Dynamic soil–structure interaction analysis via coupled finite-element–boundary-element method

In this paper, a study on the transient response of an elastic structure embedded in a homogeneous, isotropic and linearly elastic half-plane is presented. Transient dynamic and seismic forces are considered in the analysis. The numerical method employed is the coupled Finite-Element–Boundary-Element technique (FE–BE). The finite element method (FEM) is used for discretization of the near field and the boundary element method (BEM) is employed to model the semi-infinite far field. These two methods are coupled through equilibrium and compatibility conditions at the soil–structure interface. Effects of non-zero initial conditions due to the pre-dynamic loads and/or self-weight of the structure are included in the transient boundary element formulation. Hence, it is possible to analyse practical cases (such as dam–foundation systems) involving initial conditions due to the pre-seismic loads such as water pressure and self-weight of the dam. As an application of the proposed formulation, a gravity dam has been analysed and the results for different foundation stiffness are presented. The results of the analysis indicate the importance of including the foundation stiffness and thus the dam–foundation interaction.     

Stochastic dynamic response of dam–reservoir–foundation systems to spatially varying earthquake ground motions

In this paper, stochastic dynamic responses of dam–reservoir–foundation systems subjected to spatially varying earthquake ground motions are investigated using the displacement-based fluid finite elements. For this purpose, variable-number-node two-dimensional (2D) fluid finite elements based on the Lagrangian approach is programmed in FORTRAN language and incorporated into a computer program SVEM, which is used for stochastic dynamic analysis of solid systems subjected to spatially varying earthquake ground motion. The spatially varying earthquake ground motion model includes incoherence, wave-passage and site-response effects. The incoherence effect is examined by considering the Harichandran and Vanmarcke coherency model. The effect of the wave passage is investigated by using various wave velocities. Homogeneous medium and firm soil types are selected for considering the site-response effect where the foundation supports are constructed. The Sar?yar concrete gravity dam, constructed in Turkey is selected for numerical example. The ground motion is described by filtered white noise and applied to each support point of the 2D finite element model of the dam–reservoir–foundation system. The record of Kocaeli earthquake in 1999 is used in the analyses. Displacements, stresses and hydrodynamic pressures occurring on the upstream face of the dam are calculated for four cases. It is concluded that spatially varying earthquake ground motions have important effects on the stochastic dynamic response of dam–reservoir–foundation systems.     

A continuum damage concrete model for earthquake analysis of concrete gravity dam–reservoir systems

In this study, the earthquake damage response of the concrete gravity dams is investigated with considering the effects of dam–reservoir interaction. A continuum damage model which is a second-order tensor and includes the strain softening behavior is selected for the concrete material. The mesh-dependent hardening technique is adopted such that the fracture energy dissipated is not affected by the finite element mesh size. The dynamic equilibrium equations of motion are solved by using the improved form of the HHT-α time integration algorithm. Two dimensional seismic analysis of Koyna gravity dam is performed by using the 1967 Koyna earthquake records. The effects of damage on the earthquake response of concrete gravity dams are discussed. Comparison of the Westergaard and Lagrangian dam–reservoir interaction solutions is made. The effects of viscous damping ratio on the damage response of the dam are also studied.     

Nonlinear seismic response analysis of arch dam-foundation systems- part I dam-foundation rock interaction

The arch dam–foundation rock dynamic interaction and the nonlinear opening and closing effects of contact joints on arch dam are important to the seismic response analysis of arch dams. Up to date, there is not yet a reasonable and rigorous procedure including the two factors in seismic response analysis. The methods for the analysis of arch dam–foundation rock dynamic interaction in frequency domain are not suitable to the problem with nonlinear behaviors, in this paper, so an analysis method in time domain is proposed by combining the explicit finite element method and the transmitting boundary, and the dynamic relaxation technique is adopted to obtain the initial static response for dynamic analysis. Moreover, the influence of arch dam–foundation dynamic interaction with energy dispersion on seismic response of designed Xiaowan arch dam in China is studied by comparing the results of the proposed method and the conventional method with the massless foundation, and the local material nonlinear and nonhomogeneous behaviors of foundation rock are also considered. The reservoir water effect is assumed as Westergaard added mass model in calculation. The influence of the closing–opening effects of contact joints of arch dam on the seismic response will be studied in another paper.     

