Professional Documents
Culture Documents
net/publication/314223590
CITATIONS READS
10 603
5 authors, including:
All content following this page was uploaded by Bai Xiaoyu on 05 September 2018.
Abstract: The expressions of hydrodynamic pressure based on the radiation wave theory are too complex to be conveniently utilized in engi-
neering applications; therefore, a simplified extraction that is accurate and concise for calculating the hydrodynamic effect on deepwater bridge
Downloaded from ascelibrary.org by Beijing Jiaotong University on 09/04/18. Copyright ASCE. For personal use only; all rights reserved.
piers is highly desirable for seismic analyses. With a further comparison between the elastic vibration and rigid motion of cylindrical piers sub-
merged in water, the hydrodynamic pressure of the cylindrical piers can be approximately acquired in terms of the added mass caused by rigid
motion based on the linear radiation wave theory. A simplified formula for the hydrodynamic pressure is proposed by elaborate data fitting, in
which the section radius of the pier and the water height are the main parameters. The accuracy of the proposed formula is validated by the radia-
tion wave theory and the potential-based fluid-element method. Results prove that the simplified expressions are not only accurate but also con-
venient in calculating the hydrodynamic pressure of cylindrical piers with a wide range of geometries. The applicability of the Morison equation
and the Goto-Toki formula is also presented. DOI: 10.1061/(ASCE)BE.1943-5592.0001032. © 2017 American Society of Civil Engineers.
Author keywords: Deepwater cylindrical pier; Hydrodynamic pressure; Added mass; Simplified expression; Radiation wave theory;
Potential-based fluid-element method.
Introduction reaches 168 m, and the submerged depth of the pier may even
exceed 100 m. Because these areas in China are potential regions of
Deepwater bridges have become effective means across broad riv- high seismicity, these bridges will inevitably suffer from the threat
ers, straits, gulfs, and reservoirs (Feng 2009). Examples include the of earthquakes in the form of associated hydrodynamic pressures.
Rio-Antirrio Bridge (Dobry et al. 2003), the East China Sea Bridge As it is known, hydrodynamic pressure during earthquakes
(Liu et al. 2007), the Su Tong Yangtze River Highway Bridge may alter the vibration characteristics and dynamic responses of
(Bittner et al. 2007), and the Xiangshan Harbor Bridge (Gao et al. deepwater structures. Therefore, an accurate estimation of the
2008). The depths of the submerged parts of some deepwater bridge hydrodynamic force of bridge piers under earthquakes is very im-
piers exceed 100 m, and the diameters of some submerged bridge portant for the seismic design of deepwater bridges. The hydrody-
foundations reach tens of meters. Taking Japan’s Akashi Kaikyo namic effects on offshore and reservoir structures have received
Bridge as example, the maximum water depth near the midspan is considerable attention. The earliest study on the hydrodynamic
110 m. With a foundation diameter of 80 m, the submerged depth of force on deepwater structures during earthquakes began in the
the cylindrical caisson foundation below the main tower is greater 1930s. Westergaard (1933) performed experimental and analytical
than 60 m. In the past 20 years, growth has been noticed in terms studies on the hydrodynamic forces of dams and proposed an ana-
of the engineering application of deepwater bridges in China lytical expression to calculate the hydrodynamic pressure. Savage
(Feng 2009). Examples include bridges over many gigantic hydro- (1939) studied the influence of the cross-section shape on hydro-
power stations in the mountainous areas of western China, such as dynamic pressure based on scale model experiments. Morison et
the Mangjiedu Bridge, the Miaoziping Bridge, and the Yunyang al. (1950) proposed a semiempirical formula of added mass (com-
Yangtze River Bridge. The largest pier depth of these bridges monly called the Morison equation) for slender structures. Later,
Penzien and Kaul (1972) modified the Morison equation to ana-
1
Associate Professor, School of Civil Engineering, Beijing Jiaotong Univ., lyze the earthquake-induced hydrodynamic pressure. Despite the
Beijing 100044, People’s Republic of China (corresponding author). ORCID: lack of theoretical basis, added mass is still a widely used concept
http://orcid.org/0000-0003-2223-0088. E-mail: jianghui@bjtu.edu.cn because of its simplicity (Park et al. 2011). Goto and Toki (1965)
2
Assistant Engineer, CSCEC AECOM Consultants Co., Ltd., Lanzhou also developed a method to transform the hydrodynamic forces
730030, People’s Republic of China. E-mail: wbx0623@bjtu.edu.cn
3 during an earthquake into an added mass. Liaw and Chopra (1974)
Graduate Student, School of Civil Engineering, Beijing Jiaotong
Univ., Beijing 100044, People’s Republic of China. E-mail: 14121174@ developed analytical expressions for the hydrodynamic pressure
bjtu.edu.cn of cantilever towers in water, considering the surface-wave and
4
Doctoral Candidate, School of Civil Engineering, Beijing Jiaotong water-compressibility effects. Tanaka and Hudspeth (1988),
Univ., Beijing 100044, People’s Republic of China. E-mail: 14121209@ Huang and Li (2011), and Li and Yang (2013) showed that the sur-
bjtu.edu.cn face waves and water compressibility could be of little conse-
5
Associate Professor, School of Civil Engineering, Tianjin Chengjian quence and therefore ignored when analyzing the dynamic
Univ., Tianjin 300384, People’s Republic of China. E-mail: zhhuidong@ response of slender cylinders surrounded by water in earthquakes.
