaded by

Copyright:

Available Formats

J Engfailanal 2019 05 037

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

J Engfailanal 2019 05 037

Uploaded by

Copyright:

Available Formats

Engineering Failure Analysis 104 (2019) 154–176

Contents lists available at ScienceDirect

Engineering Failure Analysis

Fatigue cracking of welded railway bridges: A review

ARTICLE INFO ABSTRACT

2. A historical overview of welded railway bridges construction

iv) Welding is less time consuming than riveting.

Nevertheless, welded joints also show some drawbacks such as:

Earliest welded a) Westinghouse Truss 41 m Freight Through Truss Out-of- X

b) Westinghouse Girder 16.8 m Freight Plate girder Out-of-

c) ELiN A.G. Bridge 9.7 m Freight Plate girder Out-of-

d) Reichsbahn Trial Bridge 10 m Mixed Half-through plate Out-of-

e) Finnish State Railways 12 m Mixed Plate girder Out-of-

Commuters bridges f) Tratinska Street Bridge 18 m Commuter Curved Framed In use

g) Zdroje Szczecin Bridge 13.5 m Commuter Curved frame In use X

i) Bridge at Saint-Denis 35 m Mixed Half-through Out-of-

j) Miryang River Bridge 13.5 m Mixed Plate girder In use X

k) Söderström Bridge 33.7 m Mixed Half-through In use X

OSD l) Bridge over 36 m Mixed Ortho- tropic Steel Deck Out-of-

Weld & High m) Burdekin Rail-Road 76 m Freight Bolt-rivet-welded In use

n) Langjiang Bridge 62 m Mixed Bolt-welded through In use X

p) Shinkansen Through 60 m High-speed Through truss In use X

q) TGV Northern HSR 40 m High-speed Composite deck In use

- Steels with improved

Welding reinforcement of Resumption of welding of old metallic

First brittle failures

Fig. 1. Penetration of welding in railway bridge construction.

3. Fatigue cracking of welded railway bridges

Low fatigue strength category in terms of nominal stresses

Fatigue cracks originating from weld defects

Secondary and out-of-plane stresses

r) Zoo Train Station Viaduct 50 m Commuter Plate girder Out-of- X

s) Welded bridge at FAST 19.8 m Freight Plate girder Out-of- X

u) Potomac River Metrorail Bridge 36 m Commuter Box girder bridge In use X

3.1. Fatigue induced by weld crack-like defects

3.1.1. Examples of fatigue cracking induced by welding defects

(a) (b) Cracked

3.1.2. Designing fatigue resistant welded connections to crack-like welding defects

3.2. Load-induced fatigue

3.2.1. Examples of load-induced fatigue cracking

3.2.2. Designing to load-induced fatigue

3.3. Distortion-induced fatigue

3.3.1. Examples of distortion-induced fatigue cracking

3.3.2. Designing to distortion-induced-fatigue

Lack of fusion of the bottom-

Out-of-plane bending of the

(a) hangar (b)

3.4. Vibration-induced fatigue

3.4.1. Examples of vibration-induced fatigue cracking

3.4.2. Designing to vibration-induced fatigue

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.