THE STRUCTURAL DESIGN OF TALL AND SPECIAL BUILDINGS Struct. Design Tall Spec. Build.

2011; 20: S47S62 Published online 22 November 2011 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/tal.735

Modeling soilfoundationstructure interaction

WD Liam Finn*,, Bishnu H Pandey and Carlos E Ventura
Department of Civil Engineering, University of British Columbia, Vancouver, Canada

SUMMARY This paper offers a guided tour through the various ways, used in practice, of accounting for soilstructure interaction in design and analysis, ranging from a complete analysis of the total combined system of foundation, soil and structure to approximate models of the system. The focus is on three types of structures: bridges on pile foundations, tall buildings with several levels of underground parking and large basement slabs with shallow embedment. The paper also reports on preliminary results from the study of seismic pressures against deep basement walls commissioned by the Structural Engineers Association of British Columbia. Copyright 2011 John Wiley & Sons, Ltd.
Received 13 October 2011; Accepted 26 October 2011
KEY WORDS:

soil-structure interaction; base slab averaging; seismic earth pressures; basement walls

1. INTRODUCTION This paper offers a guided tour through the various ways, used in practice, of accounting for soilstructure interaction (SSI) in design and analysis, ranging from a complete analysis of the total combined system of foundation, soil and structure to approximate models of the system. The focus is on three types of structures: bridges on pile foundations, tall buildings with several levels of underground parking and large basement slabs with shallow embedment. The paper also reports on the preliminary results from the study of seismic pressures against deep basement walls commissioned by the Structural Engineers Association of British Columbia (SEABC). When analysis of the total soilstructure system is carried out, the effects of SSI are implicitly included in the analysis and reected in the results. No special consideration of SSI is required. However, this type of analysis, although feasible, is rarely practical in practice because the structural analysis programs used usually by structural engineers cannot handle the nonlinear soil continuum directly. There are powerful commercial programs available that can do complete analyses, but the learning curve is steep and long and the computational time is too long for the designers requirements except for special projects. Therefore, it is necessary to uncouple the computational model of a structure from the soil and to include SSI effects by appropriate springs and dashpots.

2. BRIDGE PIERS ON PILES A three-span continuous box girder bridge structure shown in Figure 1 was chosen for a fundamental study of SSI in pile foundations. A rigid base version of this bridge was used as an example in the guide to the seismic design of bridges published by the American Association of State Highway and
*Correspondence to: WD Liam Finn, Department of Civil Engineering, University of British Columbia, Vancouver, Canada. E-mail: nn@civil.ubc.ca Copyright 2011 John Wiley & Sons, Ltd.

S48

WD LIAM FINN, B.H PANDEY AND C.E VENTURA

Figure 1. Three span bridge on pile foundation. Transportation Ofcials (AASHTO, 1983). The sectional and physical properties of the superstructure and piers were taken from the AASHTO guide. Each pier is supported by a group of 16 (4 4) concrete piles. The diameter and length of each pile are 0.36 m and 7.2 m, respectively. The piles are spaced at 0.90 m, center to center. The Youngs modulus and mass density of the piles are E = 22 000 MPa and r = 2.6 Mg/m3, respectively. The soil beneath the foundation is assumed to be a nonlinear hysteretic continuum with unit weight g = 18 kN/m3 and Poissons ratio n = 0.35. The low-strain shear modulus of the soil varies based on the square root of the depth, with values of zero at the surface and 213 MPa 10 m deep. The variations of shear moduli and damping ratios with shear strain are those recommended by Seed and Idriss (1970) for sand. The surface soil layer overlies a hard stratum at 10 m. For the PILE-3D (Wu and Finn, 1997) nite element mesh, the soil deposit was divided into 10 sublayers of varying thicknesses. Sublayer thicknesses decrease toward the surface where soilpile interaction effects are stronger. Nine hundred brick elements were used to model the soil around the piles, and 64 beam elements were used to model the piles. The input acceleration record used in the study was the rst 20 s of the NS component of the free eld accelerations recorded at CSMIP Station No. 89320 at Rio Dell, CA, during the April 25, 1992, Cape Mendocino Earthquake. The power spectral density of this acceleration record shows that the predominant frequency of the record is 2.2 Hz. 2.1. Pile cap stiffness The pile cap stiffness of the pile foundation shown in Figure 1 was determined for two different column/foundation stiffness ratios, 7% and 50%. A PILE-3D analysis was conducted rst, and the spatially varying time histories of modulus and damping were stored. Then, an associated program PILIMP calculated the time histories of dynamic pile head impedances using the stored data. The dynamic impedances were calculated at any desired frequency by applying a harmonic force of the same frequency to the pile head and calculating the generalized forces for unit displacements. In this paper, the focus is on the stiffness only because these are the parameters of primary interest for current practice. However, the effects of damping are always included in the analyses. The stiffness was calculated rst without taking into account inertial interaction between the superstructure and the pile foundation. This is the usual condition in which stiffness is estimated by static loading tests, static analysis or elastic formulae. The stiffness was also calculated taking the inertial effects of the superstructure into account. In the latter case, both kinematic and inertial interactions were taken into account. Since the entire pile group was being analyzed, pilesoilpile interaction was automatically taken into account under both linear and nonlinear conditions. Therefore, the usual
Copyright 2011 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 2011; 20: S47S62 DOI: 10.1002/tal

