Engineering Structures 112 (2016) 162–171

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Movable anchorage systems for vibration control of stay-cables

in bridges
Ioannis G. Raftoyiannis ⇑, George T. Michaltsos
National Technical University of Athens, Dept. of Civil Engineering, 15780, Greece

a r t i c l e i n f o a b s t r a c t

Article history: Cables are efficient structural elements that are used in cable-stayed bridges, suspension bridges and
Received 9 June 2015 other cable structures. These cables are subjected to environmental excitations such as rain and wind
Revised 30 November 2015 induced vibrations. Rain–wind induced stay-cable vibrations may occur at different cable eigenfrequen-
Accepted 11 January 2016
cies. Therefore, external transverse dampers have to be designed for several targeted cable modes in
Available online 28 January 2016
order to decrease the oscillations amplitude and to finally diminish them. For the analytical study of such
problems, one may usually employ finite series or sinusoidal forms. In this paper, the real eigenfrequen-
Keywords:
cies and shape functions are determined and used. This paper aims to study the optimum state and char-
Cables dynamics
C–S bridges
acteristics of such a damper as well as its influence on the cable’s vibrations using the real cable modes.
Cables’ shape functions Ó 2016 Elsevier Ltd. All rights reserved.
Movable anchorages
Vibrations control
Damping systems

1. Introduction
Cable-stayed bridges have been known since the beginning of
the 18th century, but they have been of great interest only in the
last fifty years, particularly due to their special shape and also
because they are an alternative solution to suspension bridges
for long spans. The main reasons for this delay were the difficulties
in their static and dynamic analysis, the various non-linearities, the
absence of computational capabilities, the lack of high strength
materials and construction techniques. There are several studies
related to the static behavior [1–8], the dynamic analysis [9–18],
or the stability of cable-stayed bridges [19–22].
A significant problem, which arises from practical applications,
is the cables’ rain–wind induced vibrations. Large amplitude Rain–
Wind-Induced-Vibrations (RWIV) of stay cables are a challenging
problem in the design of cable-stayed bridges. Such phenomena
were first observed on the Meikonishi bridge in Nagoya, Japan
[23] and also later on other such bridges, as for instance the fully
steel Erasmus bridge in Rotterdam, the Netherlands (1996) and
the Second Severn Crossing, connecting England and Wales [24].
It was found that the cables, which were stable under wind action
only, were oscillating under a combined influence of rain and wind,
leading to large amplitude motions, even for light-to-moderate
⇑ Corresponding author.
E-mail addresses: rafto@central.ntua.gr (I.G. Raftoyiannis), michalts@central.
ntua.gr (G.T. Michaltsos).
simultaneous rain and wind action. The frequency of the observed
vibrations was much lower than the critical one of the vortex-
induced vibrations, while it was also perceived that the cable oscil-
lations took place in the vertical plane mostly in single mode; for
increasing cable length however, higher modes (up to the 4th)
appeared. Most importantly, during the oscillations a water rivulet
appeared on the lower surface of the cable, which was character-
ized by a leeward shift and vibrated in circumferential directions
[23,25,26].
The so-induced vibrations can cause reduced life of the cable
and its connection due to fatigue or rapid progress of corrosion.
Several methods, including aerodynamic or structural concepts
and tools, have been investigated in order to control the vibrations
of bridge’s stay cables. Aerodynamic methods, just as change of the
cables’ roughness were effective only for certain classes of vibra-
tion [6]. Another method is the coupling of the stays with sec-
ondary wires, in order to reduce their effective length and
thereby to avoid resonance [27,28]. This method changes unfavor-
ably the bridge’s esthetics.
In C–S bridges, the stays are generally fixed in anchorages on
the deck through an anchor head as it is shown in Fig. 1.
The most commonly applied method is the one of external dam-
pers attached transversely to a point of the stay-cable being at
some distance from the anchorage [29–31]. Many researchers have
proposed passive control of cables with the use of viscous dampers.
The last method is applied in a system of movable anchorage by
using a Friction Pendulum Bearing or an Elastomeric Bearing to

http://dx.doi.org/10.1016/j.engstruct.2016.01.014
0141-0296/Ó 2016 Elsevier Ltd. All rights reserved.
 I.G. Raftoyiannis, G.T. Michaltsos / Engineering Structures 112 (2016) 162–171 163

Fig. 1. Anchorage of stay-cables.

