Soil Structure Interaction For Building

NIST GCR 12-917-21
Soil-Structure
Interaction for
Building Structures
NEHRP Consultants Joint Venture

A partnership of the Applied Technology Council and the
Consortium of Universities for Research in Earthquake Engineering
Disclaimers
This report was prepared for the Engineering Laboratory of the National Institute of Standards and Technology (NIST) under
the National Earthquake Hazards Reduction Program (NEHRP) Earthquake Structural and Engineering Research Contract
SB134107CQ0019, Task Order 69221. The contents of this publication do not necessarily reflect the views and policies of
NIST or the U.S. Government.
This report was produced by the NEHRP Consultants Joint Venture, a joint venture of the Applied Technology Council (ATC)
and the Consortium of Universities for Research in Earthquake Engineering (CUREE). While endeavoring to provide practical
and accurate information, the NEHRP Consultants Joint Venture, the authors, and the reviewers assume no liability for, nor
express or imply any warranty with regard to, the information contained herein. Users of information contained in this report
assume all liability arising from such use.
Unless otherwise noted, photos, figures, and data presented in this report have been developed or provided by NEHRP
Consultants Joint Venture staff or consultants engaged under contract to provide information as works for hire. Any similarity
with other published information is coincidental. Photos and figures cited from outside sources have been reproduced in this
report with permission. Any other use requires additional permission from the copyright holders.
Certain commercial software, equipment, instruments, or materials may have been used in the preparation of information
contributing to this report. Identification in this report is not intended to imply recommendation or endorsement by NIST, nor is
it intended to imply that such software, equipment, instruments, or materials are necessarily the best available for the purpose.
NIST policy is to use the International System of Units (metric units) in all its publications. In this report, information is
presented in both metric units and U.S. Customary Units (inch-pound), as the inch-pound system is the preferred system of
units in the U.S. earthquake engineering industry.
Cover illustration – Rendering of the Sherman Oaks building structural and foundation system, used in an example application
of soil-structure interaction principles (courtesy of C. Haselton).
NIST GCR 12-917-21
Soil-Structure Interaction for

Building Structures
Prepared for
U.S. Department of Commerce
National Institute of Standards and Technology
Engineering Laboratory
Gaithersburg, MD 20899
By
A partnership of the Applied Technology Council and the
Consortium of Universities for Research in Earthquake Engineering
September 2012
U.S. Department of Commerce

Rebecca M. Blank, Acting Secretary

Patrick D. Gallagher, Under Secretary of Commerce
for Standards and Technology and Director
NIST GCR 12-917-21
Participants
John (Jack) R. Hayes, Jr., Director, National Earthquake Hazards Reduction Program
Steven L. McCabe, Deputy Director, National Earthquake Hazards Reduction Program
John (Jay) L. Harris III, Project Manager

Applied Technology Council Consortium of Universities for
201 Redwood Shores Parkway, Suite 240 Research in Earthquake Engineering
Redwood City, California 94065 1301 S. 46th Street, Building 420
www.ATCouncil.org Richmond, California 94804
www.CUREE.org
Joint Venture Management Joint Venture Program Committee

Committee
Jon A. Heintz (Program Manager)
James R. Harris Michael Constantinou
Robert Reitherman C.B. Crouse
Christopher Rojahn James R. Harris
Andrew Whittaker William T. Holmes
Jack Moehle
Andrew Whittaker
Project Technical Committee Project Review Panel

Jonathan P. Stewart (Project Director) Craig Comartin
C.B. Crouse Yousef Hashash
Tara Hutchinson Annie Kammerer
Bret Lizundia Gyimah Kasali
Farzad Naeim George Mylonakis
Farhang Ostadan Graham Powell
Working Group Members Project Manager

Fortunato Enriquez David Hutchinson
Michael Givens
Curt Haselton
Silvia Mazzoni
Erik Okstad
Andreas Schellenberg

Preface
The NEHRP Consultants Joint Venture is a partnership between the Applied

Technology Council (ATC) and the Consortium of Universities for Research in
Earthquake Engineering (CUREE). In 2007, the National Institute of Standards and
Technology (NIST) awarded the NEHRP Consultants Joint Venture a National
Earthquake Hazards Reduction Program (NEHRP) “Earthquake Structural and
Engineering Research” task order contract (SB1341-07-CQ-0019) to conduct a
variety of tasks. In 2009, NIST initiated Task Order 69221 entitled “Improved
Procedures for Characterizing and Modeling Soil-Structure Interaction for
Performance-Based Seismic Engineering.” The purpose of this project was to
develop consensus guidance for implementing soil-structure interaction in response
history analyses, and to identify areas of further research that were needed. The
starting point for this work was decades of available soil-structure interaction
research, and recently published engineering guidance for static analysis procedures.
Soil-structure interaction (SSI) analysis evaluates the collective response of three

linked systems: the structure, the foundation, and the soil underlying and surrounding
the foundation. Problems associated with practical application of SSI for building
structures are rooted in a poor understanding of fundamental SSI principles.
Implementation in practice is hindered by a literature that is difficult to understand,
and codes and standards that contain limited guidance. This report represents an
advancement in the state of SSI knowledge for practicing engineers. It provides a
synthesis of the body of SSI literature, distilled into a concise narrative, and
harmonized under a consistent set of variables and units. Techniques are described
by which SSI phenomena can be simulated in engineering practice, and specific
recommendations for modeling seismic soil-structure interaction effects on building
structures are provided. The resulting recommendations are illustrated and tested, in
detail, on realistic example buildings.
This work is the result of an extensive literature search and collection of available
information on soil-structure interaction, discussions with researchers and
practitioners on the state of SSI knowledge and practice, conduct of problem-focused
investigations, and analytical parametric studies. A workshop of invited experts and
other stakeholders was convened to receive feedback on the developing report and
preliminary recommendations. Input from this group was instrumental in shaping the
final product.
GCR 12-917-21 Preface iii

The NEHRP Consultants Joint Venture is indebted to the leadership of Jonathan
Stewart, Project Director, and to the members of the Project Technical Committee,
consisting of C.B. Crouse, Tara Hutchinson, Bret Lizundia, Farzad Naeim, and
Farhang Ostadan, for their significant contributions in the development of this report
and the resulting recommendations. Focused analytical studies were conducted by
Curt Haselton, Fortunato Enriquez, Michael Givens, Silvia Mazzoni, Erik Okstad and
Andreas Schellenberg. Technical review and comment at key developmental stages
on the project were provided by the Project Review Panel consisting of Craig
Comartin, Yousef Hashash, Annie Kammerer, Gyimah Kasali, George Mylonakis,
and Graham Powell. The names and affiliations of all who contributed to this project
are included in the list of Project Participants at the end of this report.
NEHRP Consultants Joint Venture also gratefully acknowledges Jack Hayes

(Director, NEHRP), Steve McCabe (Deputy Director, NEHRP), and Jay Harris
(NIST Project Manager) for their input and guidance in the preparation of this report,
Dave Hutchinson for ATC project management, and Peter N. Mork and Bernadette
Hadnagy for ATC report production services.
Jon A. Heintz
Program Manager
iv Preface GCR 12-917-21

Table of Contents
Preface ..................................................................................... iii
List of Figures ............................................................................. ix
List of Tables............................................................................. xix
1. Introduction ..................................................................... 1-1

1.1 Background and Objectives ................................................................... 1-1
1.2 Overview of Soil-Structure Interaction .................................................. 1-3
1.3 Report Organization and Content .......................................................... 1-7
2. Inertial Interaction ............................................................ 2-1

2.1 Soil-Structure System Behavior ............................................................. 2-1
2.2 Equations for Shallow Foundation Stiffness and Damping ................... 2-8
2.2.1 Models for Rigid Foundations and Uniform Soils .................... 2-8
2.2.2 Effect of Non-Uniform Soil Profiles....................................... 2-17
2.2.3 Effect of Flexible Structural Foundation Elements ................ 2-20
2.2.4 Limiting Spring Forces ........................................................... 2-23
2.3 Impedance of Vertical Pile Foundations .............................................. 2-24
2.3.1 Impedance of Single Piles....................................................... 2-25
2.3.2 Impedance of Grouped Piles ................................................... 2-28
2.3.3 Discrete Element Methods (p-y and t-z Curves) ..................... 2-32
2.4 Nonlinear Soil-Structure Interaction Models for Response History
Analysis ............................................................................................... 2-33
2.4.1 Nonlinear Structure and Equivalent-Linear Soil..................... 2-33
2.4.2 Nonlinearity in the Foundation and Soil ................................. 2-34
2.4.3 Model Comparisons and Recommendations for Response
History Analysis ..................................................................... 2-44
3. Kinematic Interaction ......................................................... 3-1

3.1 Shallow Foundations at the Ground Surface ......................................... 3-1
3.2 Embedded Shallow Foundations............................................................ 3-5
3.3 Pile Foundations .................................................................................... 3-8
3.4 Application of Transfer Functions ......................................................... 3-8
4. Implementation in Standards and Design Guidelines ................... 4-1

4.1 Force-Based Procedures......................................................................... 4-2
4.2 Displacement-Based Procedures ............................................................ 4-5
4.3 Response History Procedures ................................................................ 4-9
4.4 Nuclear Structures................................................................................ 4-10
5. Experiments and Case Studies .............................................. 5-1

5.1 Seismic Monitoring and Field Tests ...................................................... 5-1
GCR 12-917-21 Table of Contents v

5.1.1 System Studies .......................................................................... 5-1
5.1.2 Field Experiments for Foundation Impedance.......................... 5-2
5.1.3 Field Experiments for Nuclear Applications ............................ 5-5
5.2 Laboratory-Scale Tests .......................................................................... 5-6
6. State of Practice ............................................................... 6-1

6.1 Overview................................................................................................ 6-1
6.2 Observations .......................................................................................... 6-1
6.2.1 General Observations................................................................ 6-2
6.2.2 Collaboration Between Design Professionals ........................... 6-2
6.2.3 Information Needed by the Structural Engineer ....................... 6-3
6.2.4 Information Needed by the Geotechnical Engineer .................. 6-3
6.2.5 Understanding of Soil-Structure Interaction Principles ............ 6-4
6.2.6 Analysis Procedures.................................................................. 6-4
6.2.7 Implementation of Soil Springs ................................................ 6-5
6.2.8 Modeling Approaches for Common Design Situations ............ 6-6
6.3 Recommendations................................................................................ 6-10
6.3.1 Improved Collaboration Between Structural and
Geotechnical Engineers .......................................................... 6-10
6.3.2 Checklist of Information That Should be Provided to
Geotechnical Engineers .......................................................... 6-11
6.3.3 Checklist of Information Needed by Structural Engineers ..... 6-13
6.3.4 Sample Format for Soil Spring Characterization .................... 6-16
6.3.5 Guidance on Modeling Approaches for Typical Foundation
Situations ............................................................................... 6-17
6.3.6 Example Applications of Simplified Soil-Structure
Interaction ............................................................................... 6-17
7. Example Applications ......................................................... 7-1

7.1 Overview of Example Applications ....................................................... 7-1
7.1.1 Summary of Results from Prior Studies ................................... 7-1
7.1.2 Building Selection .................................................................... 7-3
7.1.3 Modeling Approaches ............................................................... 7-4
7.2 Sherman Oaks Building ......................................................................... 7-6
7.2.1 Site Characterization ................................................................. 7-7
7.2.2 Ground Motion Recordings ...................................................... 7-8
7.2.3 Foundation Conditions............................................................ 7-11
7.2.4 Development of Foundation Springs and Dashpots................ 7-12
7.2.5 Analysis Results...................................................................... 7-19
7.3 Sherman Oaks Building Parametric Studies ........................................ 7-27
7.3.1 Parametric Study Results ........................................................ 7-29
7.3.2 Parametric Study Observations............................................... 7-35
7.4 Walnut Creek Building ........................................................................ 7-36
7.4.1 Site Characterization ............................................................... 7-37
7.4.2 Ground Motion Recordings .................................................... 7-38
7.4.3 Foundation Conditions............................................................ 7-41
7.4.4 Development of Foundation Springs and Dashpots................ 7-42
7.4.5 Analysis Results...................................................................... 7-47
7.5 Example Applications Summary and Conclusions .............................. 7-51
vi Table of Contents GCR 12-917-21

8. Conclusions and Recommendations ........................................ 8-1
8.1 When is Consideration of Soil-Structure Interaction Important? .......... 8-1
8.2 Summary of Soil-Structure Interaction Analysis Procedures ................ 8-3
8.2.1 Developing Springs and Dashpots ............................................ 8-3
8.2.2 Modifying Ground Motions due to Kinematic
Interaction ................................................................................. 8-7
8.2.3 Incorporating Soil-Structure Interaction in Response History
Analyses .................................................................................... 8-7
8.3 Future Research Needs .......................................................................... 8-8
8.3.1 Theme 1: Expansion of Current Studies ................................... 8-8
8.3.2 Theme 2: Research to Address Knowledge Gaps ..................... 8-9
Appendix A: Sherman Oaks Building Model Development ...................... A-1

A.1 Baseline Model Development ............................................................... A-1
A.2 Baseline Model Calibration .................................................................. A-4
Appendix B: Sherman Oaks Building Stick Model Development ............... B-1

B.1 Stick Model Development .................................................................... B-1
B.1.1 Moment Frame Stick Model Development.............................. B-1
B.1.2 Moment Frame Stick Model Calibration ................................. B-6
B.2 Additional Moment Frame Stick Model Comparisons to Full-
Building Model Results ...................................................................... B-14
B.3 Moment Frame Stick Model Comparisons for Different Foundation
Configurations .................................................................................... B-27
B.4 Shear Wall Stick Model Development ............................................... B-32
B.5 Shear Wall Stick Model Comparisons ................................................ B-33
B.6 Moment Frame Stick Model and Shear Wall Stick Model
Comparisons ....................................................................................... B-50
Appendix C: Walnut Creek Building Model Development ....................... C-1

C.1 Walnut Creek Building Model Development ....................................... C-1
C.2 Model Validation .................................................................................. C-7
Symbols ................................................................................... D-1
References ............................................................................... E-1
Project Participants .................................................................... F-1
GCR 12-917-21 Table of Contents vii

List of Figures
Figure 1-1 Schematic illustration of a direct analysis of soil-structure interaction

using continuum modeling by finite elements ....................................... 1-4
Figure 1-2 Schematic illustration of a substructure approach to analysis of soil-
structure interaction using either: (i) rigid foundation; or (ii) flexible
foundation assumptions ......................................................................... 1-6
Figure 2-1 Schematic illustration of deflections caused by force applied to:
(a) fixed-base structure; and (b) structure with vertical, horizontal,
and rotational flexibility at its base ........................................................ 2-2
Plot of period lengthening ratio ( T T ) and foundation damping (f )

~
Figure 2-2
versus structure-to-soil-stiffness ratio for square foundations (L = B)
and varying ratios of h/B ........................................................................ 2-4
Figure 2-3 Illustration of inertial SSI effects on spectral acceleration (base shear)
associated with period lengthening and change in damping .................. 2-7
Figure 2-4 Plot of dynamic stiffness modifiers and damping ratios versus
dimensionless frequency, for rectangular footings resting on the surface
=0.33: (a) geometry; (b) x-direction; and (c) y-direction .................. 2-15
of a homogeneous half-space, with zero hysteretic damping, and
dimensionless frequency, for square footings embedded in a
=0.33: (a) geometry; and (b) x-direction (y-direction similar) .......... 2-16
homogeneous half-space, with zero hysteretic damping, and
Figure 2-6 Plot of dynamic stiffness modifiers and radiation damping ratios versus
dimensionless frequency comparing uniform half-space and non-uniform
profiles in which G varies with depth: (a) translation for circular
foundations; and (b) rotation for square foundations ........................... 2-19
rotational radiation damping ratio (rr) for circular foundations

Figure 2-7 Effect of flexible foundation elements on rotational stiffness (krr) and
supporting a rigid core and flexible perimeter walls............................ 2-20

Figure 2-8 Vertical spring distribution used to reproduce total rotational stiffness
kyy. A comparable geometry can be shown in the y-z plane (using
foundation dimension 2B) to reproduce kxx ......................................... 2-22
Figure 2-9 Schematic illustration showing replacement of piles with equivalent
springs for translational and vertical (rotational) impedance............... 2-25
GCR 12-917-21 List of Figures ix

with =0.4 and p/s =1.3: (a) geometric parameters; (b) lateral
dimensionless frequency, for single piles in a homogeneous half-space,
loading; and (c) vertical loading .......................................................... 2-26

Figure 2-11 Plots of pile group efficiency factors and damping ratios versus
dimensionless frequency for square pile groups for: (a) lateral loading at
head of pile group under zero cap rotation; (b) moment at head of pile
group, introducing rocking under zero cap translation; and (c) vertical
loading at head of pile group ............................................................... 2-31
Figure 2-12 Comparison between the impedance (stiffness and damping ratio)
of a 3x3 pile group and the impedance of a footing with equivalent
dimensions for: (a) lateral loading at head of pile group; and (b)
moment at head of pile group, introducing rocking ............................. 2-32
Figure 2-13 Schematic illustration of a Beam-on-Nonlinear Winkler Foundation
(BNWF) model: (a) hypothesized foundation-superstructure system;
(b) idealized model; and (c) variable vertical stiffness distribution..... 2-36
Figure 2-14 Cyclic response of OpenSees BNWF springs subjected to a sinusoidal
displacement: (a) q–z spring (Qzsimple2 material model); (b) p–x spring
(Pxsimple1 material model); and (c) t–x spring (Txsimple1 material
model) .................................................................................................. 2-38
Figure 2-15 Comparison of BNWF model response to centrifuge tests for a
medium aspect ratio building on dense sand, with Dr = 80%, and
FSv = 2.3: (a) moment-rotation; (b) settlement-rotation; (c) shear-
sliding; and (d) settlement-sliding ....................................................... 2-40
Figure 2-16 Conceptual illustration of a macro-element contact interface model
(CIM) ................................................................................................... 2-41
Figure 2-17 Contact interface model (CIM): (a) definition of critical contact
length; and (b) tracking of foundation position relative to soil
pressures .............................................................................................. 2-42
Figure 2-18 Comparison of CIM simulation to centrifuge tests for a medium aspect
ratio building on dense sand, with Dr = 80%, and FSv = 2.6: (a) moment-
rotation; (b) settlement-rotation; (c) shear-sliding; and (d) settlement-
sliding .................................................................................................. 2-43
Figure 3-1 Illustration of foundation subjected to inclined shear waves:
(a) schematic geometry; (b) transfer functions between FIM and free-
field motion for wave passage using a semi-empirical model for
incoherent waves ................................................................................... 3-2
Figure 3-2 Illustration of foundation subjected to vertically incident shear waves:
(a) schematic geometry; and (b) transfer functions for horizontal
foundation translation and rocking ........................................................ 3-6
x List of Figures GCR 12-917-21

Figure 3-3 Illustration of base-slab averaging and embedment effects at CSMIP
Station 23497 during the 1987 Whittier Earthquake: (a) foundation and
sensor location plan; (b) acceleration time histories; and (c) observed
and model-based transfer functions ....................................................... 3-7
Figure 3-4 Illustration of differences between transfer functions and ratios of
response spectra using data from CSMIP Station 23497 during the 1987
Whittier Earthquake: (a) E-W response spectra recorded at the site; and
(b) ratio of response spectra from recordings compared to model
prediction for RRS and transfer function, Hu ....................................... 3-10
Figure 4-1 Schematic illustration of the shape of the design response spectrum
in the NEHRP Recommended Provisions .............................................. 4-2
Figure 4-2 Plot of relationship between period lengthening and foundation
damping ................................................................................................. 4-4
Figure 4-3 Schematic illustration of a pushover analysis and development of a
pushover curve for a structure with a flexible base ............................... 4-5
Figure 4-4 Schematic illustration of procedures used to combine pushover curves
with design response spectra to estimate seismic displacements in a
structure ................................................................................................. 4-7
Figure 4-5 Ratios of response spectra (uFIM/ug) for base slab averaging using the
semi-empirical formulation adopted in FEMA-440............................... 4-8
Figure 4-6 Schematic illustration of a tall building with subterranean levels:
(a) complete system; (b) simplified model for service-level
earthquake intensity; and (c) simplified foundation model for
maximum considered earthquake intensity .......................................... 4-10
Figure 5-1 Non-normalized impedance values from experimental data compared
with theoretical predictions for a nuclear containment structure at
Hualien, Taiwan for translational (top) and rotational (bottom)
modes ..................................................................................................... 5-3
Figure 5-2 Normalized impedance values from experimental data compared with
theoretical predictions for the Garner Valley site for: (a) translational;
and (b) rotational modes ........................................................................ 5-4
Figure 5-3 Comparison of damping ratio of vibrating blocks of various sizes
predicted using impedance models for a half-space and measured
from free-vibration tests on laboratory-scale models ............................ 5-8
Figure 6-1 Illustration of an embedded building configuration with a basement
surrounded by soil and a level grade on all sides ................................... 6-7
Figure 6-2 Modeling approaches for embedded foundations .................................. 6-7
Figure 6-3 Illustration of a partially embedded building configuration with
unbalanced loading ................................................................................ 6-9
Figure 6-4 Illustration of a building configuration without a basement ................ 6-10
GCR 12-917-21 List of Figures xi

Figure 6-5 Sample format for presentation of force-displacement relationships
for soil springs ..................................................................................... 6-16
Figure 7-1 Exterior elevations of two case-study buildings: (a) 13-story Sherman
Oaks building; and (b) 10-story Walnut Creek building ....................... 7-3
Figure 7-2 Modeling approaches considered in example applications .................... 7-4
Figure 7-3 Plan view of the Sherman Oaks site showing locations of borehole
and geophysical logs used for site characterization ............................... 7-7
Figure 7-4 Subsurface characteristics of the Sherman Oaks site: (a) shear wave
velocity profile; and (b) material profile................................................ 7-8
Figure 7-5 Sherman Oaks building (CSMIP Station No. 24322) instrument
locations ................................................................................................. 7-9
Figure 7-6 Response spectra for recorded foundation input motion, uFIM, and
inferred free-field motion, ug, at the Sherman Oaks building, 1994
Northridge earthquake ......................................................................... 7-10
Figure 7-7 Computed variation of peak ground acceleration (PGA) and peak
ground velocity (PGV) with depth at the Sherman Oaks site, 1994
Northridge earthquake ......................................................................... 7-11
Figure 7-8 Sherman Oaks building foundation plan, based on construction
drawings provided by CSMIP.............................................................. 7-12
Figure 7-9 Vertical spring and dashpot intensities distributed over the footprint
of the Sherman Oaks building. Solid lines represent tributary area
boundaries and solid circles represent column nodes .......................... 7-16
Figure 7-10 Plan view, transverse section, and longitudinal section of foundation
nodes for the Sherman Oaks building, and calculation of associated
spring stiffnesses .................................................................................. 7-18
Figure 7-11 Comparison of displacements, drifts, story shears, and accelerations
between Model MB, MB.1 and MB.2 for the Sherman Oaks
building ................................................................................................ 7-22
between Model MB and Model 4 for the Sherman Oaks building ...... 7-23
Figure 7-16 Comparison of peak displacements, drift ratios, and story shears
from all model configurations, in the transverse direction .................. 7-27
Figure 7-17 Elevation an idealized stick model of the Sherman Oaks building...... 7-29
xii List of Figures GCR 12-917-21

Figure 7-18 Comparison of maximum displacement envelopes in each direction
for foundation Models 1, 2, 4a, and MB, for each structural system
variant (MF, SW, SW1, SW2, and SW3) ............................................ 7-31
Figure 7-19 Comparison of maximum story drift envelopes in each direction
Figure 7-20 Comparison of maximum story shear envelopes in each direction
Figure 7-21 Comparison of peak floor acceleration envelopes in each direction
Figure 7-22 Walnut Creek building typical floor plan, based on construction
Figure 7-23 Walnut Creek building longitudinal elevation, based on
construction drawings provided by CSMIP ......................................... 7-37
Figure 7-24 Plan view of the Walnut Creek site showing borehole and refraction
survey locations used for site characterization .................................... 7-38
Figure 7-25 Subsurface characteristics of the Walnut Creek site: (a) shear wave
velocity profile; and (b) material profile .............................................. 7-39
Figure 7-26 CSMIP Station No. 58364: Walnut Creek 10-Story commercial
building, sensor location sketch ........................................................... 7-40
Figure 7-27 Response spectra for foundation input motion, uFIM, and free-field
motion, ug, at the Sherman Oaks building, 1989 Loma Prieta
earthquake ............................................................................................ 7-41
Figure 7-28 Walnut Creek building foundation plan, based on construction
Figure 7-29 Vertical spring and dashpot intensities distributed over the mat
foundation of the Walnut Creek building ............................................ 7-46
Figure 7-30 Schematic illustration of Model 4 (bathtub); Model 2 (fixed at
foundation level); and Model 1 (fixed at ground surface) for the
Walnut Creek building ......................................................................... 7-48
Figure 7-31 Comparison of roof acceleration and displacement histories for
Model 4 (taken as Baseline Model MB), Model 1, and Model 2 for
the Walnut Creek building ................................................................... 7-49
Figure 7-32 Comparison of maximum relative displacement, peak floor
acceleration, maximum story drift, and story shear coefficients for
Model 4 (taken as Baseline Model MB), Model 1, and Model 2 for
the Walnut Creek building ................................................................... 7-50
GCR 12-917-21 List of Figures xiii

Figure 7-33 Comparison of modeled versus recorded roof displacement histories
due to base rotation of the Walnut Creek building .............................. 7-50
Figure A-1 Renderings of the Sherman Oaks building, including cut-away views
showing structural details in the first story, basement, and foundation
regions................................................................................................... A-2
Figure A-2 Schematic illustration of elastic springs connecting the framing lines
of the superstructure with the wall lines in the basement levels ........... A-2
Figure A-3 Comparison of modeled versus recorded displacement histories for
Model MB in the 1994 Northridge earthquake: roof level (a) and (b);
8th floor (c) and (d); ground level (e) and (f); foundation level (g) and
(h); and roof displacement due to base rocking (i) and (j) .................... A-6
Figure A-4 Comparison of modeled versus recorded peak displacements and
accelerations for Model MB in the 1994 Northridge earthquake ......... A-7
Model MB in the 1992 Landers earthquake: roof level (a) and (b);
(h); and roof displacement due to base rocking (i) and (j) .................... A-8
accelerations for Model MB in the 1992 Landers earthquake .............. A-9
Model MB in the 1987 Whittier earthquake: roof level (a) and (b);
(h); and roof displacement due to base rocking (i) and (j) .................. A-10
accelerations for Model MB in the 1987 Whittier earthquake............ A-11
Figure B-1 Plan view of the full-building model of the Sherman Oaks
building. ................................................................................................ B-2
Figure B-2 Longitudinal elevation (X-direction) of the full-building model of
the Sherman Oaks building. .................................................................. B-2
Figure B-3 Transverse elevation (Y-direction) of the full-building model of
the Sherman Oaks building. .................................................................. B-3
Figure B-4 Elevation the idealized stick model of the Sherman Oaks building ..... B-4
Figure B-5 Idealized force-displacement curve for the Sherman Oaks building .... B-4
Figure B-6 Shear wall and basement level node definitions, constraints, and
connectivity for the Sherman Oaks stick model ................................... B-5
Figure B-7 Comparison of roof displacement histories for the moment frame
stick model and the full-building Model MB ....................................... B-9
Figure B-8 Comparison of maximum displacement envelopes for the moment
frame stick Model MB and the full-building Model MB ................... B-10
xiv List of Figures GCR 12-917-21