MODELING SOILFOUNDATIONSTRUCTURE INTERACTION

S49

difcult problem of what interaction factors to use or what group factor to apply was avoided. The time histories of lateral and cross-coupling kinematic stiffness are shown in Figure 2. This stiffness, resulting from kinematic interaction only, was calculated for the predominant frequency of the input motions, f = 2.2 Hz. It is clearly not an easy matter to select a single representative stiffness to characterize the discrete single-valued springs often used in structural analysis programs to represent the effects of the foundation. In the absence of a complete analysis, probably, a good approach to including the effects of soil nonlinearity on stiffness is to get the vertical distribution of effective moduli by a SHAKE (Schnabel et al., 1972) analysis of the free eld and calculate the stiffness at the appropriate frequency using PILIMP with these moduli. The constant stiffness calculated in this way is shown by the horizontal lines in Figure 2. However, this is kinematic stiffness. It is shown in the latter part of this paper that inertial interaction by the superstructure may cause greater nonlinear behavior, leading to substantially reduced stiffness. The SHAKE analysis cannot capture this effect. 2.2. Seismic response of code bridge to transverse earthquake loading A three-dimensional space frame model of the bridge is shown in Figure 3. At the abutments, the deck is free to translate in the longitudinal direction but restrained in the transverse and vertical directions. Rotation of the deck is allowed about all three axes. The space frame members were modeled using two eight-node three-dimensional beam elements with 12 degrees of freedom, 6 degrees at each end. The bridge deck was modeled using 13 beam elements, and each pier was modeled by 3 beam elements. The cap beam that connects the top of the adjacent piers was modeled using a single-beam element. The sectional and physical properties of the deck and the piers
1.2 1.0
SHAKE: Lateral (MN/m) PILE-3D: Lateral (MN/m)

Stiffness (x103)

0.8 0.6 0.4 0.2 0 0

SHAKE: Cross-Coupling (MN/rad) PILE-3D: Cross-Coupling (MN/rad)

12

16

20

Time (s)

Figure 2. Time histories of kinematic lateral and cross-coupling stiffness.

Figure 3. Computational model of the bridge.

Copyright 2011 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 2011; 20: S47S62 DOI: 10.1002/tal

S50

WD LIAM FINN, B.H PANDEY AND C.E VENTURA

are those provided in the AASHTO guide (1983). The pier foundation was modeled using a set of time-dependent nonlinear springs and dashpots that simulate exactly the time histories of stiffness and damping from the PILE-3D analyses. The response of the bridge structure was analyzed for different foundation conditions to study the inuence of various approximations to foundation stiffness and damping using the computer program BRIDGE-NL. The free eld acceleration was used as the input acceleration, and the peak acceleration was set to 0.5 g. 2.3. Foundation conditions for analyses The seismic response of the bridge to transverse earthquake loading was analyzed for the four different foundation conditions listed below: 1. 2. 3. 4. Rigid foundation xed base condition Flexible foundation with elastic stiffness and damping Flexible foundation with kinematic time-dependent stiffness and damping Flexible foundation with stiffness and damping based on the SHAKE effective moduli