In-plane foundation-soil interaction for embedded circular foundations

Foundation soil interaction is studied using an analytical two-dimensional model, for circular foundations embedded in a homogeneous elastic half-space and for incident plane P- and SV- and for surface Rayleigh waves. The scattered waves are expanded in complete series of cyclindrical wave functions. A detailed analysis is presented of the foundation response to unit amplitude incident waves as a function of the type of incident waves and angle of incidence, the depth of the embedment and the foundation mass per unit length.It is shown that free-field translations and point rotation approximate well the foundation input motion only for very long incident waves. For shorter incident waves, those in general overestimate the foundation input motion. Neglecting the rotation of the foundation input motion (which is usually done in practice) may eliminate a major contribution to the base excitation of buildings and may cause nonconservative estimates of the forces in these buildings. Incident waves appear as ‘longer’ to a shallow foundation than to a deeper foundation. Therefore, deeper foundations are more effective in reflecting and scattering the short incident waves.     

Effects of rainfall on soil–structure system frequency: Examples based on poroelasticity and a comparison with full-scale measurements

In this paper, a simple two-dimensional soil–structure interaction model, based on Biot's theory of wave propagation in fluid saturated porous media, is used to explain the observed increase of the apparent frequencies of Millikan library in Pasadena, California, during heavy rainfall and recovery within days after the rain. These variations have been measured for small amplitude response (to microtremors and wind excitation), for which Biot's linear theory is valid. The postulated hypothesis is that the observed increases in frequency are due to the water saturation of the soil. The theoretical model used to explore this hypothesis consists of a shear wall supported by a circular foundation embedded in a poroelastic half-space. This rigid foundation model may be appropriate only for the NS response of Millikan library. This paper presents results for the foundation stiffness, and for the system response for model parameters similar to those for Millikan library (located on alluvium with shear wave velocity of about 300 m/s). The foundation impedance matrix, foundation input motion and system response are compared for dry and fully saturated half-space, with permeable and impermeable foundation. The results show that for embedded foundations, the effects of saturation on the horizontal foundation stiffness are as significant as for the vertical stiffness, contrary to what has been known for surface foundations investigated by other authors. Further, the results suggest a 1–2% increase in system frequency of the first two modes of vibration, depending on the drainage condition along the foundation–soil interface. Such increases agree qualitatively with the observations.     

A hybrid distinct element–boundary element approach for seismic analysis of cracked concrete gravity dam–reservoir systems

This paper proposes a new algorithm for modeling the nonlinear seismic behavior of fractured concrete gravity dams considering dam–reservoir interaction effects. In this algorithm, the cracked concrete gravity dam is modeled by distinct element (DE) method, which has been widely used for the analysis of blocky media. Dynamic response of the reservoir is obtained using boundary element (BE) method. Formulation and various computational aspects of the proposed staggered hybrid approach are thoroughly discussed. To the authors' knowledge, this is the first study of a hybrid DE–BE approach for seismic analysis of cracked gravity dam–reservoir systems. The validity of the algorithm is discussed by developing a two-dimensional computer code and comparing results obtained from the proposed hybrid DE–BE approach with those reported in the literature. For this purpose, a few problems of seismic excitations in frequency- and time-domains, are presented using the proposed approach. Present results agree well with the results from other numerical methods. Furthermore, the cracked Koyna Dam is analyzed, including dam–reservoir interaction effects with focus on the nonlinear behavior due to its top profile crack. Results of the present study are compared to available results in the literature in which the dam–reservoir interaction were simplified by added masses. It is shown that the nonlinear analysis that includes dam–reservoir interaction gives downstream sliding and rocking response patterns that are somehow different from that of the case when the dam–reservoir interaction is approximated employing added masses.     