126.com
Goyal and Chopra (1989) discussed the hydrodynamic pressure on
Note. This manuscript was submitted on May 31, 2016; approved on
December 1, 2016; published online on March 3, 2017. Discussion pe- towers of arbitrary geometry but with two axes of plane symmetry,
riod open until August 3, 2017; separate discussions must be submitted and the effects of the tower–foundation–soil interaction under
for individual papers. This paper is part of the Journal of Bridge earthquakes were included. Bhatta and Rahman (2003) analyzed
Engineering, © ASCE, ISSN 1084-0702. the wave forces by scattering and radiation for a submerged
formulas for the hydrodynamic pressures of circular hollow piers Radiation wave theory is an analytical method for solving the
and solid piers, respectively. Wei et al. (2015) originally devel- hydrodynamic pressure, and lateral vibration equations of beams
oped simplified procedures for the efficient seismic design and are used to describe the lateral deformation of cylinders submerged
analysis of water-surrounded composite axisymmetric structures, in water. To have a closed-form solution for the equation, the bridge
and effects such as higher-order vibration modes, water compres- cylinder has to be deemed as a cantilever, with the bottom fixed and
sibility, flexibility of underlying soil foundation, and weight of the the other end set free. These assumptions make it hard to consider
superstructure could be effectively taken into account. other factors simultaneously, such as the weight of the superstruc-
The hydrodynamic pressure expressions based on the radiation ture and environmental constraints, such as soil stiffness. Therefore,
wave theory are complex because they include special functions, the circular solid pier in deep water can be assumed as a flexible cyl-
integrals, and similar terms. To solve this problem, custom-made inder partially submerged in the water, with the bottom end fixed.
programs developed from specific mathematical software are gener- As shown in Fig. 1, the water depth is h (m), and the diameter and
ally in need of massive calculation, thus they are difficult to utilize radius of the pier are D (m) and a (m), respectively. Both the cylin-
in practical applications. Therefore, a simplified extraction to both drical coordinate system (o-r-u -z) and Cartesian coordinate system
accurately and concisely calculate the hydrodynamic effects on (o-x-y-z) are introduced, with the origin of the coordinate located at
deepwater piers is highly desirable. In addition, the representative the center of the pier bottom. Note that the z-axis is vertically
simplified methods for the determination of hydrodynamic pres- upward from the rigid ground, whereas the r-axis measures radially
sure, such as the Morison equation (Morison et al. 1950) and the from the z-axis, and the u -axis from the positive x-axis. The earth-
formula proposed by Goto and Toki 1965), are semiempirical for- quake excitation propagates along the x-axis, and therefore, the pier
mulas, and their application scopes need further investigation. can be characterized by translational motion in the x-axis as well as
Further, Ju and Zeng (1983) and Lai (2004) found that the total rotational motion about the y-axis.
velocity potential under the linear radiation wave could be divided The work by Liaw and Chopra (1974) showed that, when the
into two parts as the velocity potentials from rigid motion and from structural fundamental frequency is not very low and is far from the
elastic vibration. Li and Yang (2013) reported that the added masses wave period, the effect of the surface gravity waves and the water
caused by elastic vibration are approximately equal to that caused compressibility would be small enough to be ignored. This was
by the rigid motion for the submerged cylinders. These outstanding verified by Tanaka and Hudspeth (1988), Huang and Li (2011), and
works provide important enlightenment and serve as the foundation Li and Yang (2013). Therefore, water is assumed incompressible
for the simplification of hydrodynamic pressure. Taking the radia- and inviscid with irrotational motion, and surface waves are ignored
tion wave theory as the basis, this paper develops a simplified for- in the subsequent analysis. Before the occurrence of earthquake ex-
mula for calculating the hydrodynamic pressure on cylindrical piers citation, the water surface remains flat, and a single radiation wave
z y
r
θ
D =2a
Pier x
o x
Rigid ground
Earthquake wave
U ðr; u ; z; tÞ ¼ f ðr; u ; zÞeiwt (1) a large number of units, and then the higher vibration modes of
deepwater piers can be well considered. Therefore, it is feasible to
where f (r, u , z) = spatial factor of the velocity potential; and t (s) substitute MR for ME , thus Eq. (7) can be rewritten as
denotes the time.