The fundamental transverse mode frequency of the computational model of the bridge structure with a xed base was found to be 3.18 Hz. This is the frequency quoted in the AASHTO-83 guide. This agreement in fundamental frequencies indicates an acceptable structural model. For this original case, the lateral stiffness of the bridge pier is only 7% of the foundation stiffness. For this extremely low stiffness ratio, the columns control the fundamental frequency of the bridge, and the inuence of the foundation is negligible. Results from analyses in which the column/foundation stiffness ratio is 50% will be presented here. The stiffness ratio was raised by increasing the stiffness of the piers only, with no changes to the superstructure. Normally, much stiffer piers would imply a heavier superstructure and therefore higher inertial forces. For a 50% stiffness ratio, the xed base fundamental frequency of the bridge is 5.82 Hz. When the stiffness associated with low-strain initial moduli was used, the fundamental frequency was 4.42 Hz, a 24% reduction from the xed base frequency. With kinematic strain-dependent stiffness, the frequency reached a minimum value of 3.97 Hz during strong shaking, a 32% reduction from the xed base frequency. When the foundation stiffness was based on effective shear moduli from a SHAKE analysis of the free eld, the frequency was 4.18 Hz, a 28% change from the xed base frequency. Figure 4 shows the variation with time in fundamental transverse modal frequency for the different foundation conditions. 2.4. Inertial interaction of structure and pile The time-dependent stiffness used in the analyses described above was computed without taking the inertial interaction of superstructure and foundation into account. The primary effect of this interaction is to increase the lateral pile displacements and cause greater strains in the soil. This in turn leads to smaller moduli and increased damping. The preferred method of capturing the effect of superstructure interaction is to consider the bridge structure and the foundation as a fully coupled system in the nite element analysis. However, such fully coupled analysis is not possible with current commercial
First Transverse Mode Frequency (Hz)
5.5

Fixed Base Frequency = 5.82 Hz

5.0
Constant stiffness = 5.82 Fixed Base Frequency based on Hz the initial shear moduli

4.5

4.0
Variable stiffness based on shear moduli from PILE-3D analysis

Constant stiffness based on shear moduli from SHAKE analysis

3.5 0

10

15

20

Time (s)

Figure 4. Time history of transverse modal frequencies for different foundation conditions.
Copyright 2011 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 2011; 20: S47S62 DOI: 10.1002/tal

MODELING SOILFOUNDATIONSTRUCTURE INTERACTION

S51

structural software. Even if it were, it would not be feasible in practice because it would require enormous amounts of computational storage and time using the more sophisticated computer codes. An approximate way of including the effects of superstructure interaction is to use the model shown in Figure 5. In this model, the superstructure is represented by a single degree of freedom (SDOF) system. The mass of the SDOF system is assumed to be the portion of the superstructure mass carried by the foundation. The stiffness of the SDOF system is selected so that the system has the period of the mode of interest of the xed base bridge structure. This approximate approach is demonstrated by the analysis of the center pier at Bent 2. The fundamental transverse mode frequency of the xed base model was found earlier to be 5.82 Hz. The static portion of the mass carried by the center pier is 370 Mg. The superstructure can be represented by an SDOF system having a mass of 370 Mg at the same height as the pier top and frequency of 5.82 Hz. The corresponding stiffness of the SDOF system is 495 MN/m. A coupled soilpilestructure interaction analysis can be carried out using PILE-3D by incorporating the SDOF model into the nite element model of the pile foundation. The pile foundation stiffness derived from this nite element model incorporates the effects of both inertial and kinematic interactions and is called total stiffness. The time histories of stiffness with and without the superstructure are shown in Figure 6. The reduction in lateral stiffness is greater throughout the shaking when the inertial interaction was included. There is a similar reduction in the rotational and cross-coupling stiffness. When inertial

Figure 5. Pile foundation with superstructures.

Lateral Stiffness - MN/m (x103
1.0 0.8 0.6 0.4 0.2 0
SHAKE Shear Moduli from SHAKE PILE-3D: With Superstructure Analysis PILE-3D: Without Superstructure

Figure 6. Effect of inertial interaction on lateral pile cap stiffness.