Figure B-9 Comparison of maximum story drift envelopes for the moment
frame stick Model MB and the full-building Model MB.................... B-11
Figure B-10 Comparison of maximum story shear envelopes for the moment
Figure B-11 Comparison of peak floor acceleration envelopes for the moment
frame stick Model 1 and the full-building Model 1 ............................ B-15
Figure B-14 Comparison of normalized story shear envelopes for the moment
frame stick Model 4a and the full-building Model 4a ........................ B-23
frame stick Models 1, 2, 4a, and MB .................................................. B-28
GCR 12-917-21 List of Figures xv

Figure B-28 Comparison of maximum displacement envelopes for the SW stick
Models 1, 2, 4a, and MB..................................................................... B-34
Figure B-29 Comparison of maximum story drift envelopes for the SW stick
Models 1, 2, 4a, and MB..................................................................... B-35
Figure B-30 Comparison of normalized story shear envelopes for the SW stick
Models 1, 2, 4a, and MB..................................................................... B-36
Figure B-31 Comparison of peak floor acceleration envelopes for the SW stick
Models 1, 2, 4a, and MB..................................................................... B-37
Figure B-32 Comparison of maximum displacement envelopes for the SW1 stick
Models 1, 2, 4a, and MB..................................................................... B-38
Figure B-33 Comparison of maximum story drift envelopes for the SW1 stick
Models 1, 2, 4a, and MB..................................................................... B-39
Figure B-34 Comparison of normalized story shear envelopes for the SW1 stick
Models 1, 2, 4a, and MB..................................................................... B-40
Figure B-35 Comparison of peak floor acceleration envelopes for the SW1 stick
Models 1, 2, 4a, and MB..................................................................... B-41
Models 1, 2, 4a, and MB..................................................................... B-42
Models 1, 2, 4a, and MB..................................................................... B-43
Models 1, 2, 4a, and MB..................................................................... B-44
Models 1, 2, 4a, and MB..................................................................... B-45
Models 1, 2, 4a, and MB..................................................................... B-46
Models 1, 2, 4a, and MB..................................................................... B-47
Models 1, 2, 4a, and MB..................................................................... B-48
Models 1, 2, 4a, and MB..................................................................... B-49
frame stick Model MB and the shear wall stick Model MB ............... B-50
xvi List of Figures GCR 12-917-21

Figure C-1 Renderings of the Walnut Creek building, including cut-away views
showing: (a) the exterior facade; (b) structural framing; (c) core wall
elevations; (d) core wall layout; (e) details of the precast cladding
system; and (f) interior layout of a typical floor level .......................... C-2
Figure C-2 Plan views of the Walnut Creek building shear wall core:
(a) foundation level; (b) first floor; and (c) typical floor level ............. C-3
Figure C-3 Illustrations of the Walnut Creek building OpenSees model:
(a) three-dimensional model; and (b) simplified two-dimensional
model .................................................................................................... C-6
Figure C-4 Horizontal and vertical spring and dashpot locations for the Walnut
Creek building model............................................................................ C-7
Figure C-5 Comparison of relative roof displacement histories from recorded
data, two-dimensional (2D) model results, and three-dimensional
(3D) model results for the Walnut Creek building ............................... C-7
Figure C-6 5%-damped elastic response spectra of recorded motions for the
Walnut Creek building in the 1989 Loma Prieta earthquake ................ C-8
Figure C-7 Comparison of modeled versus recorded acceleration and relative
displacement histories for the Walnut Creek building in the 1989 Loma
Prieta earthquake: roof level (a) and (b); 8th floor (c) and (d); 3rd floor
(e) and (f); and ground floor (g) and (h) ............................................... C-9
Figure C-8 Comparison of modeled versus recorded maximum relative
displacements and peak accelerations for the Walnut Creek building
in the 1989 Loma Prieta earthquake ................................................... C-10
Figure C-9 Maximum story drift profile for the Walnut Creek building model
subjected to the 1989 Loma Prieta earthquake ................................... C-10
Figure C-10 High-pass filtering of vertical motions at base level showing the
effects of corner frequency on displacement and rocking histories .... C-11
GCR 12-917-21 List of Figures xvii

List of Tables
Table 2-1 Values of Shear Wave Velocity and Shear Modulus Reduction for
Various Site Classes and Shaking Amplitudes .................................... 2-10
Table 2-2a Elastic Solutions for Static Stiffness of Rigid Footings at the
Ground Surface .................................................................................... 2-11
Table 2-2b Embedment Correction Factors for Static Stiffness of Rigid
Footings ............................................................................................... 2-12
Table 2-3a Dynamic Stiffness Modifiers and Radiation Damping Ratios for
Rigid Footings...................................................................................... 2-13
Table 2-3b Dynamic Stiffness Modifiers and Radiation Damping Ratios for
Embedded Footings ............................................................................. 2-14
Table 2-4a Equations for Static Stiffness of Single Piles ...................................... 2-29
Table 2-4b Equations for Dynamic Stiffness Modifiers and Damping Ratios for
Single Piles .......................................................................................... 2-30
Table 5-1 Summary of Field Forced-Vibration Tests Used to Evaluate
Foundation Impedance Functions .......................................................... 5-2
Table 5-2 Summary of Laboratory-Scale Tests of Dynamic Response of
Footings ................................................................................................. 5-7
Table 7-1 Summary of Modeling Approaches Considered in Example
Applications ........................................................................................... 7-6
Table 7-2 Summary of Earthquake Events Recorded at the Sherman Oaks
Building ............................................................................................... 7-10
Table 7-3 Summary of Effective Profile Depths and Average Effective Profile
Velocities for the Sherman Oaks Building .......................................... 7-13
Table 7-4 Calculation of Shallow Foundation Stiffness and Damping
Parameters for the Sherman Oaks Building ......................................... 7-15
Table 7-5 Calculation of Pile Stiffness and Damping Parameters for the
Sherman Oaks Building ....................................................................... 7-17
Table 7-6 Comparison of Fundamental Periods for Alternative Foundation
Modeling Configurations for the Sherman Oaks Building .................. 7-20
Table 7-7 Summary of Sherman Oaks Building Parametric Stick Models .......... 7-28
Table 7-8 Comparison of Results for Moment Frame, Core Shear Wall, and
Perimeter Shear Wall Stick Models and Alternative Foundation
Modeling Configurations ..................................................................... 7-30
GCR 12-917-21 List of Tables xix

Table 7-9 Summary of Earthquake Events Recorded at the Walnut Creek
Building ............................................................................................... 7-39
Table 7-10 Summary of Effective Profile Depths and Average Effective Profile
Velocities for the Walnut Creek Building ........................................... 7-43
Table 7-11 Comparison of Average Effective Profile Velocities with and
without Correction for Structural Overburden Weight ........................ 7-43
Table 7-12 Evaluation of Soil-to-Foundation Stiffness Ratios for Flexible Mat
Foundation below Shear Wall Core in the Walnut Creek Building ..... 7-44
Table 7-13 Calculation of Mat Foundation Stiffness and Damping Parameters
for the Walnut Creek Building............................................................. 7-45
Modeling Configurations for the Walnut Creek Building ................... 7-48
Table A-1 Comparison of Measured and Modeled Periods for the Sherman
Oaks Building ....................................................................................... A-5
Table B-1 Axial and Rotational Properties of the Moment Frame Stick Model ... B-7
Table B-2 Lateral Force-Deformation Properties of the Moment Frame Stick
Model .................................................................................................... B-7
Table B-3 Masses in the Moment Frame Stick Model .......................................... B-8
Table B-4 Comparison of Modeled Periods for the Moment Frame Stick
Model and the Full-Building Model MB .............................................. B-8
Model and the Full-Building Model for each Alternative
Foundation Configuration ................................................................... B-14
Table B-6 Comparison of Modeled Periods for the Shear Wall Stick Models
for each Alternative Foundation Configuration .................................. B-32
xx List of Tables GCR 12-917-21

Chapter 1
Introduction
The response of a structure to earthquake shaking is affected by interactions between

three linked systems: the structure, the foundation, and the soil underlying and
surrounding the foundation (FEMA, 2009). Soil-structure interaction analysis
evaluates the collective response of these systems to a specified ground motion. The
terms Soil-Structure Interaction (SSI) and Soil-Foundation-Structure Interaction
(SFSI) are both used to describe this effect in the literature. In this report, the
foundation is considered part of the structure, and the term SSI has been adopted.
This report presents a synthesis of the body of knowledge contained in SSI literature,
which has been distilled into a concise narrative and harmonized under a consistent
set of variables and units. Specific techniques are described by which SSI
phenomena can be simulated in engineering practice, and recommendations for
modeling seismic soil-structure interaction effects on building structures are
provided.
1.1 Background and Objectives
Problems associated with the practical application of SSI for building structures are
rooted in a poor understanding of fundamental SSI principles. Soil-structure
interaction topics are generally not taught in graduate earthquake engineering
courses, so most engineers attempting SSI in practice must learn the subject on their
own. Unfortunately, practice is hindered by a literature that is often difficult to
understand, and codes and standards that contain limited guidance. Most articles rely
heavily on the use of wave equations in several dimensions and complex arithmetic
to formulate solutions and express results. Moreover, nomenclature is often
inconsistent, and practical examples of SSI applications are sparse. This gives rise to
the present situation in which soil-structure interaction is seldom applied, and when it
is, modeling protocols vary widely and are not always well conceived.
Although the NEHRP Recommended Provisions for Seismic Regulations for New
Buildings and Other Structures has provided guidance for consideration of SSI
effects in forced-based procedures for several decades, these procedures have not
found significant use in practice. Practical application of SSI gained momentum
following publication of FEMA 440, Improvement of Inelastic Seismic Analysis
Procedures (FEMA, 2005), which provided the design community with procedures
for incorporating the effects of soil-structure interaction in nonlinear static pushover-
type analyses. These procedures were eventually adopted into ASCE/SEI 41-06,
GCR 12-917-21 1: Introduction 1-1

Seismic Rehabilitation of Existing Buildings (ASCE, 2007). Because they were
developed for nonlinear static analysis, however, they were not directly applicable to
response history analysis, which requires more careful consideration of specific
components of soil-foundation interaction and the manner by which input ground
motions should be applied to structural models.
In 2009, the National Institute of Standards and Technology (NIST) initiated the Task
Order 69221 Project entitled “Improved Procedures for Characterizing and Modeling
Soil-Structure Interaction for Performance-Based Seismic Engineering.” The
purpose of this project was to develop consensus guidance for implementing soil-
structure interaction in response history analyses, such that input ground motions
accurately reflect the input at the base of structures, and that structural models
include elements that account for the geotechnical and foundation conditions
associated with the building under consideration. Work also included an extensive
review of available research on soil-structure interaction, evaluation of existing SSI
guidelines for static-type analyses, and development of recommendations for
improvement where necessary.
This report is intended to provide background for addressing three fundamental

questions that arise when contemplating the use of SSI in the seismic analysis of
building structures:
1. When is the use of foundation springs and dashpots important, and which
structural response parameters are affected?
2. Under what conditions is consideration of the differences between foundation

input motions and free-field ground motions important?
3. What field and laboratory investigations are necessary to develop foundation

springs and dashpots for SSI analysis?
Once the decision to implement SSI has been made, a basic level of understanding of
the physical phenomenon and a practical analysis methodology for simulating its
effects are needed. This report describes the principal components of SSI in a clear
and concise way, and consistent nomenclature is used throughout. Explicit
computational tools that can be used in engineering practice are provided, and
applications of SSI to force-based analysis procedures, pushover (displacement-
based) procedures, and response history analysis procedures are described.
As part of the work, soil-structure interaction procedures were applied to detailed

example applications that investigate the influence of various SSI components on
response history analysis results. The resulting recommendations are drawn from
these studies, along with the results of other studies in recent literature.
1-2 1: Introduction GCR 12-917-21

Implementation of SSI within a design setting requires close collaboration between
structural and geotechnical engineers. Neither discipline alone is likely to have
sufficient knowledge of structural, foundation, and site considerations necessary to
properly complete a meaningful analysis considering SSI effects. As such, the
intended audience for this work is expected to be structural and geotechnical
engineers experienced in earthquake engineering with a working knowledge of
structural and soil dynamics principles. The mathematical complexities are kept to a
minimum to make the content as accessible as possible to non-experts in structural
dynamics and the mathematics of wave propagation. Detailed guidance on
communication between structural and geotechnical engineers is also provided.
1.2 Overview of Soil-Structure Interaction
A seismic soil-structure interaction analysis evaluates the collective response of the

structure, the foundation, and the geologic media underlying and surrounding the
foundation, to a specified free-field ground motion. The term free-field refers to
motions that are not affected by structural vibrations or the scattering of waves at,
and around, the foundation. SSI effects are absent for the theoretical condition of a
rigid foundation supported on rigid soil. Accordingly, SSI accounts for the
difference between the actual response of the structure and the response of the
theoretical, rigid base condition.
In Part 2 of FEMA P-750, NEHRP Recommended Seismic Provisions for New

Buildings and Other Structures (FEMA, 2009), SSI effects are categorized as inertial
interaction effects, kinematic interaction effects, and soil-foundation flexibility
effects. The terms kinematic and inertial interaction were introduced in 1975 by
Robert Whitman (Kausel, 2010). In the context of engineering analysis and design,
these effects are related to:
 Foundation stiffness and damping. Inertia developed in a vibrating structure
gives rise to base shear, moment, and torsion. These forces generate
displacements and rotations at the soil-foundation interface. These displacements
and rotations are only possible because of flexibility in the soil-foundation
system, which significantly contributes to overall structural flexibility (and
increases the building period). Moreover, these displacements give rise to energy
dissipation via radiation damping and hysteretic soil damping, which can
significantly affect overall system damping. Since these effects are rooted in
structural inertia, they are referred to as inertial interaction effects.
 Variations between foundation input motions and free-field ground motions.
Foundation input motions and free-field motions can differ because of: (i)
kinematic interaction, in which stiff foundation elements placed at or below the
ground surface cause foundation motions to deviate from free-field motions due
to base slab averaging, wave scattering, and embedment effects in the absence of

structure and foundation inertia; and (ii) relative displacements and rotations
between the foundation and the free-field associated with structure and
foundation inertia.
 Foundation Deformations. Flexural, axial, and shear deformations of structural
foundation elements occur as a result of forces and displacements applied by the
superstructure and the soil medium. These represent the seismic demands for
which foundation components should be designed, and they could be significant,
especially for flexible foundations such as rafts and piles.
Methods that can be used to evaluate the above effects can be categorized as direct
and substructure approaches. In a direct analysis, the soil and structure are included
within the same model and analyzed as a complete system. In a substructure
approach, the SSI problem is partitioned into distinct parts that are combined to
formulate the complete solution.
Direct Analysis. As schematically depicted in Figure 1-1, the soil is often

represented as a continuum (e.g., finite elements) along with foundation and
structural elements, transmitting boundaries at the limits of the soil mesh, and
interface elements at the edges of the foundation.
Figure 1-1 Schematic illustration of a direct analysis of soil-structure interaction

using continuum modeling by finite elements.
Evaluation of site response using wave propagation analysis through the soil is
important to this approach. Such analyses are most often performed using an
equivalent linear representation of soil properties in finite element, finite difference,

or boundary element numerical formulations (Wolf, 1985; Lysmer et al., 1999).
Direct analyses can address all of the SSI effects described above, but incorporation
of kinematic interaction is challenging because it requires specification of spatially
variable input motions in three dimensions.
Because direct solution of the SSI problem is difficult from a computational
standpoint, especially when the system is geometrically complex or contains
significant nonlinearities in the soil or structural materials, it is rarely used in
practice.
Substructure Approach. Proper consideration of SSI effects in a substructure
approach requires: (i) an evaluation of free-field soil motions and corresponding soil
material properties; (ii) an evaluation of transfer functions to convert free-field
motions to foundation input motions; (iii) incorporation of springs and dashpots (or
more complex nonlinear elements) to represent the stiffness and damping at the soil-
foundation interface; and (iv) a response analysis of the combined structure-
spring/dashpot system with the foundation input motion applied.
The superposition inherent in a substructure approach requires an assumption of
linear soil and structure behavior, although in practice this requirement is often
followed only in an equivalent-linear sense. As depicted in Figure 1-2, the steps in a
substructure approach are as follows:
 Specification of a foundation input motion (FIM), which is the motion of the
base-slab that accounts for the stiffness and geometry of the foundation. Because
inertia is dealt with separately, the FIM applies for the theoretical condition of
the base-slab and structure having no mass (Figure 1-2b). This motion generally
differs from the free-field motion, involves both translational and rotational
components, and represents the seismic demand applied to the foundation and
structural system. The variation between free-field and foundation input motions
is expressed by a transfer function that represents the ratio of foundation/free-
field motion in the frequency domain. Since inertial effects are neglected, the
transfer function represents the effects of kinematic interaction only.
An essential first step in defining the FIM is to evaluate the free-field response of
the site, which is the spatial and temporal variation of ground motion in the
absence of the structure and foundation. This task requires that the earthquake
input motion in the free field is known, either at a specific point (e.g., ground
surface, rock-outcrop) or in the form of incident waves (e.g., oblique shear
waves) propagating up from a reference depth.
Having established the free-field motion, wave-propagation analyses are
performed to estimate the foundation input motion along the planned soil-
foundation interface, as depicted in Figure 1-2d. Equivalent linear properties for

the soil (e.g., shear modulus, material damping) can be evaluated as part of this
analysis.
(rotational and vertical

kz springs in parallel)
Figure 1-2 Schematic illustration of a substructure approach to analysis of soil-

structure interaction using either: (i) rigid foundation; or (ii) flexible
foundation assumptions.
 The stiffness and damping characteristics of the soil-foundation interaction are

characterized using relatively simple impedance function models or a series of
distributed springs and dashpots. Impedance functions represent the frequency-
dependent stiffness and damping characteristics of soil-foundation interaction.
Use of impedance function models for rigid foundations is illustrated in Figure
1-2c(i). Use of a series of distributed springs and dashpots acting around the
foundation is illustrated in Figure 1-2c(ii). The latter case of distributed springs
and dashpots is needed when foundation elements are non-rigid, or when internal

demands (e.g., moments, shears, deformations) are required outcomes of the
analysis.
 The superstructure is modeled above the foundation and the system is excited
through the foundation by displacing the ends of the springs and dashpots using
the rocking and translational components of the FIM. It should be noted that
FIM varies with depth. In the case of the distributed spring and dashpot model,
differential ground displacements should be applied over the depth of
embedment. This application of spatially variable displacements introduces a
rotational component to the FIM, which is why a rotational component does not
specifically appear in Figure 1-2d(ii).
1.3 Report Organization and Content
This report summarizes the body of knowledge on soil-structure interaction, and

provides specific recommendations for modeling seismic soil-structure interaction
effects on building structures in engineering practice.
Chapter 2 describes inertial SSI effects, beginning with the behavior of simple
structure-soil systems to provide insight into the conditions for which inertial SSI
effects are most critical. Also presented are detailed procedures for computing
foundation stiffness and damping, both for idealized conditions in classical solutions
and for more realistic conditions that may include flexible foundation systems, non-
uniform soil, and material nonlinearity. Both shallow foundation systems (e.g.,
footings, mats) and deep foundation systems (e.g., piles) are discussed.
Chapter 3 describes the manner by which transfer functions can be computed for
various foundation configurations considering kinematic interaction effects, and how
transfer function amplitudes can be used to modify response spectral ordinates.
Chapter 4 describes how SSI procedures described in Chapter 2 and Chapter 3 are
implemented in seismic design provisions contained in currently available consensus
standards and design guidelines.
Chapter 5 describes the relatively limited number of SSI experiments and case
history studies available in the literature, and summarizes some of the lessons learned
from that work. Efforts to calibrate or verify analysis procedures against laboratory
data or field performance data are noted, where applicable.
Chapter 6 presents a summary of the state-of-practice in SSI analysis, along with

recommendations for improving communications between geotechnical and
structural engineers.
Chapter 7 provides detailed example applications of the substructure analysis

approach and procedures described in this report. Studies are performed on two

buildings with available earthquake recordings and varying degrees of SSI modeling
sophistication. Recommendations for implementing SSI in response history analyses
are provided, taking into consideration the capabilities of currently available
structural engineering software.
Chapter 8 answers three fundamental questions that arise when contemplating the use
of SSI in the seismic analysis of a building structure, summarizes SSI analysis
procedures that would be followed on typical project in a step-by-step manner, and
provides short-term and long-term research needs.
Appendices A, B, and C provide additional, more detailed information on the

development of, and results from, analytical models used in the example applications
of Chapter 7.
A list of Symbols defining key notation, and a list of References cited from the body
of SSI literature are provided at the end of this report.

Chapter 2
Inertial Interaction
Inertial interaction refers to displacements and rotations at the foundation level of a

structure that result from inertia-driven forces such as base shear and moment. This
chapter describes inertial soil-structure interaction effects. Inertial displacements and
rotations can be a significant source of flexibility and energy dissipation in the soil-
structure system.
Section 2.1 discusses system behavior and highlights some of the principal effects of
inertial interaction and the conditions for which its effects are significant. The
methods focus on single degree-of-freedom systems, but they can be extrapolated to
multi-degree-of-freedom systems with a dominant first mode. Section 2.2 provides a
relatively detailed description of how foundation springs and dashpots can be
specified to represent the flexibility and damping associated with soil-foundation
interaction in translational and rotational vibration modes for shallow foundations
(e.g., footings and mats). Section 2.3 provides corresponding solutions for the
stiffness and damping characteristics of deep foundations (e.g., piles and drilled
shafts). Some of the procedures given in Section 2.2 are coded into available
computer programs such as DYNA6 (Western Engineering, 2011). This program can
also be used for pile foundations, although the results are relatively approximate.
Section 2.4 presents several models that can be used to evaluate shallow foundation
response for conditions involving nonlinear material behavior or geometric
nonlinearities (i.e. gapping).
2.1 Soil-Structure System Behavior
A rigid base refers to soil supports with infinite stiffness (i.e., without soil springs).
A rigid foundation refers to foundation elements with infinite stiffness (i.e., not
deformable). A fixed base refers to a combination of a rigid foundation elements on
a rigid base. A flexible base analysis considers the compliance (i.e., deformability) of
both the foundation elements and the soil.
Consider a single degree-of-freedom structure with stiffness, k, and mass, m, resting

on a fixed base, as depicted in Figure 2-1a. A static force, F, causes deflection, :

F
(2-1)
k
From structural dynamics, the undamped natural vibration frequency, , and period,
T, of the structure are given by Clough and Penzien (1993) as:
GCR 12-917-21 2: Inertial Interaction 2-1

2
 , T  2

k m
(2-2)
m k
By substituting Equation 2-1 into Equation 2-2, an expression for the square of
period is obtained as:
2 m
T 2   2    2 
F 
2 m
(2-3)
F
(a) (b)
Figure 2-1 Schematic illustration of deflections caused by force applied to: (a)
fixed-base structure; and (b) structure with vertical, horizontal, and
rotational flexibility at its base.
Now consider the same structure with vertical, horizontal, and rotational springs at its
base, representing the effects of soil flexibility against a rigid foundation, as depicted
in Figure 2-1b. The vertical spring stiffness in the z direction is denoted kz, the
horizontal spring stiffness in the x direction is denoted kx, and the rotational spring is
denoted kyy, representing rotation in the x-z plane (about the y-y axis). If a force, F, is
applied to the mass in the x direction, the structure deflects, as it does in the fixed-
moment ( F  h) deflects the rotational spring by . Accordingly, the total deflection

base system, but the base shear (F) deflects the horizontal spring by uf , and the base
with respect to the free-field at the top of the structure,  , is:
   u f    h
F
k
F F  F h 
   
(2-4)
h
k k x  k yy 
If Equation 2-4 is substituted into Equation 2-3, an expression for flexible base
period, T , is obtained as:
2-2 2: Inertial Interaction GCR 12-917-21

2 m  1 1 h2 
T 2   2    2  m   
 k k x k yy 
 2
 
(2-5)
F
Combining expressions in Equation 2-5 and Equation 2-2 results in:
 T   1 1 h2 
    m     
2
k
T   k k x k yy 
(2-6)
m
Equation 2-6 simplifies into a classical period lengthening expression (Veletsos and
Meek, 1974):
 1 
T k kh 2
(2-7)
T k x k yy
Equation 2-7 can be applied to multi-degree-of-freedom structures by taking the

height, h, as the height of the center of mass for the first-mode shape. This is
commonly referred to as the effective modal height, which is approximately
two-thirds of the overall structure height, and taken as 0.7 times the height in
ASCE/SEI 7-10 (ASCE, 2010). In such cases, period lengthening applies to only the
first-mode period.
In previous work by Veletsos and Nair (1975) and Bielak (1975), it has been shown
that the dimensionless parameters controlling period lengthening are:
, and 
 s 4 BLh
h h B m
, , , (2-8)
V sT B L
where h is the structure height (or height to the center of mass of the first mode
mass (or effective modal mass),s is the soil mass density, and  is the Poisson’s
shape), B and L refer to the half-width and half-length of the foundation,m is the
ratio of the soil. Previous work was applicable to circular foundations, and has been
adapted here for rectangular shapes considering the ratio, B/L.
To the extent that h/T quantifies the stiffness of the superstructure, the term h/(VsT) in
Equation 2-8 represents the structure-to-soil stiffness ratio. The term h/T has units of
velocity, and will be larger for stiff lateral force resisting systems, such as shear
walls, and smaller for flexible systems, such as moment frames. The shear wave
velocity, Vs, is closely related to soil shear modulus, G, computed as:
Vs  G /  s (2-9)
For typical building structures on soil and weathered rock sites, h/(VsT) is less than
0.1 for moment frame structures, and between approximately 0.1 and 0.5 for shear
wall and braced frame structures (Stewart et al., 1999b). Period lengthening

increases markedly with structure-to-soil stiffness ratio, which is the most important
parameter controlling inertial SSI effects.
The structure-height-to-foundation-width ratio, h/B, and foundation-width-to-length
structure system. The mass ratio, m/s4BLh, is the ratio of structure mass to the mass
ratio, B/L, in Equation 2-8 are aspect ratios describing the geometry of the soil-
of soil in a volume extending to a depth equal to the structure height, h, below the
foundation. In Equation 2-7, it can be seen that period lengthening has no
fundamental dependence on mass. The mass ratio term was introduced so that period
lengthening could be related to easily recognizable characteristics such as structural
first mode period, T, and soil shear wave velocity, Vs, rather than structural stiffness,
taken as 0.15 (Veletsos and Meek, 1974). The Poisson’s ratio of the soil, , affects
k, and soil shear modulus, G. The effect of mass ratio is modest, and it is commonly
the stiffness and damping characteristics of the foundation.
Using models for the stiffness of rectangular foundations (of half-width, B; half-
length, L; and L ≥ B) resting on a homogeneous isotropic half-space with shear wave
velocity, Vs, period lengthening ratios can be calculated with the results shown in
Figure 2-2a, which is plotted for the special case of a square footing (L = B).
Plot of period lengthening ratio ( T T ) and foundation damping (f )

~
Figure 2-2
and varying ratios of h/B. In this plot,  = 0.33, B/L = 1.0, hysteretic
versus structure-to-soil-stiffness ratio for square foundations (L = B)
soil damping s = 0, mass ratio= 0.15, and exponent n = 2.