Fig. 3. Deformation of the cable.

replace the conventional fixed support of stay cables or in a system

of springs and dampers attached to the point of anchorage [32]. ds
It must be stated at this point that this work does not aim to

mðxÞ ¼ mðsÞ ð2Þ
dx
study the aeroelastic phenomena developed due to the interaction
between the cable and rain–wind, or to propose an aerodynamic f. The studied cables are referred to the inclined axis system
model for this type of loading. x–y–z of Fig. 3.
The present work focuses on the effectiveness of movable
anchorage systems consisting of a spring and a damper, where 3. The equations of motion
optimum combinations of which are investigated. Thus, the con-
sidered loading model is a simple sinusoidal excitation of the form 3.1. Projection on xz-plane
pðx; tÞ ¼ pðxÞ  sin xt where pðxÞ is the load distribution function
and x is the frequency of the excitation. For a shallow form of the cable, the following relations are
For the analytical study of such problems, sinusoidal series valid:
solutions are usually employed. In this paper the exact eigenfre- cos qz ¼ dx=ds ffi 1
quencies and shape functions are determined and used.
sin qz ¼ dw=ds ð3Þ
2. Basic assumptions sin dqz ffi 0

a. The elastic line of the cable under study has, due to dead and
live loads, a catenary shape with transverse displacement wo 3.1.1. Equilibrium of horizontal forces
and tensile force T o (see Fig. 2). Projecting on the xz-plane (Fig. 4) and taking the equilibrium of
b. Under the action of the dynamic loads py ðx; tÞ and pz ðx; tÞ, the horizontal forces we obtain:
cable takes the shape of Fig. 3, with additional displace-
ments ud ; td , wd and tensile forces T d .

dx dx dx
c. The deformed shape of the cables (under static and dynamic T þT þ@ T _  mu
þ px ds  cuds € ds ¼ 0
ds ds ds
loads) is a catenary curve which, due to its shallow form, can
be replaced by a second-order parabola.
d. The static and dynamic tensile forces are connected with the or finally:
following relations:
@H
 cu_  mu
€ ¼ px ðx; tÞ ð4Þ
TðtÞ ¼ T o þ T d ðtÞ @s
ð1Þ
HðtÞ ¼ Ho þ Hd ðtÞ
where H, is the projection of T on x-axis. 3.1.2. Equilibrium of vertical forces


e. The mðxÞ and mðsÞ of Fig. 2, are connected through the Projecting on the xz-plane (Fig. 4) and taking the equilibrium of
relation vertical forces we obtain:

Fig. 2. Cable and reference axes. Fig. 4. Projection on xz-plane.

164 I.G. Raftoyiannis, G.T. Michaltsos / Engineering Structures 112 (2016) 162–171


@w @w @w 5. Catenary and parabola approach
T þT þ@ T

þ mgds _  mwds
þ pz ds  cwds € ¼0
@s @s @s
ð5aÞ It is necessary to determine the form wo of a cable in rest. In this
case we have: pz ðx; tÞ ¼ py ðx; tÞ ¼ 0; u ¼ t ¼ 0; w ¼ wo ðxÞ.
Since the equation of the static equilibrium is valid, that is @Ho

 Therefore, Eq. (4) gives: @s
¼ 0, that is
@ @wo @wo ðxÞ
To

¼ mg; and also: w ¼ wo þ wd and ¼0 Ho ¼ Constant ð11Þ
@s @s @t
 @wo 
@
On the other hand, Eq. (2.3a) gives: @s
or
T o @s ¼ mg
Eq. (5a) becomes:
!


@ 2 wd @ 2 wo @ 2 wd @ dx dwo @ dwo
To þ Td þ _ d  mw
 cw € d ¼ pz ðx; tÞ ð5bÞ To   ¼ Ho 
g
¼ m
@x 2 @x2 @x2 @s ds dx @s dx

3.2. Projection on xy-plane or finally:

2
d wo ds
Ho 
 g ¼ m  g 
¼ m ð12Þ
Projecting the on xy-plane (Fig. 5), taking the equilibrium of dx
2 dx
vertical forces and through a similar process like the one of
Section 3.1.2 we obtain: It is usual to employ a parabolic shape function that is very
2 2
close to the catenary one, especially for shallow forms of cables.
@ td @ td For a cable’s shallow form (i.e. ds  dx), the equation of a parabola
To þ Td  ct_ d  mt
€d ¼ py ðx; tÞ ð6Þ
@x2 @x2

passing from the points (0, 0), (L, 0) and having w00o ¼  mg (from Eq.
Ho
12), is given by the following formula:
4. The cable’s deformation

mg
wo ðxÞ ¼  x  ðL  xÞ ð13Þ
2Ho
The following relations are valid:
2 2 2 A comparison between the parabolic shape of a cable (such as
ds ¼ dx þ dwo the one used in C–S-bridges) and the corresponding catenary
2 2 2 2
ð7Þ
ðds þ DdsÞ ¼ ðdx þ DdxÞ þ ðDdtÞ þ ðdwo þ Ddwo Þ shape, shows that the differences of the two carves’ sags amount
up to 0.02%. Thus, the parabolic shape is employed herein.
From this last, neglecting the higher order terms we get:
Ddx ¼ Dds dx
ds
 Ddwo dw
dx
o
, and since Ddwo ¼ dwd , we obtain: 6. Eigenfrequencies and shape functions [33]
Dds ds ds dwd dwo
Ddx ¼   dx   dx ð8Þ For the analytical study of the dynamic behavior of cables, sinu-
ds dx dx dx dx
soidal series solutions are usually employed.
On the other hand, we have: r ¼ e  E or In order to determine a range of functions satisfying the bound-
T r ary conditions and approximating in the best possible way the
e¼ ¼ d ð9aÞ
E EA exact shape of the vibrating cable, we will study the freely vibra-
Td
 ds2 tion of the cable, neglecting the influence of its rigidity.
Due to (9a), Eq. (8) becomes: Ddx ¼ EA  dx  dx  dw
dx
d
 dw
dx
o
 dx
Z L Considering the freely vibrating cable, the dynamic deforma-
and using the condition Ddx ¼ 0, the above gives: tions wd and td will be significantly smaller than the static ones.
Z Z
0 Therefore, the equations of motion can be written as follows:
L L
Td dx dwd dwo T o w00d þ T d w00o ¼ cw _ d þ mw €d
   dx ¼ 0 ð9bÞ
EA cos3 qz dx dx
0 0
T o td ¼ ct_ d þ mt
00 €d
From this last equation, after integration by parts and with
Z L 
1
boundary condition wd ð0Þ ¼ 0 we obtain: Td ¼ w0o ðLÞwd ðLÞ  w00o wd dx ð14a; b; c; dÞ