Copyright 2011 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 2011; 20: S47S62 DOI: 10.1002/tal

10

15

20

Time (s)

S52

WD LIAM FINN, B.H PANDEY AND C.E VENTURA

interaction was included, the lateral stiffness reached a minimum of 188 MN/m, which is 22% of the initial value. When the inertial interaction was not included, the minimum was 400 MN/m. Clearly, in this case, inertial interaction has a major effect on foundation stiffness. The results of the analyses for four different foundation conditions are summarized in the displacement spectra for transverse vibrations of the bridge, shown in Figure 7. The displacement spectra clearly show the importance of including inertial interaction when calculating foundation stiffness in this case. The xed base model for estimating response is inadequate. As the ratio of superstructural stiffness to foundation stiffness is reduced, the effect of inertial interaction on system frequency is reduced and kinematic stiffness becomes adequate. The xed base model is adequate only for low stiffness ratios. For the example bridge, when effective moduli from a SHAKE analysis of the free eld are used in an elastic analysis to obtain discrete foundation stiffness for each degree of freedom, the corresponding system frequencies lead to acceleration and displacement responses very close to the responses from a PILE-3D nonlinear analysis. This is true when the complete pile foundation is included in the analysis. It may or may not be true if the effective moduli are used to get the stiffness of a single pile and the stiffness of the pile group is developed from this with the help of empirical factors for group effects. The results above suggest that kinematic stiffness may be obtained, taking nonlinear soil effects into account, by an elastic structural program that can model the pile group foundation, if the effective moduli from a SHAKE analysis are used. This needs to be veried by a few more case histories. A more detailed discussion of SSI of pile foundation and a critical review of foundation springs used in practice may be found in Finn (2004, 2005).

3. MODELS FOR TALL BUILDINGS The presentation in this section is based on the paper by Naeim et al. (2008). The paper describes the most reliable and effective way to develop reliable simplied computational models of structures that incorporate SSI effects. Unfortunately, many analyses of the type used would be necessary to develop a database that would be applicable to tall buildings in general. But valuable lessons can be learned even from one well-conducted study of a single instrumented building. The process of exploring what may be effective models of this structure starts with the construction of the most accurate (MA) computational model compatible with the current structural software. The response of this model to the Northridge Earthquake was evaluated to provide baseline response data against which the performance of various simpler models could be checked. The MA model of the 54-storey building is shown in Figure 8. The action of the foundation soil against the basement walls was modeled by appropriate horizontal springs and dashpots. The vertical and rocking stiffnesses of the base slab were modeled simultaneously by vertical springs with an appropriate distribution of stiffnesses. A detail of how these springs and dashpots were determined is described in Naeim et al. (2008).
Spectral Displacement (cm)

Response Spectra (5% Damping)

15
PILE-3D: Nonlinear Kinematic PILE-3D with Shake Moduli and Damping

10

Rigid supports PILE-3D Inertial + Kinematic Interaction

0 0 0.5 1 1.5 2

Period (sec)

Figure 7. Spectral displacements of bridge for four different foundation conditions.

Copyright 2011 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 2011; 20: S47S62 DOI: 10.1002/tal

MODELING SOILFOUNDATIONSTRUCTURE INTERACTION

S53

Spring ends constrained to the ground motion history

Foundation walls modeled with the actual stiffness and strength

Figure 8. Most accurate model of a 54-storey building. The foundation input motion (FIM) to the base slab is the recorded motion of the slab. SHAKE analysis of the free eld was conducted using FIM as the input. This analysis provided depth-dependent ground motion for application to the ends of springs acting on the basement walls. The response of the MA structural model to the Northridge Earthquake was evaluated, and the accelerations at different elevations in the building were compared with those recorded during the earthquake. The MA model was tuned to give good agreement with the recorded accelerations. Many different simplied soilstructure models were tested, but only three will be presented here to illustrate how simpler models should be evaluated. 3.1. Approximate models The simplest model (model 3c) is shown in Figure 9. The building was assumed to rest on a rigid base. There was no interaction between the soil and the basement walls. The performance of this model was compared with that of the MA model in terms of several different response parameters in Naeim et al. (2008), but owing to space limitations, the

ug(z=0)

Figure 9. Simplest model of building with no soilstructure interaction.

