Experimental Investigation of Rain-Wind-Induced Vibration of Cables in Cable-Stayed Bridges and Its Mitigation
Experimental Investigation of Rain-Wind-Induced Vibration of Cables in Cable-Stayed Bridges and Its Mitigation
Experimental Investigation of Rain-Wind-Induced Vibration of Cables in Cable-Stayed Bridges and Its Mitigation
Abstract
Keywords: Cable-stayed bridge; Stay cable; Rain–wind-induced vibration; Wind tunnel test; Mitigation
0167-6105/$ - see front matter r 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jweia.2004.09.003
ARTICLE IN PRESS
80 M. Gu, X. Du / J. Wind Eng. Ind. Aerodyn. 93 (2005) 79–95
1. Introduction
There have been many reports in the past 20 years on excessive and unanticipated
vibration of cables in cable-supported bridges under the simultaneous occurrence of
wind and rain [1–5]. The authors of the present paper also observed strong
rain–wind-induced cable vibrations from two cable-stayed bridges recently built in
Shanghai and in Nanjing, China [6]. The bridge located in Shanghai has a main span
of 602 m with two inclined cable planes, and the other bridge in Nanjing has a main
span of 628 m. Under a moderate rainfall condition and a wind of mean speed
between 10 and 17 m/s, some cables of the bridge in Shanghai exhibited vibrations so
strong that the steel tubes protecting the cable at the bridge deck level were broken.
New rubber rings of high energy absorbing capacity had to be installed between the
cable and steel tube to prevent damage to the tube, but cable vibration of limited
amplitude can still be observed under the wind and rain conditions. Rain–wind-
induced cable vibration has become a great concern to bridge engineering and wind
engineering communities.
To find out the reason why stay cables have severe vibration under wind and rain
conditions, field measurements [1,4,5], wind tunnel simulation tests [7–14] and
theoretical analyses [15–17] were conducted by researchers and engineers around the
world. In wind tunnel simulation tests, there are two approaches for simulating
rivulets on the cable section model: one is to spray water appropriately onto the
surface of the cable model to form rivulets [1,7,8,12,13] and the other is to stick
artificial rivulets on the cable surface [9,15,17]. It is now believed that the rivulet
formed along the upper surface of a cable plays an important role in developing
rain–wind-induced cable vibration.
In this paper, the phenomenon of rain–wind-induced vibration of a cable model is
successfully reproduced under conditions of spraying water onto the cable model to
form the upper and lower real rivulets in the TJ-1 boundary layer wind tunnel in
Tongji University. The effects of several main factors on the characteristics of
rain–wind vibration are experimentally investigated. Moreover, countermeasures of
spiral wires and dampers for the mitigation of the vibration are experimentally
studied in detail. The results obtained from the tests may be used as valuable
references for further studies on the mechanisms and mitigation countermeasures of
the vibration.
The present test is conducted in the efflux section of the TJ-1 boundary layer wind
tunnel in Tongji University. This tunnel is a straight-through boundary layer wind
tunnel with an original working section of 1.8 m (width) 1.8 m (height). The final
diffuser of the wind tunnel is replaced by a new contraction section, which is
specially designed and manufactured to improve the quality of the wind field of the
new testing section. The exit of the contraction section is a round one with a
diameter of 2.4 m and the maximum wind speed available is about 20 m/s. In fact,
ARTICLE IN PRESS
M. Gu, X. Du / J. Wind Eng. Ind. Aerodyn. 93 (2005) 79–95 81
even if there is a contraction section at the end of the wind tunnel the uniformity of
mean wind speed of the present testing section is not comparable to that of the
original working section of the wind tunnel. Fig. 1 shows the distributions of mean
wind speeds of two sections. The distance between Section 1 and the wind tunnel exit
is 1.5 m, and that between Section 2 and the exit is 3 m. The experimental set-up is
located between Sections 1 and 2. From the figure it can be seen that the wind speed
at the central point of Section 2 is 9.6 m/s, while the maximum wind speed at a
measuring point of the same section is 12 m/s.
The cable model for the test has a length of 2.5 m and a diameter of 120 mm. The
model consists of a wood cylinder as its core and PE rubber as the surface material,
which is the same as real cables. The weight of the model is 15 kg, and the original
damping ratio is 0.14%. Scruton number (¼ 4pmx=rD2 ; where x is the damping
ratio, r is the air density, m is the mass of the cable model per unit length, D is the
diameter of the cable model) is thus only about 5.86. For the prototype cable the
structural density is about 7800 kg/m3. Thus Scruton number of the prototype cable
is about 86.22, 14.7 times as large as that of its model.