The governing equation for the water can be written in terms of ðM þ MR Þ€x þ Cx_ þ Kx ¼ ðM þ MR Þ€x g (8)
its hydrodynamic pressure in cylindrical coordinates. To satisfy
Laplace’s equation, it becomes Now Eq. (8) can be solved with the help of standard finite-element
software because the added masses on both sides remain the same. In
∂2 f 1 ∂ f 1 ∂ f ∂2 f
þ þ 2 2þ 2 ¼0 (2) addition, the effects of other factors, such as the weight of the super-
∂r 2 r ∂r r ∂u ∂z structure and soil–structure interactions, can be considered by using
the semianalytical and seminumerical approach, with a refined finite-
For a linear radiation wave, the total velocity potential of the element model of the entire bridge structure.
radiation wave [U (r, u , z, t)] can be divided into two parts: the ve- The hydrodynamic pressure applied on a cylindrical pier can
locity potential from rigid motion [U R(r, u , z, t) = f R(r, u , z, t)eiwt] now be represented by the added mass (MiR ) of node i caused by
and from elastic vibration [U E(r, u , z, t) = f E(r, u , z, t)eiwt] (Ju and rigid motion (Fig. 2) (Li and Yang 2013). For the sake of brevity,
Zeng 1983; Lai 2004) the detailed derivation procedures have been omitted
Ð
U ¼ UR þ UE (3) X1 K ðk aÞsinðk hÞ
1 n n Li cosðkn zÞdz
Mi ¼ 2 r ap
R
(9)
n¼1
K1 ðkn aÞkn2 h
0
added mass coefficient. The added mass coefficient related to tions are not reasonable in actual engineering projects, and the pur-
rigid motion (CaRi ) can be calculated as shown in Eq. (10), and pose of this paper is to make sure that the achieved simplified
the corresponding coefficient related to elastic vibration (CaEi ) method is widely applicable by means of constructing a broad-
can also be deduced based on the radiation wave theory. To scope pier database. As previously mentioned, the water depth (h)
avoid a tedious pile-up of equations, the detailed derivation pro- around the bridge pier is assumed equal to the pier height (H) in the
cedures are not listed here, but can be inferred from the work of process of curve fitting.
Lai (2004). The chi-square minimization algorithm is used to fit the data.
Six groups of piers with heights of H = 20, 30, 60, 90, 120, and The best-matching function is found by minimizing the square of
150 m are selected. The water depth (h) is set equal to H for all cases the errors. The fitting process consists of the following procedures:
here for the convenience of comparison. Coefficients CaRi and CaEi Step 1: The added mass coefficients of piers listed in Table 1 are
for different piers are calculated with MATLAB software, and their first calculated by MATLAB software without any simplification
distribution curves as functions of h and D/h are shown in Fig. 3. using Eq. (14). According to the distribution of Cai , the first curve
The subscripts RM and EV represent rigid motion and elastic vibra- fitting is performed using the data statistics software OriginPro 9.0.
tion, respectively. The figure shows that the CaRi and CaEi curves are Because the distribution shape of Cai is very close to that of the
highly similar to each other over the whole water depth, which con- built-in regressive function MnMolecular in OriginPro 9.0, this
sequently verifies the consistency of the added mass caused by rigid function is then chosen for the first step of regression [Eq. (15)]. As
motion and that caused by elastic vibration. Obviously, the calcula- is known, each of the intermediate parameters used in the data
tion of the added mass caused by elastic vibration is much more regression have no clear physical significance
complex and time-consuming than that caused by rigid motion.
y ¼ A1 ½1 ekðzi A2 Þ
Therefore, it is meaningful to make the substitution as described in
Eq. (8). A1 ¼ f ða; hÞ
However, because of the special functions and integrals included
A2 ¼ gða; hÞ; k ¼ sða; hÞ (15)
in Eq. (10), specialized computing abilities are required, which
inevitably limit the engineering application of radiation wave
where A1, k, and A2 = constants to be determined; zi = height of node
theory. Given that
i, ranging from zero to h; and f, g, and s = functions of variables a
MiR ¼ Miwater Cai (11) and h. The result of this step implies that A2 is almost equivalent to
the water depth (h), and therefore, A2 is set equal to h.