Copyright 2011 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 2011; 20: S47S62 DOI: 10.1002/tal

S54

WD LIAM FINN, B.H PANDEY AND C.E VENTURA

comparison was limited here to the interstorey drift ratios shown in Figure 10. For such a crude model, the drift ratios were reasonably good except in the basement levels where the interstorey drift ratios were overestimated and near the roof where the drift ratios were underestimated. The next model (model 3b, Figure 11) rested on a rigid base, but some passive lateral restraint was imposed on the basement walls by springs. Naiem et al. (2008) described this model as follows: Neglect entirely soil exibility at the level of the base slab (i.e. the base slab is xed vertically and horizontally), and simulate soil exibility along the basement walls with horizontal springs with ends xed to match the free eld ground motion. Seismic demand consists only of horizontal motions (equivalent free-eld condition) at the base slab level and at the ends of foundation springs. This simulates a condition commonly used in structural design ofces. The drift ratios of this model were compared with those of the MA model in Figure 12. Despite the introduction of some restraint on the basement walls to model the effects of the soil, the drift ratios predicted by this model compared very poorly with the ratios from the MA model. The last model considered is model 3d shown in Figure 13. In this model, the structure was assumed

52 47 42 37 32
MA 3c

Story

27 22 17 12 7 2 -3 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

Story Drift Ratio (x10-3)

Figure 10. Drift ratios for models MA and 3c. MA, most accurate.

Figure 11. Model 3b.

Copyright 2011 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 2011; 20: S47S62 DOI: 10.1002/tal

MODELING SOILFOUNDATIONSTRUCTURE INTERACTION

52 47 42 37 32

S55

Story

27 22 17 12 7 2 -3 0 0.5 1.0 1.5 2.0 2.5 3.0

Story Drift Ratio (x10-3)

Figure 12. Drift ratios for models MA and 3b. MA, most accurate.

Figure 13. Model 3d. to be on a rigid base at the ground surface and the structures below that were ignored. According to Naeim et al. (2008), seismic demand consists only of horizontal motions (equivalent free-eld condition) applied at ground level. The response of this model to the Northridge Earthquake compared surprisingly well with the MA model as shown in Figure 14, although the drift ratios were overestimated by model 3d at the ground level and underestimated near the roof. This study explores the effectiveness of simplied models of the tall buildings commonly used in practice by comparing the response of each model with the response of MA model. The MA was carefully calibrated using data recorded on the structure during Northridge Earthquake. Two models commonly used in practice, models 3c and 3d, performed reasonably well. The immediate effect of choosing a simpler model is to alter the dynamic characteristics of the computational model such as periods and mode shapes. The effects of these changes are never explored on a job-to-job basis. Therefore, studies such as that of Naeim et al. (2008) are crucial in providing some reliability covered for simpler models. However, many more studies on different building
Copyright 2011 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 2011; 20: S47S62 DOI: 10.1002/tal

S56

WD LIAM FINN, B.H PANDEY AND C.E VENTURA

52 47 42 37
MA

32

Story

27 22 17 12 7 2 -3 0 1.0 2.0 3.0

3d

4.0

Story Drift Ratio (x10-3)

Figure 14. Drift ratios for models MA and 3d. MA, most accurate. congurations would be required to develop a database to guide reliably the selection of simpler model for structural analysis. Since the SSI effects are primarily related to the underground section of a building, it would seem logical that the height of the building would have considerable impact on how well simpler models may behave. The development of appropriate springs and dashpots to represent SSI effects for simplied models is not an easy task, especially when nonlinear soil behavior is a factor. As demonstrated in the analysis of the pile foundation of the bridge with nonlinear soil response earlier, the foundation stiffness would vary with time and vary with relative impact of inertial and kinematic responses. Selecting appropriate single-valued springs and dashpots without knowing the time variation of stiffness and damping with time can be a risky business. It is more appropriate to nd alternative methods for evaluation of stiffness that takes into account in a reasonable way the nonlinear behavior of the soil. The use of elastic effective moduli from SHAKE analysis proved to be an effective procedure for estimating kinematic stiffness for the bridge pier.

4. SEISMIC PRESSURES ON BASEMENT WALLS Structural and geotechnical engineers have long relied upon the use of the MononobeOkabe (MO) method for determining seismic lateral pressures acting on retaining walls. This limit equilibrium-based method was originally developed for rigid retaining walls with sufcient rigid body displacements to mobilize the active wedge in the backll soil. In reality, however, certain types of retaining walls, such as basement walls, have variable degrees of exibility and deformation at different depths. Recently, the SEABC initiated a task force to review current design procedures for basement walls. A series of dynamic numerical analyses, using computer program FLAC 2D, is being conducted to study the response of basement walls to seismically induced lateral earth pressures, taking into account the exibility and potential yielding of the wall components (Ahmadnia et al., 2011). Basement walls in British Columbia (BC) are designed for seismic pressures based on the MO method (Mononobe and Matsuo, 1929; Okabe, 1924). Before 2005, the peak ground acceleration (PGA) used in the MO method was 0.24 g, which had an exceedance rate of 10% in 50 years. Since 2005, PGA = 0.46 g is used for design. This has an exceedance rate of 2% in 50 years. The rst step in the study of basement wall was to evaluate how walls designed using PGA = 0.24 g would behave when subjected to the new hazard of PGA = 0.46 g.
Copyright 2011 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 2011; 20: S47S62 DOI: 10.1002/tal