Z L  Lo 0
1 Z L
Td ¼ w0o ðLÞwd ðLÞ  w00o wd dx dx
Lo with: Lo ¼
0 EA cos qz
0 3
Z L ð10a; bÞ
dx
with: Lo ¼
0 EA cos qz
3
6.1. The vertical motion

Eq. (14a), due to Eq. (14c) gives:

Z
w00o  w0o ðLÞ w002 L
_ d  mw
T o w00d  cw €d ¼   wd ðLÞ þ o wd dx ð15Þ
Lo Lo 0

Taking into account Eq. (13), the above becomes:

00 Z


mgLw w002 L
_ d  mw
T o w00d  cw €d ¼ o
 wd ðLÞ þ o wd dx ð16Þ
2Ho Lo Lo 0

We are seeking a solution of the form:

wd ðx; tÞ ¼ WðxÞUðtÞ ð17Þ
Introducing Eq. (17) into Eq. (16) we get:

00
w00o RL
T o W 00  mgLw o
WðLÞ  W dx €þcU
U _
2Ho Lo Lo 0
¼ m
¼ x2w
Fig. 5. Projection on xy-plane.
mW U
 I.G. Raftoyiannis, G.T. Michaltsos / Engineering Structures 112 (2016) 162–171 165

00 Z


mgLw w002 L
WðxÞ ¼ c1 sinkw x þ c2 cos kw x þ o
WðLÞ þ o
W dx
2Ho Lo T o k2w T o Lo k2w 0

mx2w
with: k2w ¼
To
ð19Þ
For x = L, the above Eq. (19), becomes:

00 Z


mgLw w002 L
WðLÞ ¼ c1 sin kw L þ c2 cos kw L þ o
WðLÞ þ o
W dx
2Ho Lo T o k2w T o Lo k2w 0

Fig. 6. Cable with a damping system.

which finally gives:

Table 1
Eigenfrequencies of cables without damping. Z !
2Ho Lo T o k2w w002 L

m = 7 kg/m, T o = 300,000 dN WðLÞ ¼ 00

 c1 sin kw L þ c2 cos kw L þ o
W dx
2


2Ho Lo T o kw  mgLw o T o Lo k2w 0
L = 150 m L = 250 m L = 350 m
ð20Þ
x1 (sec1) 4.4697 2.8186 2.1507
x2 (sec1) 8.6716 5.2029 3.7164 Due to Eqs. (20) and (19) becomes:
x3 (sec1) 13.0125 7.8130 5.5868
x4 (sec1) 17.3432 10.4059 7.4328 WðxÞ ¼ c1 ðsin kw x þ G1 sin kw LÞ þ c2 ðcos kw x þ G1 cos kw LÞ
x5 (sec1) 21.6801 13.0093 9.2936 Z L
þ ðG2 þ G3 Þ þ W dx
0
00


mgLw w002
which yields the following uncoupled equations: where: G1 ¼ o
; G2 ¼ o
;
00
00 Z 2Ho Lo T o k2w 

mgLw Lo T o k2w
mx2w

mgLw w002 L o
W 00 þ W¼ o
WðLÞ þ o W dx G3 ¼ G1  G2 ð21Þ
To 2Ho Lo T o T o Lo 0 ð18a; bÞ
€ c _ The above integral–differential of Hammerstein type equation,
U þ U þ xw U ¼ 0
2
m finally gives:
The solution of Eq. (18a) is:

Fig. 7. Oscillations wd (m) and t (sec) of the midpoint of a cable of length L = 150 m, without a damping system.
166 I.G. Raftoyiannis, G.T. Michaltsos / Engineering Structures 112 (2016) 162–171

wd
wd

0.6
1
0.4
0.5
0.2

t t
5 10 15 20 25 5 10 15 20 25
-0.2 -0.5
-0.4
-1
-0.6
p=20*Sin(1*t) p=20*Sin(2*t)
wd
wd
1
4

0.5
2

t t
5 10 15 20 25 5 10 15 20 25

-2 -0.5

-4

p=20*Sin(3*t) p=20*Sin(4*t)
In this case, we have big
deformations, since the eigenfrequency
of the external load is near to the first one
of the cable (that is ω1 =2.82 sec-1).