Step 2: Parameters A1 and k are fitted according to the water
where Miwater can be defined as depth (h). Based on the distribution of A1 and k along the water
depth for different pier radii (Fig. 4), the fitting functions are chosen
Miwater ¼ r p a2 (12)
to conduct the second round of data fitting
then the added mass coefficient (Cai Þ of the hydrodynamic force is P1 h1:5
A1 ¼
ð P2 þ h1:5
2X 1
1 K1 ðkn aÞ
Cai ¼ sinðkn hÞ cosðkn zÞdz (13) k ¼ k1 =h þ k2 (16)
a n¼1 kn2 h K10 ðkn aÞ Li
where P1, P2, k1, and k2 = functions of the pier radius (a). After this
Eq. (12) defines the mass of water whose volume is equiva- step, the values of P1 and k1 are coincidentally close to 1.0 and 2.0
lent to that of a circular pier with unit height. kn is the wave num- by data regression. The correlation coefficients for P1 and k1 signify
ber, and kn equals ð2n 1Þp =2h (n = 1, 2…) according to the the reliable regression process as shown in Table 2.
boundary conditions. Substituting kn into Eq. (13) results in the Step 3: Parameters P2 and k2 are fitted according to the pier ra-
simplified expression of the added mass coefficient caused by dius (a). In view of the distribution curves of P2 and k2 along the
rigid motion pier radius (Fig. 5), the fitting functions are chosen to conduct the
third round of data fitting
K1 ð2n 1Þp a ð2n 1Þp ð2n 1Þp z P2 ¼ l1 þ l2 a þ l3 a2
sin sin
8h X 1
2h 2 2h Li k2 ¼ l4 =a (17)
Cai ¼
ap 2 n¼1 ð2n 1Þ2 K 0 ð2n 1Þp a=ð2hÞ
1
where l1, l2, and l3 in the parabolic function = functions of variable a;
(14) and parameter l4 = coefficient of the inverse function of variable a.
Fig. 3. Comparison of CaRi and CaEi : (a) h = 20 m; (b) h = 30 m; (c) h = 60 m; (d) h = 90 m; (e) h = 120 m; (f) h = 150 m
11 22–200
10 0.9999 0.9960
12 24–200
11 0.9999 0.9964
13 26–200
12 0.9999 0.9968
14 28–200
13 0.9999 0.9972
15 30–200
14 0.9998 0.9957
15 0.9999 0.9978
12
10
8
6
4
2
0
0 2 4 6 8 10 12 14 16
pier radius (m)
(a)
1.0
0.8
0.6
k2
0.4
0.2
0.0
0 2 4 6 8 10 12 14 16
pier radius(m)
(b)
Mi ¼ Miwater Cai
Fig. 4. Distribution curves of A1 and k along water depth with different
Miwater ¼ r p a2
pier radiuses
h1:5 2a þ h
Cai ¼ 1 exp ðz i h Þ
Step 4: A simplified formula can be arrived at to calculate the ð0:1a2 þ 1:9a 1:7Þ þ h1:5 ah
hydrodynamic pressure of circular piers submerged in deep water, (18)
as shown in Eq. (18), which takes the water height (h) and the pier
radius (a) as the principal parameters. It should be noted that, if dif- The accuracy and convenience of the regression are the key indi-
ferent fitting regressive functions are used, different fitting expres- cators of its applicability. For least-squares nonlinear curve fitting,
sions might be achieved the regression quality can be evaluated by using the standard
Fig. 6. Comparison of added mass coefficients (Ca ) by Eqs. (18) and (14): (a) h = 10 m; (b) h = 20 m; (c) h = 30 m; (d) h = 60 m; (e) h = 90 m; (f) h =
120 m
r p D2 1
Test Based on Radiation Wave Theory FP ¼ Ca €x ðtÞ CD r D_x ðtÞj_x ðtÞj (20)
4 2
Test of Simplified Formula where r , Ca; and CD = mass density of water, inertial-force coeffi-
Downloaded from ascelibrary.org by Beijing Jiaotong University on 09/04/18. Copyright ASCE. For personal use only; all rights reserved.
Fig. 8. Comparison of added mass coefficients: (a) h = 20 m; (b) h = 30 m; (c) h = 60 m; (d) h = 90 m; (e) h = 120 m; (f) h = 150 m
lf
Fluid-structure interface
ls
Solid element
Downloaded from ascelibrary.org by Beijing Jiaotong University on 09/04/18. Copyright ASCE. For personal use only; all rights reserved.