MODELING SOILFOUNDATIONSTRUCTURE INTERACTION

S57

Figure 15 depicts the design lateral earth seismic pressure distribution for the basement wall for a friction angle of 33 and PGA = 0.24 g. The required moment resistance of the wall along its height is shown on the right in Figure 15. In this design, no overstrength or ductility factor is applied to the seismic pressures on the wall. The wall has uniform properties of I = 0.0013 m4, A = 0.25 m2 and E = 2.74107 kN/m2. Ground motions for the analyses were selected from the Pacic Earthquake Engineering and Research Center (PEER, 2011) strong ground motion database. Based on the results of de-aggregation of the Uniform Hazard Spectrum (UHS), Site Class C for Vancouver (NRCC, 2005) candidate input motions were selected in the magnitude range M = 6.57.5 and the distance range 1030 km using the program Design Ground Motion Library (PEER, 2011). Table 1 shows the list of three ground motions selected for this initial part of the study. The selected ground motions are spectrally matched to UHS for Vancouver in the period range of 0.021.7 s using the computer program SeismoMatch (2011). The time history of resultant force against the wall due to the 1979 Imperial Valley Earthquake is shown in Figure 16. More recent results from nine different earthquakes show that the total MononobeOkabe force is in the range of 10% to 2% of the maximum dynamic force. The dynamic pressure distribution at the moment of maximum force is very different from the distribution assumed for the MO force in the current design as shown in Figure 17. This is due to the very different displacement patterns of the wall in the dynamic analysis compared with that assumed for development of the active MO force. Envelopes of moments and shears are shown in Figure 18. Yielding occurs where the seismic moment envelope touches the yield moment frame. Average drift ratios for three ground motions used in analysis are shown in Figure 19. The denition of drift ratio is also illustrated in the gure. Acceptable drift ratios have not been presented for basement walls, and discussions with structural engineers on the issue were inconclusive.

Figure 15. Distribution of the design lateral pressure along the height of the wall based on the current practice for a seismic event with PGA = 0.24 g and a backll soil with friction angle of 33 ; the gure on the right shows the moment resistance distribution along the height of the wall.

Table 1. List of selected ground motions. Ground motion

G1 G2 G3

Record no.
162 987 778

Event
Imperial Valley Northridge Loma Prieta

Year
1979 1994 1989

Station
Calexico Fire LA-Centinela Hollister

Magnitude
6.53 6.69 6.93

Copyright 2011 John Wiley & Sons, Ltd.

Struct. Design Tall Spec. Build. 2011; 20: S47S62 DOI: 10.1002/tal

S58

WD LIAM FINN, B.H PANDEY AND C.E VENTURA

Figure 16. Time history of maximum force against the wall compared with MO seismic forces for 1/475 and 1/2475 rates of exceedance. MO, MononobeOkabe.

Mononobe- Okabe

Figure 17. Pressure distribution at time of maximum force on the wall compared with linear MononobeOkabe maximum pressure. MO, MononobeOkabe.

Figure 18. Envelopes of positive and negative bending moments and shears (yield limits for moments and shear are shown by the black frames.)

Figure 19. Average drift ratios for three earthquakes over the depth of the wall.
Copyright 2011 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 2011; 20: S47S62 DOI: 10.1002/tal