All other factors being equal, period lengthening increases with the structure-height-
rotation, . This implies that inertial SSI effects would be more significant in tall
to-foundation-width ratio, h/B, due to increased overturning moment and foundation
buildings, but this is not the case. Tall buildings typically have low h/(VsT) ratios,
which is more important for controlling inertial SSI effects. Hence period
lengthening in tall buildings is near unity (i.e., little or no period lengthening). For a
fixed ratio of h/B, period lengthening is observed to decrease modestly with
foundation-width-to-length ratio, B/L, due to increased foundation size (and therefore
stiffness) normal to the direction of loading.
associated with soil-foundation interaction, referred to as foundation damping, f.

In addition to period lengthening, system behavior is also affected by damping
This damping is composed of two parts: (1) contributions from soil hysteresis
(hysteretic damping); and (2) radiation of energy away, in the form of stress waves,
to the flexible-base system damping, 0:

from the foundation (radiation damping). Foundation damping is a direct contributor
0   f  i
T T 
1
n
(2-10)
where i is the structural damping in the superstructure assuming a fixed base, which
is generally taken as 5% for typical structural systems. More refined estimates of i
are possible based on structural system type and configuration, as described in
PEER/ATC-72-1, Modeling and Acceptance Criteria for Seismic Design and
al., 1999b) have shown that f ranges from approximately 0% to 25%. The exponent,
Analysis of Tall Buildings (ATC, 2010). Observations from case studies (Stewart et
n, on the period lengthening term in Equation 2-10 is taken as 3 for linearly viscous
structural damping, and 2 otherwise (e.g., for hysteretic damping) (Givens, 2013).
Analytical models for foundation damping have been presented by Veletsos and Nair
(1975), Bielak (1975 and 1976), Roesset (1980), Wolf (1985), Aviles and Perez-
Rocha (1996), Maravas et al. (2007), and Givens (2013), among others. The classical
solution of Veletsos and Nair accounts for the frequency dependence of foundation
circular foundation resting on a half-space. The equation for f provided by Veletsos

damping terms. It assumes structural damping to be purely viscous, and applies for a
and others is complex-valued (i.e., composed of real plus imaginary values), which
complicates the interpretation of its physical meaning. Bielak’s work utilizes the
same conditions except that the foundation is assumed to be a cylinder penetrating a
half-space to an embedment depth, D, and the resulting expressions are real-valued.
The value of exponent n in Equation 2-10 is taken as 3 for the Veletsos and Bielak
solutions because structural damping is assumed to be viscous.

The procedure given by Wolf (1985) neglects the frequency dependence of
foundation stiffness terms, and assumes foundation radiation damping to be linearly
viscous (i.e., constant dashpot coefficients for translation and rotation, cx and cyy), and
applies for a circular foundation resting on a half-space. Considering frequency
dependence, the form of Wolf’s damping expression (similar to Roesset, 1980) can
 T T  n s  1 
be re-written as:
f  s  x   yy
 T T    T    T  n yy
1 1
(2-11a)
 
ns nx
T x T yy
Where s is soil hysteretic damping, x and yy are damping ratios related to radiation
damping from translational and rotational modes (described further in Section 2.2),
and Tx and Tyy are fictitious vibration periods, calculated as if the only source of the
vibration was foundation translation or rotation, as follows:
T x  2 T yy  2
m mh 2
(2-11b)
kx k yy
Exponents ns, nx, and nyy depend on the specific form of damping associated with the
respective components of the foundation damping, and all other terms are as
previously defined. However, because none of these terms would be expected to be
Note that for n = ns, the period lengthening terms in front of the i term in Equation
linearly viscous, it is recommended to take these exponents as 2 (Givens, 2013).
2-10 and the s term in Equation 2-11a are weight factors that together sum to unity.
Accordingly, Equation 2-11a can be viewed as a “mixing rule” for damping in
different vibration modes and sources. Because Wolf’s results were produced
neglecting the frequency dependence of foundation stiffness terms, Equation 2-11a
can provide more accurate results if those effects are included in the period
lengthening calculation.
Soil hysteretic damping, s, is strain-dependent, and can typically be evaluated from
information in the literature. Classical models are summarized in Kramer (1996).
More contemporary empirical models by Darendeli (2001) and Menq (2003) account
for overburden pressure and shear strain in a consistent manner across multiple soil
types.
The Wolf solution for foundation damping in Equation 2-11a, along with the classical
Veletsos, Bielak, and Roesset solutions, neglect contributions from terms involving
the product of two damping ratios. Maravas et al. (2007) presents exact solutions in
which those terms are included. Like Wolf, Maravas et al. (2007) utilizes hysteretic
damping so exponents n = 2, and if terms involving the product of two damping
ratios are excluded, Equation 2-11a is recovered.

foundation damping f increases strongly with structure-to-soil-stiffness ratio,
As was the case for period lengthening shown in Figure 2-2a, Figure 2-2b shows that
h/(VsT). In Figure 2-2b, all exponents were taken as 2. Damping f decreases with
increasing values of h/B, indicating that lateral movements of the foundation (which
rocking (which dominates at high h/B ratios). Radiation damping terms (x and yy)
dominate at low h/B ratios) dissipate energy into soil more efficiently than foundation
are reduced significantly when a stiff bedrock layer is encountered at moderate or

shallow depths, as described further in Section 2.2.2.
Analysis procedures for T / T and f similar to those described above have been
validated relative to observations from instrumented buildings shaken by earthquakes
(Stewart et al., 1999a; 1999b). These studies show that the single most important
parameter controlling the significance of inertial interaction is h/(VsT), and that
inertial SSI effects are generally negligible for h/(VsT) < 0.1, which occurs in flexible
structures (e.g., moment frame buildings) located on competent soil or rock.
Conversely, inertial SSI effects tend to be significant for stiff structures, such as shear
wall or braced frame buildings, located on softer soils.
The effect of inertial SSI on the base shear of a building is illustrated in Figure 2-3.
Because base shear for elastic response is commonly computed based on pseudo-
spectral acceleration in the first mode, the figure depicts the variation in pseudo-
spectral acceleration versus period in both linear and log scales. The pseudo-spectral
acceleration for a flexible-base structure, Sa , is obtained by entering the spectrum
drawn for effective damping ratio, 0, at the corresponding elongated period, T .
Figure 2-3 Illustration of inertial SSI effects on spectral acceleration (base shear)
associated with period lengthening and change in damping.
The effect of SSI on base shear is related to the slope of the spectrum. Base shear
tends to increase when the slope is positive and decrease when the slope is negative.

For the common case of buildings with relatively long periods on the descending
portion of the spectrum, use of Sa (flexible base) in lieu of Sa (fixed base) typically
results in reduced base shear demand. Conversely, inertial SSI can increase the base
shear in relatively short-period structures.
The period at which the spectral peak occurs, referred to as the predominant period of
ground motion, Tp, is generally controlled by the tectonic regime, earthquake
magnitude, and site-source distance (Rathje et al., 2004), and will only match the site
period in certain cases involving large impedance contrasts due to soil layering. In
the absence of unusual site effects, typical values of Tp range from approximately 0.2
to 0.5 seconds for shallow crustal earthquakes in tectonically active regions, such as
California.
2.2 Equations for Shallow Foundation Stiffness and Damping
2.2.1 Models for Rigid Foundations and Uniform Soils
Impedance functions represent the frequency-dependent stiffness and damping

characteristics of soil-foundation interaction. Classical solutions for the complex-
valued impedance function can be written as (Luco and Westman, 1971; Veletsos and
Wei, 1971):
k j  k j  i c j (2-12a)
where k j denotes the complex-valued impedance function; j is an index denoting

modes of translational displacement or rotation; kj and cj denote the frequency-
 is the circular frequency (rad/s). A dashpot with coefficient cj represents the

dependent foundation stiffness and dashpot coefficients, respectively, for mode j; and
effects of damping associated with soil-foundation interaction. An alternative form
 
for Equation 2-12a is:
k j  k j 1  2i  j (2-12b)
where:
c j
j  (defined for kj > 0) (2-13a)
2k j
An advantage of the expression for j in terms of cj in Equation 2-13a is that, at

resonance of the SSI system, j can be interpreted as a fraction of critical damping in
that, as kj approaches zero, j goes to infinity.

the classical sense (Clough and Penzien, 1993). A drawback of Equation 2-13a is
harmonic excitation and response at a given frequency. The phase difference, j,
The imaginary part of the complex impedance represents a phase difference between
between force and (lagged) displacement is (Clough and Penzien 1993; Wolf 1985):

 j  tan 1  2 j  (2-13b)
Angle j is also known as a loss angle. For example, if j is 10%, peak harmonic
displacement will lag peak force by 0.197 radians (11.3 degrees). When j goes to
infinity, j is bounded by /2.
Many impedance function solutions are available for rigid circular or rectangular
foundations located on the surface of, or embedded within, a uniform, elastic, or
visco-elastic half-space. In the case of a rigid rectangular foundation resting on the
surface of a half-space with shear wave velocity Vs, Pais and Kausel (1988), Gazetas
(1991), and Mylonakis et al. (2006) review impedance solutions in the literature and
present equations for computing the stiffness and damping terms in Equation 2-12.
Solutions describe translational stiffness and damping along axes x, y, and z, and
rotational stiffness and damping about those axes (denoted xx, yy, and zz). Stiffness
Poisson’s ratio of the soil, , dynamic stiffness modifiers, j, and embedment
is denoted kj, and is a function of foundation dimensions, soil shear modulus, G,
modifiers, j:
k j  K j   j  j (2-14a)
K j  GB m f  B L ,  ,  j  f  B L , a 0  (2-14b)
 j  f  B L , D B , d w B , Aw B L  (2-14c)
where Kj is the static foundation stiffness at zero frequency for mode j, and m = 1 for
translation, and m = 3 for rotation. Shear modulus, G, should reflect the effects of
modulus reduction with increasing shear strain amplitude. ASCE/SEI 7-10,
Minimum Design Loads for Buildings and Other Structures (ASCE, 2010), and
FEMA P-750, NEHRP Recommended Seismic Provisions for New Buildings and
Other Structures (FEMA, 2009), provide the information presented in Table 2-1 for
adjusting the shear modulus and shear wave velocity for large strain levels.
as G 0  V s2  s , where Vs is based on geophysical measurements in the field, and s is

Maximum (or small strain) shear modulus, G0, can be calculated from Equation 2-9
the soil mass density. An average effective value of Vs is generally computed across
modifiers, j, are related to the dimensionless frequency a0:

an effective profile depth, zp, as described in Section 2.2.2. Dynamic stiffness
B
a0  (2-15)
Vs
sixth of the seismic wavelength for frequency . This frequency parameter is

which has the physical interpretation of being the ratio of B to approximately one-
essentially unique for half-space conditions, but may not be so in presence of a stiff

a single frequency  is usually selected for the purpose of evaluating foundation
stratum at shallow depth (Anoyatis and Mylonakis, 2012). For time domain analysis,
spring and dashpot coefficients that depend on a0. This can be taken as the frequency
corresponding to the period associated with the dominant response of the structure.
In most cases, this will be the first-mode, flexible-base period.
Table 2-1 Values of Shear Wave Velocity and Shear Modulus

Reduction for Various Site Classes and Shaking
Amplitudes (ASCE, 2010; FEMA, 2009)
Reduction Factor (Vs) Reduction Factor (G/G0)
SDS/2.5 (1) SDS/2.5 (1)
Site Class ≤0.1 0.4 ≥0.8 ≤0.1 0.4 ≥0.8
A 1.00 1.00 1.00 1.00 1.00 1.00
B 1.00 0.97 0.95 1.00 0.95 0.90
C 0.97 0.87 0.77 0.95 0.75 0.60
D 0.95 0.71 0.32 0.90 0.50 0.10
E 0.77 0.22 (2) 0.60 0.05 (2)
F (2) (2) (2) (2) (2) (2)
Notes: (1) SDS is the short period spectral response acceleration parameter defined in
ASCE/SEI 7-10; use straight line interpolation for intermediate values of SDS/2.5.
(2) Value should be evaluated from site-specific analysis.
Table 2-2a lists expressions for static foundation stiffness, Kj, for three translational
and three rotational degrees of freedom for a rigid rectangular footing located at the
ground surface. These equations are similar for Pais and Kausel (1988), Gazetas
(1991), and Mylonakis et al. (2006).
stiffness. Factors, j, to increase Kj for the effects of embedment are provided in
Embedment of foundations below the ground surface increases static foundation
Table 2-2b. The Pais and Kausel (1988) equations are most often used in practice.
The equations by Gazetas (1991) and Mylonakis et al. (2006) are more general,
accounting for embedment effects resulting from gapping between the soil and
foundation side walls.
Equations for dynamic stiffness modifiers, j, and radiation damping ratios, j, for
rigid footings located at the ground surface are provided in Table 2-3a. Dynamic
stiffness modifiers and radiation damping ratios for embedded footings are provided
in Table 2-3b.

Table 2-2a Elastic Solutions for Static Stiffness of Rigid Footings at the Ground Surface
Degree of Freedom Pais and Kausel (1988) Gazetas (1991); Mylonakis et al. (2006)
  L  0.75  2GL  B 

K z , sur  3.1   1.6 K z , sur  0.73  1.54   
0.75
1  1  
GB
  B   L 
Translation along z-axis
  L  0.65 L   B 

K y , sur  6.8    0.8    1.6  K y , sur      
0.85
2  2 
GB 2GL
  B  B    L  
Translation along y-axis 2 2.5
  L  0.65   B
K x , sur  6.8    2.4  K x , sur  K y , sur  GL 1  
2  0.75 
GB 0.2
  B    L
Translation along x-axis
 L    B 
K zz , sur  GB 3  4.25    4.06  K zz , sur  GJ t0.75  4  111   
2.45 10
 B    L  
Torsion about z-axis
 I y  3   
 L  0.75   L  
 3.73    0.27  
2.4 0.15
GB 3
1  1 
G
 B    B  
Rocking about y-axis K yy , sur K yy , sur
 L    B 
K xx , sur  3.2  B   0.8   I x     2.4  0.5  L  
0.25
GB 3
1  1 
G L
    B   
0.75
Rocking about x-axis K xx , sur
Notes: Axes should be oriented such that L ≥ B.

area moment of inertia of soil-foundation contact,
i denotes which axis to take the surface around.
polar moment of inertia of soil-foundation
contact surface.
G = shear modulus (reduced for large strain effects, e.g., Table 2-1).
GCR 12-917-21 Inertial Interaction 2-11

Table 2-2b Embedment Correction Factors for Static Stiffness of Rigid Footings
Degree of Freedom Pais and Kausel (1988) Gazetas (1991); Mylonakis et al. (2006)
  0.25  D    D  B    Aw  
 z  1.0   0.25      z  1  1  1.3   1  0.2   
0.8 2/3
  L / B  B    21B  L     4 BL  
  1.34  D    D  z A  
 y  1.0   0.33      y  1  0.15  1  0.52  2  
0.8 0.4
  1  L / B  B       
w w
Translation along y-axis
B BL
Translation along x-axis x  y Same equation as for y, but Aw term changes for B  L
  1.32  D    B  d w 
 zz  1  1.3      zz  1  1.4 1  
L  B 
0.9 0.9
  L / B  B   
 D   D  2  dw   dw   B 
0.6
 yy  1.0       yy  1  0.92      D  L 

B  0.35   L / B  4
0.6 1.9
 B  
1.6
    B       
Rocking about y-axis 1.5
 D   D   dw  dw dw  B
0.2
 xx  1.0       xx  1  1.26 1  
B  D 
2
1.6
 B  0.35  L / B  B   B  L 
Rocking about x-axis
Notes: dw = height of effective side wall contact (may be less

than total foundation height)
zw = depth to centroid of effective sidewall contact
Aw = sidewall-solid contact area, for constant
effective contact height, dw, along perimeter.
For each degree of freedom, calculate Kemb = ηKsur
D
Coupling Terms: K emb, rx    K emb, x
3
D
K emb , ry    K emb, y
3
2-12 Inertial Interaction GCR 12-917-21

Table 2-3a Dynamic Stiffness Modifiers and Radiation Damping Ratios for Rigid Footings (adapted from Pais and Kausel, 1988)
Degree of Freedom Surface Stiffness Modifiers Radiation Damping
 0.2  2 
  0.4  
  L/B
a0
  4  L / B    a 
 z  1.0   z  
  K z , sur / GB    2 z 
  0 
 
 2 
 
  1  3  L / B  1  
10
a 0
 4 L / B   a 0 
  K y , sur / GB    2 y 
Translation along y-axis  y  1.0 y    
 4 L / B   a 
 x  1.0 x  
  K x , sur / GB    2 x 
Translation along x-axis  0 
 
 
   

 0.33  0.03 L / B  1 a 0 
2 
  4 / 3  L / B    L / B  a 02  a 
 zz  1.0    zz    0 
3

  2 zz 
  K zz , sur / GB  
      
 2 
  
  1  0.33  L / B  1 
  0 
 1  3  L / B  1 

0.8
 
1.4
  
a 0
3
a 2
0.7
   
   
    4 / 3 L / B  a   a0 
 yy  1.0     yy    
3 2
0.55a 02
 1.4  2    K yy , sur      2 yy 
2 
0
Rocking about y-axis
  0.6    a0      a 0  
  1  1.75  L / B  1 

 L / B  
1.8
     GB  
3 3
   
 
   
 0.55  0.01 L / B  1 a 0    4 / 3 L / B  a 02  a 
 xx  1.0     xx    0 
2
   2 xx 
  K xx , sur / GB   2.2 
  0.4     0.4 
  2.4    a0    
0 

  L / B 
    L / B    
2

3 2
a

3 3
Notes: Orient axes such that L ≥ B.

Soil hysteretic damping, s, is additive to foundation radiation damping, j.
a 0 B / V s ;   2 1    / (1  2 ) ;   2.5

Table 2-3b Dynamic Stiffness Modifiers and Radiation Damping Ratios for Embedded Footings (adapted from Pais and Kausel, 1988)
Degree of Freedom Radiation Damping
 4   L / B    D / B 1  L / B     a 
 z  
 K z ,emb / GB 
 0 
   2 z 
 4  L / B   D / B 1   L / B     a 
 y   
 K y,emb / GB 
 
   2 y 
0
 
Translation along y-axis
 4  L / B   D / B   L / B     a 
 x  
 K x,emb / GB 
 0 
   2 x 
 
 
  4 / 3 3  L / B  D / B     L / B   D / B   3  L / B   D / B     D / B    L / B   ( L / B)  a 02   a 
 zz     0 
3 2 3
  zz ,emb        2 zz 
  3 
   a 02  
 GB   1  3  L / B  1 
K 1.4

 
0.7
  L  3  D  L  3
  4 / 3               3         a 02  4   L     D  
D  L D  D  L 
3 3 2 3
  B   B   B B  B  B  B   B         a 
 yy      B 
 

3 B
          
 yy 

0
 3    1  1.75  L / B  1   a 0   3 
K 2
 
K 1.8
 
yy , emb
  
  
yy , emb

2
 
GB GB
  D   D  3  L  3
  4 / 3            3         a 02  4   L  1  D  
 L  D   D  L 
3
 B   B   B  B   B  B   B        
 B  B    a0 
 xx      
  2 xx 
3
  K xx , emb      K xx , emb 
      
 GB   1  1.75  L / B  1 
  3   
1.8
  GB 
2
 
a 0 3
Notes: Soil hysteretic damping, s, is additive to foundation radiation damping, j.
; from Table 2-3a
a 0 B / V s ;   2 1    / (1  2 ) ;   2.5

The frequency dependence of quantities provided in Table 2-3a reflects the effect of
condensation of infinite degrees of freedom in soil having mass and associated
for massless soil, as a0 would become zero (because V s  G max    in

dynamic effects (i.e., hidden mass effect). Frequency dependence would disappear
Equation 2-15), causing j = 1, and j = 0. Dynamic stiffness modifiers provided in

Table 2-3b are insensitive to embedment, so values for embedded foundations are the
j, emb = j, sur).

same as values given in Table 2-3a for footings located at the ground surface (i.e.,
Figure 2-4 shows the variation in dynamic stiffness modifiers versus frequency for
rigid footings located at the ground surface. In the case of translational stiffness,
dynamic stiffness modifiers (x, y) are essentially unity, regardless of frequency or
foundation aspect ratio. For rotational stiffness, however, dynamic stiffness
modifiers for rocking (xx, yy) degrade markedly with frequency, but are relatively
insensitive to aspect ratio.
a 0  B / V s a 0  B / V s
Figure 2-4 Plot of dynamic stiffness modifiers and damping ratios versus dimensionless frequency,
hysteretic damping, and =0.33: (a) geometry; (b) x-direction; and (c) y-direction.
for rectangular footings resting on the surface of a homogeneous half-space, with zero
Because soil hysteretic damping, s, is taken as zero, Figure 2-4 also shows the
variation in radiation damping ratios for translation (x, y) and rotation (xx, yy)
versus frequency. Translational radiation damping is only modestly affected by the
direction of shaking or the aspect ratio of the foundation. The modest increase of
translational damping with aspect ratio is a result of the increased foundation size
(i.e., larger wave source).

In contrast, rotational radiation damping is strongly sensitive to the direction of
shaking and the aspect ratio of the foundation. Rotational damping is largely
controlled by vertical cyclic displacements at the edges of the foundation (without
separation between soil and footing). As aspect ratio increases, the ends of the
foundation are located further apart, and energy radiating into the soil from each end
of the foundation experiences less destructive interference, thus increasing damping.
At low frequencies (a0 < 1 to 2), damping from rotation is generally smaller than
damping from translation, although the trend reverses as frequency increases and
foundations become relatively oblong. The practical significance of this effect is that
translational deformation modes in the foundation, while often relatively unimportant
from the perspective of overall structural system flexibility, can be the dominant
j term should be taken as the sum of radiation damping for the appropriate vibration
source of foundation damping. When used to calculate the dashpot coefficient, cj, the
mode (from Equation 2-13a and Table 2-3) and soil hysteretic damping, s, provided
by a geotechnical engineer.
Figure 2-5 shows the variation in dynamic stiffness modifiers and radiation damping
ratios versus frequency for embedded foundations. In equations provided by Pais and
Kausel (1988), dynamic stiffness modifiers are unaffected by embedment, and this
a 0  B / V s
dimensionless frequency, for square footings embedded in a
=0.33: (a) geometry; and (b) x-direction (y-direction similar).

homogeneous half-space, with zero hysteretic damping, and

result is shown in Figure 2-5b. Apsel and Luco (1987), however, found a somewhat
different result, indicating some sensitivity to embedment. Apsel and Luco observed
a more rapid decay of stiffness with frequency for embedded foundations, which is
not reflected in the figure.
The elasto-dynamic analyses upon which Figure 2-5 is based assume perfect contact
between soil and basement walls. Accordingly, the solutions indicate much higher
damping levels than those for shallow foundations (Gazetas, 1991). These damping
levels may not be reliable when gaps form between foundations and the adjacent soil,
which reduces the potential for radiation damping from basement walls. In studies
performed by Stewart et al. (1999b), buildings shaken by earthquakes generally do
not exhibit damping levels consistent with such models. As a result, the impedance
of embedded foundations can be conservatively estimated from the equations for
static stiffness in Table 2-2a and adjusted by dynamic stiffness modifiers for surface
foundations from Table 2-3b.
2.2.2 Effect of Non-Uniform Soil Profiles
In most cases, Vs profiles are evaluated away from foundations (i.e., in the free-field)
and reflect a variation of shear modulus with depth. Variation in soil shear modulus
with depth, and the presence of additional weight from a structure, complicates the
selection of an appropriate shear wave velocity in the calculation of static foundation
stiffnesses. To evaluate a single effective Vs value for use in computations, it is
necessary to: (1) correct Vs values measured in the free-field to account for
overburden pressures associated with the added weight of the structure; and (2)
calculate an average effective Vs value over an appropriate depth range.
 m , as follows:
Soil shear modulus, G0, is known to increase with mean effective confining stress,
 
G0  G a  m 
n
 pa 
(2-16)
where Ga is the shear modulus at atmospheric pressure, m is the effective confining
stress, pa is taken as approximately 100 kPa, and n varies from approximately 0.5 for
granular soils (Hardin and Black, 1968; Marcuson and Wahls, 1972) to 1.0 for
cohesive soils with plasticity index (PI) greater than 6.5 (Yamada et al., 2008).
Recognizing that Vs is proportional to the square root of shear modulus, free-field
measurements of shear wave velocity (at depth) can be corrected to account for
overburden pressures due to the added weight of the structure as follows:
   (z )   v (z ) 
Vs , F (z )  Vs (z )  v 
n/ 2
  v (z ) 
(2-17)

denotes the shear wave velocity measured in the free-field at depth z,  v ( z ) is the
where Vs,F (z) is the overburden-corrected shear wave velocity at depth z, Vs (z)
effective stress from the self-weight of the soil at depth z, and v (z) is the increment
of vertical stress due to the weight of the structure at depth z, which can be computed
using classical Boussinesq stress distribution theory (e.g., Fadum, 1948). The
overburden correction in Equation 2-17 is typically significant only at shallow depths
(i.e., 50% to 100% of the foundation dimension).
Using overburden-corrected values of shear wave velocity, an average effective

profile velocity must be calculated over an appropriate depth interval. The depth
interval necessary for computing an average effective profile velocity was
investigated by Stewart et al. (2003). By matching the static stiffnesses for a uniform
half-space to those for computed for a non-uniform profile using the solutions of
Wong and Luco (1985), this work determined that shear wave velocities should be
averaged over an effective profile depth, zp, below the foundation bearing level. The
profile depth is discretized into layers having thickness zi and velocity Vs,F (z)i.
Designated as Vs, avg, the average effective profile velocity should be calculated as the
ratio of the profile depth, zp, to the summation of shear wave travel time through each
depth interval (Equation 2-18a). The depth interval, zp, necessary for computing an
average effective profile velocity can be taken as the half-dimension of an equivalent
square foundation matching the area of the actual foundation, BeA , or the half-
dimension of an equivalent square foundation matching the moment of inertia of the
actual foundation, BeI , computed as follows:
Vs , avg 
zp
 z 
(2-18a)
  i
Vs ,F (z ) i 
n
i 1