A graduated disc is fixed at the lower end of the model to record approximately
the positions of the upper rivulet.
A new experimental set-up (see Fig. 2) is specially designed for the tests, which
includes two main parts: one is the base for supporting the other part—the frame,
and the other is the frame composed of steel tubes, on which the cable model is hung
with springs. The base of the experimental set-up has four wheels and thus can be
easily rotated for simulating the required wind yaw angles. The frame can be easily
adjusted to different angles with regard to the base for required inclination angles of
the cable model. The inclination angle of the cable model, a, and the wind yaw angle,
b, are schematically shown in Fig. 3. The cable model is supported with springs to
Fig. 3. Schematic diagram of inclination angle of cable model and wind yaw angle.
the frame. The original natural frequency and damping ratio of the cable model are
measured to be 1 Hz and 0.14%, respectively. The rain-simulating unit of the
experimental set-up includes a water pool, a lift pump, a valve and a sprinkler with
16 sprinkling heads. The required rainfall and direction can be archived using the
rain-simulating unit. The testing wind speeds range from 5 to 13 m/s with the
increment of 1 m/s. Two accelerometers are mounted, respectively, on both ends of
the cable model to pick up its vibration signals.
The testing conditions are listed in Table 1.
M. Gu, X. Du / J. Wind Eng. Ind. Aerodyn. 93 (2005) 79–95
Table 1
ARTICLE IN PRESS
Testing conditions
83
ARTICLE IN PRESS
84 M. Gu, X. Du / J. Wind Eng. Ind. Aerodyn. 93 (2005) 79–95
As mentioned above, several wind tunnel tests have been carried out by the
other researchers under the condition of simulated real rain to investigate
ARTICLE IN PRESS
M. Gu, X. Du / J. Wind Eng. Ind. Aerodyn. 93 (2005) 79–95 85
30
25 With rain
Without rain
20
Amplitude (cm)
15
10
0
5 6 7 8 9 10 11 12 13
Wind speed (m/s)
Fig. 4. Comparison of vibration amplitudes of cable model with and without simulated rain (a=251,
b=351, x=0.14%).
ARTICLE IN PRESS
Parameters of cable model Hikami Matsumoto Verwiebe Flamand Larose Cosentino
et al. [1] et al. [9] et al. [12] [7] et al. [13] et al. [14]
Diameter Length (cm) 14 260 14,16 950 10 200 16 700 25 600 16 560
Weight (kg/m) — — — 16.0 14.0
Frequency (Hz) 1,2,3 — 2,4 1 0.66,1.2 1.08
Damping ratio (%) — — — 0.1% 0.025, 0.49, 1.31, 1.62, 1.94 0.08–0.25
Inclination angle (1) 45 40 30 25 30 —
Wind yaw angle (1) 45 35,45 0,45,90 0,30,45,60,90 20,30 0,20,30,45,50,60
Wind speed (m/s) 6–18 0–15 2–30 6–13 9–12 5–17
Maximum amplitude (peak–peak) (cm) 36 25 - 32 50 26
ARTICLE IN PRESS
M. Gu, X. Du / J. Wind Eng. Ind. Aerodyn. 93 (2005) 79–95 87
35
o o
α=25 ,β =35
30 o o
α=30 , β =35
o o
α=35 ,β =35
25 o o
α=40 ,β =35
Amplitude (cm)
o o
α=45 ,β =35
20
15
10
0
5 6 7 8 9 10 11 12 13 14
Wind speed (m/s)
parameters for the investigation of the effect of the wind yaw angle on the vibration
are selected to be 301 inclination angle and wind yaw angles of 01, 251, 301, 351, 401
and 451. Fig. 6 presents the testing results. In the test, for the wind yaw angles of 301,
351 and 401, the upper rivulet can form and vibrate on the surface of the cable model,
but for the wind yaw angles of 01, 251 and 451, the upper rivulet cannot form. As a
result, the cable model has severe vibrations for the wind yaw angles of 301, 351 and
401, while it remains in a static state for the wind yaw angles of 01, 251 and 451
throughout all the testing wind speeds. Moreover, the vibration amplitudes of the
cable model are largest for the wind yaw angles of 301 and 351 compared with the
other wind yaw angles.
3.2.4. Effects of inclination angle and wind yaw angle on position of upper rivulet
As mentioned above, the lower rivulet remains almost static when the rain–wind-
induced vibration occurs. The vibration ranges of the upper rivulet can be found by
ARTICLE IN PRESS
88 M. Gu, X. Du / J. Wind Eng. Ind. Aerodyn. 93 (2005) 79–95
25
o o
α =30 , β =25
o o
α =30 , β =30
20 o o
α =30 , β =35
o o
α =30 , β =40
o o
α =30 , β =45
Amplitude (cm)
15
10
0
5 6 7 8 9 10 11 12 13
Wind speed (m/s)
Fig. 6. Effect of wind yaw angle (a=301, x=0.14%).