MODELING SOILFOUNDATIONSTRUCTURE INTERACTION

S59

The denition of drift ratio given above is identical to the denition of hinge rotation given by the Task Committee on Blast Resistant Design (TCBRG, 1997). This committee related hinge rotation to structural performance. They specied two performance categories that may apply to basement walls: low and medium response categories. The low response category is dened as follows: . . . localized building/component damage. Building can be used; however repairs are required to restore integrity of structural envelope. Total cost of repairs is moderate. The medium response category is dened as follows: . . . widespread building/component damage. Building cannot be used until repaired. Total cost of repairs is signicant. The hinge rotations (and hence the drift ratios) associated with these two response states are 2% and 4%, respectively. According to these criteria, only the response of the top basement wall needs careful consideration. 5. BASE SLAB AVERAGING It appears to be generally accepted that large foundation slabs reduce the free eld ground motions for period up to 0.5 s. FEMA-440 (ATC, 2005) has developed reduction factors for spectral values due to the action of foundation slabs as shown in Figure 20 for slab foundation with shallow embedment. We have examined 98 case histories of free eld and slab motions, and 68 pair of these motion pairs of free eld and slab motions showed a reduction in slab motions from free eld motions. Typically, examples are shown in Figures 21 and 22. However, in 30 cases, we found that the slab motions are greater than the free eld as shown in Figures 23 and 24. Poland et al. (2000) also found that there were cases of motion amplied instead of reduced. The fact that the 30% of cases we investigated showed signicant increases in slab motion suggests that caution is warranted in relying on a reduction. It is interesting to note that a recent series of centrifuge tests on a structure resting on a foundation slab consistently showed an increase in slab motion over the free eld (Rayhani and Nagger, 2008). In this test, the foundation was on soft clay. Analysis of the test using the computer program FLAC 3D and Mohr Column failure criterion conrmed the increase in slab motion.

6. CONCLUDING REMARKS The coupled analysis of structures and foundations is not a feasible option in engineering practice at present because of practical difculties with the analysis. Therefore, the analysis was conducted on simpler structural models with add-ons, usually linear or nonlinear springs, to simulate SSI effects. The uncertainties inherent in these uncoupled systems needs to be more fully documented. The sensitivity of response to spring characteristics is especially important.
1

Foundation/free-fieldRRS from base slabaveraging(RRSbsa)

0.9 0.8 0.7

Simplified Model
0.6 0.5 0.4 0
be = 65 ft be = 130 ft be = 200 ft be = 330 ft

0.2

0.4

0.6

0.8

1.2

Period (s)

Figure 20. Spectral reduction from base slab averaging (RRSbsa) as a function of period in FEMA-440.
Copyright 2011 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 2011; 20: S47S62 DOI: 10.1002/tal

S60

WD LIAM FINN, B.H PANDEY AND C.E VENTURA

0.7 0.6 0.5 Free Field Basement

PSA (g)

0.4 0.3 0.2 0.1 0 0 0.5 1 1.5 2

T (s)

Figure 21. Spectral accelerations at the site of Pomona two-storey commercial building in the 1990 Upland Earthquake (EW direction).
0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

Free Field Basement

PSA (g)

0.2

0.4

0.6

0.8

T (s)

Figure 22. Spectral accelerations at the site of Rancho Cucamonga four-storey Justice Centre building in the 1990 Upland Earthquake (EW direction).

Figure 23. Spectral accelerations at the site of El Centro Imperial County Service building in the 1979 Imperial Valley Earthquake (NS direction). Naeim et al. (2008) have used the recorded response of a tall building to evaluate the reliability of simpler structural models used in practice in which soilstructural interaction effects are simulated by single-valued springs and dashpots. They found that some models were not reliable and should not be used, whereas other models gave various levels of satisfactory performances. A database on the performance of simple models for different building congurations and heights and for various levels of shaking is essential for providing a reliable basis for selection of simpler computational models. The behavior of basement walls during earthquakes and the seismic pressures for which they should be designed are important aspects of SSI in tall buildings. A major study of this problem is being
Copyright 2011 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 2011; 20: S47S62 DOI: 10.1002/tal

MODELING SOILFOUNDATIONSTRUCTURE INTERACTION

S61

Figure 24. Spectral accelerations at the site of Sylmar six-storey County Hospital building in the 1994 Northridge Earthquake (EW direction). conducted at the University of British Columbia at the request of SEABC. In exploring the capacity of the walls to absorb seismic demand, the exibility and yield moments of the wall structure are taken into account. Early results suggest the behavior of the top basement is critical. This basement is higher than the other basements (4 m versus 3 m in BC) and is not as restrained at the top as the lower basements are. Typically, the drift ratio in the top basement is more than twice the drift ratios in the lower basements. There is, at present, no performance criteria for basement walls. In the study, the standard adopted for blast loading would suggest that the behavior of a wall designed for a PGA = 0.24 g would potentially show unsatisfactory performance only in the top basement when subjected to PGA = 0.46 g. A major study of the effects of shallow foundation slabs with shallow embedment on free eld motions was conducted involving 98 pairs of free eld and slab motions. Sixty-eight pairs show that the motions of the slabs were smaller than the free eld motions in the period range below 0.5 s. But in 30 cases, the motions of the slabs were greater than the free eld motions. These ndings conrm earlier ndings of Poland et al. (2000) who also showed many of the slab motions were amplied. These ndings raised some concerns about the FEMA-440 recommendations for universal reduction in slab motions from free eld values. So far, no clear reason for the different slab behavior in the amplied cases has been advanced.
ACKNOWLEDGMENTS