Fig. 8. Oscillations wd (m) and t (sec) of the midpoint of a cable of length L = 250 m, without a damping system.

wd wd
6
1
4
0.5 2

t
t 5 10 15 20 25
5 10 15 20 25
-2
-0.5
-4

-1 -6

p=20*Sin(1*t) p=20*Sin(2*t)
In this case, we have big
deformations, since the eigenfrequency
of the external load is near to the first one
of the cable (that is ω1 =2.15 sec-1).

wd wd
0.75
1.5 0.5
1
0.25
0.5
t
t 5 10 15 20 25
5 10 15 20 25 -0.25
-0.5
-0.5
-1
-0.75
-1.5
p=20*Sin(3*t) p=20*Sin(4*t)
Fig. 9. Oscillations wd (m) and t (sec) of the midpoint of a cable of length L = 350 m, without a damping system.
 I.G. Raftoyiannis, G.T. Michaltsos / Engineering Structures 112 (2016) 162–171 167

Table 2
Eigenfrequencies of cables with damping.

L (m) 150 250 350

ks (kN/m) 3000 9000 20,000 1 3000 9000 20,000 1 3000 9000 20,000 1

x1 (sec1) 3.105 3.728 4.069 4.470 2.129 2.480 2.628 2.817 1.687 1.909 1.989 2.151
x2 (sec1) 6.934 7.486 7.972 8.672 4.320 4.707 4.951 5.203 3.192 3.467 3.607 3.716
x3 (sec1) 11.089 11.520 12.021 13.012 6.744 7.101 7.406 7.813 4.876 5.172 5.371 5.587
x4 (sec1) 15.356 15.700 16.156 17.343 9.295 9.589 9.899 10.406 6.693 6.947 7.161 7.433
x5 (sec1) 19.654 19,926 20.334 21.680 11.847 12.098 12.402 13.009 8.499 8.727 8.953 9.294

Z L
1 VðxÞ ¼ d1 sin kt x þ d2 cos kt x
W dx ¼ 
0 1  ðG2 þ G3 ÞL mx2t ð29Þ


 where: k2t ¼
1  cos kw L sin kw L To
c1 þ G1 L sin kw L þ c2 þ G1 L sin kw L
kw kw
The boundary conditions are:
Therefore, Eq. (21) becomes:
Vð0Þ ¼ 0
ð30a; bÞ
WðxÞ ¼ c1 ðsin kw x þ D1 Þ þ c2 ðcos kw x þ D2 Þ T o V 0 ðLÞ þ ks VðLÞ ¼ 0


G2 þ G3 1  cos kw L Introducing (29) into the above we get d2 ¼ 0 and the equation
where : D1 ¼ G1 sin kw L þ þ G1 L sin kw L
1  ðG2 þ G3 Þ kw ð22Þ of the eigenfrequencies is:


G2 þ G3 sin kw L T o kt cos kt L þ ks sin kt L ¼ 0 ð31Þ
D2 ¼ G1 cos kw L þ þ G1 L cos kw L
1  ðG2 þ G3 Þ kw
Finally the shape functions are:
The boundary conditions are: V n ðxÞ ¼ d1 sin kt x ð32Þ
Wð0Þ ¼ 0 with kt from Eq. (29).
0
ð23a; bÞ
T o W ðLÞ þ ks WðLÞ ¼ 0
7. The vertically forced vibrating cable
where ks is the spring constant.
Introducing Eq. (22) into the above conditions, we obtain: According to the assumptions of Section 1, the equations of verti-
c1 D1 þ c2 ð1 þ D2 Þ ¼ 0 cal motion of a cable subjected to external dynamic loadings and
supported by a damper system (see Fig. 6), can be written as follows:
c1 ½kw T o cos kw L þ ks ðsin kw L þ D1 Þ þ c2 ½kw T o sin kw L
þks ðcos kw L þ D2 Þ ¼ 0 ð24a; bÞ _ d  mw
T o w00d þ T d w00o  cw € d ¼ pz ðxÞf ðtÞ þ cd w
_ d  dðx  LÞ

Z L 
1 ð33a; bÞ
In order for the above system to have a non-trivial solution, the Td ¼ w0o ðLÞwd ðLÞ  w00o wd dx
determinant of the coefficients of the unknowns must be zero. This
Lo 0

condition yields the following eigenfrequencies equation: where cd is the damper’s coefficient, and dðxÞ the Dirac’s delta func-
tion. We are seeking a solution of the form:
D2 ½D1  ð1 þ D2 Þks  þ ½D1 ks  ð1 þ D2 ÞT o kw  cos kw L X
wd ðx; tÞ ¼ W n ðxÞUn ðtÞ ð34Þ
 ½ð1 þ D2 Þks þ D1 T o kw  sin kw L ¼ 0 ð25Þ n

where W n ðxÞ from Eq. (26) and Un ðtÞ unknown functions under
Finally from Eqs. (24), (25) and (22), one can determine the follow-
determination.
ing form of the shape functions:

Introducing Eq. (34) into Eq. (33a), and taking into account
kw T o cos kw L þ ks ðsin kw L þ D2 Þ

Eq. (33b) keeping in mind that w0o ðLÞ ¼  mgL we obtain:
W n ðxÞ ¼ c1 ðsin kw x þ D1 Þ þ  ðcos kw x þ D2 Þ 2Ho
kw T o sin kw L  ks ðcos kw L þ D2 Þ
X 00 X Z
ð26Þ

mgLw w002 L X
 W 00n Un þ o
W n ðLÞUn þ o W n Un dx
n
2Ho T o n Lo o n
c X _nþm
X
€n
6.2. The lateral motion þ WnU W nU
To n To n
Eq. (14b) is written: p ðxÞf ðtÞ cd X _ n dðx  LÞ
¼ z  W nU ð35aÞ
To To n
T o t00d  ct_ d  mt
€d ¼ 0 ð27Þ
Following the same as above procedure and seeking a solution Since W n satisfies Eq. (18a), the above becomes:
X cX X
of the form td ðx; tÞ ¼ VðxÞRðtÞ, we conclude to the following x2n W n Un þ _nþ
W nU W nU€n
equations: n
m n n

T o V 00 þ mx2t V ¼ 0 p ðxÞf ðtÞ cd X _ n dðx  LÞ

¼ z  W nU ð35bÞ
ð28a; bÞ m m n
€ þ c R_ þ x2 R ¼ 0
R t
m Multiplying by W k , and integrating the outcome from 0 to L, we
Eq. (28a) has the solution: finally obtain:
168 I.G. Raftoyiannis, G.T. Michaltsos / Engineering Structures 112 (2016) 162–171

Z L
w
f ðtÞ pz ðxÞW k dx P _n
€k þ c U
_ k þ x2 Uk ¼ cd W k ðLÞ n W n ðLÞU
U k
0
Z L
 Z L 0.4
m
m W 2k dx m W 2k dx
0 0
0.2
for: k ¼ 1 to n ð35cÞ
In order to solve the above differential system, we use the t
Lagrange transformation. 2 4 6 8 10
Therefore we set:
-0.2
LUk ðtÞ ¼ uk ðsÞ; Lf ðtÞ ¼ FðsÞ ð36a; bÞ
-0.4
_ k ð0Þ ¼ 0
with initial conditions: Uk ð0Þ ¼ U ð37Þ
Fig. 10. Oscillations w (m) and t (sec) of the middle of a cable with L = 150 m,
Hence, we conclude to the following system: ks = 3 dN/mm. ___ Cable with fixed anchors, . . . cd = 5, - - - cd = 10 and _ _
cd = 20 dN sec/mm.
ak1 u1 þ ak2 u2 þ   þ akn un ¼ bk
where :
c W ðLÞW ðLÞ  s
akq ¼ d kZ L q
w
m W 2k dx
0
Z L 0.4
pz ðxÞW k dx
c cd  s
akk ¼ s2 þ  s þ x2k þ Z L  W 2k ðLÞ ; bk ¼ 0 Z L  FðsÞ
m 0.2
m W 2k dx m W 2k dx
0 0
k ¼ 1 to n; q ¼ 1 to n t
2 4 6 8 10
ð38a;b;c;dÞ
-0.2
Solving the above system of Eq. (38a) we determine uk ðsÞ and
finally: -0.4

1
Uk ðtÞ ¼ L uk ðsÞ ð39Þ
Fig. 11. Oscillations w (m) and t (sec) of the middle of a cable with L = 150 m,
ks = 9 dN/mm. ___ Cable with fixed anchors, . . . cd = 5, - - - cd = 10 and _ _
cd = 20 dN sec/mm.
8. Numerical results and discussion

8.1. The cables

We consider a C–S-bridge with dense distribution of cables and

w
we study an individual cable with tension T o = 300,000 dN/cable,
cross-sectional area F = 7.5 ⁄ 103 m2, diameter D = 0.13 m, weight
per unit length G = 70 dN/m and mass per unit length m = 7 kg/m. 0.4
The deformed shape of the cable is assumed to be parabolic.
0.2

8.2. The rain–wind combination

t
2 4 6 8 10
It has been observed that the rain–wind-induced vibration in
bridge cables usually occurs in the frequency range from 0.5 to -0.2
4 sec1, [23,25]. For the study of the vibration of a cable under
the action of a rain–wind combination we choose the following -0.4
loading:
p ¼ pðxÞ  f ðtÞ ¼ 20  sin xt, where x = 1, 2, 3, 4 sec1 (for the
Fig. 12. Oscillations w (m) and t (sec) of the middle of a cable with L = 150 m,
study of the above cables without a damping system) and ks = 20 dN/mm. ___ Cable with fixed anchors, . . . cd = 5, - - - cd = 10 and _ _
x = 3 sec1 (for the study of the above cables with damping cd = 20 dN sec/mm.
system).