Horizontal  x and y  : z p  BeA , BeA  A 4  BL (2-18b)
Rocking ( xx) : z p  B eI , xx : B eI = 4 0.75 I x  4 B 3 L (2-18c)
Rocking ( yy ) : z p  B eI , yy : B eI = 4 0.75I y  4 BL3 (2-18d)
This approach can, in principle, be applied to effectively rigid foundation systems by

considering B and L as the half-dimensions of the entire foundation plan. This
requires consideration of foundation flexibility and connectivity of the discrete
elements comprising the foundation system. Guidance for non-rigid (i.e., flexible)
foundation elements is provided in Section 2.2.3.
When soil shear modulus increases with depth, some of the seismic energy radiating
from the foundation reflects upward towards the foundation, hence it is not “lost” as
occurs in a uniform half-space. Impedance solutions for this phenomenon are

available for circular foundations (vertical translation: Guzina and Pak, 1998;
horizontal translation and rotation, approximate cone solution: Gazetas, 1991) and
rectangular foundations (vertical translation, rocking: Vrettos, 1999) for different
types of heterogeneity.
Figure 2-6 shows a plot of dynamic stiffness modifiers and radiation damping ratios
comparing results for a uniform half-space and non-uniform profiles in which G
varies with depth, as shown. The effect on radiation damping is more pronounced in
rotational stiffness modifier, yy, for square foundations (Figure 2-6b) is modest.
rotation (Figure 2-6b) than in translation (Figure 2-6a). Also, the effect on the
Hence, the effect of variation in soil shear modulus with depth is most critical for
static stiffness and radiation damping associated with foundation rocking. Because
rocking is often an insignificant contributor to overall foundation damping, the
practical impact of soil non-homogeneity is primarily related to its effect on static
stiffness.
a 0  r / v s 0 a 0  B / v s 0
Figure 2-6 Plot of dynamic stiffness modifiers and radiation damping ratios
versus dimensionless frequency comparing uniform half-space and
non-uniform profiles in which G varies with depth: (a) translation for
circular foundations (from Gazetas, 1991); and (b) rotation for square
foundations (from Vrettos, 1999).
In the extreme case of a rigid material at depth in a soil profile, radiation damping
from body wave propagation disappears at frequencies lower than the fundamental
frequency of the soil column. While no geologic materials are actually rigid, this can

be of concern when the ratio of the shear wave velocity of the firm layer to that of the
soil exceeds approximately two. In such cases, the static stiffness of foundations is
increased (solutions available in Mylonakis et al., 2006).
2.2.3 Effect of Flexible Structural Foundation Elements
Classical impedance function solutions, such as those presented in Table 2-2 and
Table 2-3, strictly apply for rigid foundations. As illustrated in Figure 2-1, soil-
foundation interaction for rigid foundations can be represented by individual springs
for each foundation degree of freedom. Actual foundation slabs and basement walls,
however, are non-rigid structural elements. The few theoretical solutions that exist
apply to circular foundations supporting a rigid core (Iguchi and Luco, 1982),
flexible perimeter walls (Liou and Huang, 1994), or rigid concentric walls (Riggs and
rotational stiffness, krr, and rotational radiation damping ratio, rr, for the cases of a
Waas, 1985). Figure 2-7 shows the effect of flexible foundation elements on
circular foundation supporting a rigid core or flexible perimeter wall.
a 0  r f / V s a0
rotational radiation damping ratio (rr) for circular foundations supporting a

Figure 2-7 Effect of flexible foundation elements on rotational stiffness (krr) and
rigid core (Iguchi and Luco, 1982) and flexible perimeter walls (Liou and
Huang, 1994).
stiffness ratio,  , taken from plate theory as:

The flexibility of the foundation is represented by a relative soil-to-foundation

E 12(1  )  
 
Gr f3
(2-19)
3 2
ft f f
where rf is the foundation radius, tf is the foundation thickness, and Ef and f are the
Young’s modulus and Poisson’s ratio of the foundation concrete. The case of  = 0
corresponds to a rigid foundation slab.
Liou and Huang (1994) showed that foundation flexibility does not significantly
affect translational stiffness and damping terms for the case of flexible perimeter
walls. For rotational stiffness and radiation damping, Figure 2-7 shows that
foundation flexibility effects are relatively modest for the case of flexible perimeter
walls, and most significant for the case of a rigid core.
Typical practice does not adjust the impedance function for non-rigid foundations as
shown in Figure 2-7. Instead, foundations springs are distributed across the extent of
the foundation, as illustrated in Figure 1-2c. Distributed springs allow the foundation
to deform in a natural manner given the loads imposed by the superstructure and the
spring reactions. For vertical springs, this can be accomplished by calculating the
vertical translational impedance, as described above, and normalizing it by the
foundation area to compute stiffness intensity, k zi (also known as coefficient of
subgrade reaction), with dimensions of force per cubic length:
k zi 
kz
(2-20a)
4 BL
A dashpot intensity can be similarly calculated as:
c zi 
cz
(2-20b)
4 BL
As illustrated in Figure 2-8, the stiffness of an individual vertical spring in the

interior portion of the foundation can be taken as the product of k zi and the spring’s
tributary area dA. If this approach were used across the entire length, the vertical
stiffness of the foundation would be reproduced, but the rotational stiffness would
generally be underestimated. This occurs because the vertical soil reaction is not
uniform, and tends to increase near the edges of the foundation. Using a similar
process with czi would overestimate radiation damping from rocking. This occurs
because translational vibration modes (including vertical translation) are much more
effective radiation damping sources than rocking modes.
To correct for underestimation of rotational stiffness, strips along the foundation edge
(of length ReL) are assigned stiffer springs. When combined with springs in the
interior, the total rotational stiffness of the foundation is reproduced. Harden and
Hutchinson (2009) present expressions for end length ratios and spring stiffness
increases as a function of L/B using static stiffnesses from Gazetas (1991).

Figure 2-8 Vertical spring distribution used to reproduce total rotational stiffness
kyy. A comparable geometry can be shown in the y-z plane (using
foundation dimension 2B) to reproduce kxx.
More generally, the increase in spring stiffness, Rk, can be calculated as a function of
foundation end length ratio, Re, as:
 3k yy 
 i 3   1  R e 
 z 
3
1  1  R e 
4 k BL
Rocking ( yy ) : R k , yy 3
(2-21a)
 3k xx 
 i 3   1  R e 
 z
4k B L 
3
1  1  R e 
Rocking ( xx) : R k , xx 3
(2-21b)
Equations 2-21 were derived by matching the moment produced by the springs for a
unit foundation rotation to the rotational stiffness kyy or kxx. In these equations, a
value of Re can be selected (typically in the range of 0.3 to 0.5), which then provides
a unique Rk. This correction for rotational stiffness, however, does not preserve the
original vertical stiffness kz. This is considered an acceptable approximation, in
general, because rocking is the more critical foundation vibration mode in most
structures.
To correct for overestimation of rotational damping, the relative stiffness intensities

and distribution are used (based on the stiffness factor Rk and end length ratio Re), but
dashpot intensities over the full length and width of the foundation are scaled down
by a factor, Rc, computed as:
3c yy
   1  R 
Rocking ( yy) : R c , yy 
R k , yy 1  1  R e 
4c zi BL3
3 3
(2-21c)
e

3c xx
   1  R 
Rocking ( xx) : R c , xx 
R k , xx 1  1  R e 
4c zi B 3 L
3 3
(2-21d)
e
Use of the above procedures for modifying vertical spring impedances will reproduce
the theoretical rotational stiffness and damping through distributed vertical springs
and dashpots. While this allows foundation flexibility to be accounted for, in the
sense that foundation structural elements connected to springs and dashpots are non-
rigid, a question that remains is whether or not the rotational impedance computed
using a rigid foundation impedance function is an appropriate target for calibration.
For the case of a rigid core illustrated in Figure 2-7 it is not, but solutions for more
practical situations are not available.
In the horizontal direction, the use of a vertical distribution of horizontal springs

depends largely on whether the analysis is two-dimensional or three-dimensional, and
whether or not the foundation is embedded. Current recommendations are as
follows:
 For two-dimensional analysis of a foundation on the ground surface, the
horizontal spring from the impedance function is directly applied to the
foundation, as shown in Figure 2-8 (i.e., no distributed springs).
 For two-dimensional analysis of an embedded foundation, the component of
the embedment modifier, kx/x) can be applied to the spring at the base slab
the embedded stiffness attributable to the base slab (i.e., the stiffness without
basement walls with a cumulative stiffness equal to kx (1–1/x).

level. Distributed springs are then positioned along the height of the
 For three-dimensional analysis, springs are distributed in both horizontal

directions uniformly around the perimeter of the foundation. The sum of the
spring stiffnesses in a given direction should match the total stiffness from
the impedance function.
2.2.4 Limiting Spring Forces
Previous sections have discussed linear spring stiffness, but have not addressed
foundation capacity. Pushover procedures for SSI analysis typically utilize elastic-
perfectly-plastic force-deflection relationships for springs. Hence, limiting spring
forces (i.e., capacities) are needed for vertical and horizontal springs.
In the case of vertical springs, the capacity is the unfactored bearing capacity of the
foundation distributed over the tributary area of the spring (dA). Bearing capacity
should be calculated considering the foundation geometry, drained or undrained shear
strength parameters as appropriate, soil unit weight, and simultaneous presence of
both horizontal and vertical loads on the foundations. These concepts are discussed

at length in foundation engineering textbooks (e.g., Coduto, 1994; Salgado, 2008),
and are not reviewed here. Bearing capacity solutions provided by Soubra (1999) are
commonly used in the nuclear industry. Limiting spring forces should not be
calculated using allowable bearing pressures that have been derived based on long-
term foundation settlement limitations, nor should factors of safety be applied.
In the case of horizontal springs, the capacity of springs located at the level of a
footing or mat should reflect the unfactored sliding resistance at the slab-soil
interface. The capacity of springs along basement walls should reflect the unfactored
passive earth pressure.
Shear strength parameters used for computation of bearing capacity, sliding

resistance, and passive earth pressure should be selected with due consideration of
soil type, level of soil saturation, possible cyclic degradation effects, and the rapid
loading rate applied during earthquakes. Selection of appropriate strength parameters
is beyond the scope of this report, but good general guidelines can be found in soil
mechanics textbooks, and in SCEC (2002).
Limiting lateral and vertical capacities of foundations are usually not simultaneously
realizable. This is especially important in the presence of geometric nonlinearities
such as soil-foundation gapping, described further in Section 2.4.
2.3 Impedance of Vertical Pile Foundations
Buildings founded on soft soils may have pile-supported footings or mats, especially
when the foundation is not embedded (i.e., no basement levels). This section
discusses the effective stiffness and damping of pile-supported foundations. Only the
case of vertical piles is considered, as battered piles are seldom used in building
structures.
Impedance of single pile foundations in the translational and vertical modes of

vibration using elasto-dynamic solutions is covered. Rotational (or cross rotational-
translational) impedance of single piles is omitted because rotational stiffness is
typically derived from groups of piles supporting a footing or mat, which is based on
vertical response parameters.
The impedance of pile groups for lateral and rotational vibration modes using elasto-
dynamic solutions is also covered. Pile-to-pile interaction effects are considered and
the manner by which vertical responses of piles are combined to develop rotational
impedance is discussed. Pile stiffnesses from elasto-dynamic solutions are contrasted
with the discrete element modeling typically performed in practice.
An important consideration when piles are combined with shallow spread footings or
a mat foundation is whether or not lateral resistance is provided by the shallow
foundation elements in combination with the piles. Soil might be expected to settle

away from shallow foundation elements in cases involving clayey foundation soils
and end-bearing piles, particularly when there are surface fills at the site. In such
cases, lateral load resistance would derive solely from the piles and basement walls.
On the other hand, when soil settlement is not expected, a hybrid impedance model
can be used in which lateral load resistance is provided by both shallow and deep
foundation elements (e.g., in the case of sandy soils and friction piles).
2.3.1 Impedance of Single Piles
Springs and dashpots effectively replace a single pile in the numerical modeling of a
pile-supported foundation, as schematically illustrated in Figure 2-9. The impedance
of single piles, represented by k jp and c jp , can be described in the notation used for
shallow foundations (Equations 2-12 to 2-14a). The dynamic stiffness for a
particular vibration mode is denoted k jp , and the corresponding dashpot, representing
the effects of damping, is denoted c jp . Subscript j represents the vibration mode,
which is taken as x (horizontal) and z (vertical).
Figure 2-9 Schematic illustration showing replacement of piles with equivalent

springs for translational and vertical (rotational) impedance.
stiffness K jp and a dynamic modifier  jp :

The dynamic stiffness of a single pile can be represented as the product of static
k jp  K jp   p
j (2-22a)
where:
K jp   j E s d
 j   w pj  w sj  w bj  f  E p E s , L p d 
 jp  f  E p E s ,  p  s , w sj , , a 0p 
(2-22b)
In Equation 2-22b, j is a dimensionless constant for vibration mode j; d is pile
s and p are the mass densities for soil and pile materials, respectively;  is the
diameter; Es and Ep are the Young’s moduli for soil and pile materials, respectively;

Poisson’s ratio of the soil; wpj, wsj, and wbj represent weight factors that together sum
to unity for pile, soil, and pile tip stiffness contributions, respectively, for vibration
mode j (details below); and a 0p is a dimensionless frequency for piles (Kaynia and
Kausel, 1982):
d
a 0p  (2-23)
Vs
Because a 0p uses pile diameter in lieu of foundation half-width B, it is typically more
frequency . A fundamental aspect of pile response to lateral head loading is that a

than an order of magnitude smaller than a0 (Equation 2-15) under the same cyclic
long pile does not deflect over its entire length, but only to a certain depth, termed the
active pile length, La (shown in Figure 2-10).
a 0p  d / V s a 0p  d / V s
with =0.4 and p/s =1.3: (a) geometric parameters;

dimensionless frequency, for single piles in a homogeneous half-space,
(b) lateral loading; and (c) vertical loading.
The active pile length is on the order of 10 to 20 pile diameters, depending on pile-
soil stiffness contrast, soil non-homogeneity, and fixity conditions at the pile head
(Randolph, 1981; Gazetas, 1991; Pender 1993; Syngros 2004; Karatzia and
Mylonakis 2012). Piles with lengths Lp > La essentially behave as infinitely long
beams, and the actual length does not affect flexural response. Active lengths tend to
be greater for dynamic loading than for static loading, due to the ability of elastic
waves to travel further down the pile than a static stress field.

Axially loaded piles tend to respond to much greater depths (in excess of 50 pile
diameters), and tip reaction is almost always mobilized. Accordingly, an axially-
loaded pile cannot usually be approximated by an infinite rod.
Expressions for active pile length for lateral deformations can be cast in the form:
 Ep 

La     d
 Es 
(2-24)
where  and μ are dimensionless constants, and all other terms are as previously
(1981) and Fleming et al. (1992) recommend  = 1.8 and μ = 0.25. For dynamic
defined. For fixed-head piles in homogeneous soil under static loading, Randolph
loading, Gazetas (1991) recommends  = 2 and μ = 0.25. Based on a more accurate

set of finite-element analyses, Syngros (2004) recommends  = 2.4 and μ = 0.25. A
discussion about the differences in the various formulas for active length can be
found in Karatzia and Mylonakis (2012). Approximate values of active pile lengths,
La, are 10d to 20d for lateral loading, and the actual pile length, Lp, for axial loading.
Once an appropriate active length, La, is selected, an average effective profile

velocity between the ground surface and depth, La, can be computed. These
guidelines are approximate, and further research is needed to better define effective
G0, is then evaluated from shear wave velocity, Vs, as G0   sVs2 (from Equation
values of Vs for use in the analysis of pile impedance. Small-strain shear modulus,
2-9). Soil shear modulus, G, should be reduced relative to G0 for large strain effects
E s  2 1   G
(e.g., Table 2-1). Young’s modulus for soil can then be related to shear modulus as:
(2-25)
As an alternative to the use of Table 2-1 for evaluating modulus reduction effects,
soil strains adjacent to a horizontally loaded pile can be evaluated as (Kagawa and
1    u  z 
Kraft, 1981):
  z  (2-26)
2.5d
where γ(z) denotes an average soil shear strain at depth z, and u(z) the corresponding
horizontal pile displacement. On the basis of this equation, a strain-compatible soil
shear modulus can be obtained through conventional modulus reduction curves (e.g.,
Darendeli, 2001;Vucetic and Dobry, 1991).
Weight factors in Equation 2-22b (wpj, wsj, and wbj) represent the relative
contributions of the pile structural stiffness, pile-soil interaction through side-load
transfer, and pile-soil interaction through toe resistance for vibration mode j. These
weight factors always sum to unity (i.e., wpj + wsj + wbj = 1.0), and are not required

for analysis of static stiffness. Individual weight factors are used, however, for
dynamic modifiers and damping coefficients, as described below.
The equations are used to determine dimensionless parameters j and weight factors
Equations related to the static stiffness of single piles are provided in Table 2-4a.
(wpj, wsj, and wbj). Note that the equations for j depend, in turn, on a series of
variable is the dimensionless modulus of subgrade reaction, x, which is related to

additional dimensionless parameters. For lateral vibration (j=x), the additional
are  and Lp (both related to Ep/Es), and the corresponding modulus, z. Equations
Ep/Es as indicated in the table. For vertical vibration (j=z), the additional variables
for dynamic stiffness modifiers (  jp terms) and damping ratios (  jp terms) for single
piles are provided in Table 2-4b.
Damping ratios reflect material damping in the pile and soil materials (p and s,
respectively) as well as radiation damping (rj terms). The dynamic stiffness
modifiers and radiation damping ratios for single piles obtained from these
expressions are plotted in Figure 2-10.
2.3.2 Impedance of Grouped Piles
When piles are used as part of a building foundation system, they are usually
configured in groups to support continuous mat foundations or discrete pile caps for
individual load-bearing elements. The impedance of a pile group cannot be
determined by simple addition of individual pile impedances because grouped piles
interact through the soil by “pushing” or “pulling” each other through waves emitted
from their periphery. This is called a group effect, and it can significantly affect the
impedance of a pile group as well as the distribution of head loads among individual
piles in the group. Group effects depend primarily on pile spacing, frequency, and
number of piles. They are more pronounced in the elastic range, and dynamic group
effects decrease in the presence of material nonlinearity.
The ratio of the pile group impedance in any oscillation mode, k Gj , to the sum of the
individual static pile impedances in the same oscillation mode, N piles  k Pj , static , is the
efficiency factor of the pile group (Kaynia and Kausel, 1982). Efficiency factors are
generally less than unity for low frequencies, but can increase significantly at higher
frequencies under low strain conditions. Negative efficiency factors are also
possible, which suggests a phase difference of over 90 degrees between oscillations
of a single pile and oscillations of the pile group at the same frequency. Note that
these factors strictly refer to dynamic compliance of the group and are different from
the familiar efficiency factors for group bearing capacities in foundation engineering.

Table 2-4a Equations for Static Stiffness of Single Piles
Degree of Freedom Surface Stiffness Modifiers Reference
E 
Poulos and Davis (1980)
 x   1/ 4 x 3/ 4  p 
1/ 4
1
 Es 
Translation along x-axis Scott (1981), Mylonakis (1995)
2
E 
3/ 40
 x  2 p 
Dobry et al. (1982)
 Es 
Syngros (2004)
w px  1/ 4 

Dobry et al. (1982)
w sx  3 / 4  Long pile ( L p / d  L active ) Mylonakis (1995)
w bx  0  Mylonakis and Roumbas (2001)
   Ep  Ω  tanh( L p )
Randolph and Wroth (1978)
z  z   
1/ 2 1/ 2
 2   Es  1  Ω tanh( L p )
Translation along z-axis Scott (1981)
 Ep 
1/ 2
  1    E s 
Ω  
2
 z 2
Mylonakis and Gatezas (1998),
Randolph (2003), Salgado (2008)
 4  2  E   Lp 

L p   z   p 
1 1
 
    Es 
2
 d 
 z  0.6 ; L / d  10, E p / E s  100 

Blaney et al. (1975)
p Roesset (1980)
w pz  1   w sz  w bz 
Thomas (1980)
2   L p    2  1     2 cosh  2 L p   1   2  sinh  2 L p 

4 cosh 2   L p     tanh   L p   1   tanh   L p  
w sz 
2
2 cosh   L p  1   2  sinh   L p 
w bz 

Table 2-4b Equations for Dynamic Stiffness Modifiers and Damping Ratios for Single Piles
Degree of Freedom Static Stiffness Modifier Reference
a o 
3   p / s  p 2
 xp  1  
32 x  1  
Translation along x-axis Mylonakis and Roumbas (2001)
a o   a o  
    p /  s  p 2 1 p 1/ 2 
 zp  1  w sz  
 8 x   1   
Translation along z-axis Mylonakis (2011)
2

Degree of Freedom Damping Ratio Reference
 xp   p   s   rx
1 3 3 Dobry et al. (1982)
4 4 4 Mylonakis and Roumbas (2001)
 a o 
  p 3/ 4
 rx  
 2 x 1     x 
3
Gazetas and Dobry (1984a,b)
Translation along z-axis  zp  w pz  p   w sz  w pz   s   rz Mylonakis (2011)
1 
 a op   w bz 0.21a op 
1.2 
 rz  p  sz
 z  4 1    z
3/ 4

w Gazetas and Dobry (1984b)
Results for horizontal and rocking oscillations are provided in Figure 2-11 for pile
groups in square configurations computed using the solution by Mylonakis and
Gazetas (1999). Peaks and valleys observed in the plots are due to destructive and
constructive interference of the waves between piles, which tend to increase and
decrease in dynamic impedance, as first identified by Wolf and Von Arx (1978), and
explained by Kaynia and Kausel (1982), Nogami (1983), and Dobry and Gazetas
(1988).
The above effects tend to decrease with non-homogeneity and nonlinearity in the soil,
as the waves emitted from the periphery of the piles become less coherent (El-Naggar
and Novak, 1994 and 1995; Michaelides et al., 1998). At the low normalized
frequencies of interest in most practical problems ( a0p  0.4) , efficiency factors are
less than one and saturate as the number of piles, Npiles, increases. Hence, for large
piles groups, low-frequency efficiencies would not be significantly lower than those
shown for the 44 pile group in Figure 2-11. Efficiency factors above unity for the
rocking mode are due to the intrinsic out-of-phase movement of the piles located on
opposite sides of the rocking axis.

a 0p  d / V s a 0p  d / V s a 0p  d / V s
Figure 2-11 Plots of pile group efficiency factors and damping ratios versus dimensionless
frequency for square pile groups for: (a) lateral loading at head of pile group
under zero cap rotation; (b) moment at head of pile group, introducing rocking
and rocking results are for Ep/Es = 1000, Lp/d = 20, p/s = 1.3,  =0.4, (pile
under zero cap translation; and (c) vertical loading at head of pile group. Lateral
spacing)/d = 5, p = 0, and s = 0.05. Vertical resutls are for Ep/Es = 100, Lp/d =
15, p/s = 1.4,  = 0.4, (pile spacing)/d = 5, p = 0, and s = 0.05.
In Figure 2-12, the group impedance of a 33 pile group is compared to that of a
footing of equivalent dimensions. Results are presented relative to a common
normalized frequency using the conversion:
a 0  a 0P
B
(2-27)
d
Results show that the effect of the piles is dramatic for rotational stiffness, increasing
kyy by factors up to 50 relative to a shallow footing alone. Translational stiffness is
also increased, but to a lesser degree. Interestingly, damping ratios are on the same
order of magnitude, with somewhat greater variation with frequency caused by pile-
to-pile interaction effects.

(a): Lateral Loading (b): Rocking
a 0  a 0p (B / d ) a 0  a 0p (B / d )
Figure 2-12 Comparison between the impedance (stiffness and damping ratio) of
a 3x3 pile group and the impedance of a footing with equivalent
dimensions for: (a) lateral loading at head of pile group; and (b)
Lp/d = 20, p/s = 1.3,  =0.4, (pile spacing)/d = 5, p = 0, and

moment at head of pile group, introducing rocking. Results are for
s = 0.05.
2.3.3 Discrete Element Methods (p-y and t-z Curves)
In engineering practice, piles are typically modeled using discrete element methods,
as prepared in common commercial computer programs such as APILE (Ensoft, Inc.
program for analysis of axial capacity of piles) and LPILE (Ensoft, Inc. program for
analysis of piles under lateral loads). In such programs, it is straightforward to
extract head load-deflection relationships from the analysis, which are often used in
lieu of elasto-dynamic solutions described above.
Unfortunately, the p-y (nonlinear lateral spring) and t-z (nonlinear vertical spring)
relations used in these programs were developed to represent large-deformation
response in static or cyclic loading, and they do not accurately capture stiffness or
damping. In fact, many spring formulations have an initial stiffness that is infinite
(i.e., the backbone curve is vertical at the origin). When the pile-soil system is not
expected to yield, the elasto-dynamic solutions presented above provide a superior

representation of stiffness and damping. Research is needed to develop next-
generation p-y and t-z curves that capture these stiffness formulations at small
displacements and soil nonlinearity.
2.4 Nonlinear Soil-Structure Interaction Models for Response

History Analysis
Nonlinear soil-structure interaction phenomena can involve geometric and material

nonlinearities in the superstructure, foundation, and soil. Specific sources of
nonlinearity may include: (1) yielding of seismic-force-resisting elements in the
superstructure; (2) yielding of soil, potentially exacerbated by strength loss induced
by pore-pressure (e.g., liquefaction, cyclic softening); (3) gapping between the
foundation and the soil, such as base uplift or separation of foundation sidewalls from
the surrounding material; and (4) yielding of foundation structural elements. These
nonlinear effects must be computed with response history analyses (RHA) performed
in the time domain. Addressing all these issues in RHA is a formidable task, even
with modern computational capabilities. Despite recent progress, knowledge in the
area is incomplete and the subject is evolving.
Most of the research performed on nonlinear SSI has been related to structural
yielding with linear, or equivalent-linear, soil (Case 1, above) or soil
yielding/gapping with a linear structure (Case 2/Case 3, above). A brief overview of
this work is presented in the following sections.
2.4.1 Nonlinear Structure and Equivalent-Linear Soil
Because of difficulties associated with modeling the constitutive behavior of soil in

three dimensions and wave propagation in a finite volume of geologic material under
the structure, without spurious wave reflections at fictitious model boundaries, most
studies focus on nonlinearities in the superstructure. If structural yielding develops at
relatively low intensity input motions, or if the foundation is over-designed,
significant material nonlinearities in the foundation and soil may not occur. This
justifies the use of equivalent-linear representations of subsurface material properties
in such analyses.
In a series of studies involving response of yielding simple oscillators, supported on

foundations resting on linear or equivalent-linear soil, subjected to strong earthquake
excitations, Ciampoli and Pinto (1995), Mylonakis and Gazetas (2000), Perez-Rocha
and Aviles (2003) among others, concluded that nonlinear SSI generally reduces
ductility demand in the superstructure. This effect can be rationalized using
principles for linear SSI presented in Figure 2-3. When the period is lengthened on
the descending branch of the spectrum (i.e., T/Tp > 1), the seismic demand is reduced,
regardless of whether the structure yields or not.