The study is nancially supported by a grant for SSI studies from the Canadian Seismic Research Network. The section on seismic pressure against basement walls is a part of an ongoing study for the SEABC.
REFERENCES
AASHTO. 1983. Guide specications for seismic design of highway bridges. American Association of State Highway and Transportation Ofcials, Washington, DC, USA. Ahmadnia, A, Taiebat, M, Finn, WDL, Ventura, C, 2011. Seismic evaluation of existing basement walls. Proceedings of the Third International Conference on Computational Methods in Structural Dynamics & Earthquake Engineering. Corfu, Greece. ATC. 2005. Improvement of nonlinear static seismic analysis procedures. Rep. No. FEMA-440, Washington, DC. Finn WDL. 2004. Characterizing pile foundations for evaluation of performance based seismic design of critical lifeline structures. Invited keynote lecture, 13th WCEE, Vancouver, BC, Canada. Finn WDL. 2005. A study of piles during earthquakes: issues of design and analysis. Bulletin of Earthquake Engineering 3 141-234. Mononobe N. Matsuo H. 1929. On determination of earth pressure during earthquakes. Proceedings of the World Engineering Congress. Tokyo, 9, 275. Naeim F, Tileylioglu S, Alimoradi A, Stewart AP. 2008. Impact of foundation modeling on the accuracy of response history analysis of a tall building. SMIP08 Seminar Proceedings. (http://www.consrv.ca.gov/cgs/smip/docs/seminar/SMIP08/Documents/ Z4_Paper_NaeimStewart.pdf) NRCC. 2005. National Building Code of Canada. National Research Council of Canada. Canadian Commission on Building and Fire Codes, Ottawa, Canada.

Copyright 2011 John Wiley & Sons, Ltd.

Struct. Design Tall Spec. Build. 2011; 20: S47S62 DOI: 10.1002/tal

S62

WD LIAM FINN, B.H PANDEY AND C.E VENTURA

Okabe S. 1924. General theory on earth pressure and seismic stability of retaining walls and dams. Journal of the Japanese Society of Civil Engineers 10(6): 12771323. PEER. 2011. Pacic Earthquake Engineering and Research Center. University of California, Berkley, California. Poland C, Soulages J, Sun J, Meija L. 2000. Quantifying the effect of soilstructure interaction for use in building design. Data Utilization Report, CSMIP/00-02(OSMS 0004), Ofce of Strong Motion Studies, Division of Mines and Geology, California Department of Conservation, CA. Rayhani MHT. El Naggar MH. 2008. Numerical modeling of seismic response of rigid foundation on soft soil. International Journal of Geomechanics (ASCE) 8, No. 6, December 2008. Schnabel PB, Lysmer J, Seed HB. 1972. SHAKE: a computer program for earthquake response analysis of horizontally layered sites. Report No. EERC72-12, Earthquake Engineering Research Center, University of California, Berkeley, CA. Seed HB, Idriss IM. 1970. Soil moduli and damping factors for dynamic response analysis. Report No. EERC70-10, Earthquake Engineering Research Center, University of California, Berkeley, CA. SeismoMatch. 2011. Educational version, Seismosoft Company. (http://www.seismosoft.com) TCBRG. 1997. Design of blast resistant buildings in petro chemical facilities. Report of Task Committee on Blast Resistant Design, Energy Division, ASCE. Wu G, Finn WDL. 1997. Dynamic nonlinear analysis of pile foundations using the nite element method in the time domain. Canadian Geotechnical Journal 34: 144152.

Copyright 2011 John Wiley & Sons, Ltd.

Struct. Design Tall Spec. Build. 2011; 20: S47S62 DOI: 10.1002/tal