35
30 f = 1.0 Hz
f = 1.7 Hz
25 f = 2.1 Hz
f = 2.6 Hz
Amplitude (cm)
20
15
10
0
5 6 7 8 9 10 11 12 13 14
Wind speed (m/s)
Fig. 7. Vibration amplitudes of cable model for different natural frequencies (a=301, b=351, x=0.14%).
reading the data shown in the graduated disc fixed at the lower end of the cable
model. The definition of the upper rivulet is schematically shown in Fig. 8, and the
ranges of the vibration angle of the upper rivulet for different inclination angles of
ARTICLE IN PRESS
M. Gu, X. Du / J. Wind Eng. Ind. Aerodyn. 93 (2005) 79–95 89
upper rivulet
lower rivulet
the cable model and wind yaw angles are presented in Fig. 9. The results are almost
the same as those reported by Hikami et al.[1].
In the test, double-spiral wires of diameters 0.5, 1, 3 and 7 mm are clockwise and
anti-clockwise twined along the cable model in 30 and 60 cm pitches, respectively, for
the investigation of mitigation efficiency of rain–wind-induced vibration. The
clockwise and anti-clockwise senses of the spirals are marked from the view of the
top of the cable model.
The effect of the diameter of the spiral wire on the mitigation efficiency can be seen
from Fig. 10. In this figure, the vibration amplitudes of the cable model with spiral
wires in a pitch of 30 cm and in clockwise sense are drawn for the different wire
diameters. It is observed in the test that the spiral wires with diameter of 1 mm can
stop the formation of the upper rivulet to some extent, and the vibration amplitude
of the cable model thus becomes much smaller, as indicated in Fig. 10. For the spiral
wires of diameters of 3 and 7 mm, no upper rivulet is observed on the cable model’s
surface, and thus the vibration completely vanishes correspondingly. Unexpectedly,
it is found in the present test that the vibration amplitude of the cable model with
double-spiral wires of diameters 0.5 mm is even larger than that without spiral wires.
In the test, the upper rivulet and its motion around the surface of the cable model
with double-spiral wires of diameters of 0.5 mm can clearly be seen. The spiral wires
ARTICLE IN PRESS
90 M. Gu, X. Du / J. Wind Eng. Ind. Aerodyn. 93 (2005) 79–95
100
α =2 5 ° , β =3 5 °
95
α =3 0 ° , β =3 5 °
α =3 5 ° , β =3 5 °
90
α =4 0 ° , β =3 5 °
α =4 5 ° , β =3 5 °
85
80
θ( )
o
75
70
65
60
55
50
6 7 8 9 10 11 12 13
100
α =30 ° , β =2 5 °
95
α =30 ° , β =3 0 °
α =30 ° , β =3 5 °
90
α =30 ° , β =4 0 °
α =30 ° , β =4 5 °
85
80
θ( )
o
75
70
65
60
55
50
5 6 7 8 9 10 11 12 13
Fig. 9. Effects of inclination angle and wind yaw angle on position of upper rivulet.
seem too thin to prevent the rivulet from forming. To ensure the reliability of the
unexpected testing result, the cable model of the same parameters and spiral wires is
repeatedly tested under the same testing conditions several times. The results from
the repeated tests are almost the same.
The cable models twined with spiral wires with diameter of 1 mm in clockwise and
anti-clockwise senses are then tested to investigate the effect of twine direction on the
mitigation efficiency of the vibration. The results are presented in Fig. 11. From the
figure it can be seen that for the cable model twined with double-spiral wires in
ARTICLE IN PRESS
M. Gu, X. Du / J. Wind Eng. Ind. Aerodyn. 93 (2005) 79–95 91
24
Without rain, without spiral wires
With rain, without spiral wires
20 ϕ0.5mm@30cm in clockwise
ϕ1 mm@30cm in clockwise
16
Amplitude (cm)
12
0
6 7 8 9 10 11 12 13
Wind speed (m/s)
Fig. 10. Effect of diameter of the spiral wire on mitigation efficiency (a=301, b=351, x=0.14%).
clockwise sense and in a pitch of 30 cm, the vibration amplitude of the cable model is
small, that is to say, the rain–wind-induced vibration of the cable can be mitigated
efficiently. But when the spiral wires are anti-clockwise twined also in a pitch of
30 cm, the upper rivulet can obviously be observed, and the vibration amplitudes are
much larger than that for the clockwise-twined spiral wires. Only when the pitch of
the spiral wires becomes smaller than 15 cm the vibration vanishes. It is obvious that
the twine direction has great effect on the mitigation efficiency of the vibration.