In the case of long-period input motions that potentially place the structure on the
ascending branch of the spectrum (i.e., T/Tp < 1), SSI-induced period lengthening
may lead to an increase in ductility demand. This can be viewed as progressive
resonance, when the effective fundamental period of the yielding structure, T,
approaches the predominant period of the foundation input motion, Tp. Evaluation of
the damping effects is more complex as nonlinearity tends to reduce radiation
damping and increase material damping in both the soil and the structure.
global or system ductility demand, s; or (2) the conventional member ductility
The performance of yielding structures is typically represented in terms of: (1) the
demand,  (Priestley and Park, 1987; Paulay and Priestley, 1992). For the structure
in Figure 2-1, these factors are defined as:

 s  max
y
(2-28)
 max

y
(2-29)
where  max ,  y , max ,  y are the maximum earthquake-induced displacement at the

top of a structure, the corresponding displacement at yield, the maximum earthquake-
induced column displacement relative to a rotated foundation, and the corresponding
column displacement at yield, respectively.
A key difference between these factors is that member ductility demand, , refers
exclusively to structural deformations, whereas global or system ductility demand, s,
encompasses rigid body movements associated with translation and rotation of the
foundation, which do not reflect strains in the superstructure. Equation 2-28 and
Equation 2-29 are geometric relations, and the former always provides smaller
numerical values than the latter for a given set of structural response values
(Ciampoli and Pinto, 1995; Mylonakis and Gazetas, 2000). Because of this, the use
of Equation 2-28 for assessing SSI effects on structural response might be of limited
value.
2.4.2 Nonlinearity in the Foundation and Soil
There is mounting analytical and experimental evidence that material and geometric
nonlinearities in the soil may be beneficial to the seismic response of a structure.
This has led some authors (e.g., Gazetas, 2006; Gajan and Kutter, 2008) to propose
revising the foundation design philosophy by allowing significant yielding in the soil
close to the foundation, or the foundation itself, to dissipate energy and protect the
superstructure. This requires control of settlement and tilting of the structure.
Hence, the analysis and design process considering soil nonlinearity involves
optimization of the trade-offs between the potentially beneficial effects of soil

yielding (especially with regard to energy dissipation) and the detrimental effects of
settlement or residual tilt. In this section, several methods are described by which
calculations of this type can be carried out.
Soil-structure interaction studies with nonlinear soil and foundation behavior can be
classified into three approaches: (1) continuum models, (2) beam-on-nonlinear
winkler foundation (BNWF) models, and (3) plasticity-based macro-models. The
first approach is by far the most computationally demanding, and has been employed
to a limited extent (Borja and Wu, 1994; Jeremic et al., 2009). Available findings
suggest the creation of stress-induced inhomogeneities under the foundation, which
may limit wave radiation away from the structure and cause wave reflections leading
to resonance effects.
The second and third approaches for nonlinear soil modeling are briefly described
below. Although both approaches can consider material nonlinearities, only macro-
element approaches are currently configured for material nonlinearities (gapping).
The basic description of the models, their input parameters, and a brief comparison
with experimental data are provided. The reader is referred to related literature for
additional information. The emphasis in the following discussion is on two specific
models that are implemented in OpenSees, Open System for Earthquake Engineering
Simulation (McKenna, 1997; OpenSees, 2011). Their implementation could be
reasonably extended to other computational platforms in the future. Much of the
content is adapted from Gajan et al. (2010).
Beam-on-Nonlinear Winkler Foundation (BNWF) Models
Starting with the pioneering work of McClelland and Focht (1958), beam-on-
nonlinear winkler foundation (BNWF) models have been used for many years for
analyzing the response of foundations, most notably piles, for static loads (Matlock,
1970; Cox et al., 1974) and dynamic loads (Penzien, 1970; Nogami et al., 1992;
Boulanger et al., 1999). Key advantages of these models over continuum
formulations lies in their ability to describe soil-structure interaction phenomena by
one-dimensional nonlinear springs distributed along the soil-foundation interface. It
is well-known that the modulus of the springs (also known as modulus of sub-grade
reaction) is not uniquely a soil property, but also depends on foundation stiffness,
geometry, frequency, response mode, and level of strain. A limitation of the
approach relates to its one-dimensional nature. A spring responds only to loads
acting parallel to its axis, so loads acting in a perpendicular direction have no effect
on the response of the spring. Accordingly, the concept of plastic potential and flow
rule cannot be explicitly incorporated. Nevertheless, the BNWF approach is popular
because of its simplicity and predictive abilities on a variety of problems.

The impedance models described in Section 2.2 are associated with linear springs
that can be coupled with gapping and damper elements (e.g., implemented by Chopra
and Yim, 1985). Nonlinear springs for shallow foundations have been used in
conjunction with gapping and damper elements by Allotey and Naggar (2003 and
2007) as well as Raychowdhury and Hutchinson (2009).
The BNWF model implemented into OpenSees by Raychowdhury and Hutchinson

(2009) consists of elastic beam-column elements that capture the structural footing
behavior with independent zero-length soil elements that model the soil-footing
behavior. Currently it is developed for two-dimensional analysis only. Therefore,
the one-dimensional elastic beam-column elements used for the footing have three
degrees-of-freedom per node (i.e., horizontal, vertical, and rotation). As illustrated in
Figure 2-13, one-dimensional uniaxial springs are used to simulate the vertical load-
displacement behavior (q-z), horizontal passive load-displacement behavior against
the side of a footing (p-x), and horizontal shear-sliding behavior at the base of a
footing (t-x). Moment-rotation behavior is captured by distributing vertical springs
along the base of the footing.
Figure 2-13 Schematic illustration of a Beam-on-Nonlinear Winkler Foundation (BNWF)

model: (a) hypothesized foundation-superstructure system; (b) idealized model;
and (c) variable vertical stiffness distribution (Raychowdhury and Hutchinson,
2009, with permission).

The mathematical model of the q-z, p-x, and t-x nonlinear springs in Figure 2-13 was
adapted from a model for pile foundations by Boulanger et al. (1999). Within the
OpenSees framework, the materials describing these springs are the QzSimple2,
PxSimple1, and TxSimple1 material models, respectively. The QzSimple2,
PxSimple1and TxSimple1 material models differ from their parent (pile-calibrated)
models (QzSimple1, PySimple1, and TzSimple1) only in the backbone shape
parameters. Each material captures the “far-field” elastic behavior and “near-field”
permanent displacements.
The material models are mechanistic, based on an arrangement of various linear and
nonlinear springs, gap elements, and dashpots. Radiation damping can be accounted
for using a dashpot that is placed in parallel with the far-field elastic component. The
backbone curves are thus characterized by a linear-elastic region, followed by an
increasingly growing nonlinear region. The QzSimple2 material has an asymmetric
hysteretic response, with a backbone curve defined by an ultimate load on the
compression side and a reduced strength in tension to account for the low strength of
soil in tension. The PxSimple1 material is envisioned to capture the passive
resistance, associated stiffness, and potential gapping of embedded shallow footings
subjected to lateral loads. This material model is characterized by a pinched
hysteretic behavior, which can more suitably account for the phenomena of gapping
during unloading on the opposite side of a footing. The TxSimple1 material is
intended to capture the frictional resistance along the base of a shallow foundation.
This material is characterized by a large initial stiffness and a broad hysteresis, as
anticipated for frictional behavior associated with foundation sliding.
The functional forms and parameters describing the p-x, t-x, and q-z springs are
similar, so only the q-z model is described here. The backbone curve has linear and
nonlinear regions. The linear-elastic portion of the backbone curve is described by
the initial stiffness kz:
q  k zs (2-30)
where q represents the spring force, and s represents the spring deflection. The upper
limit of the linear-elastic region, defined as q0, is taken as a fraction of the ultimate
load qult as follows:
q 0  C r q ult (2-31)
where Cr is a parameter specified in OpenSees. The nonlinear (post-yield) portion of

the backbone is described by:
 
q  q ult   q ult  q0   (for |s| > s0)
n
cs 50
 cs 50  s  s 0 
(2-32)

where s50 is the displacement at which 50% of the ultimate load is mobilized, s0 is the
displacement at load q0, and both c and n are constitutive parameters controlling the
shape of the post-yield portion of the backbone curve.
The unload-reload rules that operate with the backbone curve are relatively simple,
generally consisting of the familiar Masing rules (i.e., the shape of the unload and
reload portion of the cyclic loop matches twice the backbone curve). The drag and
gap component is parameterized by a bilinear closure spring in parallel with a
nonlinear drag spring. The cyclic response of each of the material models, when
subjected to a sinusoidal displacement, is demonstrated in Figure 2-14.
Figure 2-14 Cyclic response of OpenSees BNWF springs subjected to a sinusoidal

displacement: (a) q–z spring (Qzsimple2 material model); (b) p–x spring
(Pxsimple1 material model); and (c) t–x spring (Txsimple1 material model)
(Raychowdhury and Hutchinson, 2009, with permission).
User-defined parameters for the q-z element can be synthesized based on two
physical parameters obtained from the results of a typical high-quality geotechnical
site investigation (i.e., bearing capacity, qult, and elastic stiffness, kz) and several
relatively subtle parameters defining the details of the elements described above. The
need for several relatively unfamiliar parameters presents a barrier to implementation
of this type of model in practice. These parameters include:
 Radiation damping (cz). This dashpot coefficient is considered to be a physical
parameter that is well documented in the literature (e.g., Section 2.2). The
parameter is sensitive to soil stiffness, footing shape, aspect ratio and
embedment.
 Tension capacity (TP). The tension capacity parameter, TP, determines the
maximum magnitude of the drag force in Component 1 of the nonlinear springs.
It is the ratio of tension capacity to bearing capacity with typical selected values
of 0 to 0.10 (as suggested in Boulanger et al., 1999), although, more recently
some experts (e.g., Kutter) have recommended using a TP value of zero.
 Distribution and magnitude of vertical stiffness. As illustrated in Figure 2-8
and Figure 2-13, two parameters are necessary to account for the distribution and

magnitude of the vertical stiffness along the length of a footing: (1) the stiffness
intensity ratio, Rk (where, Rk=Kend/Kmid); and (2) the end length ratio, Re (where,
Re=Lend/2L). A variable stiffness distribution along the length is used to force the
distributed BNWF spring model to match the overall rotational stiffness
(Equations 2-21). The end region, Lend, is defined as the length of the edge
region over which the stiffness is increased. ATC-40, Seismic Evaluation and
Retrofit of Concrete Buildings (ATC, 1996) suggests the use of Lend = B/6 from
each end of the footing. This expression of end length ratio is independent of the
footing aspect ratio. Harden and Hutchinson (2009) suggest an expression that is
a function of the footing aspect ratio.
 Spring spacing (S). The spring spacing is input by the user as a fraction of the
footing half-length L (S = le/L), where le is the non-normalized spring spacing. A
maximum element length equal to 8% of the footing half-length (i.e., a minimum
number of 25 springs along the full length of the footing) is recommended to
provide numerical stability and reasonable accuracy.
 Shape parameters (Cr, c, n). These parameters are hard-wired into the
OpenSees implementation of the material models, meaning that they are not
specified by users. The recommended values are soil-type dependent, and were
developed based on comparisons of model predictions to test data as described by
Raychowdhury and Hutchinson (2008).
Figure 2-15 shows a comparison of the BNWF model of Raychowdhury and

Hutchinson (2009) to results from centrifuge test SSG04–06 conducted at the
a slow, cyclic test conducted on a medium aspect ratio ( M / (V  L) = 2.4) shearwall

University of California at Davis (summarized in Thomas et al., 2005). This test was
building, with a vertical factor of safety, FSv = 2.3, resting on dense sand with Dr =
80%. In these tests, M is the moment applied by the horizontal force, V, at the level
of the soil-foundation interface. Figures 2-15a through 2-15d show moment versus
rotation, settlement versus rotation, shear force versus sliding, and settlement versus
sliding histories, as predicted by the model and compared with the experiment. The
model compares reasonably well with the experimental results in terms of capturing
the shapes of the hysteresis loops and rotational and lateral stiffnesses. Additional
comparisons are provided in Raychowdhury (2008).
Plasticity Based Macro-Element (PBM) Models
Plasticity Based Macro-Element (PBM) theory is a recent development (Nova and

Montrasio, 1991) with application to nonlinear response of rigid foundations. This
family of models combines elements from both continuum and BNWF formulations.
Hence, it aims at bridging the gap between the two approaches. The models are
based on the following main assumptions: (1) the foundation is considered
sufficiently rigid, so its response can be described by three or six degrees of freedom

Figure 2-15 Comparison of BNWF model response to centrifuge tests for a medium aspect
ratio building on dense sand, with Dr = 80%, and FSv = 2.3: (a) moment-rotation;
(b) settlement-rotation; (c) shear-sliding; and (d) settlement-sliding
(Raychowdhury and Hutchinson, 2009, with permission).
in two and three dimensions, respectively; (2) an ultimate surface, corresponding to a

general bearing capacity failure in vertical force–shear force–moment (Q-V-M)
space, exists corresponding to the state of stress beyond which plastic flow occurs;
plastic approach or a simple G- and s- correction based on a characteristic shear
(3) a mechanism for describing plastic flow is incorporated in the form of a hypo-
strain; and (4) uplift behavior can be modeled by means of a nonlinear model
allowing for separation between the footing and soil (Pecker and Chatzigogos, 2010).
While PBM models are rational and can capture plastic effects, available
formulations possess a number of drawbacks, notably an inability to incorporate
flexible foundation behavior, the effects of stress-induced inhomogeneity on
radiation damping, and failure modes other than general shear failure. In addition,
there is limited experimental validation of the models. More details are given in
Cremer et al. (2001), Houlsby and Cassidy (2002), Chatzigogos et al. (2009) and
Pecker and Chatzigogos (2010). A variant of these models, focusing on uplift
phenomena in compliant soil, is examined in some detail below.
Macro-element models describe a single element placed at the interface between a

rigid foundation and free-field soil to represent the flexibility and energy dissipation

associated with soil-structure interaction. From a numerical modeling standpoint, the
macro-element is directly located at the footing-soil interface, replacing the rigid
foundation and surrounding soil. When incremental displacements are given to the
macro-element as input, it returns the corresponding incremental loads, and vice
versa. Figure 2-16 shows an example of a contact interface model (CIM) developed
by Gajan and Kutter (2009).
Figure 2-16 Conceptual illustration of a macro-element contact interface model (CIM)

(Gajan and Kutter, 2009, with permission from ASCE).
The model differs from other macro-element models in that its constitutive relations
are obtained by tracking the geometry of gaps and contacts of the soil-foundation
interface. To this end, the CIM provides nonlinear relations between cyclic loads and
displacements of the soil-foundation system during combined cyclic loading (i.e.,
vertical, shear, and moment).
Soil-foundation contact is tracked in the CIM using a parameter called the critical
contact area ratio, A/Ac, where A is the area of the footing, and Ac is the area of the
footing required to have contact with the soil to support the vertical and shear loads.
The ratio A/Ac can be considered to be an alternate definition of the factor of safety
with respect to bearing capacity. For a two-dimensional shear wall structure loaded
in the plane of the wall, the ratio A/Ac equals the footing length ratio 2L/Lc, as shown
in Figure 2-17a. In Figure 2-17b, the foundation position is tracked relative to the
underlying soil surface, which is pushed as far as the “soil_max” surface, but which
rebounds to the “soil_min” surface when unloaded. Zero stress transfer between the
soil and foundation occurs in the gap region.
With seven user-defined input parameters, the CIM is intended to capture the
essential features (load capacities, stiffness degradation, energy dissipation, and
permanent deformations) of the cyclic load-deformation behavior of shallow
foundations. One advantage of the CIM relative to the BNWF model is that the

moment, shear, and vertical load capacities are coupled. Coupling between the
vertical and moment capacities is caused by gap formation (i.e., the moment capacity
typically occurs after a gap has formed, causing the vertical capacity to drop). The
coupling between shear and moment capacity is accounted for using an interaction
diagram as described by Gajan and Kutter (2009).
(a) (b)
Figure 2-17 Contact interface model (CIM): (a) definition of critical contact length; and (b)
tracking of foundation position relative to soil pressures (Gajan and Kutter, 2009,
with permission from ASCE).
Input parameters for the CIM are the ultimate vertical load, QULT, the length of
elastic rotation limit, elastic, the rebound ratio, Rv, and the internal node spacing, DL.
footing, 2L, the initial vertical stiffness, Kv, the initial horizontal stiffness, Kh, the
stiffness and footing geometry. Of those parameters, elastic, Rv, and DL are relatively
The initial rotation stiffness is calculated by the CIM based on the specified vertical
subtle, and are briefly described below:

 Elastic rotation limit (elastic). The elastic rotation limit is defined as the
maximum amplitude of rotation for which no settlement occurs. A value of
0.001 radians is recommended based on comparisons with centrifuge
experiments (Gajan et al., 2008).
 Rebound ratio (Rv). The rebound ratio, Rv, is an empirical factor that accounts
for the elastic rebound and bulging of soil into the gap below an uplifted footing.
It is the ratio of the soil uplift displacement (from rebound) to the total soil
settlement, and significantly affects the length of the transition zone between the
gap and the full foundation pressure zone shown in Figure 2-17b. The default
value is 0.1, which reasonably fits data from available centrifuge model tests.
Increasing Rv reduces cyclic foundation settlement, and improves numerical
convergence in most cases.
 Footing node spacing (DL). The node spacing, DL, specifies the distance
between footing nodes internally created within the model. This user-defined
parameter affects numerical stability and accuracy as well as computation time.
Node spacing should be selected based on the vertical factor of safety, FSv;

foundations with large FSv require small node spacing, DL. The suggested range
for DL varies from 1.0% of the half-length, L, of the footing (for FSv  2) to 0.1%
(for FSv  40).
A number of additional parameters describing the shape of the moment-shear

interaction diagram and the shape of transition curves between elastic and yielding
states are hard-wired into the OpenSees source code for this element, as described in
Gajan and Kutter (2009).
Figure 2-18 shows a comparison of the CIM model to results from centrifuge test
SSG02–05 conducted at the University of California at Davis (summarized in Gajan
( M / (V  L) = 3.44) shearwall building, with a vertical factor of safety, FSv = 2.6,

et al., 2003). This test was a static cyclic test conducted on a medium aspect ratio
resting on dense sand with Dr = 80%. Figures 2-18a through 2-18d show moment
versus rotation, settlement versus rotation, shear force versus sliding, and settlement
versus sliding histories, as predicted by the model and compared with the experiment.
The CIM demonstrates good comparison with the experimental results in terms of
capturing the shapes of the hysteresis loops and rotational and lateral stiffnesses.
Additional comparisons are presented in Gajan (2006) and Gajan et al. (2008).
800 150
Shear force (KN)
Moment (KNm)
100
400
50
0
0
-400
-50
-800 -100
Experiment
40 40
CIM Simulation
Settlement (mm)
Settlement (mm)
0 0
-40 -40
-80 -80
-120 -120
-160 -160
-0.04 -0.02 0 0.02 0.04 -40 -20 0 20 40
Rotation (rad) Sliding (mm)
Figure 2-18 Comparison of CIM simulation to centrifuge tests for a medium aspect ratio
building on dense sand, with Dr = 80%, and FSv = 2.6: (a) moment-rotation;
(b) settlement-rotation; (c) shear-sliding; and (d) settlement-sliding (Gajan
and Kutter, 2009, with permission from ASCE).

2.4.3 Model Comparisons and Recommendations for Response History
Analysis
Gajan et al. (2010) compared the performance of the BNWF and CIM nonlinear SSI
models for hypothetical site conditions and experimental datasets. Based on those
comparisons, recommendations for model selection are as follows:
 If the simulations are to be used for structural design of footing elements, or the
footing flexibility is anticipated to contribute to the foundation response, the
BNWF model should be chosen. This model can be used to more directly
evaluate internal moments and shears used for section design.
 If the normalized moment to shear ratio M / (V  L) is less than approximately
3.0, and sliding is not restrained by slabs and grade beams, then the moment
capacity of the footing will be sensitive to shear load, and vice versa. In this
between the moment, shear, and axial responses. For cases with M / (V  L)
case, the CIM model is preferred because of its ability to account for coupling
greater than 3.0, rocking will tend to dominate, and both models should produce
may also be neglected for very small M / (V  L) ratios where sliding is known
similar results if the parameter selection protocols herein are followed. Coupling
to dominate. Note: M, V, and L are as defined previously.

 In other analytical platforms (other than OpenSees), implementation of the
BNWF model will be more easily accomplished if bilinear spring, gap, and
damping elements are available. Although feasible, implementation of the CIM
in another platform would require implementation of a new element, and access
to the source code for the host platform.

Chapter 3
Kinematic Interaction
Kinematic interaction results from the presence of stiff foundation elements on or in

soil, which causes motions at the foundation to deviate from free-field motions. One
cause of these deviations is base-slab averaging, in which spatially variable ground
motions within the building envelope are averaged within the foundation footprint
due to the stiffness and strength of the foundation system. Another cause of
deviation is embedment effects, in which foundation-level motions are reduced as a
result of ground motion reduction with depth below the free surface. If the
foundation is pile-supported, the piles interact with wave propagation below the base
slab, which can further modify foundation-level motions at the base of a structure.
This chapter describes the phenomena of base-slab averaging, embedment effects,

and kinematic pile response, and presents available models for analysis of these
effects. Models for kinematic interaction effects are expressed as frequency-
dependent ratios of the Fourier amplitudes (i.e., transfer functions) of foundation
input motion (FIM) to free-field motion. The FIM is the theoretical motion of the
base slab if the near-surface foundation elements (i.e., base slabs, basement walls)
and the structure had no mass, and is used for seismic response analysis in the
substructure approach described in Section 1.2.
3.1 Shallow Foundations at the Ground Surface
Base-slab averaging results from adjustment of spatially variable ground motions that
would be present within the envelope of the foundation, which are averaged within
the foundation footprint due to the stiffness and strength of the foundation system.
Base-slab averaging can be understood by recognizing that the motion that would
have occurred in the absence of the structure is spatially variable. Placement of a
foundation slab across these variations produces an averaging effect in which the
foundation motion is less than the localized maxima that would have occurred in the
free-field. Torsional rotations, referred to as the “tau effect” (Newmark, 1969), can
also be introduced.
Motions of surface foundations are modified relative to the free-field when seismic
waves are incoherent. Incoherence of the incident waves at two different points
means that they have variations in their phase angle. Some incoherence is
deterministic (i.e., predictable), because it results from wave passage. For example,
as illustrated in Figure 3-1a, the presence of a non-zero vertical angle causes waves to
arrive at different points along the foundation of a building at different times. This is
GCR 12-917-21 3: Kinematic Interaction 3-1

referred to as the wave passage effect. Investigation of wave passage effects in dense
seismic arrays at soil sites indicates apparent propagation velocities, Vapp, of
approximately 2.0 km/sec to 3.5 km/sec, which appears to be controlled by wave
propagation in crustal rock beneath the soil column (Ancheta et al., 2011).
Incoherence that remains when waves are aligned to have common arrival times is
stochastic, and is quantified by lagged coherency models. Stochastic incoherence
results from source-to-site heterogeneities in the seismic path of travel, which scatters
seismic waves. Lagged coherency is also well-documented in array studies (e.g.,
Abrahamson et al., 1991; Ancheta et al., 2011). As a practical matter, incoherence
from wave passage and lagged coherency is always present in earthquake ground
motions to some degree.
Figure 3-1 Illustration of foundation subjected to inclined shear waves: (a) schematic
passage using a semi-empirical model for incoherent waves (parameter a is

geometry; (b) transfer functions between FIM and free-field motion for wave
defined in Equation 3-5).
In the presence of incoherent wave fields, translational base-slab motions are reduced
relative to the free-field, and rotational motions are introduced. The reduction in
translational motion is generally the more important result. Reductions of base-slab
translation and the introduction of torsion and rotation in the vertical plane are effects
that tend to become more significant with increasing frequency. The frequency-
dependence of these effects is primarily associated with: (1) the increased effective
size of the foundation relative to the seismic wavelengths at high frequencies; and (2)
significant reductions in lagged coherency with increasing frequency (Abrahamson et
al., 1991).
There are numerous theoretical models for predicting the relationship between
foundation input and free field ground motions in the presence of inclined, but
3-2 3: Kinematic Interaction GCR 12-917-21

otherwise coherent, shear waves (i.e., wave passage effects). Mylonakis et al. (2006)
synthesized these models with the following expressions:
u FIM  H u u g (3-1a)
  V 
sin  a 0k  s  
  V app    V app
Hu    
, a 0k 
 V 
(3-1b)
a 0k  s 
V 
2 Vs
 app 
 V app
Hu  , a 0k 

2
(3-1c)
2 Vs
In the above expressions, a 0k is similar to a0 defined in Equation 2-15, except that the
foundation dimension is related to base contact area, as follows:
 B eA
a  k
0 (3-2)
Vs
where BeA is related to foundation area, as defined in Equation 2-18b. At a 0k = 0, the

expression in Equation 3-1b becomes indeterminate, and Hu should be taken as unity.
If, as indicated in array studies, Vapp ranges from approximately 2.0 km/s to 3.5 km/s,
then for a typical soil site, a reasonable estimate of the velocity ratio, Vapp/Vs, is
approximately 10. In Figure 3-1b, the result labeled “wave passage only” shows the
transfer function between uFIM and ug based on Equations 3-1. Using this model,
wave passage alone causes relatively modest base-slab reductions in ground motion
across the frequency range of engineering interest.
Transfer functions of recorded foundation input and free-field motions are generally
significantly lower at high frequencies than predicted by wave passage models. This
occurs because wave passage is a relatively modest contributor to the spatial
variation in ground motion that drives base-slab averaging. Additional sources of
variability include stochastic phase variability (quantified by lagged coherency) and
stochastic variations in Fourier amplitudes. Two approaches for capturing these
effects in the analysis of transfer functions are: (1) continuum modeling of the soil
and foundation system subject to input motions with a defined coherency function
(Computer Program SASSI2000, Lysmer et al., 1999; Ostadan et al., 2005); and (2)
application of a semi-empirical simplified model (Veletsos et al., 1997; Kim and
Stewart, 2003).
In SASSI, a site-specific and foundation-specific model is generated in three

dimensions. The foundation and soil material properties are equivalent-linear
without iteration on strain-dependent properties. Empirical coherency models can be

used that include wave passage and lagged coherency (e.g., Ancheta et al., 2011, for
soil sites; EPRI, 2007, for hard rock sites).
The semi-empirical model is based on a theoretical formulation of the kinematic

interaction problem by Veletsos and Prasad (1989) and Veletsos et al. (1997), who
apply spatially variable ground motions to a rigid foundation perfectly bonded to the
soil. Models evaluate the response of rigid, massless, circular and rectangular
either vertically or at an angle v to the vertical, as shown in Figure 3-1a. The results
foundations on the surface of an elastic half-space to incoherent S-waves propagating
are a relationship between transfer function amplitude and a 0k . This relationship is
parameter, a, related to lagged coherency and wave inclination that scales the
essentially independent of foundation shape, but is strongly dependent upon a
frequency axis of the theoretical transfer function. For vertically propagating waves
this transfer function (adapted from Veletsos and Prasad, 1989) can be written as:
H u   2 1  exp  2b0 2  I 0 (2b0 2 )  I1 (2b0 2 )   

 1 
12
 b0 
 
(3-3)
where b0  4   a a0k and I0 and I1 are modified Bessel functions, zero and first
order, respectively. Equation 3-3 was developed for circular foundations. The
4  term adapts a 0k (for rectangles) to a0 defined for an equivalent radius that
preserves foundation area. For small and large values of the argument (2b02), the
Bessel function summation in Equation 3-3 can be written in terms of power series
and exponential functions, respectively (Watson, 1995); for routine application, these
approximations can be expressed as:

      for b0  1
 I 0 (2b02 )  I1 (2b02 )   
b0 6 b08 b010

2 4
1 b b
 exp  2b 2   1 1  1   for b  1

0 0
2 4 12
 
(3-4)
   b0  16b0  

0 2 0
Note that the exponential terms in Equation 3-3 and Equation 3-4 cancel for b0 > 1.
The two functions in Equation 3-4 have a misfit of 0.0073 at b0 = 1, which is accurate
enough for practical purposes.
By matching model predictions to observed variations between foundation input and
developed a semi-empirical model for a that can be written as:

free-field ground motions from instrumented buildings, Kim and Stewart (2003)
 a  0.00065  V s , 200  V s  500m / s (3-5)
where Vs is a representative small-strain shear wave velocity for the soil beneath the
foundation, which can be calculated as described in Section 2.2.2.