(a)
B B
Free surface interface element Free surface interface element
∞ ∞
∞ ∞
∞ ∞
∞ ∞
Infinite interface element
Fig. 10. Finite-element model of piers and surrounding water: (a) discretization of elements; (b) finite-element model
Table 3. Comparison of Vibration Property under Different Water motions are characterized by significant long-period pulses in both
Widths the acceleration and velocity time histories.
Frequency (Hz)
Setting of Harmonic Loads
Water width [B (m)] First order Third order Fifth order Sixth order
According to Table 4, the predominant periods (T) of near-fault and
2.5 (1a) 1.561 10.03 12.13 17.34 far-field earthquake waves are close to 0.7 and 0.4 s, respectively.
7.5 (3a) 1.590 10.40 12.27 17.33 Accordingly, two harmonic excitations are defined using the formula
12.5 (5a) 1.599 10.55 12.33 17.32 y ¼ sin ð2p ftÞ=5, with f = 1.43 Hz (T = 0.7 s) and f = 2.5 Hz (T =
17.5 (7a) 1.601 10.59 12.35 17.31 0.4 s), respectively. The duration time is set to 5T, which is 3.5 and
25.0 (10a) 1.602 10.60 12.35 17.31 2.0 s for the two harmonic excitations. The peak acceleration is also
27.5 (11a) 1.602 10.61 12.36 17.31 normalized to 0.2g. The acceleration time histories and response
32.5 (13a) 1.602 10.61 12.36 17.31 spectra of the two harmonic loads are shown in Fig. 13. Clearly, the
37.5 (15a) 1.602 10.61 12.36 17.31 harmonic load can be regarded as relative low-frequency earthquake
waves with several pulses.
Selected Earthquake Ground Motions
The earthquake ground motion is a complex random vibration
Test of Formulas by Vibration Properties
process where significant difference exists in spectrum character-
istics between near-fault and far-field earthquake ground motions. To assess the accuracy of the methods in calculating the fundamental
To test the accuracy and applicability of the proposed method under
vibration period of the piers, results of the three added mass methods
different types of earthquakes, six representative near-fault and far-
are compared with that of the potential-based finite-element method
field earthquake records are selected as inputs. These records are
using the equation
obtained from the Consortium of Organizations for Strong-Motion
Observation Systems (COSMOS 2013) and the Pacific Earthquake TAM TPBF
RT ¼ 100% (23)
Engineering Research Center Ground Motion Database (PEER TPBF
2013). In Table 4, MW is the moment magnitude of an earthquake
event, and PGA and PGV are the peak ground acceleration and peak where RT = error of the fundamental periods calculated by the three
ground velocity, respectively. All of the records are normalized by added mass methods (the proposed simplified formula, the Morison
having the same PGA equal to 0.20g. Figs. 11 and 12 show the accel- equation, and the Goto-Toki formula); TAM = fundamental natural
eration and velocity time histories of the fault-normal component of periods determined from these methods; and TPBF = fundamental
the near-fault and far-field ground motions. The near-fault ground period given by the potential-based finite-element method.
0.2 45
35
0.1 25
15
0.0
5
-5
-0.1
-15
-0.2 -25
0 5 10 15 20 0 4 8 12 16 20
(a) (b)
0.2 40
30
0.1 20
10
0
0.0
-10
-20
-0.1 -30
-40
-0.2 -50
0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40
(c) (d)
0.2 30
20
0.1
10
0.0 0
-10
-0.1
-20
-0.2 -30
0 4 8 12 16 20 24 28 32 36 0 4 8 12 16 20 24 28 32 36
(e) (f)
Fig. 11. Near-fault ground motions: (a) RRS228 acceleration; (b) RRS228 velocity; (c) H-EMO000 acceleration; (d) H-EMO000 velocity;
(e) H-E07140 acceleration; (f) H-E07140 velocity
Table 5 shows comparisons of the fundamental vibration periods agreement with that of the potential-based finite-element method.
under the four methods, and Fig. 14 illustrates the error distribution The maximum error is less than 0.9%, which demonstrates the va-
curves of the different methods. Clearly, the fundamental periods lidity of the proposed simplified method for a wide range of piers.
achieved by the proposed simplified formulas are in very close For the Morison equation, RT is less than 2% when D/H is less than
0.00 0
-0.05
-5
-0.10
-0.15 -10
0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35
Downloaded from ascelibrary.org by Beijing Jiaotong University on 09/04/18. Copyright ASCE. For personal use only; all rights reserved.