Some analytical results on lateral dynamic stiffness for footings supported on hysteretic soil medium

The nonlinearity of the soil affects soil–structure interaction to a considerable extent. For a reliable and safe analysis of soil interaction effects on the dynamic response of structures, a more realistic and relatively straightforward method incorporating the nonlinear hysteretic nature of the underlying soil–foundation system needs to be developed. The present paper models the soil–foundation system as a single degree of freedom spring–dashpot system with nonlinear hysteresis in form of elasto-perfectly plastic behavior. Analytical results for the lateral dynamic stiffness on footing have been presented. An example study has been carried out in case of circular footings. It is shown how the analytical results can be used to get a preliminary idea of the lateral dynamic stiffness of footings on a soil medium prior to a detailed computational geo-mechanics analysis provided the static nonlinear load–deformation characteristic of the soil medium is known and can be modeled by a hysteretic elasto-plastic behavior. The corresponding results are presented in a graphical form. The results have been computed showing parametric variations with the change in the amplitude and dimensionless frequency of the non-dimensional excitation force. Analytical results are also presented for the asymptotic cases at low and very high values of dimensionless frequency parameter.     

Investigation of seismic behavior of fluid–rectangular tank–soil/foundation systems in frequency domain

Main purpose of this study is to evaluate the dynamic behavior of fluid–rectangular tank–soil/foundation system with a simple and fast seismic analysis procedure. In this procedure, interaction effects are presented by Housner's two mass approximations for fluid and the cone model for soil/foundation system. This approach can determine; displacement at the height of the impulsive mass, the sloshing displacement and base forces for the soil/foundation system conditions including embedment and incompressible soil cases. Models and equations for proposed method were briefly explained for different tank–soil/foundation system combinations. By means of changing soil/foundation conditions, some comparisons are made on base forces and sloshing responses for the cases of embedment and no embedment. The results showed that the displacements and base shear forces generally decreased, with decreasing soil stiffness. However, embedment, wall flexibility, and soil–structure interaction (SSI) did not considerably affect the sloshing displacement.     

Seismic analysis of base-isolated liquid storage tanks using the BE–FE–BE coupling technique

A three-dimensional soil–structure–liquid interaction problem is numerically simulated in order to analyze the dynamic behavior of a base-isolated liquid storage tank subjected to seismic ground motion. A dynamic analysis of a liquid storage tank is carried out using a hybrid formulation, which combines the finite shell elements for structures and the boundary elements for liquid and soil. The system is composed of three parts: the liquid–structure interaction part, the soil–foundation interaction part, and the base-isolation part. In the liquid–structure interaction part, the tank structure is modeled using the finite elements and the liquid is modeled using the internal boundary elements, which satisfy the free surface boundary condition. In the soil–foundation interaction part, the foundation is modeled using the finite elements and the half-space soil media are modeled using the external boundary elements, which satisfy the radiation condition in the infinite domain. Finally, above two parts are connected with the base-isolation system to solve the system's behavior. Numerical examples are presented to demonstrate the accuracy of the developed method, and an earthquake response analysis is carried out to demonstrate the applicability of the developed technique. The properties of a real LNG tank located in the west coast of Korea are used. The effects of the ground and the base-isolation system on the behavior of the tank are analyzed.     

Dam–water–sediment–rock systems: Seismic analysis

A computational procedure for two-dimensional finite-element analysis of dam–water–sediment–rock systems subjected to seismic excitations is reviewed. In particular, the semidiscrete approximation of the water–sediment–rock region on the upstream side of the dam by means of a hyperelement is described in detail. The sediment is represented in the hyperelement as a fluid-filled porous solid on the basis of the Biot theory of wave propagation in poroelastic media while the water is taken as a compressible, inviscid fluid and the rock as a viscoelastic solid. An application of the procedure to a study of the effects of sediment porosity and thickness on the response of a model dam to horizontal and vertical ground motions is presented and discussed.     

Scattering of plane SV waves by cylindrical canyons in saturated porous medium

On the basis of Biot dynamic theory, an analytic solution of two-dimensional scattering and diffraction of plane SV waves by circular cylindrical canyons in a half space of saturated porous media is presented in this paper for the first time. The solution is obtained by employing the Fourier–Bessel series expansion technique. Parametric studies had been carried out, which includes: the angle of incidence, the frequency of the incident SV wave, the porosity of saturated porous medium and the stiffness and Poisson's ratio of the solid-skeleton. All the outcomes are useful for the seismic analysis of the surface topography conditions.     