8.3. Behavior of the cables without a damping system

Using Eq. (25) with ks ¼ 1, we determine the eigenfrequencies
of the above considered cables, as they are shown in the following
Table 1.
In Figs. 7–9 one can see the oscillations of the midpoint of each
cable with L = 150, L = 250 and L = 350 m, subjected to loadings act-
ing with frequencies x = 1, 2, 3, 4 sec1.
8.4. The damping system
In the followings, we will use a system like the one shown in
Fig. 6, consisting of a spring with constant ks (3, 9 and 20 dN/
mm), and a damper with coefficient cd (5, 10 and 20 dN sec/mm).
In the following Table 2, we see the eigenfrequencies of the consid-
ered cables for each one of the used springs.
 I.G. Raftoyiannis, G.T. Michaltsos / Engineering Structures 112 (2016) 162–171 169

wL w
4
0.15

0.1
2
0.05

t t
2 4 6 8 10 2 4 6 8 10
-0.05

-0.1 -2

-0.15

Fig. 13. Oscillations wL (m) and t (sec) of movable anchorage of a cable with
Fig. 16. Oscillations w (m) and t (sec) of the middle of a cable with L = 250 m,
L = 150 m, ks = 3 dN/mm and . . . cd = 5, - - - cd = 10 and _ _ cd = 20 dN sec/mm.
ks = 3 dN/mm. ___ Cable with fixed anchors, . . . cd = 5, - - - cd = 10 and _ _
cd = 20 dN sec/mm.

wL w
4
0.15

0.1
2
0.05
t
2 4 6 8 10 t
-0.05 2 4 6 8 10
-0.1
-2
-0.15

Fig. 14. Oscillations wL (m) and t (sec) of movable anchorage of a cable with
L = 150 m, ks = 9 dN/mm and . . . cd = 5, - - - cd = 10 and _ _ cd = 20 dN sec/mm. Fig. 17. Oscillations w (m) and t (sec) of the middle of a cable with L = 250 m,
ks = 9 dN/mm. ___ Cable with fixed anchors, . . . cd = 5, - - - cd = 10 and _ _
cd = 20 dN sec/mm.

wL w
0.15 4
0.1

2
0.05

t
2 4 6 8 10 t
2 4 6 8 10
-0.05

-0.1 -2

Fig. 15. Oscillations wL (m) and t (sec) of movable anchorage of a cable with
L = 150 m, ks = 20 dN/mm. and . . . cd = 5, - - - cd = 10 and _ _ cd = 20 dN sec/mm. Fig. 18. Oscillations w (m) and t (sec) of the middle of a cable with L = 250 m,
ks = 20 dN/mm. ___ Cable with fixed anchors, . . . cd = 5, - - - cd = 10 and _ _
cd = 20 dN sec/mm.

In the plots of Figs. 10–12 we see the oscillations of the middle

of a cable of length 150 m, tension 300,000 dN, for various values of
springs and dampers.
In the plots of Figs. 13–15 we see the oscillations of the movable
anchorage of a cable of length 150 m, tension 300,000 dN, for var-
ious values of springs and dampers.
From the above plots (of Figs. 10–15), we ascertain that stronger
springs are not always effective and thus, their combination with
dampers should be considered.
In general, stronger dampers are more effective than the softer
ones.
The above, are valid for both the deformations of the cables and
the anchorage motion.
In the plots of Figs. 16–18 we see the oscillations of the middle
of a cable of length 250 m, tension 300,000 dN, for various values of
springs and dampers.
170 I.G. Raftoyiannis, G.T. Michaltsos / Engineering Structures 112 (2016) 162–171

wL w
0.4
1.5

1
0.2
0.5
t t
2 4 6 8 10 2 4 6 8 10
-0.5
-0.2
-1

-0.4 -1.5

Fig. 19. Oscillations wL (m) and t (sec) of movable anchorage of a cable with Fig. 23. Oscillations w (m) and t (sec) of the middle of a cable with L = 350 m,
L = 250 m, ks = 3 dN/mm. and . . . cd = 5, - - - cd = 10 and _ _ cd = 20 dN sec/mm. ks = 9 dN/mm. ___ Cable with fixed anchors, . . . cd = 5, - - - cd = 10 and _ _
cd = 20 dN sec/mm.

wL
0.4

w
0.2
1.5

t 1
2 4 6 8 10
0.5
-0.2 t
2 4 6 8 10
-0.5
-0.4
-1
Fig. 20. Oscillations wL (m) and t (sec) of movable anchorage of a cable with
L = 250 m, ks = 9 dN/mm. and . . . cd = 5, - - - cd = 10 and _ _ cd = 20 dN sec/mm. -1.5

wL Fig. 24. Oscillations w (m) and t (sec) of the middle of a cable with L = 350 m,
0.4 ks = 20 dN/mm. ___ Cable with fixed anchors, . . . cd = 5, - - - cd = 10 and _ _
cd = 20 dN sec/mm.

