Journal of Constructional Steel Research 65 (2009) 269–277

Contents lists available at ScienceDirect

Journal of Constructional Steel Research

journal homepage: www.elsevier.com/locate/jcsr

Hydrogen embrittlement and corrosion fatigue of corroded bridge wires

Shun-ichi Nakamura a,∗ , Keita Suzumura b
a Department of Civil Engineering, Tokai University, 1117 Kitakaname, Hiratsuka, 259-1292, Japan
b Engineering Support Center, Nippon Steel Engineering Ltd., 5-9-1 Nishihashimoto, Sagamihara, Japan

article info a b s t r a c t

Article history: Cables of old suspension bridges and stays of cable-stayed bridges often suffer from steel corrosion.
Received 11 October 2007 Corroded galvanized steel wires on different corrosion levels were produced at laboratories, and their
Accepted 25 March 2008 mechanical properties and remaining strength were investigated. Actual tensile strength of corroded
wires did not decrease with corrosion levels, whereas elongation decreased sharply after the zinc
Keywords: layer was partly depleted and the steel started to corrode. As the accumulated amounts of diffusive
Galvanized steel wires
hydrogen of corrosion level-2 wires – with and without added tension – were almost the same
Corrosion fatigue
Cables
and less than 0.2 ppm, the applied tension of steel wires did not affect the amount of diffusive
Suspension bridges hydrogen which was well below the critical concentration of 0.7 ppm to cause brittleness. This
Cable-stayed bridges indicates that hydrogen embrittlement is unlikely to occur. Fatigue tests showed that fatigue strength
did not change when only the galvanized layer was corroded, but it significantly decreased after
corrosion of the steel below the galvanized layer progressed. Fatigue strength further lowered when the
steel wire was cyclically stressed under wet environments when compared with the fatigue strength
under dry environments. The broken wires of an old suspension bridge were also investigated. The
fracture surface was similar to that caused by corrosion fatigue rather than hydrogen embrittlement.
It was estimated that the wires were fractured by the mixed effects of corrosion, cyclic stresses and
hydrogen.
© 2008 Elsevier Ltd. All rights reserved.

1. Introduction progress. Some researchers believe that hydrogen is accumulated

while steel wires corrode and brittles and breaks wires [1,3]. Others
Bridge cables usually consist of high tensile galvanized steel believe that corrosion fatigue causes wire breakage, in particular,
wires. Old suspension bridges often suffer from deteriorated for hanger ropes that have high stress fluctuation [3,4].
cables. Some of steel wires of the main-cables are heavily corroded In this paper fracture mechanism and remaining properties of
and fractured. Wide rehabilitation works of corroded wires were corroded bridge wires are studied from three points of views. First,
carried out on the Brooklyn Bridge, the Williamsburg Bridge and effect of hydrogen is studied. Amount of accumulated hydrogen is
other bridges [1–3]. Hanger ropes of suspension bridges and stays
measured for steel wires on different levels of corrosion. Then, it is
of cable-stayed bridges have also been severely corroded on many
compared with the critical hydrogen amount to cause brittleness.
bridges [3].
Emphasis is placed to clarify the difference between stressed
Long-span suspension bridges are relatively new in Japan.
corroded wires and un-stressed wires.
However, it was found by field surveys that galvanized steel wires
were partly corroded on several suspension bridges within ten Second, corrosion fatigue is investigated. The fatigue strength
years after the bridges were completed [4]. Wide surveys were of corroded wires under wet environments can be lower than
carried out on these bridges. Water was present inside the cable that in dry environments because of corrosion fatigue. Fatigue
and made the environment inside the cable highly humid, which tests are therefore conducted in both dry and wet environments
caused the corrosion of wires. However, the steel corrosion was to clarify the corrosion fatigue strength of galvanized steel wires.
not deep and no broken wire was observed. Scanning electron microscopy of the fracture surface is shown and
Although many studies have been conducted so far to clarify analyzed.
the mechanism to cause wire breakage, discussions are still in Third, considering the above studies on hydrogen embrittle-
ment and corrosion fatigue, the fracture mechanism of corroded
bridge wires is estimated. Investigation on the broken wire cut
∗ Corresponding author. Tel.: +81 463 58 1211; fax: +81 463 50 2045. from an old suspension bridge is referred in the estimated fracture
E-mail address: [email protected] (S. Nakamura). mechanism.

0143-974X/$ – see front matter © 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jcsr.2008.03.022
270 S. Nakamura, K. Suzumura / Journal of Constructional Steel Research 65 (2009) 269–277

Fig. 2. Tensile strength of corroded wires.