Moreover, the effect of the pitch of the spiral wires on the mitigation efficiency is
also experimentally studied, and some test results are presented in Fig. 12. In the test,
the upper rivulet can still form on the surface of the cable model for the clockwise-
twined spiral wires in a pitch of 60 cm and a vibration of a rather large amplitude
exhibits accordingly. Only when the pitch decreases to 30 cm, no vibration can be
recorded. The result indicates that only a proper pitch of the spiral wires can destroy
the upper rivulet and further suppress the rain–wind-induced vibration of cables.
The above discussions suggest that one should be careful when determining the
diameter, pitch and twine direction of the spiral wire for the suppression of
rain–wind-induced vibration of cables in practical applications.
To provide damping to the cable model, thin steel wire rings are attached to the
springs of the cable model system. As the cable model vibrates and the springs
oscillate correspondingly, the wire rings deform and absorb vibration energy of the
cable model and thus provide damping to the cable model system [14]. The cable
ARTICLE IN PRESS
92 M. Gu, X. Du / J. Wind Eng. Ind. Aerodyn. 93 (2005) 79–95
16
Without rain, without spiral wires
14 With rain, without spiral wires
ϕ1mm@30cm in reverse-clockwise
12 ϕ1mm@15cm in reverse-clockwise
ϕ1mm@30cm in clockwise
10
Amplitude (cm)
0
6 7 8 9 10 11 12 13
Wind speed (m/s)
Fig. 11. Effect of twine direction on mitigation efficiency (a=301, b=351, x=0.14%).
14
6 7 8 9 10 11 12 13
Wind speed (m/s)
Fig. 12. Effect of pitch of spiral wires on mitigation efficiency (a=301, b=351, x=0.14%).
model’s damping ratio can be achieved at different levels by changing the diameter of
the ring or the size of steel wires. The values of the damping ratio are measured by
free-decay vibration test technique in the test. It is found that the measured damping
ratios are repeatable even for large vibration amplitudes of the cable model. Besides
ARTICLE IN PRESS
M. Gu, X. Du / J. Wind Eng. Ind. Aerodyn. 93 (2005) 79–95 93
30
ξ = 0.14%
25 ξ = 0.19%
ξ = 0.29%
ξ = 0.33%
20 ξ = 0.50%
Amplitude (cm)
15
10
0
5 6 7 8 9 10 11 12
Wind speed (m/s)
Fig. 13. Vibration amplitudes of cable model for different damping levels (a=351, b=351).
the original damping ratio of the cable model, five kinds of damping ratios of the
cable model are achieved in the test, which are 0.19%, 0.29%, 0.33%, 0.50% and
0.6%.
The vibration amplitudes of the cable model for the different damping levels are
shown in Fig. 13. It can be found from the figure that with the increase of the
structural damping ratio the vibration amplitude decreases. Several rounds of
repeated tests show that when the damping ratio of the cable model achieves 0.6%
the vibration of the cable model vanishes completely. However, field measurements
[18] of mitigation of rain–wind-induced vibration of cables of No. 2 Nanjing Bridge
over the Yangtze River, a cable-stayed bridge with a steel box girder and a main span
of 628 m, indicated that some cables with even 0.8% damping ratio of the first mode
could not stop their vibrations completely. The field measurement results are
somewhat inconsistent with the present test ones, which may mainly be due to the
difference between the Scruton numbers of the cable model and its prototype cable.
The relationship between the minimum damping ratio required for suppressing the
rain–wind-induced vibration of real cables and that from wind tunnel tests needs to
be further studied.
5. Concluding remarks
simultaneous actions of wind and rain. Parametric studies on the effects of the main
factors on the rain–wind vibration are performed. Countermeasures of spiral wires
twined on the surface of the cable model and dampers for mitigation of the vibration
are then experimentally investigated. Some conclusions can be made as follows:
(1) The upper rivulet and its motion around the surface of the cable seem to be the
prerequisite for the appearance of rain–wind-induced vibration of the cable
model. The rain–wind-induced vibration of cables exhibits a kind of velocity- and
amplitude-restricted vibration.
(2) The rain–wind-induced vibration is most severe for the inclination angle of 301
and wind yaw angles between 301 and 351.
(3) The vibration amplitude of rain–wind-induced vibration decreases obviously
with the increase in frequency and damping of the cable.