Values of a identified through calibration reflect the combined effects of
incoherence from wave passage and stochastic processes, as well as Fourier
semi-empirical approach near the upper and lower limits of a. Two key
amplitude variability. Figure 3-1b shows transfer function Hu calculated using the
observations from this figure are: (1) as a increases (indicating increasingly spatially
variable motions), Hu decreases significantly; and (2) for the range of a supported by
case history data, Hu from the semi-empirical procedure is much lower than from
wave passage models. This can be attributed to significant contributions of stochastic
phase and amplitude variability to base-slab averaging.
The data set considered by Kim and Stewart (2003) consists of buildings with mat
foundations, footing and grade beam foundations, and grade beam and friction pile
foundations, generally with base dimensions, BeA , in the range of 15 m to 40 m.
Although the Veletsos models strictly apply to rigid foundations, the semi-empirical
model applies to the more realistic foundation conditions present in the calibration
data set.
Errors could occur, however, when the model is applied to conditions beyond the
calibration data set. In particular, the effects of incoherence in the Veletsos models is
taken as proportional to wavelength, thus implying strong scaling with frequency and
distance. Array data indicate that distance scaling is much weaker than the frequency
scaling (Abrahamson et al., 1991; Ancheta et al., 2011), so the model would be
expected to over-predict the effects of incoherence (under-predict Hu) for very large
foundations. The opposite would be true for small foundations. Even within the
parameter range of the calibration data set, it should be recognized that the empirical
model fits the data in an average sense, and should not be expected to match any
particular observation.
3.2 Embedded Shallow Foundations
If the base slab of a building is embedded below the ground surface (i.e., the structure
has a basement), foundation-level motions are further reduced as a result of ground
motion reduction with depth below the free surface. The available solutions apply to
rigid cylinders embedded in a uniform soil of finite or infinite thickness (half-space).
Analytical solutions by Kausel et al. (1978) and Day (1978) describe foundation
input motions at the base of embedded cylinders as a function of free-field surface
ground motion ug. When subjected to vertically propagating coherent shear waves,
embedded cylinders experience a reduction in base translational motion, relative to
ug, due to ground motion reductions with depth and wave scattering effects.
Rotations in the vertical plane are also introduced as a result of differential
displacements imposed upon the cylinders over their embedded depth. These transfer
functions can be adapted to rectangular foundation shapes as:

 D   D  D
Hu   cos  a 0k   cos  ,  1.1
u FIM
 Be   Vs 
(3-6a)
ug Vs
D
H u  0.45,  1.1 (3-6b)
Vs
L   D   D 
H yy     0.26 1  cos   , 
  Vs  
(3-6c)
ug Vs 2
D 
H yy    0.26,  (3-6d)
Vs 2
where D is the embedment depth, as shown in Figure 3-2a. Velocity, Vs, in this case
should be interpreted as the average effective profile velocity, Vs, avg, defined in
Section 2.2.2. The transfer functions for rotation in the xx direction have the same
B/ug.
form as Equation 3-6(c) and Equation 3-6(d), except that it expresses the rotation
Figure 3-2 Illustration of foundation subjected to vertically incident shear waves:

(a) schematic geometry; and (b) transfer functions for horizontal
foundation translation and rocking.
Figure 3-2b plots the transfer functions for translation and rotation. The reduction of
translation is substantial at high frequencies, saturating at about 70% of fE, which is
the fundamental frequency of the soil column between the surface and depth, D.
With a high-frequency de-amplification level of 0.45, the embedment effect is often
more important than the base-slab averaging effect for building structures.

Conversely, foundation rotation increases with frequency and saturates at fE. Stewart
and Tileylioglu (2007) checked the predictions of rigid cylinder models against
records from nuclear reactor structures and embedded buildings. Away from
frequencies strongly influenced by inertial interaction, the comparison was favorable,
indicating that these functions can be applied to realistic conditions encountered in
engineering practice.
To illustrate the significance of kinematic interaction from base-slab averaging and

embedment effects, transfer functions were computed using motions recorded from
the 1987 Whittier earthquake at the Ranch Cucamonga Law and Justice Center
building (California Strong Motion Instrumentation Program, CSMIP, Station
Number 23497). Results are shown in Figure 3-3.
Figure 3-3 Illustration of base-slab averaging and embedment effects at CSMIP

Station 23497 during the 1987 Whittier Earthquake: (a) foundation
and sensor location plan; (b) acceleration time histories; and (c)
observed and model-based transfer functions.
Reference recording locations in the east-west (E-W) direction, at an embedment

depth of 5.9 m below the foundation, and in the free-field, are shown in Figure 3-3a.
Acceleration histories shown in Figure 3-3b illustrate the general reduction of ground
motions at the foundation, ufndn, relative to the free field, ug. Figure 3-3c compares
the transfer function of recordings, ufndn/ug, computed using procedures given in
Mikami et al. (2008) to model predictions for base-slab averaging alone and base-
slab averaging combined with embedment. Model predictions are based on
Equations 3-3 to 3-6, using Vs = 390 m/sec, B = 16.8 m, L = 63.1 m, and D = 5.9 m.
The data have significant scatter, but trend from unity at zero frequency to about 0.2

to 0.4 at high frequencies. The models capture these general trends, but are clearly
not a perfect fit to the data.
3.3 Pile Foundations
When building foundations are pile-supported, the kinematic interaction problem is

complicated by the influence of the piles on wave propagation below the foundation,
and also by the potential for the soil to settle away from the pile-supported base of
the structure, forming a gap. This is a complex kinematic soil-structure interaction
problem for which there are no well-calibrated engineering models.
Flores-Berones and Whitman (1982), Fan et al. (1991), and Nikolaou et al. (2001)
describe the kinematic response of vertical piles and pile groups in elastic soil
subjected to vertically propagating coherent shear waves. Similar solutions for
inclined (coherent) waves have been presented by Barghouthi (1984), Mamoon and
Banerjee (1990), and Kaynia and Novak (1992). Because the incident motions
assumed in the development of these models were coherent, these models do not
adequately incorporate base-slab averaging effects. Consequently, model predictions
do not compare favorably to data (Kim and Stewart, 2003). In particular, transfer
function ordinates, Hu, are significantly over-predicted by the Fan et al. (1991)
models at high frequencies.
Kim and Stewart (2003) found that observed variations between foundation input and
free-field ground motions at building sites in California could be adequately
represented with the models for shallow foundations in Section 3.1 and Section 3.2.
The pile-supported buildings considered were generally founded on alluvial soils, and
the piles were not end-bearing. At building sites in Japan, however, Mikami et al.
(2006) found that model predictions for base-slab averaging (SASSI, and the semi-
empirical method in Section 3.1) overestimated Hu relative to data. The pile
foundations in the Japanese data set have relatively high flexural stiffness, and were
more nearly end-bearing, in comparison to the California data set. Development of
analytical solutions for the kinematic interaction problem for pile-supported
foundations of varying flexural rigidity subjected to realistic (incoherent) input
motion remains an important research need.
3.4 Application of Transfer Functions
Design-basis response spectra generally apply to free-field ground motion conditions.

Suites of acceleration time histories are also developed when response history
analyses are to be performed. This section discusses how the transfer function from a
kinematic interaction analysis should be used to modify free-field ground motions
represented by a response spectrum, or ground motion suite, to account for the
interaction effects described above.

The response spectrum for uFIM (Sa-FIM) differs from the spectrum for ug (Sa) because
of a reduction in high frequency ground motion components due to kinematic
interaction. For moderate to low frequencies, the ratio of spectral ordinates can be
estimated as:
 H u ( f ), f  fL
S a  FIM ( f )
(3-7a)
Sa( f )
At higher frequencies, the approximation from Equation 3-7 does not hold because
high frequency spectral ordinates and PGA are controlled by lower-frequency
components of the ground motion. At high frequencies, a conservative
approximation of spectral ordinates can generally be obtained from:
 H u ( f L ), f  fL
S a  FIM ( f )
(3-7b)
Sa( f )
The limiting frequency, fL, depends on the frequency content of ug. For typical stiff
soil or rock ground motions having mean periods in the range of 0.2 sec to 0.5 sec, fL
has been found to be approximately 5 Hz, as documented in Appendix E of FEMA
440, Improvement of Nonlinear Static Seismic Analysis Procedures (FEMA, 2005).
Long-period ground motions resulting from near-fault directivity pulses or soft soil
effects, however, can have much lower limiting frequencies, fL, such that no
significant spectral ordinate reductions from kinematic interaction are realizable.
Further research is needed to identify factors controlling fL, and to develop more
reliable recommendations for engineering application.
Differences between transfer functions and ratios of response spectra are illustrated in
Figure 3-4 using the Rancho Cucamonga data presented in Figure 3-3. East-west
(E-W) response spectra are shown in Figure 3-4a. Reductions in foundation motions
relative to free-field motions are apparent for periods less than approximately 0.7 sec
(of frequencies greater than 1.4 Hz). Figure 3-4b shows the ratio of response spectra
(RRS) for the E-W direction along with: (1) the transfer function model, Hu, for both
base-slab averaging and embedment effects; and (2) the RRS model derived from Hu
using Equation 3-7. The RRS model captures the general trends of the data, although
there are significant period-to-period variations in the data (and even some RRS
ordinates greater than unity). The RRS data does not show the strong decrease in
spectral ratios with decreasing period that is evident in the Hu function, which
illustrates the saturation effect described above.
Acceleration histories representing the FIM can be modified from free-field motions
using the following procedure:
1. Calculate the Fourier transforms of ug.

2. Multiply the amplitude of ug by Hu. As an optional step, the phase of uFIM could
be randomized from that of ug using the procedures in Chapter 3 of Ancheta
(2010). If the phase is not randomized, then ug and uFIM will be coherent.
3. Perform a reverse Fourier transform to estimate uFIM(t).
For most practical situations, this procedure could be avoided by merely selecting
and modifying ground motions for compatibility with Sa-FIM in lieu of Sa, in which
case no further modification is needed.
(a)
(b)
Figure 3-4 Illustration of differences between transfer functions and ratios of

response spectra using data from CSMIP Station 23497 during the
1987 Whittier Earthquake: (a) E-W response spectra recorded at the
site; and (b) ratio of response spectra from recordings compared to
model prediction for RRS and transfer function, Hu.

Chapter 4
Implementation in Standards and
Design Guidelines
Soil-structure interaction provisions exist in the following engineering standards and

design guidelines:
 ATC-40, Seismic Evaluation and Retrofit of Concrete Buildings (ATC, 1996).
 ASCE 4-98, Seismic Analysis of Safety-Related Nuclear Structures and
Commentary (ASCE, 1998).
 FEMA 440, Improvement of Nonlinear Static Seismic Analysis Procedures
(FEMA, 2005).
 ASCE/SEI 41-06, Seismic Rehabilitation of Existing Buildings (ASCE, 2007)
 FEMA P-750, NEHRP Recommended Seismic Provisions for New Buildings and
Other Structures (FEMA, 2009).
 ASCE/SEI 7-10, Minimum Design Loads for Buildings and Other Structures
(ASCE, 2010).
 PEER Report No. 2010/05, Guidelines for Performance-Based Seismic Design of
Tall Buildings (PEER, 2010).
Despite the availability of these resources, soil-structure interaction is seldom

considered in typical U.S. design practice. This is driven in part by challenges in
understanding, learning, and implementing fundamental SSI principles, but also in
the way that SSI provisions are characterized in seismic design provisions. In
general, accounting for SSI is handled through optional procedures, some of which
can only reduce base shear demands. Under such conditions, ignoring SSI is not only
easier, it is conservative.
However, trends in practice are tending toward increased use of SSI. This has been
driven principally by seismic retrofit projects in which SSI analysis is used to gain
better insight into structural performance and to improve accuracy in the analytical
simulation of important structural response quantities. This chapter describes the
implementation of SSI procedures in currently available engineering standards and
guidelines. Limitations of those procedures are also briefly discussed.
GCR 12-917-21 4: Implementation in Standards and Design Guidelines 4-1

4.1 Force-Based Procedures
Implementation of soil-structure interaction into the equivalent lateral force

procedure for seismic design is specified in Chapter 19 of ASCE/SEI 7-10, which
was developed based on information contained in the NEHRP Recommended
Provisions (FEMA, 2009). The seismic base shear considering SSI effects is
defined as:
V  C sW (4-1)
where Cs is a seismic coefficient, taken as the design response spectral ordinate, at

building period T, normalized by the acceleration of gravity, g, and W is the effective
seismic weight of the structure (taken as 70% of the total weight). The shape of the
design response spectrum is schematically illustrated in Figure 4-1.
Figure 4-1 Schematic illustration of the shape of the design response spectrum in the
NEHRP Recommended Provisions.
ASCE/SEI 7-10 (and the NEHRP Recommended Provisions) neglect kinematic

interaction effects but account for inertial interaction effects related to period
lengthening and damping ratio. The change in base shear is calculated as:
  0.05  
 V  C s  C s   W
0.4
  0  


(4-2)
The change in base shear is related to the change in seismic coefficient (or spectral
obtained from the design spectrum at an elongated period, T . The term  0.05  0 
acceleration). The C s term in Equation 4-2 represents the seismic coefficient
0.4
from the fixed-base structural damping value of i = 0.05, to the flexible-base value
represents the reduction in spectral ordinate associated with a change in damping
of 0.
slope with respect to period. Coupled with the requirement that 0 must exceed i
It is important to note that the shape of the design spectrum is flat, or has a negative
4-2 4: Implementation in Standards and Design Guidelines GCR 12-917-21

(discussed further below), this ensures that utilizing SSI provisions will always
reduce base shear.
Modification of design base shear for SSI effects in equivalent lateral force
procedures has a potentially significant shortcoming. There is no link between base
shear reduction factors intended to represent structural ductility (i.e., R factors) and
soil-structure interaction. Crouse (2001) noted that existing R factors may already
reflect the beneficial effects of soil-structure interaction, and modifying the base
shear to account for both SSI and ductility may be unconservative in some cases.
Accordingly, there is a need to revisit the definition of R factors with respect to SSI
effects, and define values that represent structural ductility effects alone.
Period lengthening is calculated using an equation similar to Equation 4-3 (originally

presented as Equation 2-7 in Chapter 2 of this report):
 1 
T k kh 2
(4-3)
T k x k yy
however, ASCE/SEI 7-10 does not specify how lateral stiffness, kx, or rotational
stiffness, kyy, are to be evaluated. The Commentary to the NEHRP Recommended
Provisions provides some guidance related to circular foundations, but the
rectangular foundation models contained in Chapter 2, Section 2.2 are more useful
for practical engineering applications.
The shear modulus, G, used in conjunction with equations for static foundation
stiffness, must be reduced from the shear modulus at small strain levels, G0, to
account for large strain effects. ASCE/SEI 7-10 (and the NEHRP Recommended
Provisions) provides values for adjusting the shear modulus and shear wave velocity
for large strain levels based on site class and spectral response acceleration levels.
These values are provided in Table 2-1.
Foundation damping is calculated using an equation similar to Equation 4-4

(originally presented as Equation 2-10 in Chapter 2 of this report):
0   f  i
T T 
1
n
(4-4)
with fixed-base structural damping, i = 0.05, and exponent, n = 3 (for ideally viscous
the foundation damping factor, f, is not evaluated directly from controlling variables
material damping). In ASCE/SEI 7-10 (and the NEHRP Recommended Provisions),
shown in Figure 4-2, in which the period lengthening ratio is related to f as a

as in Equation 2-11a (Chapter 2 of this report), but is taken from a plot like the one
function of structure aspect ratio, h/r. Note that r is an equivalent foundation radius,
which is calculated to match the foundation area for squat structures and the

foundation moment of inertia for slender structures. The relationship in Figure 4-2 is
calculated from Veletsos and Nair (1975), using the absolute value of the complex-
valued damping relationship, and including both radiation damping and hysteretic
soil damping.
Figure 4-2 Plot of relationship between period lengthening and foundation

damping (FEMA, 2009; ASCE, 2010).
Soil-structure interaction provisions for the equivalent lateral force procedure in

ASCE/SEI 7-10 (and the NEHRP Recommended Provisions) are written such that
base shear demand is only reduced through consideration of SSI. Therefore, ignoring
SSI effects in the design process is analytically conservative. In practice, the
beneficial effects of period lengthening and foundation damping are negligible for
tall, flexible structures. Use of SSI procedures yields the most benefit for short-
period, stiff structures with stiff, interconnected foundation systems (i.e., mats or
interconnected footings) founded on soil.
Consideration of soil-structure interaction is also permitted when modal response

spectrum analysis is used for seismic design. Implementation of SSI in modal
response spectrum analysis is similar to the implementation for equivalent lateral
force analysis. Period lengthening and modification of damping, however, are only
applied in the fundamental lateral mode of response. Higher mode vibration periods
and damping ratios are not modified for the effects of SSI.
The principal limitations of force-based procedures in ASCE/SEI 7-10 are: (1) use of
simplified spectra that can only result in a decrease in base shear as period lengthens;

(2) use of relatively simplified models, applicable to circular foundation geometries,
for soil-foundation springs and foundation damping; and (3) lack of consideration of
kinematic interaction effects on foundation-level ground motions. These limitations
were considered necessary to make the procedures sufficiently simple for broad use
in practice.
A potentially important consideration associated with the use of the SSI procedures
in Chapter 19 of ASCE/SEI 7-10 is the value of the fixed-base fundamental period, T.
Chapter 12 of ASCE/SEI 7-10 contains approximate methods for evaluation of T, and
limiting values (i.e., CuTa), which bias the estimate of T to intentionally produce
conservative values of design base shear. In SSI procedures, T should be taken as the
best estimate value of period, without deliberate bias.
Chapter 12 of ASCE/SEI 7-10 (Section 12.13) also contains procedures for

incorporation of foundation flexibility (i.e., soil springs) into structural models for
linear analysis. The use of an elongated period from Chapter 19 with a structural
model containing foundation springs (per Section 12.13) would overestimate the
effects of foundation flexibility, so the simultaneous use of both sets of procedures is
not permitted in ASCE/SEI 7-10.
4.2 Displacement-Based Procedures
In displacement-based procedures, system behavior is represented by a force versus

displacement relationship that is calculated through nonlinear static (i.e., pushover)
analyses. A pushover analysis involves the application of static lateral loads
distributed over the height of the structure, and calculation of the resulting
displacements in a model of the SSI system. A pushover analysis of a structure with
a flexible base is schematically illustrated in Figure 4-3.
Figure 4-3 Schematic illustration of a pushover analysis and development of a pushover curve
for a structure with a flexible base.

The cumulative lateral load resultant, H, is related to a reference displacement, ,
forming the nonlinear pushover curve. In some applications, the pushover curve is
modified to an acceleration-displacement response spectrum (ADRS) by converting
H to an equivalent spectral acceleration, and converting  to an equivalent spectral
displacement (e.g, Chopra and Goel, 1999; Powell, 2006). At each point on the
pushover curve, the deformations of all components in the structural system are
related to the reference displacement.
Powell (2006) describes common ways by which the pushover curve is combined
with a design response spectrum to estimate the seismic displacement in a structure.
Three such methods are known as the Capacity Spectrum Method (ATC, 1996), the
Coefficient Method (FEMA, 1997; FEMA, 2000; and ASCE, 2007), and Equivalent
Linearization (FEMA, 2005). These methods are illustrated in Figure 4-4.
Soil-structure interaction is considered in displacement-based analysis procedures

through: (1) foundation springs used in the pushover model; (2) reduction of the free-
response spectrum for flexible-base damping ratios, 0, that are greater than the
field response spectrum for kinematic interaction effects; and (3) reduction of the
fixed-base structural damping ratio, i. The manner by which these components are
evaluated in displacement-based analysis procedures is described below. Slight
modifications to the notation contained in reference engineering standards and
guidelines have been made for consistency with the notation adopted in this report.
In general, soil-foundation springs used in pushover analyses are similar to those

described in Chapter 2, Section 2.2, except that dynamic stiffness modifiers are
neglected. The Pais and Kausel (1988) static stiffness equations listed in Table 2-2
are used. Distributed vertical springs are evaluated in a manner similar to that
described in Chapter 2, Section 2.2.3. Horizontal springs are not distributed but are
concentrated at the end of the foundation as shown in Figure 2-8 (in Chapter 2).
Kinematic interaction effects are represented in terms of ratios of response spectra

(RRS) between the foundation and free-field motions. Equations for RRS as a
function of period are given for the effects of base slab averaging and embedment are
1  2  B e  0.3048  
adapted from FEMA 440 as follows:
 
1.2
1 T
A
14100  
1
 
RRS bsa (4-5a)
T fL
RRS bsa  1 
1

2  B eA  0.3048 f L  T
1
1.2
(4-5b)
14100 fL

Figure 4-4 Schematic illustration of procedures used to combine pushover
curves with design response spectra to estimate seismic
displacements in a structure (Powell, 2006, with permission).

 2 D 
RRS emb  cos   T
1
 TV sr 
(4-6a)
fL
 2 Df L 
RRS emb  cos   T
1
 V sr 
(4-6b)
fL
where Vsr is the strain-reduced shear wave velocity evaluated using the reduction
factors in Table 2-1. In Equations 4-5, the equivalent foundation dimension BeA is
expressed in units of meters. These equations are a curve-fit of the semi-empirical
base-slab averaging transfer function described in Equation 3-3 (in Chapter 3). A
shear wave velocity term does not appear in Equations 4-5 because the Vs terms
cancel in the expression for b0 in Equation 3-3. The resulting RRS curves for base-
slab averaging are shown in Figure 4-5.
Figure 4-5 Ratios of response spectra (uFIM/ug) for base slab averaging using
the semi-empirical formulation adopted in FEMA 440.
re-writing in terms of period, T, instead of angular frequency, . In FEMA 440, the

For embedment, the RRS in Equations 4-6 match Equation 3-4 (in Chapter 3) after
limiting period, fL, is taken as 5 Hz (0.2 sec). As of this writing, these equations are
in the process of being revised in the next version of ASCE/SEI 41 (ASCE, 2013) to
reflect the recommendations contained in Chapter 3 of this report.
damping ratio, f, which is then combined with the fixed-base structural damping
In FEMA 440, the objective of the damping analysis is to estimate the foundation
ratio, i, to estimate 0 using Equation 4-4 (with n = 3). The principal challenge is to
extract f from the results of the pushover analysis of the structure in both its fixed-
foundation flexibility can significantly reduce radiation damping (yy) from rotational
base and flexible-base condition. As described in Chapter 2, Section 2.2.3,
vibration modes, which is considered in the FEMA 440 procedures.

First, the period lengthening ratio at small displacements is estimated using the initial
stiffness of capacity diagrams for the fixed-base and flexible-base structures.
Table 2-2 (the dynamic stiffness modifier, x, is assumed as unity). The effective
Assuming shaking in the x-direction, stiffness Kx is then evaluated using equations in
rotational stiffness of the foundation system is then evaluated from a manipulation of

Equation 2-7 as follows:
K yy 
K *fixed h 2
 T 
(4-7)
   
2
K *fixed
T 
1
Kx
*
where K fixed is the equivalent fixed-base stiffness of the structure evaluated from:
 2 
M 
2
 T 
*
K fixed (4-8)
Note that dynamic stiffness modifier, yy, is also taken as unity. The value of Kyy
estimated from Equation 4-7 reflects the stiffness of the foundation structural
elements as implemented in the pushover analysis, so no assumptions of foundation
rigidity are required.
The next step is to reduce the period lengthening ratio from the small-displacement
periods). Taking  as the expected ductility demand for the system (including
condition to the large-displacement (i.e., post-yield) condition (with elongated
structure and soil effects), the effective period lengthening in the post-yield state is
computed as:
 T   1  T  2  
0.5
   1     1 
 T  eff    T   
(4-9)
the foundation damping ratio, f.