Plain strain soil–structure interaction model for a building supported by a circular foundation embedded in a poroelastic half-space

A simple theoretical model for soil–structure interaction in water saturated poroelastic soils is presented, developed to explore if the apparent building–foundation–soil system frequency changes due to water saturation. The model consists of a shear wall supported by a rigid circular foundation embedded in a homogenous, isotropic poroelastic half-space, fully saturated by a compressible and inviscid fluid, and excited by in-plane wave motion. The motion in the soil is governed by Biot's theory of wave propagation in fluid saturated porous media. Helmholtz decomposition and wave function expansion of the two P-wave and the S-wave potentials is used to represent the motion in the soil. The boundary conditions along the contact surface between the soil and the foundation are perfect bond (i.e. welded contact) for the skeleton, and either drained or undrained hydraulic condition for the fluid (i.e. pervious or impervious foundation). For the purpose of this exploratory analysis, the zero stress condition at the free surface is relaxed in the derivation of the foundation stiffness matrix, which enables a closed form solution. The implications of this assumption are discussed, based on published comparisons for the elastic case. Also, a closed form representation is derived for the foundation driving forces for incident plane (fast) P-wave or SV wave. Numerical results and comparison with the full-scale measurements are presented in the companion paper, published in this issue.     

Response of L-shaped rigid foundations embedded in a uniform half-space to traveling seismic waves

A simplified indirect boundary element method is applied to compute the impedance functions for L-shaped rigid foundations embedded in a homogeneous viscoelastic half-space. In this method, the waves generated by the 3D vibrating foundation are constructed from radiating sources located on the actual boundary of the foundation. The impedance functions together with the free-field displacements and tractions generated along the soil–foundation interface are used to calculate the foundation input motion for incident P, S and Rayleigh waves. This is accomplished by application of Iguchi's averaging method which, in turn, is verified by comparison with results obtained rigorously using the relation between the solutions of the basic radiation (impedance functions) and scattering (input motions) problems. Numerical results are presented for both surface-supported and embedded foundations. It is shown how the seismic response of L-shaped foundations with symmetrical wings differs from that of enveloping square foundations. The effects of inclination and azimuth of the earthquake excitation are examined as well. These results should be of use in analyses of soil–structure interaction to account for the traveling wave effects usually overlooked in practice.     

Displacement solutions for dynamic loads in transversely-isotropic stratified media

Solutions for the displacements caused by dynamic loads in a viscoelastic transversely-isotropic medium are derived. The medium extends horizontally to infinity, but is bounded below by a rigid base. Stratification of the medium presents no difficulties. The medium is discretized in the vertical direction only; discretization in the horizontal direction is obviated by use of analytical solutions to the equations of motion. Application of the displacement solutions to soil-structure interaction is illustrated. A soil flexibility matrix (and hence, a stiffness matrix) for a surface foundation follows directly from the displacement solutions. A simple modification to obtain the soil stiffness for an embedded foundation of arbitrary geometry is described. Stiffnesses of rigid surface and embedded foundations are computed and compared with previously published results. In addition, the dynamic stiffness of a rigid surface foundation on a soil layer with linearly increasing shear modulus is compared to that for a homogeneous soil layer. A reduction in radiation damping is found to result from the inhomogeneity.     

无破碎带断层场地对SH波的散射

在层状半空间精确动力刚度矩阵和斜线荷载动力格林函数的基础上建立间接边界元方法,在频域内求解无破碎带断层场地对入射平面SH波的散射。为方便求解,将总波场分解为自由波场和散射波场,自由波场由直接刚度法求得,断层两侧的散射波场通过在断层面上分别对两侧施加均布斜线荷载产生的动力响应来模拟,虚拟荷载的密度可通过引入断层表面的边界条件确定,最后叠加自由波场和散射波场求得总波场。以有落差断层和无落差断层模型为例进行数值计算,分析断层落差、断层倾角以及断层两侧介质的刚度比对散射效应的影响。研究表明,断层落差与波长相当时,断层对SH波的放大作用最大;地表位移幅值随着断层倾角的增大逐渐增大;若断层无落差且其两侧刚度不同时,一般刚度较小一侧地表位移幅值较大且振荡更为剧烈,波从刚度较小一侧入射时位移幅值放大尤为显著。     

Diffraction of SH-waves by topographic features in a layered transversely isotropic half-space