Fig. 1. Specimens of corroded wires (wire diameter 5 mm).

2. Mechanical properties of corroded wires and hydrogen

Fig. 3. Actual tensile strength of corroded wires.
embrittlement

2.1. Corroded specimens of about 100 g/m2 corresponds to the level-1 corrosion (zinc
corrosion), that of about 300 g/m2 to the level-2 corrosion (steel
Nakamura, Suzumura and Tarui investigated the remaining corrosion), and that of about 600 g/m2 to the level-3 corrosion
strength of corroded steel wires [5,6]. Some results are briefly (severe steel corrosion).
reviewed in Sections 2.1 and 2.2.
Suspension bridge cables are under severe corrosion environ- 2.2. Tensile strength and elongation
ments; wires are wet at high temperature. To simulate this envi-
ronment, galvanized steel wires are wrapped with wet gauze and Tension tests were performed with the corroded specimens [5].
kept in an enclosed box at a temperature of 40 ◦ C. Water and oxy- Tensile strength of the specimens, the breaking force divided
gen are needed to corrode steel, and the wet gauze supplies both. by the original cross sectional area, is shown in Fig. 2 with the
It was proved by the past study that this simulation test with wet mass loss that corresponds to the corrosion level. The tensile
gauze produced the steel wires corroded in an actual corrosion en- strength decreases with the mass loss due to corrosion both in
vironment of suspension bridge cables. The wire specimen is 5 mm the galvanized and bare steel specimens. Fig. 3 shows the actual
in diameter, its tensile strength is 1570 MPa and the attached zinc tensile strength, the breaking force divided by the reduced cross
mass is 350 g/m2 , which is equivalent to 50 µm in thickness. sectional area due to corrosion. The actual tensile strength does not
The specimens were taken out from the enclosed box after decrease with the corrosion level both for the galvanized and bare
90 days, 250 days and 360 days, producing the corroded wires steel specimens.
on different corrosion levels. Their appearances are shown in
On the other hand, elongation of the galvanized steel wire
Fig. 1. The galvanized steel wires after 90 days were covered with
specimens decreases sharply after the corrosion level-1 (the
white zinc corrosion substance (corrosion level-1), ferrous rust
reduced mass of about 100 g/m2 ) and that of the bare steel
occurred locally in the specimens after 250 days (corrosion level-
decreases linearly with the reduced mass due to corrosion, as
2), and ferrous rust spread widely in the specimens after 350 days
shown in Fig. 4. Elongation does not change when only the
(corrosion level-3).
galvanized layer is corroded (level-1 corrosion) but it decreases
The same simulation tests were carried out for bare steel wires.
when the steel part starts to corrode. Elongation corresponds to
The galvanized layer was removed by immersing galvanized steel
ductility. When the ductility of galvanized steel wires lowers,
wires in H2 SO4 to make the bare steel wire specimens. The bare
safety factor of the bridge itself is also jeopardized and should be
steel specimen is expected to clarify the effect of the galvanized
considered carefully. It is noted that each point in Figs. 2–4 is an
layer. The bare steel specimens after 180 days were covered with
average of several test data.
ferrous rust, and after 350 days were covered with severe ferrous
rust.
The corrosion substances on the specimen surface were 2.3. Hydrogen embrittlement of wires corroded under tension
removed by a cloth containing 10% H2 SO4 , and the mass loss
was obtained by subtracting the final mass from the initial mass. It is important to measure an amount of absorbed hydrogen
The mass loss due to corrosion increased exponentially with because occurrence of hydrogen embrittlement depends on the
the elapsed days both for the galvanized specimens and the amount of hydrogen. The absorbed hydrogen emits from the steel
bare specimens. For the galvanized specimens, the reduced mass wire when it is heated. The specimen is therefore placed in a tube.
 S. Nakamura, K. Suzumura / Journal of Constructional Steel Research 65 (2009) 269–277 271

Fig. 4. Elongation of corroded wires.

Fig. 7. Emission rate of hydrogen with heated temperature.

Fig. 8. Volume of emitted hydrogen versus corrosion levels.