0.2
wL

0.3
t
2 4 6 8 10 0.2

0.1
-0.2
t
2 4 6 8 10
-0.1

Fig. 21. Oscillations wL (m) and t (sec) of movable anchorage of a cable with -0.2
L = 250 m, ks = 20 dN/mm. and . . . cd = 5, - - - cd = 10 and _ _ cd = 20 dN sec/mm.
-0.3
w
Fig. 25. Oscillations wL (m) and t (sec) of movable anchorage of a cable with
1.5 L = 350 m, ks = 3 dN/mm. and . . . cd = 5, - - - cd = 10 and _ _ cd = 20 dN sec/mm.

1
In the plots of Figs. 19–21 we see the oscillations of the movable
0.5 anchorage of a cable of length 250 m, tension 300,000 dN, for var-
ious values of springs and dampers.
t From the above plots (of Figs. 16–21) we observe that although
2 4 6 8 10
the oscillations’ amplitude is remarkable large when the frequency
-0.5
of the external loading approaches the first eigenfrequency of the
-1 cable, the effectiveness of the system is not always satisfactory
since the motion amplitude has a low reduction. In this case, one
-1.5 has to search for the most effective combination of spring–damper.
In the plots of Figs. 22–24 we see the oscillations of the middle
Fig. 22. Oscillations w (m) and t (sec) of the middle of a cable with L = 350 m,
of a cable of length 350 m, tension 300,000 dN, for various values of
ks = 3 dN/mm. ___ Cable with fixed anchors, . . . cd = 5, - - - cd = 10 and _ _
cd = 20 dN sec/mm. springs and dampers.
 I.G. Raftoyiannis, G.T. Michaltsos / Engineering Structures 112 (2016) 162–171
wL
0.2
0.1
t
2 4 6 8 10
-0.1
-0.2
-0.3
Fig. 26. Oscillations wL (m) and t (sec) of movable anchorage of a cable with
L = 350 m, ks = 9 dN/mm, and . . . cd = 5, - - - cd = 10 and _ _ cd = 20 dN sec/mm.
wL
0.2
0.1
t Eng Struct Dyn 1980;8(1):1–16.
2 4 6 8 10
-0.1
-0.2
Fig. 27. Oscillations wL (m) and t (sec) of movable anchorage of a cable with
L = 350 m, ks = 20 dN/mm. and . . . cd = 5, - - - cd = 10 and _ _ cd = 20 dN sec/mm.
In the plots of Figs. 25–27 we see the oscillations of the movable
anchorage of a cable of length 350 m, tension 300,000 dN, for var-
ious values of springs and dampers.
From the above plots (of Figs. 22–27) we observe that as the
cable length increases, the stronger springs are more effective for
both the oscillations of the cables and for the motion of the
anchorages.
9. Conclusions
In this paper, a mathematical model is presented for the study
of a damping system applied at the movable anchorage of a cable
consisting of a spring and a damper. The exact eigenfrequencies
and shape functions of such a cable are determined and used.
From the above models studied herein, one can draw the fol-
lowing conclusions:
1. Stronger springs are not always effective due to the combined
eigenfrequencies of the cable–spring system and thus, a combi-
nation with dampers should always be considered and studied.
2. In general, strong dampers are more effective than soft dampers
since the maximum amplitude of the motion if strong dampers
are applied is always smaller compared to the maximum ampli-
tude for soft dampers. This is valid for both the cables’ deforma-
tions and the anchorages’ motion.
3. It is observed that as the cable length increases, strong springs
are more effective for both the oscillations of the cables and the
motion of the anchorages. This is due to the fact that the motion
frequency decreases as the cable length increases, while it
increases with higher rate as the spring becomes stronger.
4. When the frequency of the external loading approaches the first
eigenfrequency of the cable, the oscillations’ amplitude is
remarkably large and the effectiveness of the movable anchor-
171
age (spring–damper) system is not satisfactory since the motion
amplitude has a low reduction. In such a case, one must search
for the most effective combination of spring–damper values in
order to achieve reduced motion amplitudes for both the cable
and the anchorage system.
References
[1] Fleming JF. Non-linear static analysis of cable-stayed bridge structures.
Comput Struct 1979;10:986–1000.
[2] Kollbruner CF, Hajdin N, Stipanic B. Contribution to the analysis of cable-
stayed bridges, Inst. For Engineering Research ed., vol. 48. Zürich: Schulthess
Verlag; 1980.
[3] Bruno D, Grimaldi A. Non-linear behavior of long-span cable-stayed bridges,
Meccanica. J Ital Assoc Theor Appl Mech (A/META) 1985;20(4):303–13.