The scattering of plane SH-waves by topographic features in a layered transversely isotropic (TI) half-space is investigated by using an indirect boundary element method (IBEM). Firstly, the anti-plane dynamic stiffness matrix of the layered TI half-space is established and the free fields are solved by using the direct stiffness method. Then, Green’s functions are derived for uniformly distributed loads acting on an inclined line in a layered TI half-space and the scattered fields are constructed with the deduced Green’s functions. Finally, the free fields are added to the scattered ones to obtain the global dynamic responses. The method is verified by comparing results with the published isotropic ones. Both the steady-state and transient dynamic responses are evaluated and discussed. Numerical results in the frequency domain show that surface motions for the TI media can be significantly different from those for the isotropic case, which are strongly dependent on the anisotropy property, incident angle and incident frequency. Results in the time domain show that the material anisotropy has important effects on the maximum duration and maximum amplitudes of the time histories.     

Impedance functions for three-dimensional foundations supported on an infinitely-long canyon of uniform cross-section in a homogeneous half-space

A direct boundary element procedure is presented to determine the impedance matrix for a three-dimensional foundation supported on an infinitely-long canyon of uniform cross-section cut in a homogeneous half-space. The uniform cross-section of the canyon permits analytical integration along the canyon axis leading to a series of two-dimensional boundary problems involving Fourier transforms of the full-space Green's functions. Solution of these two-dimensional boundary problems leads to a dynamic flexibility influence matrix which is inverted to determine the impedance matrix. The accuracy of the procedure is demonstrated by comparison with previous solutions for a surface-supported, square foundation and results obtained by a three-dimensional boundary element method (BEM) for a foundation of finite-width supported on an infinitely-long canyon. Compared with the three-dimensional BEM, the present method requires less computer storage and is more accurate and efficient. The foundation impedance matrix determined by this procedure can be incorporated in a substructure method for earthquake analysis of arch dams.     

CALCULATING THE DYNAMIC STIFFNESS MATRIX OF 2-D FOUNDATIONS BY DISCRETE WAVE NUMBER INDIRECT BOUNDARY ELEMENT METHODS

Apart from some special cases, calculating the dynamic stiffness matrix of foundations on a layered half-space, especially for embedded foundations, is computationally expensive. An efficient method for two-dimensional foundations in a horizontally layered soil media is presented in this paper. This method is based on indirect boundary element methods and uses discrete wave number solution methods for calculating Green's functions for displacements and analytical methods for the integrations over the boundary. For surface foundations, the present method applies at all frequencies. For embedded foundations or for constructing energy transmitting boundaries, because the free-field part is modelled by boundary elements and the excavated part is modelled by finite elements, the present method applies only at low frequencies for the spring coefficients (the real parts of the dynamic stiffness matrix) but applies at all frequencies for the damping coefficients (the imaginary part of the dynamic stiffness matrix) for undamped sites. The novelty of the method can be used for three-dimensional foundations. © 1997 by John Wiley & Sons, Ltd.     

Dynamic analyses of multilayered poroelastic media using the generalized transfer matrix method

Since the generalized transfer matrix method overcomes the intrinsic instability of the Thomson–Haskell transfer matrix method for both high frequencies and/or thick layers, it can produce stable and accurate solutions for the dynamic analysis of viscoelastic media as well as anisotropic media. This paper extends the generalized transfer matrix method to the dynamic analysis of multilayered poroelastic media. Main improvements include the presentation of the concisely explicit general solutions for the dynamic analysis of multilayered poroelastic media and the derivation of an analytical inversion of 8×8 order layer matrix corresponding to the general solutions. The explicit solutions are valid for the dynamic analysis of one-dimensional, two-dimensional and three-dimensional poroelastic medium problems. In addition, an efficient recursive scheme is proposed for accurate determination of the equivalent interface sources for multilayered poroelastic media due to excitation by a source at an arbitrary depth. For the dynamic analysis of multilayered poroelastic media, the generalized transfer matrix method recursively transfers both the interface stiffness matrix and equivalent source starting from the bottom half space to the top layer, without resort to the numerical solution of the assembled global equations as the exact stiffness matrix method does. While keeping the simplicity of the Thomson–Haskell transfer matrix method, the generalized transfer matrix method is both accurate and stable. The related numerical examples have demonstrated that the generalized transfer matrix method is an alternative approach to conducting the dynamic analysis of multilayered poroelastic media.