Fig. 5. Typical curve of hydrogen emission rate versus heated temperature.

wires [7]. It is known that only diffusive hydrogen, the hydrogen

near the first peak about 100–150 ◦ C, affects the brittleness [7].
It is understood from Fig. 6 that hydrogen embrittlement does
not occur when the accumulated amount of absorbed diffusive
hydrogen is less than 0.7 ppm for the test wires whose tensile
strength is 1570 MPa.
Fig. 7 shows hydrogen evolution curves of the galvanized
steel wires on four different corrosion levels. In the hydrogen
evolution curves of the new wires and the corrosion levels 1 and
2 specimens, the peak appears at a temperature near 150 ◦ C. It is
understood from these figures that hydrogen is absorbed during
the manufacturing process and not absorbed during the corrosion
process of levels 1 and 2. However, the absorbed hydrogen of
the corrosion level 3 specimen is different from others. Hydrogen
is absorbed much more and the peak moved to the higher
temperature of about 300 ◦ C.
Fig. 6. Critical diffusive hydrogen to cause embrittlement. The accumulated amount of diffusive hydrogen between 0 and
200 ◦ C is shown in Fig. 8. It is about 0.15 ppm and does not depend
When it is gradually heated with a rate of 100 ◦ C/h and hydrogen on the corrosion level. This accumulated amount is well below 0.7
is released. At the same time, argon gas is continuously flown into ppm, indicating that hydrogen embrittlement is unlikely to occur.
the tube and sends the emitted hydrogen into a chromatography The accumulated amount of non-diffusive hydrogen, the hydrogen
where the hydrogen quantity is measured. Before the corroded emitted between 200 and 400 ◦ C, sharply increases on corrosion
specimens were measured in this process, they were kept in a level 3, as shown in Fig. 8. It is known that non-diffusive hydrogen
fridge box with dry ices so that hydrogen did not emit. does not affect hydrogen embrittlement, although a further study
Fig. 5 shows a typical hydrogen evolution curve, relations be- is required to clarify how non-diffusive hydrogen affects fatigue
tween the amount of emitted hydrogen and a heated tempera- strength.
ture. There are two peaks in the curve. The first peak is at around The aforementioned corrosion simulation was carried out,
100–150 ◦ C and the second peak is at around 250–300 ◦ C. The first keeping wire specimens un-stressed. However, it is important to
peak associates with diffusive hydrogen and the second peak asso- find the effects of the hydrogen absorption for the wire specimens
ciates with non-diffusive hydrogen which is trapped into the lattice while they are stressed. Fig. 9 shows the method to make wire
structure strongly and cannot move. specimens under tension. The wire was tensioned to 500 MPa,
Fig. 6 shows a critical diffusive hydrogen concentration that which was checked every day by load-cells. Wet gauze covered the
causes brittleness with different strength of galvanized steel specimen that was placed in a plastic box. Temperature in a box
272 S. Nakamura, K. Suzumura / Journal of Constructional Steel Research 65 (2009) 269–277

Fig. 12. Fatigue strength of corroded wires (dry condition).

tension and that without tension are almost the same and less than
0.2 ppm. Summarizing Figs. 10 and 11, introduced tensions of steel
Fig. 9. Method to produce corroded wires with introduced tension. wires do not affect the amount of diffusive hydrogen which is well
below the critical concentration to cause brittleness.