(4) Spiral wire twined on the surface of the cable is effective in mitigating the
rain–wind-induced vibration through carefully selecting the diameter, pitch and
twine direction of the spiral wire.
Acknowledgements
References
[1] Y. Hikami, N. Shiraishi, Rain–wind-induced vibrations of cables in cable stayed bridges, J. Wind
Eng. Ind. Aerodynam. 29 (1988) 409–418.
[2] K. Ohshima, M. Nanjo, Aerodynamics stability of the cables of a cable-stayed bridge subject to rain
(a case study of the Aji River Bridge), in: Proceedings of US–Japan Joint Seminar on Natural
Resources, 1987, pp. 324–336.
[3] B.M. Pacheco, Y. Fujino, Keeping cables calm, Civ. Eng. ASCE 63 (10) (1993) 56–58.
[4] J. A. Main, N. P. Jones, Full scale measurements of stay cable vibration, in: A. Larsen, G.L. Larose,
F.M. Livesey (Eds.), Proceedings of the 10th International Conference On Wind Engineering,
Copenhagen, Denmark, 21–24 June, 1999, pp. 963–970.
[5] A. J. Persoon, K. Noorlander, Full scale measurements on the Erasmus Bridge after rain/wind
induced cable vibration, in: A. Larsen, G.L. Larose, F.M. Livesey (Eds.), Proceedings of the 10th
International Conference On Wind Engineering, Copenhagen, Denmark, 21–24 June, 1999,
pp. 1019–1026.
[6] M. Gu, C.J. Liu, G.Q. Lou, et al., Rain–wind-induced vibration of cables on cable-stayed bridges and
its control, Shanghai J. Mech. 21 (1) (1998) 281–288 (in Chinese).
[7] O. Flamand, Rain–wind-induced vibration of cables, J. Wind Eng. Ind. Aerodynam. 57 (1995)
353–362.
[8] M. Matsumoto, T. Saitoh, M. Kitazawa, et al., Response characteristics of rain–wind-induced
vibration of stay-cables of cable-stayed bridges, J. Wind Eng. Ind. Aerodynam. 57 (1995) 323–333.
[9] M. Matsumoto, N. Shirashi, H. Shirato, Rain–wind-induced vibration of cables of cable-stayed
bridges, J. Wind Eng. Ind. Aerodynam. 41–42 (1992) 2011–2022.
ARTICLE IN PRESS
M. Gu, X. Du / J. Wind Eng. Ind. Aerodyn. 93 (2005) 79–95 95
[10] Bosdogianni, D. Olivari, Wind- and rain-induced oscillations of cables of stayed bridges, J. Wind
Eng. Ind. Aerodynam. 64 (1996) 171–185.
[11] M. Matsumoto, T.Yagi, D. Tsushima, Vortex-induced vibration of inclined cables at high wind
speed, in: A. Larsen, G.L. Larose, F.M. Livesey (Eds.), Proceedings of the 10th International
Conference On Wind Engineering, Copenhagen, Denmark, 21–24 June, 1999, pp. 979–986.
[12] C. Verwiebe, H. Rucheweyh, Recent research results concerning the exciting mechanisms of
rain–wind-induced vibrations, J. Wind Eng. Ind. Aerodynam. (1998) 1005–1013.
[13] G.L. Larose, L.W. Smitt, Rain/wind induced vibrations of parallel stay cables, IABSE Conference,
Sweden, 1999, pp. 301–310.
[14] N. Coesentino, O. Flamand, C. Ceccoli, Rain–wind-induced vibration of inclined stay cables. Part I:
experimental investigation and physical explanation, Wind Struct. 6 (6) (2003) 471–484.
[15] H. Yamaguichi, Analytical study on growth mechanism of rain vibration of cables, J. Wind Eng. Ind.
Aerodynam. 33 (1990) 73–80.
[16] C.P.W. Geurts, P.C.V. Staalduinen, Estimation of the effects of rain–wind-induced vibration in the
design stage of inclined stay cables, in: A. Larsen, G.L. Larose, F.M. Livesey (Eds.), Proceedings of
the 10th International Conference On Wind Engineering, Copenhagen, Denmark, 21–24 June, 1999,
pp. 885–892.
[17] M. Gu, Q. Lu, Theoretical analysis of wind-rain induced vibration of cables of cable-stayed bridges,
APCWE-5, J. Wind Eng. Ind. Aerodynam. 89 (2001) 125–128.
[18] J.J. Shi, A study on countermeasures for mitigation of wind-rain induced vibration of cables of No. 2
Nanjing Bridge over Yangtze River, Research Report, Department of Bridge Engineering, Tongji
University, 2000. (in Chinese).