This effective period lengthening ratio can then be used with Figure 4-2 to estimate
4.3 Response History Procedures
Most of the resources listed at the beginning of this chapter (e.g., ATC-40, FEMA
440, FEMA P-750, ASCE/SEI 41-06, and ASCE/SEI 7-10) are silent on the
implementation of SSI effects in response history analyses. Similar to ASCE/SEI
7-10 (Section 12.13), they permit the use of soil springs in principal, but offer no
specific guidance on how the springs should be selected or utilized in a response
history analysis.
The Pacific Earthquake Engineering Research Center (PEER) Guidelines for

Performance-Based Seismic Design of Tall Buildings (PEER, 2010) recommends a

response history substructure analysis procedure similar to the approach described in
this report. However, the specification of input motions and the distribution of
springs and dashpots, as depicted in Figure 1-2 (in Chapter 1), are simplified to
streamline response history analysis, as shown in Figure 4-6. Two idealizations of
the SSI system are recommended in the PEER Guidelines, depending on the level of
earthquake shaking intensity, identified as the: (1) service level earthquake (SLE);
and (2) maximum considered earthquake (MCE).
Figure 4-6 Schematic illustration of a tall building with subterranean levels:

(a) complete system; (b) simplified model for service-level earthquake intensity;
and (c) simplified foundation model for maximum considered earthquake
intensity.
Subterranean levels are modeled in both the SLE and MCE analyses, including the
mass, stiffness, and structural capacities of structural elements such as walls,
columns, and slabs. Response history analysis for the SLE (Figure 4-6b), is
performed with a relatively simple model that omits the surrounding soil and does not
include soil springs. Response history analysis for the MCE (Figure 4-6c), is
performed with springs and dashpots representing soil-foundation interaction along
basement walls and below the base slab. In this case, ground motions are applied to a
rigid “bathtub” surrounding the subterranean portions of the structure. In both the
SLE and MCE analyses, the motion applied at the base of the model can be either the
free-field motion (ug) or the foundation input motion (uFIM). These recommendations
are derived largely from the recommendations of Naeim et al. (2008).
Procedures for calculating spring stiffnesses and capacities are not specified in the
PEER Guidelines, but can be taken as those from Chapter 2 of this report. Similarly,
foundation input motions can be modified from free-field motions using the
procedures in Chapter 3 of this report.
4.4 Nuclear Structures
Guidance for the seismic analysis of nuclear safety-related facilities in the United
States is provided in ASCE 4-98, Seismic Analysis of Safety-Related Nuclear

Structures and Commentary (ASCE, 2000). This standard was developed mainly for
non-reactor nuclear structures, and at the time of this writing, is currently under
revision. ASCE 4-98 is a companion to ASCE/SEI 43-05, Seismic Design Criteria
for Structures, Systems, and Components in Nuclear Facilities (ASCE, 2005), which
is used as a reference standard for submittal of combined operating license
applications (COLAs) for new reactors.
Both the direct analysis and substructure approaches for SSI described in Chapter 1
are permitted under ASCE 4-98. Provisions related to response history analyses as
well as equivalent lateral force-based analyses are included. Referring to Figure 1-2
(in Chapter 1), consideration of kinematic interaction effects, foundation flexibility,
and damping are included. ASCE 4-98 does not consider base slab averaging effects,
but it does consider embedment effects. The foundation input motion adjusted for
embedment effects is referred to as the Foundation Input Response Spectra (FIRS),
which is computed for the free-field conditions at the foundation level using wave
propagation analysis (de-convolution). The procedure in the upcoming version of
ASCE 4 (ASCE, in preparation) will also allow for base-slab averaging through
response history analysis, in which the input motion is specified with a defined
coherency function, as described in Section 3.1, using software such as SASSI
(Ostadan et al., 2005).
Soil-foundation flexibility is represented using spring-based solutions (CLASSI,

Wong, 1979) or equivalent-linear elastic solid finite elements (SASSI, Lysmer et al.,
1999). As such, the procedures for nuclear structures have generally been based on
elasto-dynamic solutions of the soil-structure interaction problem.

Chapter 5
Experiments and Case Studies
Testing has played a relatively minor role in the research and evolution of soil-
structure interaction. For the most part, research has been dominated by numerical
analysis exercises. The models developed from numerical and theoretical studies,
however, apply for idealized conditions, so testing and seismic monitoring play a
vital role in guiding the implementation of idealized models in practice.
This chapter identifies and reviews experiments and case studies available in the
literature. Information is separated into field-scale and laboratory-scale tests, which
typically have different objectives and different applications. Tests involving
shallow foundations and dynamic loading are emphasized. Results from tests
identified in this chapter, where applicable, have been interpreted for use in SSI
modeling in other chapters of this report.
5.1 Seismic Monitoring and Field Tests
Seismic monitoring and field testing of structures is generally performed with one of
two objectives: (1) evaluation of system properties such as the fundamental mode
period and damping of an SSI system; or (2) estimation of foundation impedance
ordinates representing the stiffness and damping characteristics of soil-foundation
interaction.
5.1.1 System Studies
In system studies, seismic data from well-instrumented buildings are used in system
identification analyses to estimate modal vibration parameters. To evaluate SSI
effects, input-output pairs must be selected that isolate the system behavior
associated with the structure alone (fixed-base properties) and the full system
(flexible-base properties), as described in Stewart and Fenves (1998). Results can
then be compared to period lengthening and foundation damping models of the type
described in Section 2.1. Stewart et al. (1999a; 1999b), for example, describe the
results of such analyses and lessons learned regarding the practical application of
impedance functions for rigid circular foundations.
System studies can also be undertaken using forced-vibration tests (e.g., Yu et al.,
2005). The input-output pairs that should be used for evaluation of fixed- and
flexible-base properties are given by Tileylioglu (2008). No major study applying
these techniques to evaluate period lengthening and foundation damping has been
completed to date.
GCR 12-917-21 5: Experiments and Case Studies 5-1

5.1.2 Field Experiments for Foundation Impedance
Experimental investigations of impedance functions typically seek to evaluate

stiffness and damping terms for horizontal translation (j = x or y) and rotation within
the vertical plane (j = xx or yy). Cyclic excitation is generally provided by a shaker
installed on the roof or foundation of a structure. Table 5-1 summarizes field forced-
vibration experiments of this type. Analysis procedures for evaluation of foundation
impedance from recordings of structural vibrations are provided in Tileylioglu et al.
(2011).
Table 5-1 Summary of Field Forced-Vibration Tests Used to Evaluate Foundation Impedance
Functions
Excitation Results
Foundation Embed. Vs Freq. Range Impedance Freq. Range

Dimensions (m) (1) f1 (Hz) (2) (m/s)(3) Source (Hz) (4) T / T (5)
Obtained (Hz) (4) Reference
Lin and
shaker on 1.3 (D=1.5); modal freq.
33m 0 to 1.5 17.5 305 7 to 70 kx, cx, kyy, cyy Jennings
ground 1.5 (D = 0) only
(1984)
Luco et al.
NS: 2.16; shaker on NS: 0.8 to 2.5; NS: 1.06; NS: 0.8-2.5;
2525m 4 to 5.5 300 kx, cx, kyy, cyy (1988); Wong
EW: 1.26 roof EW: 0.8 to 1.75 EW: 1.1 EW: 0.8-1.75
et al. (1988)
1.33m; shaker on Crouse et al.
0 n/a 120; 75 10 to 60 n/a kx, cx, kyy, cyy 0 to 60
1.21.1m fndn. (1990)
cyykx-yy
shaker on kx, cx, kyy,
de Barros and
diam. = 10.8m 5.2 9.37 300 roof and 2 to 20 2 5 to 14
Luco (1995)
fndn. cx‐yy, kz, cz
shaker on Tileylioglu et
4.14.1m 0 6.0, 13 198 5 to 15 1.15, 1.29 kx, cx, kyy, cyy 5 to 15
roof al. (2011)
Notes: (1) Foundation embedment depth; (2) Fundamental mode, fixed-base frequency; (3) Vs=Shear wave velocity of soil;
(4) Frequency range; and (5) Period lengthening.
Symbols: D=embedment depth; n/a=not available; NS=North-South building axis; EW=East-West building axis;
diam=diameter; and fndn=foundation.
The first field investigations of foundation impedance provided results over a limited
range of frequencies (Lin and Jennings, 1984; Luco et al., 1988; Wong et al., 1988)
or for very small structures representative of strong motion instrument huts (Crouse
et al., 1990). More recently, de Barros and Luco (1995) tested a relatively large
model structure (of a nuclear reactor) and provided impedance ordinates over a
relatively wide frequency range (approximately 4 Hz to 20 Hz). Figure 5-1 shows
impedance ordinates evaluated by de Barros and Luco. Results are shown in non-
normalized form due to uncertainty in the shear modulus of the foundation soils, and
illustrate the noisy character of the data, especially at frequencies under 4 Hz or
greater than 14 Hz. Also shown in Figure 5-1 are three model predictions for
5-2 5: Experiments and Case Studies GCR 12-917-21

stiffness and damping, caused by uncertainty in the appropriate value for shear wave
velocity, Vs.
Figure 5-1 Non-normalized impedance values from experimental data compared with theoretical
predictions for a nuclear containment structure at Hualien, Taiwan for translational (top)
and rotational (bottom) modes (adapted from de Barros and Luco, 1995).
Two practical difficulties associated with field testing for impedance ordinates and
comparison to model predictions have been encountered in the previous work. First,
limited resolution of the data acquisition system with respect to analogue-to-digital
signal conversion and time-stamping contribute significantly to noise in the results.
Most previous studies have not formally evaluated noise effects, which can lead to
spurious results (e.g., impedance ordinates in Figure 5-1 for frequencies outside the
4 Hz to 14 Hz range). Second, shear wave velocity profiles have generally been
established using downhole or suspension logging methods in the free-field.
Use of downhole or suspension logging methods can have limited resolution very
near the ground surface (e.g., Andrus et al., 2004). Because the soil materials
immediately below the foundation exert the greatest influence on foundation
stiffness, this introduces uncertainty in the selection of an appropriate value of Vs for

use with numerical solutions. Also, seismic velocities measured in the free field
neglect the effect of confinement provided by the weight of the structure. Although
this can be corrected using the procedures given in Section 2.2.2, this has seldom
been done in prior field experiments.
The final study listed in Table 5-1 overcame many of the practical difficulties
encountered in previous work through the use of high-fidelity, modern sensors and
data acquisition equipment available through the NEES@UCSB equipment site
(http://nees.ucsb.edu/). Seismic velocities were measured immediately adjacent to
the foundation so that they incorporated overburden effects. Sample results are
shown in Figure 5-2, and the data are observed to be numerically stable across the
range of tested frequencies. Stronger damping was evident in the translational versus
the rotational vibration modes, and radiation damping was a significant contributor in
each case.
(a) (b)
Figure 5-2 Normalized impedance values from experimental data compared with
theoretical predictions for the Garner Valley site for: (a) translational; and
(b) rotational modes (adapted from Tileylioglu et al., 2011). Velocity, Vsm,
and modulus, Gm, denote median values from test data.

5.1.3 Field Experiments for Nuclear Applications
Field tests have been supported by nuclear regulatory agencies in the United States
and abroad to provide data that can be used to validate analysis procedures, such as
CLASSI (Wong, 1979) and SASSI (Lysmer et al, 1999), commonly used for nuclear
reactor structures. In one case (Hualien site in Taiwan), these experiments can be
used to infer impedance functions. Such cases were presented above and are not
repeated here (de Barros and Luco, 1995). Following is a discussion of three similar
experiments:
 Lotung, Taiwan. This experiment utilized a ¼-scale containment model
constructed in Lotung, Taiwan. The objective of the study was to validate SSI
analysis methodologies. The Lotung model was a cylindrical structure with a
total height of 15 m and embedment depth of 4.5 m. The free-field, structure,
and internal components were instrumented to record motions and pressures at
the soil-foundation interface. Earthquakes recorded in 1986 were used for
subsequent analyses. The model reactor was installed in a soft soil site with
measured shear wave velocities, shear moduli, and damping curves with shear
strain. The data were used in a blind prediction exercise (Ostadan et al., 1991).
Because the data consisted of earthquake recordings and not controlled
experiments, it was not practical to directly identify impedance functions across a
significant frequency range, as described in Section 5.1.2.
Findings from the Lotung experiments included information on the relative

accuracy of different software packages (SASSI provided favorable results), the
effects of soil nonlinearity as manifested in the fundamental frequency of the
structure (which is dominated by rocking), and the application of one-
dimensional wave propagation analyses to evaluate the free-field motions at the
foundation depth through de-convolution.
 Aomori, Japan. In 1994, the Nuclear Power Engineering Corporation (NUPEC)
initiated a multi-year test program in Aomori Prefecture in northern Japan. This
experiment investigated the effects of foundation-soil-foundation interaction of
proximate structures. The test consisted of three building conditions: (1) a single
reactor building as a reference for comparison; (2) closely spaced twin reactor
buildings; and (3) a reactor and turbine building in close proximity. With due
consideration of soil property uncertainty, SASSI was able to replicate the
principal test results (Xu et al., 2003).
 Blast Testing at Gillette, Wyoming. In 2004 and 2005, the Japan Nuclear
Safety Organization (JNES) performed a series of blast tests in the Black
Thunder mines in Gillette, Wyoming (Nie et al., 2008). The experiments
included a model of a 3-story concrete structure in close proximity to the mine.
Recorded motions at the site ranged from peak ground acceleration (PGA) of less

than 0.1g to more than 1.0g. Analyses were conducted by JNES using the soil
spring approach, SASSI, and LS-DYNA (Livermore Software Technology
Corporation). The objective of the study was to create uplift and study the effect
of uplift on SSI results. The main findings were as follows:
a. Uplift was initiated with a horizontal PGA of about 0.15g and vertical PGA
of about 0.12g.
b. The contact ratio (the foundation area in contact with soil over the total
foundation area) dropped to 65% at a horizontal PGA of 0.45g and vertical
PGA of 0.41g.
c. Linear analysis over-predicted the extent of the uplift.
d. Analyses that included soil nonlinearity resulted in lesser uplift than linear
analyses.
A number of additional tests have been performed with information relevant to

seismic earth pressures on retaining walls. Further reports on this topic are presented
in Ostadan and White (1998) and Ostadan (2005).
5.2 Laboratory-Scale Tests
Laboratory-scale investigations of soil-foundation interaction have been performed

on various configurations using shake-table, free vibration, and centrifuge excitation.
Several of these tests are summarized in Table 5-2. Laboratory-scale tests are useful
for examining frequency-dependent foundation stiffness terms and nonlinear soil-
foundation interaction. In recent years centrifuge tests have been used to guide the
development of macro-element models for nonlinear soil-foundation interaction of
the type described in Section 2.4.
However, laboratory tests are limited in their ability to reproduce certain field
conditions (e.g., Novak, 1987). For example, the finite size of a laboratory test
container precludes radiation damping of waves with quarter-wavelengths on the
order of the container dimension. This is illustrated in Figure 5-3, which shows
theoretical model-based damping ratios increasing with foundation size, whereas the
experimental data show essentially constant damping ratio at the hysteretic (material)
damping level. The difference is due to radiation damping that is present in the
theoretical model, but not in the experiments, due to the small size of the laboratory
container.

Table 5-2 Summary of Laboratory-Scale Tests of Dynamic Response of Footings
Parameters and Details Excitation Testing Information
Foundation Soil Depth and

Dimensions Emb. Container
(model Structure Depth g- Soil Dimensions Motion Data
scale) Description (cm) Level Parameters (cm) Source Analyzed Availability Reference
h=45.7cm
square/rec.
applied static vertical, Dobry et al.
=0.33;
B=5.6 to moist sand; level ground
force at top of torsion, (1986)
=121 pcf;
13.3 cm concrete D=152.4
0 1 block and coupled Report Erden (1974)
s=0.025
L=B-6B block 2B=304.8
suddenly swaying- Stokoe and
circle 2L=304.8
released rocking Erden (1985)
radius=14.9cm
h=45.9 to
46.7cm applied static Gazetas and
=0.33;
moist sand; level ground vertical,
square/rec. force at top of Stokoe (1991)
=121 pcf;
concrete ~7.5 to D=152.4 coupled
B=5.6-13.3 cm 1 block and Report Erden (1974)
s=0.025
block 18.8 2B=304.8 swaying-
L=B-6B suddenly Stokoe and
2L=304.8 rocking
circle released Erden (1985)
radius=14.9cm
vertical,
aluminum level ground base/sidewall
h=2.8cm Dry, No. 120 torsion,
block D=35.6 shear and Downloadable Gadre and
B=1.9cm 3.8 30 Nevada sand coupled
1cm below 2B=61.0 passive/active data; report Dobry (1998)
L=1.9cm Dr=75% swaying-
surface 2L=91.4 force
rocking
Nevada sand applied slow
level ground
h=50.8cm double Dr=60 to 80% sinusoidal Rosebrook and
D=53.0 Downloadable
B=1.73cm aluminum 0 to 1.5 20 and cyclic force rocking Kutter (2001a,
2B=90.0 data; report
L=6.68cm shear walls bay mud and dynamic 2001b, 2001c)
2L=175.0
Cu=100 kPa base shaking
applied slow
h=48.6cm level ground
sinusoidal
B=2.5 to single wall Nevada sand D=53.0 Downloadable Chang et al.
2.5 20 cyclic force rocking
1.73cm with frame Dr=80% 2B=90.0 data; report (2007)
and dynamic
L=2.5 to 7.1cm 2L=175.0
base shaking
applied slow
sinusoidal
B=13.5 to Nevada sand D=53.0 Downloadable
bridge pier 8.6 20 cyclic force rocking Ugalde (2007)
17.75cm Dr=80% 2B=90.0 data; report
and dynamic
L=13.5 to 2L=175.0
base shaking
17.75cm
saturated; applied vertical,
level ground
steel base dense sand sinusoidal torsion, Ghosh and
D=22.0
radius=1.5cm plate with 0 50 (Dr=85%) over cyclic force coupled Report Madabhushi
2B=23.5
dual dome loose sand (freq=50Hz swaying- (2007)
2L=56.0
(Dr=45%) for 500ms) rocking
single and Nevada sand applied slow
double rigid Dr=60 to 80% sinusoidal
B=1.63cm D=53.0 Downloadable Gajan and
steel or 0 to 7 20 and cyclic force rocking
L=6.75 to 2B=90.0 data; report Kutter (2008)
aluminum bay mud and dynamic
7.0cm 2L=175.0
shear wall Cu=100 kPa base shaking
2L=total length; 2B=total width; h=height; g-level=gravity load multiplier applied while testing; D=embedment depth; =Poisson's ratio;
=unit weight of soil; s=soil damping ratio; Dr = relative density of sand; and Cu = undrained shear strength of clay.
Symbols:

Figure 5-3 Comparison of damping ratio of vibrating blocks of various sizes
predicted using impedance models for a half-space and measured
from free-vibration tests on laboratory-scale models (Dobry et
al.,1986, with permission from ASCE).

Chapter 6
State of Practice
Substantial research has been available for many years, yet there is relatively limited
implementation of soil-structure interaction in engineering practice. There is often a
lag between the state of knowledge and the state of practice, and this has proven to be
particularly true in the case of soil-structure interaction. Even among relatively
experienced practitioners, it is clear that there is room for improvement in the
technical approaches used in SSI modeling, as well as the manner in which structural
engineers, geotechnical engineers, engineering seismologists, and other design and
construction professionals interact with each other on projects involving soil-
structure interaction.
This chapter summarizes general observations on the state of the practice with regard
to modeling soil-structure interaction effects on building structures, and makes
specific recommendations for the process of communication and collaboration
between design professionals. As a result, the intended audience for this chapter is
wider than it is for other chapters in this report, and includes owners, project
managers, architects, construction managers, and others who are involved in
managing, designing, and constructing the built environment.
6.1 Overview
The state of the practice with regard to SSI was discussed informally and anecdotally
with selected structural engineers, geotechnical engineers, and members of the
project team. Design professionals engaged in these discussions represented small
and large firms with different backgrounds, levels of experience, and geographic
locations. Collectively their experience covered a wide range of building sizes,
occupancy types, and structural materials. Comments attributed to geotechnical
engineers may also be related to engineering seismologists that are providing
recommendations as part of the geotechnical engineering scope of services, either as
a member of the geotechnical engineering firm, as a subconsultant, or as an
independent consultant on projects.
6.2 Observations
This section synthesizes observations of the state of practice related to general issues,
collaboration between design professionals, information needed and shared,
understanding of SSI principles, and implementation in terms of SSI analyses, soil
springs, and common approaches to design problems.
GCR 12-917-21 6: State of Practice 6-1

6.2.1 General Observations
Key general observations include:

 The dialogue between structural and geotechnical engineers varies widely, both
in extent and sophistication. It appears that an increase in the amount of
collaboration would be beneficial, as would better understanding of what each
discipline does, and needs, and why. Many geotechnical engineers, for example,
are not sure how their recommendations are ultimately being used, and often do
not know whether or not their recommendations are being properly implemented.
 In many cases, static and dynamic springs for modeling soil properties are not
being consistently or properly developed by geotechnical engineers, nor are they
being consistently or properly implemented by structural engineers.
 Understanding of SSI principles is fairly limited among structural engineers, and
is usually limited to application of vertical foundation springs. A broader
implementation of SSI techniques is rare, and there is virtually no use of
foundation damping in any explicit way.
 For typical foundation situations, there is no consensus among structural
engineers on the best modeling approaches to use.
 SSI has been used in design of new buildings and in seismic retrofit of existing
buildings, although there appears to be greater implementation of SSI in projects
involving existing buildings, often as part of performance-based designs or
assessments.
6.2.2 Collaboration Between Design Professionals
There is significant variation in the amount of communication and collaboration

between structural engineers and geotechnical engineers on typical projects. On
some projects, there are few, if any, meetings or direct conversations. On other
projects, there is substantial discussion, with several meetings to explore issues and
possible approaches and solutions. Some geotechnical engineers express frustration
that they provide information and recommendations, but they do not have any role in
confirming how their work is used, or whether the final approaches implemented by
the design team are appropriate.
On most projects, structural engineers are engaged by the architect, but geotechnical
engineers are engaged by the building owner. This arrangement is primarily the
result of a perceived increase in liability for geotechnical engineering, and the
reluctance of architects, and their professional liability insurers, to engage
geotechnical engineers as subconsultants. As a result, geotechnical engineers are not
directly managed by the architect as lead design professional, and geotechnical
6-2 6: State of Practice GCR 12-917-21

engineers are typically not part of formal design team meetings arranged by the
architect.
There is also an issue of timing. In some cases, geotechnical engineering work is

substantially complete by the time design work begins in earnest. Most importantly,
however, is an issue related to cost. The fee structure on a typical design project does
not allow for multiple design iterations or the use of unfamiliar analytical techniques.
This discourages the use of SSI because of a potential increases in design time, even
when it could result in significant savings in construction costs.
A need for increased communication and collaboration between engineering

disciplines seems particularly important in light of limited experience in SSI
implementation, and lack of clear standards of practice.
6.2.3 Information Needed by the Structural Engineer
Structural engineers routinely seek a common set of information from geotechnical

engineers on most projects. This includes a description of the soil and rock
characteristics at the site, geotechnical hazards that need to be mitigated, and
recommendations on appropriate foundation systems. Specific information requested
often includes: design bearing pressures under footings; estimates of predicted
settlements; at rest, active and passive lateral pressures; coefficients for sliding;
vertical and lateral capacities for deep foundations; expected site seismicity; and soil
profile type. Depending on the size, nature and sophistication of the project,
additional information is often needed, including shoring and underpinning
recommendations, force-displacement relationships or springs to represent vertical
and horizontal soil properties, site specific spectra, and response histories.
Traditionally, base-slab averaging and embedment effects have not been considered,
and this information is not regularly requested. With the publication of FEMA 440,
Improvement of Nonlinear Static Seismic Analysis Procedures (FEMA, 2005) and
other subsequent engineering resources, structural engineers are beginning to take
advantage of these effects on new projects, and are beginning to request related
design information.
Although there are consistent needs on many projects, few structural engineers report
the availability of checklists, or other standardized lists, to assist in organizing and
requesting the geotechnical information necessary for foundation design or SSI
modeling. A sample checklist of geotechnical information needed by the structural
engineer would be considered a useful aid.
6.2.4 Information Needed by the Geotechnical Engineer
Geotechnical engineers can provide better recommendations when they have more
detailed information on which to base them. In addition to information from soil

borings and other critical soil data, necessary basic information includes structural
design information such as the location of the building on the site, the governing
building code, the anticipated height and footprint, the presence of a basement, and
the potential structural and foundation systems being considered. A plan of column
locations along with anticipated column loading, if available, is useful. Tolerances
for differential settlement in the superstructure are valuable, and it is important to
know if uplift on foundations is anticipated. The need for site-specific spectra versus
response histories must be communicated so the geotechnical scope of work can
include their preparation.
If consideration of SSI effects is anticipated, necessary additional information

includes details on what type of springs will be used and the approximate building
fundamental period. To assist in preparing an appropriate scope of work and in
improving dialogue between disciplines, a sample checklist of structural design
information needed by the geotechnical engineer is considered a valuable aid.
6.2.5 Understanding of Soil-Structure Interaction Principles
Understanding of SSI principles varies widely across both the structural and
geotechnical engineering disciplines. Most structural engineers can appreciate that
SSI effects are more pronounced in soft soils, and many are aware that foundation
input motions can differ from free-field ground motions. Although SSI effects are
known to be significant on stiff, squat, shorter-period buildings, many practitioners
mistakenly believe that SSI would have a larger effect on taller, flexible, longer-
period buildings.
Structural engineers associate several benefits with modeling of SSI effects. By

including the soil in the modeling process, engineers can gain a better understanding
of the distribution of forces and displacements in the structure, and additional insight
into the foundation design. Some structural engineers are taking advantage of the
potential for lowering structural design forces through period lengthening, base-slab
averaging, and embedment effects, however, use of foundation damping, and the
corresponding reduction in design forces, is typically not being considered.
6.2.6 Analysis Procedures
Analysis procedures used to address seismic loading include equivalent lateral force
procedures, modal response spectrum analysis procedures, nonlinear static
(pushover) procedures, and nonlinear response history procedures. Although there is
increasing use of nonlinear procedures in design, response history analyses are
performed by a minority of structural engineers, and among that minority, response
history analyses are performed on a relatively small percentage of projects. Currently
available codes, standards, and guidelines provide SSI provisions for force-based
procedures (e.g., equivalent lateral force and response spectrum analysis) and

displacement-based procedures (e.g., static pushover-type analyses), however, little
or no guidance is provided for implementing SSI in response history analysis
procedures (see Chapter 4).
Typically, SSI is limited to implementing springs as boundary conditions

representing flexibility of the soil-foundation interface in the model. Finite element
modeling of the soil continuum is almost never performed in routine building design
practice.
Most software packages commonly used in building design have the ability to
include uniaxial springs representing support flexibility. However, some are not able
to model compression-only properties that might be appropriate for non-embedded
(i.e., surface) foundations. More sophisticated programs can vary the compression
and tension properties of the springs, and permit bilinear or tri-linear force-
displacement relationships. Multi-support excitation is difficult to implement and is
not widely available in commonly used structural engineering software, though it is
available in programs such as OpenSees (McKenna, 1997).
6.2.7 Implementation of Foundation Springs
Vertical springs are almost always used in the design of mat foundations in order to
properly understand the distribution of design forces in the mat. This is typically
done with a single modulus of subgrade reaction representing long-term settlement
properties of the supporting soil. Use of springs under gravity load-carrying grade
beams is also relatively frequent. Springs are occasionally placed beneath spread
footings, strip footings, grade beams, piles, and piers. To a lesser extent, they are
used to model the combination of soil and foundation vertical flexibility (e.g.,
individual piles or pile groups).
In comparison, horizontal springs are much less frequently used. In rare cases, they
are used on piers and piles to understand the distribution of lateral forces to the
foundation elements, particularly if the foundation consists of different systems with
different stiffnesses. Horizontal springs are infrequently used to represent the passive
pressure developed on the sides of pile caps, grade beams, or retaining walls.
Most structural engineers rely on geotechnical engineers to provide soil spring

properties. When values are not provided, or when structural engineers to confirm
values that are provided, they refer to engineering resources such as FEMA 356,
Prestandard and Commentary for the Seismic Rehabilitation of Buildings (FEMA,
2000) or its successor document, ASCE/SEI 41-06, Seismic Rehabilitation of
Existing Buildings (ASCE, 2007), which provide guidance on default soil properties
and equations for static spring stiffnesses. A few structural engineers also use rule-
of-thumb checks based on practical experience. Most structural engineers, however,

do not know the basis of spring values provided by the geotechnical engineer, and the
basis is often not stated in the geotechnical report.
Often the modulus of subgrade reaction, given for long-term gravity loading, is used
as the basis for dynamic loading conditions. This can lead to significant
underestimation of soil stiffness during dynamic loading, and significant
overestimation of the amount of displacement and rotation caused by the deflecting
soil. The cause for this appears to be a lack of understanding about the differences
between long-term and short-term loading effects on soil, a lack of clarity on what is
needed by the structural engineer, and a lack of understanding of the limits of the
information the geotechnical engineer is providing.
When foundation springs are used, practice varies on the extent to which uncertainty
is used to bound expected soil properties. Resources like FEMA 356 and ASCE/SEI
41-06 specify a range of two times the expected values to one-half of the expected
values, but this is dependent on the variability in site conditions and scatter of
material properties. Some structural engineers use a simple, conservative upper
bound assuming a fixed base (or an essentially infinitely rigid spring). Others modify
the spring values provided by the geotechnical engineer in some way. The best result
is obtained when the geotechnical engineer provides upper- and lower-bounds around
expected values. This usually narrows the bounds recommended by FEMA 356 and
ASCE/SEI 41-06, but is dependent on the variability in site conditions and scatter
present in measured soil properties.
When force-displacement relationships are provided for dynamic loading conditions,

a maximum or ultimate capacity is frequently provided, allowing an elasto-plastic
relationship to be developed. When this is not directly provided, safety factors are
occasionally given, permitting back-calculation of the plateau value. Some
geotechnical reports, however, remain fully in the allowable stress design format in
which safety factors are used but not reported. This situation prevents a rational
investigation of SSI effects at higher seismic force levels.
For deep foundations such as piles or piers, different stiffnesses are occasionally
provided for tension and for compression. Lateral springs are typically developed
using common commercial software programs that provide the displacement, shear,
and moment profile considering the variation in soil properties over the depth of the
element.
6.2.8 Modeling Approaches for Common Design Situations
Modeling approaches vary depending on the configuration of the building, the

presence of a basement, and the level of the supporting grade.