3. Corrosion fatigue of bridge wires

3.1. Corrosion fatigue tests

Fatigue strength of corroded galvanized steel wires may

decrease while corrosion progresses, but this has not been proved
in the past studies. Fatigue tests were, therefore, conducted for the
new wires and corroded wires at corrosion levels-1, 2 and 3 [5]. A
specimen was held with a distance of 100 mm between the grips
and cyclically loaded at 64 Hz. Cyclic tests were conducted keeping
the minimum stress at 500 MPa and varying the stress range from
200 MPa to 600 MPa. The tests were conducted with a relative
humidity under 50%.
Fig. 12 shows the relations of stress amplitudes and numbers of
Fig. 10. Emitted hydrogen rate with and without tension. cycles (the S-N curve) of the new wires and the corrosion levels-1, 2
and 3 specimens. There is not much difference between the fatigue
strength of the new and corrosion level-1 specimens. However,
the fatigue strength of the level-2 specimens is lower than that
of the new and corrosion level-1 wires, and that of the level-3
specimens is further lower than that of the level-2 specimen. It is
therefore clear that the fatigue strength does not change when only
the galvanized layer is corroded, but it significantly decreases after
the steel corrosion below the galvanized layer progresses.
It is known that the fatigue strength of high tensile strength
steel plate further decreases when they are kept in wet condi-
tions [8]. This is called the corrosion fatigue. It is reported that there
is no fatigue limit for the corrosion fatigue of high tensile steel
plates [8]. If there is no fatigue limit, even a small stress fluctu-
ation could initiate a fatigue crack. A further study with more test
data is required to prove if this is true for corroded galvanized steel
Fig. 11. Accumulated diffusive hydrogen with and without tension. wires.
The bridge wires are exposed to wet environments inside the
was kept at 20 ◦ C. It took about two years to make specimens in cable while they are cyclically stressed due to temperature change
level-2 corrosion. and live loads. Fatigue tests were, therefore, conducted in wet
Fig. 10 shows comparison of the hydrogen evolution curves of environments to clarify the corrosion fatigue of galvanized steel
corrosion level-2 wires with and without tension. It is understood wires. The fatigue test method in wet environments is shown
that both curves are almost the same and induced tension does in Figs. 13 and 14. A corroded wire is wrapped with wet gauze,
not affect the amount of emitted hydrogen. Fig. 11 shows the inserted into a silicon tube, sealed with vinyl tapes at the upper
accumulated amount of diffusive hydrogen of corrosion level-2 and bottom ends, and cyclically loaded. A specimen was held at
wires with and without tension. The amount of bare steel wires a distance of 100 mm between the grips and cyclically loaded
is less than that of galvanized steel wires. This is because that at 64 Hz, which was the same as the dry cyclic tests. The tests
hydrogen is absorbed during the manufacturing process of zinc under wet environments were conducted for the levels-2 and
coating. On the other hand the amount of corroded wires with 3 specimens. Figs. 15 and 16 show the S-N relations of the
 S. Nakamura, K. Suzumura / Journal of Constructional Steel Research 65 (2009) 269–277 273

Fig. 15. Fatigue strength of corrosion level-2.

Fig. 13. Fatigue tests under wet environment.

Fig. 14. Fatigue tests under wet environment.

Fig. 16. Fatigue strength of corrosion level-3.

levels-2 and 3 specimens under dry and wet environments. These

comparisons show that the fatigue strength in wet environments
is clearly lower than that under dry environments. initiate from the deepest corroded parts. It seems that stresses
concentrate at these points and cause fatigue failure. In the new
and level-1 corrosion specimens, as the steel corrosion does not
3.2. SEM photos of broken sections
exist and the surface is flat, stresses are unlikely to concentrate
The broken parts of the new and level-1 corrosion specimens
were photographed by an optical microscope and are shown in and fatigue strength is higher than those of the level-2 and level-
Fig. 17. Those of the level-2 corrosion specimens both under dry 3 corrosion specimens. There is no clear difference in appearances
and wet conditions are shown in Fig. 18. Also, those of the level-3 between the corroded specimens under dry and wet conditions.
corrosion specimens both under dry and wet conditions are shown The fractured sections are shot by scanning electron micro-
in Fig. 19. All of these specimens show no ductile phenomenon at scopes (SEM). Those of the new and corrosion level-1 specimens
the broken parts. are shown in Fig. 20, those of the corrosion level-2 specimens
The broken positions of the levels-2 and level-3 corrosion under dry and wet conditions are shown in Fig. 21, and those
specimens agree with those of the corroded parts. Fatigue cracks of the corrosion level-3 specimens under dry and wet conditions

(a) Initial. (b) Corrosion Level-1.

Fig. 17. Appearance of broken parts of new and corrosion level -1 wires (wire diameter 5 mm) (Upper & Lower: before & after corrosion substance is removed).
274 S. Nakamura, K. Suzumura / Journal of Constructional Steel Research 65 (2009) 269–277

(a) Fatigue tests under dry condition. (b) Fatigue tests under wet condition.

Fig. 18. Appearance of broken parts of corrosion level -2 wires (wire diameter 5 mm) (Upper & Lower: before & after corrosion substance is removed).

(a) Fatigue tests under dry condition. (b) Fatigue tests under wet condition.

Fig. 19. Appearance of broken parts of corrosion level -3 wires (wire diameter 5 mm) (Upper & Lower: before & after corrosion substance is removed).

Fig. 20. SEM photo (a: crack initiation point, b: crack propagation area).

Fig. 21. SEM photo of level -2 corrosion (a: crack initiation point, b: crack propagation area).

are shown in Fig. 22. In all of these SEM photos, cracks initiate whole section, the rest of the wire section was suddenly broken.
from the corroded parts, the top part of the wire in these pho- This shows a typical fatigue fracture section.
The width L and depth d of the corroded parts of specimens
tos (point a). Then the cracks propagate downwards, where the are summarized in Table 1. For the level-2 specimens L is about
area is flat (area b). When the crack reached about one third of the 5 mm and d is about 0.1 mm. For the level-3 specimens L is about
 S. Nakamura, K. Suzumura / Journal of Constructional Steel Research 65 (2009) 269–277 275

Fig. 22. SEM photo of corrosion level -3 (a: crack initiation point, b: crack propagation).