[4] Gimsing N. Cable supported bridges – concept and design. 2nd
ed. Chichester: John Wiley & Sons; 1997.
[5] Khalil MS. Non-linear analysis of cable-stayed bridges at ultimate load level.
Can J Civ Eng 1999;23(5):1111–7.
[6] Virgoreux M. Recent evolution on cable-stayed bridges. J Eng Struct
1999;21:737–55.
[7] Michaltsos GT, Ermopoulos JC, Konstantakopoulos TG. Preliminary design of
cable-stayed bridges for vertical static loads. J Struct Eng Mech 2003;16(1):1–15.
[8] Freire AMS, Negräo JHO, Lopez AV. Geometrical non-linearities on the static
analysis of highly flexible steel cable-stayed bridges. Comput Struct 2006;84
(31–32):2128–40.
[9] Fleming JF, Egeseli EA. Dynamic behavior of a cable-stayed bridge. Int J Earthq
[10] Nazmy AS, Abdel-Ghaffar AM. Non-linear earthquake response analysis of
long-span cable-stayed bridges. Earthq Eng Struct Dyn 1990;19:45–62.
[11] Abdel-Ghaffar AM, Khalifa MA. Importance of cable vibration in dynamics of
cable-stayed bridges. J Eng Mech 1991;117(11):2571–89.
[12] Chatterjec PK, Datta TK, Suruna CS. Vibration of cable-stayed bridges under
moving vehicles. J Struct Eng Int 1994;2:116–21.
[13] Bruno D, Colotti V. Vibration analysis of cable-stayed bridges. J Struct Eng Int
1994;1:23–8.
[14] Achkire Y, Preumont A. Active tendon control of cable-stayed bridges. J Earthq
Eng Struct Dyn 1996;25(6):585–97.
[15] Michaltsos GT. A simplified model for the dynamic analysis of cable-stayed
bridges. Facta Univ 2001;3(11):185–204.
[16] Konstantakopoulos TG, Michaltsos GT, Sophianopoulos DS. A simplified model
for the study of the lateral–torsional vibration of cable-stayed bridges. In:
Eurodyn 2002, Munich, 2002.
[17] Wang PH, Liu MY, Huang YT, Lin LC. Influence of lateral motion of cable-stayed
bridges. Struct Eng Mech 2010;34(6):719–38.
[18] Raftoyiannis IG, Michaltsos GT. Curved-in-plane cable-stayed bridges – a
mathematical model. Int J Struct Stab Dyn 2012;12(3):1–41.
[19] Ermopoulos J, Vlahinos A, Wang Yang-Cheng. Stability analysis of cable-stayed
bridges. Comput Struct 1992;44(5):1083–9.
[20] Bosdogianni A, Olivari D. Wind-reduced and Rain-induced oscillations of
cable-stayed bridges. J Wind Eng Ind Aerodyn 1997;64(2–3):171–85.
[21] Michaltsos GT. Stability of a cable-stayed bridge’s pylon, under time-depended
loading. Collection of papers in memory of Academician P.S. Theocharis. Inst.
of Mechanics Problems and Academy of Sciences, Russia, Inst. of Mechanics of
Nat. Acad. of Sciences, Armenia, 2005.
[22] Michaltsos GT, Raftoyiannis IG, Konstantakopoulos TG. Dynamic stability of
cable-stayed bridge pylons. Int J Struct Stab Dyn 2008;8(4):627–43.
[23] Hikami Y, Shiraishi N. Rain–wind-induced vibrations of cables in cable stayed
bridges. J Wind Eng Ind Aerodyn 1998;29(1–3):409–18.
[24] Macdonald JHG, Dagless EL, Thomas BT, Taylor CA. Dynamic measurements of
the second severn crossing. Proc Inst Civ Eng – Transp 1997;1233(4):241–8.
[25] Wilde K, Witkowski W. Simple model of rain–wind-induced vibrations of
stayed cables. J Wind Eng Ind Aerodyn 2003;91(7):873–91.
[26] Ibrahim RA. Non-linear vibrations of suspended cables – Part III: Random
excitation and interaction with fluid flow. Appl Mech Rev 2004;57(6):515–49.
[27] Verwiebe C, Ruschweyh H. Recent research results concerning the exciting
mechanisms of rain–wind-induced vibrations. J Wind Eng Ind Aerodyn
1998;74–76:1005–10138.
[28] Cao DQ, Tucker RW, Wang C. A stochastic approach to cable dynamics with
moving rivulets. J Sound Vib 2003;268(2):291–304.
[29] Dieng L, Helbert G, Chirani SA, Lecompte T, Pilvin P. Use of shape memory
alloys damper device to mitigate vibration amplitudes of bridge cables. Eng
Struct 2013;56:1547–56.
[30] Zhou HJ, Sun LM. Damping of stay cable with passive-on magnetorheological
dampers: a full-scale test. Int J Civ Eng 2013;11(3):154–9.
[31] Boston C, Weber F, Guzzella L. Optimal semi-active damping of cables with
bending stiffness. Smart Mater Struct 2011;20(5):055005.
[32] Hwang I, Lee JS, Lee H. Movable anchorage system with FPB for vibration
control of stay cables. KSCE J Civ Eng 2001;15(5):841–7.
[33] Avraam TP, Konstantakopoulos TG. A mathematical research on cables with
fixed or movable anchorages. Int J Bridge Eng 2014;2(1):73–91.