(a) (b)
0.2 20
15
0.1 10
5
0.0 0
-5
-10
-0.1
-15
-20
-0.2
0 4 8 12 16 20 24 28 32 0 4 8 12 16 20 24 28 32
(c) (d)
0.2 20
15
0.1 10
5
0.0 0
-5
-0.1 -10
-15
-0.2 -20
0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35
(e) (f)
Fig. 12. Far-field ground motions: (a) BAD000 acceleration; (b) BAD000 velocity; (c) H-C06000 acceleration; (d) H-C06000 velocity; (e) BRC000
acceleration; (f) BRC000 velocity
0.1. As D/H increases, the error reaches its maximum value of of 1.5 m and a height of 30 m as an example, the dynamic
nearly 8%. For the Goto-Toki formula, RT has a maximum value of responses [moment (M), shear force (Q), and displacement
approximately 2.2% for D/H = 0.01. The accuracy of the Goto-Toki (d)] under the record of H-EMO000 are shown in Fig. 15. It
formula improves with an increase in D/H, reaching good agree- reveals that the results from the proposed simplified method
ment at D/H = 1.0. These results show that, for the calculation of clearly overlap with those from the potential-based fluid-
vibration property, the Morison equation is most reliable for slender element method.
piers, whereas the Goto-Toki formula is more appropriate for piers
with a larger D/H (such as 0.5 and above), which is consistent with Peak Values of Dynamic Response of the Piers
Fig. 9. under Earthquakes
For the four calculation methods of the hydrodynamic force, the
Test of Simplified Formulas by Dynamic Responses relationships between the distribution curves of the peak dynamic
response and D/H are shown in Figs. 16 and 17. NFE and FFE
It is also necessary to assess the proposed formula with the use refer to near-fault earthquake motions and far-field earthquake
of dynamic-response analysis. Taking the pier with a diameter motions, respectively. For piers with heights of 30 and 60 m, the
3 3
0.0 MAMi MPBFi
3 i¼1 i¼1
-0.1 RME ¼ 100%
1X 3
-0.2 MPBFi
3 i¼1
-0.3
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
(a)
9
0.3 1
y = sin(2π ft )
0.2 5 6
0.1
3
0.0
–0.1 0
–0.2
-3
–0.3 0.0 0.2 0.4 0.6 0.8 1.0
0.0 0.5 1.0 1.5 2.0
(a)
(b)
8
1.8 7
6
1.5
5
1.2 4
3
0.9 2
1
0.6 0
-1
0.3
-2
0.0 -3
0 1 2 3 4 0.0 0.2 0.4 0.6 0.8 1.0
(c) (b)
Fig. 13. Time histories and acceleration spectra of harmonic loads: Fig. 14. Comparison of errors for fundamental natural periods: (a) H =
(a) T = 0.7 s; (b) T = 0.4 s; (c) acceleration spectra 30 m; (b) H = 60 m
2 2750
2250
MN·m
MN·m
0
1750
-2
1250
-4 750
-6 250
0 5 10 15 20 25 30 35 40
-250
(a) 0.0 0.2 0.4 0.6 0.8 1.0
Downloaded from ascelibrary.org by Beijing Jiaotong University on 09/04/18. Copyright ASCE. For personal use only; all rights reserved.
0.3 (a)
0.2 200
(MN)
0.1
150
0.0
(MN)
-0.1 100
-0.2
50
-0.3
0 5 10 15 20 25 30 35 40
(b) 0
20
5
0
15
(cm)
-5
-10 10
-15
0 5 10 15 20 25 30 35 40
5
(c)
Fig. 15. Time histories of (a) moment; (b) shear force; (c) displace-
0
ment of pier (D1.5-H30) from H-EMO000
0.0 0.2 0.4 0.6 0.8 1.0
(c)
!
1 X
3 X
3
QAM i QPBFi Fig. 16. Distribution of peak values of pier (H = 30 m): (a) moment;
3 i¼1 i¼1 (b) shear force; (c) displacement
RQE ¼ 100%
1X 3
QPBFi
3 i¼1 or far-field earthquake motions; and MPBFi, QPBFi, and dPBFi = peak
! values of the bending moment, shear force, and displacement from
1 X3 X3
the potential-based fluid method under the ith earthquake motions.
dAMi dPBFi
3 i¼1 i¼1
Tables 6 and 7 show the errors of the dynamic responses of the
RdE ¼ 100% (24) piers of H = 30 m and H = 60 m under earthquakes, respectively.