Fig. 23. Broken wire of an old suspension bridge main cable (A: crack initiation point, wire diameter 5 mm).

Table 1 the lower part followed. The top part is the inside of the curved
Size of pitting at broken sections specimen, which has a tensile residual stress. After the corrosion
Specimens Width Depth substance was removed from the fractured cross section, it was
L (mm) d (mm) photographed by a scanning electric microscope (SEM). Fig. 24
Corrosion level 2 3 0.12 shows the top part of the specimen. A locally corroded part is
Dry condition 6 0.32 observed at point A in Fig. 24, from where a crack initiates. Then the
3 0.10 crack gradually progresses from that point, and the black parallel
Corrosion level 2 4 0.10 lines can be seen in this photo. These lines resemble beech marks
Wet condition 5 0.20
5 0.17
in appearance, which usually associate with fatigue.
Fig. 25 shows the specimen that was purposely fractured
6 0.27
Corrosion level 3 6 0.33 by hydrogen embrittlement at a laboratory. The outer layer of
Dry condition 7 0.22 the specimen was circumferentially notched about 0.5 mm to
11 0.37 accelerate fracture. Hydrogen was charged into a galvanized steel
5 0.25 wire in the liquid of 3% NaCl and 0.3% NH4 SCN with an electric
Corrosion level 3 6 0.38 current of 0.04 mA/cm2 for 18 h. It was then statically tensioned
Wet condition 10 0.30
with 90% of the tensile strength. The specimen fractured after 4.5 h.
10 0.35
Fig. 25 shows a fractured section of the specimen that is very rough
and clearly different from the actually fractured surface of Fig. 24.
10 mm and d is over 0.3 mm. In general, the width and depth of the Mechanism of hydrogen absorption into the corroded steel is
corroded parts increase with the corrosion level. shown in Fig. 26. When the galvanized layer is consumed and the
steel layer starts to corrode, both anode and cathode reactions
4. Mechanism of broken wires take place. At the anode reaction steel becomes steel ion emitting
electrons. At the cathode reaction hydrogen ion in water receives
A broken wire cut from an old suspension bridge was these electrons and produces hydrogen. Then the hydrogen is
investigated [5]. The wire is about 200 mm long and 5.0 mm in absorbed in the corroded part of steel.
diameter. The wire was curved when it was released from the There are three possible mechanisms for broken bridge wires:
bridge. Wires were wound on a reel when they were manufactured hydrogen embrittlement, corrosion fatigue and stress corrosion
at a factory, and a reel was transported to the construction site. cracking. Stress corrosion cracking (SCC) is defined as the delayed
They were then unwound from the reel and installed between the failure of alloys by cracking when exposed to certain environments
towers and the anchorages by the air spinning method. They were in the presence of static tensile stress. This environment for high
naturally straightened when they were tensioned by dead and live strength steel is usually in sodium nitrate solutions and caustic
loads, producing residual stresses in the wires. NaOH solutions. Therefore, SCC is not considered to be the cause
Fig. 23 shows the fractured part of the broken wires. Ferrous of wire fracture.
rust is visible near the fractured part. It was estimated from Hydrogen embrittlement is defined by harmful effect of
observation that the top part of these photos was first broken and hydrogen penetration of steel. The hydrogen comes from the
276 S. Nakamura, K. Suzumura / Journal of Constructional Steel Research 65 (2009) 269–277

Fig. 24. Broken Section (A: crack initiation point).

Fig. 25. Broken section due to hydrogen embrittlement.