1X 3
The dynamic response from the proposed simplified formula is in
dPBFi
3 i¼1 very close agreement with that from the potential-based finite-
element method. The error of the simplified expression is less
where RM-E, RQ-E, and Rd-E = errors of the bending moment, shear than 1% when D/H = 1.0, although it is not a practical size for
force, and displacement, respectively, for the three added mass bridge engineering. These results demonstrate that the proposed
methods under earthquake motions; MAMi, QAMi, and dAMi = peak simplified formula is valid for the dynamic analysis of deepwater
values of the bending moment, shear force, and displacement from piers with a wide range of dimensional proportions. In Tables 6
the added mass methods under the ith (i = 1, 2, 3) selected near-fault and 7, the Morison equation gives the most conservative
where RM-H, RQ-H, and Rd-H = errors of the bending moment, shear
force, and displacement, respectively, for the three added mass meth-
Downloaded from ascelibrary.org by Beijing Jiaotong University on 09/04/18. Copyright ASCE. For personal use only; all rights reserved.
ods under harmonic loads; MAM, QAM, and dAM = peak values of the
bending moment, shear force, and displacement, respectively, from
the three added mass methods; and MPBF, QPBF, and dPBF = peak val-
ues of the bending moment, shear force, and displacement, respec-
tively, from the potential-based fluid method. The errors for piers with
heights of 30 and 60 m are calculated to yield the following results.
Tables 8 and 9 show the errors of the dynamic response of the
piers of H = 30 m (T = 0.4 s) and H = 60 m (T = 0.7 s), respectively,
under the function of harmonic excitations. The results show that the
simplified formula produces a stable trend with a small error of less
than 3.0%. Although the predominant period (T) of the harmonic
load may affect the dynamic response of the piers, it does not appear
to disturb the accuracy of the proposed simplified formula. Thus, the
proposed formula is applicable to dynamic-response analysis under a
wide range of loads. For the Morison equation, the dynamic response
is greater than that from the potential-based fluid method, and the
error is larger for higher D/H ratios. The calculated values given by
the Goto-Toki formula are less than those by the potential-based fluid
method. Similar to the piers excited under earthquakes, the results are
in agreement with those from the potential-based method only for a
large D/H, with a corresponding error of less than 0.5%. The afore-
mentioned conclusions are consistent with those obtained from Fig. 9
and Tables 6 and 7. In brief, comparisons and analyses indicate that
the proposed simplified method is valid for describing the hydrody-
namic pressure of submerged piers with a wide range of geometries.
Therefore, finite-element software can be directly used to analyze the
vibration properties and dynamic responses of deepwater bridges
without the need for custom-made programs.
Fig. 17. Distribution of peak values of pier (H = 60 m): (a) moment;
(b) shear force; (c) displacement
Summary and Conclusions
predictions among the three methods. The error of the Morison The expressions of hydrodynamic pressure based on the radiation
equation increases with D/H. The dynamic responses calculated wave theory have remained a significant challenge for the acquisi-
by the Goto-Toki formula are smaller than those calculated by tion of computational solutions due to their complexity. This paper
the potential-based fluid method, especially for piers with a D/H developed and proposed a simplified formula obtained through data
less than 0.5. The results are close to those of the potential-based fitting for cylindrical bridge piers. The accuracy of the simplified
method only for D/H near 1.0. These trends are consistent with formula was evaluated, and the applicability of the Morison equation
the results observed from the added mass coefficient analysis and Goto-Toki formula was also systematically explored. Based on
and the structural modal analysis. As shown in Tables 6 and 7, the results presented and discussed herein, the following conclusions
the distribution characters of errors under near-fault and far-field can be drawn:
earthquakes are identical. 1. Based on the substitution of the added mass from elastic vibra-
tion for that from rigid motion, the equivalent added mass of
the hydrodynamic pressure from rigid motion is simplified by
data fitting, resulting in a simplified expression without the
Errors of Dynamic Response of Piers under
inclusion of special functions or integrals.