cathode reaction of steel corrosion in normal environments. Some

of the hydrogen diffuses and others are strongly trapped into Fig. 26. Mechanism of hydrogen emission.
the lattice structure and cannot move. It is known that only the
diffusive hydrogen affects the failure and, when the accumulated
amount of diffusive hydrogen is within the critical value, it does wires under wet conditions might be significantly lower than the
not cause wire fracture as mentioned in 2.3. corresponding limit under dry condition. Therefore, even a small
The fractured section is more like fatigue failure than hydrogen stress fluctuation may cause corrosion fatigue induced fracture in
the wire under small number of loading cycles.
embrittlement, as shown in Fig. 24. On the other hand, stress
In summary, fracture process of the broken wire is estimated
fluctuation in suspension bridge wires is usually low and does not
as follows (Fig. 27). First, the zinc layer is depleted and the steel
reach a level to cause fatigue failure. However, fatigue strength
layer starts to corrode, as shown in (b). For the sharper locally
can be very low because the broken wires have high residual corroded shape of wires, stress concentrates more intensely. Then
stresses. This is because the wires are wound around a reel a crack occurs at the locally corroded part and propagates, as
during transportation and are bent. When they are set in the shown (c). In this process corrosion accelerates the crack progress,
final positions on the bridge, wires are straightened and cause cyclic stresses also accelerate the crack progress, and the emitted
residual stresses. In addition, fatigue limit for corrosion fatigue of hydrogen embrittles the steel. Finally, the wire fractures, as

Fig. 27. Mechanism of corrosion fatigue fracture.

 S. Nakamura, K. Suzumura / Journal of Constructional Steel Research 65 (2009) 269–277 277

shown in (d). Therefore, fracture mechanism of bridge wires is estimated that the wires were fractured by the mixed effects of
a mixed phenomenon of corrosion, cyclic stress fluctuation and corrosion, cyclic stresses and hydrogen.
hydrogen.
Acknowledgments
5. Conclusion
The authors are grateful to Mr. Makoto Kitagawa and Mr.
Bridge cables in suspension bridges and cable-stayed bridges Kazuhiko Furuya (Honshu Shikoku Bridge Authority, Japan), Mr.
are under severe corrosion environments due to water and Sante Camo (Weidlinger Associates, USA), Prof. Alan Pense,
high temperature inside the cables. Mechanical properties and Prof. John Fisher and Dr. Eric Kaufman (ATLSS, Lehigh University,
remaining strength of corroded galvanized steel wires at dif- USA), and Mr. Toshimi Tarui (Nippon Steel Corporation, Japan) for
ferent corrosion levels were investigated. Corroded galvanized their useful advice on this research.
steel wires on different corrosion levels were produced at
References
laboratories.
Actual tensile strength of corroded wires did not decrease with [1] Betti R, Yanev B. Conditions of suspension bridge cables. In: Proc., The New
corrosion levels, whereas elongation strength decreased sharply York city cable study, Workshop on safety appraisal of suspension bridge main
after the zinc layer was partly depleted and the steel started cables. New Jersey: NCHRP; 1998. p. 1–13.
[2] Mayrbaurl R. Corrosion in suspension bridge cables. In: Proc., congress,
to corrode. The accumulated amounts of diffusive hydrogen of international association of bridge and structural engineers, Paper no. 199.
corrosion level-2 wires with and without induced tension were 2000.
almost the same with the amounts of less than 0.2 ppm that was [3] Stahl FL, Gagnon CP. Cable corrosion. New York: ASCE Press; 1996.
[4] Furuya K, Kitagawa M, Nakamura S, Suzumura K. Corrosion mechanism and
well below a critical value of 0.7 ppm. Therefore, induced tensions new protection methods of suspension bridge cables, structural engineering
of steel wires did not affect the amount of diffusive hydrogen and international. Journal of International Association of Bridge and Structural
below the critical concentration to cause brittleness. Engineers, Zurich, Switzerland 2000;10(3):189–93.
[5] Nakamura S, Suzumura K, Tarui T. Mechanical properties and remaining
Fatigue strength of corroded wires was obtained by cyclically strength of corroded bridge wires, structural engineering international. Journal
stressing them under dry and wet environments. It was found that of International Association of Bridge and Structural Engineers, Zurich,
fatigue strength did not change when only the galvanized layer was Switzerland 2004;14(3):50–4.
[6] Suzumura K, Nakamura S. Environmental factors affecting corrosion of
corroded, but it significantly decreased after the steel corrosion galvanized steel wires. Journal of Materials in Civil Engineering, ASCE 2004;
below the galvanized layer progressed. The fatigue strength under 16(1):1–7.
wet conditions was lower than that under dry conditions. [7] Tarui T, Maruyama N, Eguchi T, Konno S. Development of high strength
galvanized steel wires for bridge cables. In: IABSE conference on cable
The broken wire of an old suspension bridge was also
supported bridges. 2001.
investigated. The fracture surface was similar to that caused by [8] Uhlig HH, Revie RW. Corrosion and corrosion control. New York: John Wiley &
corrosion fatigue rather than by hydrogen embrittlement. It was Sons; 1963.