Harmonic Loads
2. The accuracy of the proposed simplified formula is evaluated
The applicability of the three simplified methods is examined by by conducting an analysis of the added mass coefficient, pier
comparing the dynamic responses of the simplified methods with vibration property, and dynamic responses. The calculated
that of the potential-based fluid method under the function of two added mass has desirable agreement with that from the
T = 0.4 s T = 0.7 s
Error D/H Proposed expression Morison equation Goto-Toki formula Proposed expression Morison equation Goto-Toki formula
RM-H (%) 0.05 0.76 0.96 –1.91 0.45 0.98 –3.28
0.10 1.32 1.72 –3.16 1.04 5.96 –8.46
0.20 1.60 22.11 –7.29 0.46 7.76 –3.65
0.50 0.02 9.29 –1.77 1.10 14.03 –0.40
1.00 0.29 17.16 0.45 0.02 9.26 0.46
RQ-H (%) 0.05 1.36 1.33 –3.02 1.36 1.33 –3.02
0.10 0.56 3.50 –8.06 0.56 3.50 –8.06
0.20 0.50 7.94 –3.83 0.50 7.94 –3.83
0.50 0.95 11.69 –0.99 0.95 11.69 –0.99
1.00 0.09 10.54 0.03 0.09 10.54 0.03
Rd-H (%) 0.05 0.46 0.67 –2.64 0.41 1.35 0.17
0.10 –0.26 –1.65 –0.93 0.97 6.46 –8.33
0.20 2.62 26.05 –6.79 0.38 7.30 –3.38
0.50 0.04 11.35 –1.01 0.70 14.03 –0.50
1.00 –0.80 18.61 0.05 0.72 15.44 0.17
T = 0.4 s T = 0.7 s
Error D/H Proposed expression Morison equation Goto-Toki formula Proposed expression Morison equation Goto-Toki formula
RM-H (%) 0.05 –1.52 –0.85 2.38 2.00 3.35 –6.93
0.10 1.84 4.99 –4.10 0.09 8.79 –3.09
0.20 –1.32 –1.91 –1.73 0.66 9.18 –4.67
0.50 1.84 26.40 0.20 0.49 11.11 –0.78
1.00 0.09 14.84 0.58 0.12 18.32 0.93
RQ-H (%) 0.05 –1.99 –3.51 6.56 2.28 3.00 –6.93
0.10 1.20 5.92 –4.56 1.11 4.76 –3.12
0.20 1.51 3.08 –2.51 0.25 6.70 –5.31
0.50 0.68 22.91 –1.25 0.31 8.84 –1.40
Downloaded from ascelibrary.org by Beijing Jiaotong University on 09/04/18. Copyright ASCE. For personal use only; all rights reserved.
coupled with the liquid, Earthquake Press, Beijing, 1–9 (in Chinese).
Lai, W. (2004). “Dynamic response study of bridge in deep water under pressure on axisymmetric offshore structures.” Earthquake Eng. Struct.
earthquake and wave action.” D.E. thesis, Tongji Univ., Shanghai, Dyn., 20(5), 429–440.
China, 20–48 (in Chinese). Tanaka, Y., and Hudspeth, R. T. (1988). “Restoring forces on vertical circu-
Lai, W., Wang, J. J., and Hu, S. D. (2004). “Earthquake induced hydrody- lar cylinders forced by earthquakes.” Earthquake Eng. Struct. Dyn.,
namic pressure on bridge pier.” J. Tongji Univ., 32(1), 1–5 (in Chinese). 16(1), 99–119.
Li, Q., and Yang, W. L. (2013). “An improved method of hydrodynamic Wei, K., Bouaanani, N., and Yuan, W. C. (2015). “Simplified methods for
pressure calculation for circular hollow piers in deep water under earth- efficient seismic design and analysis of water-surrounded composite
quake.” Ocean Eng., 72(1), 241–256. axisymmetric structures.” Ocean Eng., 104, 617–638.
Liaw, C. Y., and Chopra, A. K. (1974). “Dynamics of towers surrounded by Wei, K., Yuan, W. C., and Bouaanani, N. (2013). “Experimental and nu-
water.” Earthquake Eng. Struct. Dyn., 3(1), 33–49. merical assessment of the three dimensional modal dynamic response of
Liu, S. X., Li, Y. C., and Li, G. W. (2007). “Wave current forces on the pile bridge pile foundations submerged in water.” J. Bridge Eng., 10.1061
group of base foundation for the east sea bridge, China.” J. Hydrodyn., /(ASCE)BE.1943-5592.0000442, 1032–1041.
19(6), 661–670. Westergaard, H. M. (1933). “Water pressures on dams during earthquakes.”
Liu, Z. Y., Li, Q., Zhao, C. H., and Zhuang, W. L. (2008). “Earthquake Trans. ASCE, 98(2), 418–433.
induced added hydrodynamic pressure on circular hollow piers in deep Zhou, D., and Liu, W. Q. (2007). “Bending-torsion vibration of a partially
water.” J. Southwest Jiaotong Univ., 43(2), 200–205 (in Chinese). submerged cylinder with an arbitrary cross section.” Appl. Math. Model,
MATLAB [Computer software]. MathWorks, Natick, MA. 31(10), 2249–2265.