Embedded Buildings
Figure 6-1 illustrates a common situation of a building with a basement that is

surrounded by soil with a level grade on all sides. Several potential modeling
strategies that can be used to analyze such a building are shown in Figure 6-2.
Figure 6-1 Illustration of an embedded building configuration with a basement

surrounded by soil and a level grade on all sides.
Figure 6-2 Modeling approaches for embedded foundations.

Of the modeling approaches illustrated in Figure 6-2, there was no consensus on the
best approach, or even the most common approach in design practice. Observations
include the following:
 Model 1. In this approach, the building is assumed to be fixed at the ground
level, and no SSI effects are considered. Reactions are calculated at the base of
the model at grade, and then applied to the foundation in a separate model. This
approach is used occasionally by some engineers, and frequently by others.
Some believe this model is more appropriate for moment frame and braced frame
buildings, and less appropriate for shear wall buildings. Others would only use
this model if there were no “back-stay effects.” Back-stay effects occur when the
lateral system of the superstructure does not align with the full footprint of the
foundation. An example of such a configuration is a superstructure with a
concrete shear wall core system founded on a wider basement podium structure
with perimeter walls. Backstay effects involve the transfer of lateral forces in the
superstructure into additional elements that exist within the basement, typically
through one or more floor diaphragms. This force transfer helps tall buildings
resist overturning effects, and is referred to as the backstay effect because of its
similarity to the back-span of a cantilever beam.
The added stiffness of the retaining walls, and the relative rigidity of the walls
and diaphragms comprising the basement, can lead to larger shear forces in the
basement levels than in the superstructure. Modeling of the in-plane flexibility
of the basement diaphragms can provide more rational results for this effect.
When the lateral force-resisting system of the superstructure covers the full
footprint of the basement, the backstay effect is reduced or eliminated.
 Model 2. In this approach, the soil on the sides of the retaining walls is ignored,
but the basement structure is explicitly included in the model. The lowest
basement level is taken as the base of the model. It is argued that the amount of
movement required to develop passive pressure of any significance far exceeds
the amount of movement anticipated in the basement retaining walls, so the
retained earth can be conservatively ignored. One variation of this model
(Model 2A) has a fixed base. A second variation of this model (Model 2B) has
vertical springs under the foundation. Model 2A is occasionally used, though
limited by some to the preliminary design phases of their projects. Model 2B is
used more frequently, though some use both Model 1 and Model 2B to bound
their analyses for design.
 Model 3. In this approach, horizontal springs are used to capture the effect of the
surrounding soil. The ends of the springs are fixed against translation, and
ground motion is input at the base of the model. When this modeling approach is
used, it is used in pushover analyses. One variation of this model (Model 3A) is
fixed vertically. A second variation of this model (Model 3B) includes vertical

springs. Model 3A is used occasionally by some, while Model 3B is used more
frequently. Some engineers never use either version of this modeling approach.
 Model 4. In this approach, horizontal springs are attached to rigid walls (referred
to as a “bathtub”), and the ground motion is applied to the bathtub so that the
ends of the horizontal springs all move together with the input motion. This
modeling approach is rarely used, although it is recommended in the recent
Guidelines for Performance-Based Seismic Design of Tall Buildings (PEER,
2010).
 Full Substructure Model (MB). This modeling approach is designated the
Baseline Model (or MB). It represents a comprehensive modeling approach in
which dashpots are used to address soil damping variation, foundation rotation is
considered, and multi-support excitation is applied through the horizontal springs
so that the inputs vary up the height of the basement walls. This model is not
currently used in practice.
Partially Embedded Buildings with Unbalanced Loading
Figure 6-3 illustrates a building with a basement that does not have retained earth on
one or more sides. In this situation, the loading caused by the soil pressure on one
side of the building is typically analyzed separately and added to the building inertial
loading by linear superposition. Structural engineers report the use of Models 2A,
2B, 3A, and 3B, as shown in Figure 6-2, for modeling partially embedded buildings.
Figure 6-3 Illustration of a partially embedded building configuration with

unbalanced loading.
Buildings without a Basement
Figure 6-4 shows a building without a basement. Structural engineers report the use
of different modeling approaches for buildings without a basement, depending on the
type of foundation system:

 For mat foundations, Model 2B is typically used. Note that Model 4 is
equivalent to Model 2B when the foundation is not embedded.
 For spread footings, both Models 1 and 2B (equivalent to Model 4) are used.
 For grade beams, both Models 1 and 2B (equivalent to Model 4) are used.
 For drilled piers or piles, Model 2B (equivalent to Model 4) is frequently used.
Figure 6-4 Illustration of a building configuration without a basement.
6.3 Recommendations
Based on the above observations, recommendations for improving the

implementation of SSI include:
 Enhanced techniques for improving dialogue and collaboration between
structural engineers and geotechnical engineers.
 A sample checklist of information that the structural engineer should consider
providing to the geotechnical engineer.
 A sample checklist of information needed by the structural engineer, which
should be requested from the geotechnical engineer.
 Sample formats for presentation of static and dynamic spring parameters.
 Guidance on modeling approaches applied to typical foundation design
situations, with recommendations for practice.
 Example applications showing the benefits and drawbacks associated with
simplified SSI modeling approaches.
6.3.1 Improved Collaboration Between Structural and Geotechnical

Engineers
Better clarity on the part of structural engineers with regard to the modeling approach
for the soil and foundation will help geotechnical engineers provide better
recommendations. Having a defined number of meetings between the geotechnical
and structural engineers will improve communication and help ensure proper
development and implementation of foundation recommendations.

Some owners require geotechnical reports to be written in stages, including an initial
draft to help foster communication, and then later drafts or a final report after the
design is further along and needs are better defined. For example, an initial draft
memo might broadly define the various possible foundation systems, and give
preliminary design values. Once the final system is selected (or narrowed down to
final candidates), then detailed recommendations can be presented.
Owners can level the playing field in the competitive market place by clearly
requesting these additional scope items in the proposal. Owners can also coordinate
the timing of consultant selection and engagement so that geotechnical and structural
engineers are on a concurrent project delivery schedule. Some projects might benefit
from a foundation design charrette involving a foundation contractor to discuss
project specific installation issues that can help in refining and improving the design.
6.3.2 Checklist of Information That Should be Provided to

Geotechnical Engineers
Certain information is needed by the geotechnical engineer to develop foundation

design recommendation and provide the necessary information for modeling SSI
effects. This information is often provided by the structural engineer, though other
members of the design team (e.g., architect or owner’s representative) can supply
much of the information. A sample checklist of items includes the following:
1. Location. Identify the site and the proposed location of the building on the site.
2. Building Code. Identify the building code, standard, or design guideline that will
be used on the project.
3. Performance Objectives. If performance-based design will be used, identify the
guidelines that are proposed and the specific performance objectives, acceptance
criteria, and hazard levels that are being considered.
4. Description of the Building. Provide a description of the building, including the
number of stories, anticipated height, footprint size, overall square footage,
occupancies, and basement depth. Describe the potential structural and
foundation systems under consideration. Provide drawings if they are available,
particularly a foundation plan or floor plan showing column locations, and a
building section.
5. Differential Settlement Tolerance. Provide special restrictions on differential
settlement, if any, as these can impact foundation system selection.
6. Approximate Fundamental Building Period. If potential structural systems are
known, provide the approximate fundamental building period in each direction.
This is useful in developing recommendations for SSI modeling parameters.

7. Loads to Foundation. If preliminary loads are known, provide values, or a range
of potential values for the following load cases (assuming earthquake governs
over wind). The load combinations listed below are an example list; others
should be provided as relevant for the particular project. Consistent with historic
practice, and much of current practice, allowable stress design combinations are
shown; however, factored loads should be provided when performance-based
design and SSI will be used. Associated capacities are needed, particularly at the
maximum or ultimate load level. Note that both factored level load combinations
and “ultimate” level load combinations are listed, as the associated capacities will
be different at each level.
a. Allowable stress design:
i. D: average and maximum.
ii. D + L: average and maximum.
iii. D + 0.75L + 0.75(0.7E): average and maximum compression.
iv. 0.6D + 0.7E: average and maximum tension.
b. Strength design:
i. 1.2D + 1.6L: average and maximum.
ii. 1.2D + E + L: average and maximum compression.
iii. 1.2D + Em + L: maximum ultimate compression.
iv. 0.9D + E: average and maximum tension.
v. 0.9D + Em: maximum ultimate tension.
8. Soil-Structure Interaction. Indicate if SSI is likely to be included in the design
process, along with the level of sophistication of modeling that is anticipated.
Indicate if recommendations on ground motion modification for foundation
embedment and base-slab averaging are needed. Indicate if recommendations for
foundation damping are needed.
9. Site Specific Spectra. Indicate if site specific spectra are needed. Indicate
whether uniform hazard spectra or conditional mean spectra are desired. If the
site is in a known near-fault region, clarify if spectra are needed for fault normal
and fault parallel orientations. If site is known to be on soft soil, particularly Site
Class F, indicate whether ground response analysis of the soil/rock column is
anticipated. If known, indicate whether shear wave velocity measurements for
the site will be needed.
10. Response History Analysis. Indicate if response history analysis is anticipated,
and, if so, how many ground motions are required or desired. Indicate the

preferred scaling technique, such as spectrally matched or amplitude scaled,
where known.
6.3.3 Checklist of Information Needed by Structural Engineers
Certain information is needed by the structural engineer to develop preliminary and

final foundation designs, to perform structural analysis, and to model SSI effects.
Ideally, this information should be requested from the geotechnical engineer prior to
commencement of the geotechnical site investigation. Some of the following items
comprise basic information that is typically included as part of a geotechnical
investigation report, while others will likely be excluded unless specifically
requested. A sample checklist of items includes the following:
1. Site Description. Request a description of the soil and rock conditions at the site.
2. Boring Reference Elevations. Where boring logs summarizing the encountered
soil conditions are presented in terms of depth from the ground surface, request
the ground surface elevation at each boring location with respect to a common
benchmark or datum. A plan of boring locations should also be requested,
preferably showing the existing and new building footprint.
3. Field and Laboratory Testing Results. Request a summary and interpretation of
field and laboratory testing, including shear wave velocities, if they were
measured.
4. Site Seismicity Information. Request information needed for calculation of the
code base shear and site specific response spectrum. Indicate the assumptions
and methodology used to develop the response spectra, including uniform hazard
versus conditional mean spectra, scaling requirements, geomean versus
maximum rotated component assumptions, assumed shear wave velocities,
whether or not site response analysis was conducted, and the type of hazard
model used. Request spectra for fault-normal and fault-parallel orientations, as
appropriate.
5. Response Histories. If response history analysis is required or desired, request
appropriate ground motion acceleration histories. Indicate the assumptions and
methodology used to develop the acceleration histories, such as the scaling
technique (spectrally matched or amplitude scaled, for example), and fault
normal/fault parallel orientation approaches.
6. Seismic Site Hazards. Request assessment of potential for liquefaction, lateral
spreading, and other types of seismic hazard that would affect the structural
design.
7. Groundwater Table. Request information on the design groundwater elevation.

8. Foundation Type Recommendations. Request recommendations for suitable
foundation types given the geotechnical conditions at the site. The following
information is needed for some of the most common foundation types:
a. Mat Foundation:
i. Bearing pressures. Allowable soil bearing pressures under dead, dead
plus live, and dead plus live plus wind/seismic load cases. Request
associated safety factors for each case. Note whether the weight of soil
above the foundation and the equivalent weight of soil displaced by the
concrete foundation can be neglected when evaluating soil bearing
pressures.
ii. Modulus of subgrade reaction. This information is needed in order to
define vertical springs for computer analysis of a mat foundation. It will
be used for long-term gravity load cases.
iii. Vertical soil springs for earthquake loading. In addition to the modulus
of subgrade reaction, request expected, upper-, and lower-bound force-
displacement relationships for short-term dynamic loading for vertical
springs. Request similar relationships for horizontal springs.
iv. Coefficient of friction at the bottom of mat/soil interface, plus associated
safety factor.
v. Allowable passive lateral earth pressure on vertical faces of mat, plus the
associated safety factor. Indicate how passive and sliding resistance can
be combined.
vi. Heave and settlement estimates.
b. Spread/Strip Footings/Grade Beams. Request information similar to that
required for the mat foundation, including an estimate of anticipated long-
term differential settlement between adjacent footings. Request expected,
upper-, and lower-bound force-displacement relationships for short-term
dynamic loading for vertical springs. Request similar relationships for
horizontal springs.
c. Drilled Piers or Piles:
i. Request skin friction and end bearing capacities at allowable levels and
at ultimate levels, suitable for use with code overstrength load
combinations for both tension and compression loading.
ii. Force-deflection plots for axial loading. Request expected, upper-, and
lower-bound force-displacement relationships for short-term dynamic
loading for vertical springs.

iii. Plots of lateral load versus deflection and lateral load versus maximum
moment for both fixed head and pinned head conditions. Request
expected, upper-, and lower-bound force-displacement relationships.
iv. Request information on how group effects may control vertical or lateral
capacity of piles.
v. Request estimates of long-term differential settlement between pile caps
and spread footings when the two systems are mixed.
vi. Request recommendations on minimum depth of embedment.
vii. Request identification of special construction/inspection requirements.
9. Miscellaneous Site Items:
a. Excavations:
i. Maximum slope angles and soil treatment necessary to provide stable
slopes during excavations.
ii. Underpinning and shoring recommendations if required.
iii. Underpinning settlement expectations.
iv. Tie back recommendations.
b. Excavated material/engineered fill. Request information on backfill and
compaction requirements, and whether or not native soils will be acceptable
as fill.
c. Retaining wall recommendations. Request design values for cantilevered
and restrained site retaining walls, including allowable and ultimate bearing
pressures, active and passive pressures, coefficient of friction, seismic and
non-seismic surcharge criteria, appropriate combinations of resisting
elements, and methods of surcharge loading. Request expected, upper-, and
lower-bound force-displacement relationships for short-term dynamic
loading for horizontal springs, if requested.
d. Existing building considerations:
i. Zone of influence of existing building foundations.
ii. Existing footing surcharge loads and other lateral earth pressures.
iii. Deflection requirements of existing foundations, soldier piles, secant
piles, and monitoring recommendations.
e. Recommendations for slabs-on-grade.
f. Assessment of soil corrosivity.

10. Additional SSI Recommendations. In addition to the spring parameters noted
above, request the following:
a. Base slab averaging. Base slab averaging recommendations and limitations
for spectral modifications, if needed.
b. Embedment effects. Recommendations on embedment effects for spectral
modifications, if needed.
c. Damping. Recommendations on soil damping and dashpots if they will be
used in SSI modeling.
6.3.4 Sample Format for Soil Spring Characterization
Soil springs are often not properly, or adequately, implemented in practice. In some
cases, the modulus of subgrade reaction developed for long-term settlement is being
used for dynamic loading situations. In other cases, a single value or single force-
displacement relationship is being used for the soil properties. Figure 6-5 provides a
sample format for presentation of soil spring data.
25
qc(UB) = 22.5 ksf (Short‐term: Upper Bound)
kUB=1920 kcf
qc(UB) = 20 ksf (Short‐term: Expected)
20
k=960 kcf qc(UB) = 18 ksf (Short‐term: Lower Bound)
Bearing Pressure (ksf)
15
kLB=480 kcf
qall = 10 ksf (Short‐term D+L+EQ at ASD)
10
qall = 7.5 ksf (Long‐term D+L at ASD)
qall = 5 ksf (Long‐term D at ASD)
5
kLT= 180 kcf
0
0 0.5 1 1.5 2
Settlement (inches)
Figure 6-5 Sample format for presentation of force-displacement relationships for

soil springs.
The sample format includes both the relationship for long-term settlement, which
needs only a lower-bound, and recommendations for the expected (or target)
properties under dynamic loading, including both an upper-bound and lower-bound.

Per ASCE/SEI 41-06 and other engineering resources, these bounds are typically two
times and one-half times the target, unless narrower bounds are provided by the
geotechnical engineer based on the variability of the site soil conditions. For
simplicity, Figure 6-5 uses an elasto-plastic format. Actual properties would have a
positive post-yield slope and a more rounded transition, which could be
approximated by a tri-linear curve.
6.3.5 Guidance on Modeling Approaches for Typical Foundation

Situations
Guidance on modeling approaches applied to typical foundation design situations,

and recommendations for practice, is provided throughout this report. Examples are
provided in Chapter 7, and recommended procedures are provided in Chapter 8.
6.3.6 Example Applications of Simplified Soil-Structure Interaction
Case studies illustrating the level of effort, relative benefits, and potential drawbacks
associated with simplified SSI modeling approaches are provided in example
applications presented in Chapter 7 for a variety of different building configurations.

Chapter 7
Example Applications
Building seismic response analysis including soil-structure interaction involves

characterization of the soil-foundation interface in the analytical model. Typical
methods for idealizing the soil-foundation interface in current engineering practice
are described in Chapter 6. While some of these models are simple, others require
significant effort to capture linear or nonlinear SSI effects. Example applications in
this chapter are used to demonstrate different foundation modeling techniques and
investigate the relative accuracy of the resulting building response predictions.
Methods for developing soil springs and dashpots described in this report are
implemented on case-study buildings using several models with different
idealizations at the soil-foundation interface. Examples are used to illustrate the
corresponding level of effort for different modeling approaches. Overall results and
selected structural response quantities from the models are compared and contrasted.
Analyses are performed in the context of a substructure approach to modeling SSI
effects.
7.1 Overview of Example Applications
The overall approach for development of example applications involved: (1) the
selection of suitable instrumented buildings that have experienced (and recorded)
earthquake ground motions; (2) development of baseline models for comprehensive
substructure-based analysis of seismic response; (3) calibration of baseline models to
approximately match the recorded response of the buildings through variation of
structural parameters; and (4) systematic variation in the idealization of the soil-
foundation interface to evaluate the impact of different modeling approaches on the
predicted response of the buildings.
Work included a review of the results from similar studies performed by others and
published in the literature. Information from these studies was used to guide the
example applications presented herein, and corroborate the resulting observations.
7.1.1 Summary of Results from Prior Studies
Prior studies reported by Naeim et al. (2008) and Tileylioglu et al. (2010) utilized
typical structural engineering software packages, such as ETABS, Extended Three
Dimensional Analysis of Building Systems (Computers and Structures, Inc.), and
SAP2000, Integrated Software for Structural Analysis and Design (Computers and
GCR 12-917-21 7: Example Applications 7-1

Structures, Incorporated), to model the soil-foundation interface of two buildings in
Los Angeles, California. One building is a 54-story building, and the second
building is a 6-story building. Both buildings have embedded foundations, and both
experienced earthquakes with recorded ground motions. Only the results of the 54-
story building, however, have been published to date.
Analyses initially employed the full substructure modeling approach, designated the
Baseline Model (or MB) as described in Chapter 6, except that kinematic base
rocking was applied to the base of vertical foundation springs along with depth-
variable ground motions. This results in double-counting of kinematic rotation
effects. Fortunately, these effects were minor for the structures investigated, and
double-counting did not have a significant impact on results.
Springs were elastic, with no compression capacity limit, and zero tension capacity.
Kinematic effects were evaluated in a manner similar to that described in Chapter 3
(except that kinematic base rocking was applied). Significant difficulties were
encountered in implementing multi-support excitation using typical engineering
production software packages. Results were considered reliable for displacement
response, but not reliable for forces. In general, however, good matches between
computed and observed responses were reported using the baseline modeling (MB)
approach.
Selected elements of the full substructure modeling approach were then omitted from
the models to investigate their impact on the computed response. The following
factors did not have a significant impact on the results: (1) consideration of multi-
support excitation along basement walls; and (2) application of a zero-tension
condition in the foundation springs. Consideration of kinematic interaction effects
had a significant impact on the distribution of interstory drifts, particularly below
grade. Consideration of foundation springs had a significant impact on building
vibration periods and distribution of interstory drifts.
Two modeling approaches commonly used in practice were shown to provide poor
results: (1) fixing the structure at the ground surface, truncating the embedded
portions of the structure, and applying the free-field translation as the input motion
(Chapter 6, Model 1); and (2) modeling the embedded base of the structure, using
horizontal and vertical springs with ends fixed against translation, and applying free-
field motions as input at the base slab level (Chapter 6, Model 3).
A fixed-base approach considering embedded portions of the structure (Chapter 6,

Model 2) provided better results, while the “bathtub” approach (Chapter 6, Model 4)
was not considered in this work.
7-2 7: Example Applications GCR 12-917-21

7.1.2 Building Selection
The criteria considered in building selection included: (1) embedded foundations, so

that kinematic effects associated with embedment and depth-variable ground motions
could be considered; (2) recordings of rocking at the foundation level, which requires
two vertical instruments; (3) relatively regular structural configurations, so that the
results obtained are not peculiar to any one building in particular; and (4) structural
configurations and site conditions that would tend to be conducive to significant
inertial SSI effects.
Only buildings with seismic instrumentation and available earthquake recordings in

California were considered, using archives provided by the Center for Engineering
Strong Motion Data (CESMD). The CESMD provides strong-motion data for
earthquake engineering applications. It was established by the United States
Geological Survey (USGS) and the California Geological Survey (CGS) to integrate
and disseminate earthquake strong-motion data from the CGS California Strong
Motion Instrumentation Program (CSMIP), the USGS National Strong Motion
Project, and the Advanced National Seismic System (CESMD, 2011).
Two case-study buildings were selected from CSMIP, consisting of: (1) a 13-story,
reinforced concrete moment frame structure (designed in 1964; retrofitted in 1994),
with two basement levels, located in Sherman Oaks, California; and (2) a 10-story,
reinforced concrete shear wall core and perimeter moment frame structure (designed
in 1970), without any basement levels, located in Walnut Creek, California. The
exterior elevations of the buildings are shown in Figure 7-1.
(a) (b)
Figure 7-1 Exterior elevations of two case-study buildings: (a) 13-story Sherman
Oaks building; and (b) 10-story Walnut Creek building (CESMD,
2011, with permission).

The Sherman Oaks building foundation consists of grade beams and friction pile
foundations on alluvial sediments. The building was shaken strongly by the 1994
Northridge earthquake and less-strongly by other, more distant events. The Walnut
Creek building foundation consists of a shallow concrete mat foundation below the
shear wall core, and a combination of spread footings and belled caissons with
shallow embedment, resting on weathered shale bedrock, below the perimeter
moment frames. The Walnut Creek building was shaken by the 1989 Loma Prieta
earthquake and other smaller events.
No single building was found to meet all of the selection criteria. The Sherman Oaks
building satisfied criteria (1) and (3) above, while the Walnut Creek building satisfied
criteria (2), (3), and (4) above.
7.1.3 Modeling Approaches
Modeling approaches considered in the example applications are shown in Figure

7-2. They are based on the approaches implemented in typical engineering practice,
as described in Chapter 6. They include the Baseline Model (designated MB), and
various simplified idealizations of the soil-foundation interface, designated Model 1,
Model 2, Model 3, and Model 4.
Figure 7-2 Modeling approaches considered in example applications.
 Baseline Model (MB). A baseline model (designated MB) is a three-

dimensional model of the building, foundation, and soil-foundation interface.
Baseline models are not intended to be the most accurate models that could be
developed (a direct SSI analysis in a finite element platform could provide

improved results). Instead, they are intended to represent a comprehensive
modeling approach that includes soil springs to consider translational and
rotational degrees of freedom, dashpots to address soil damping effects, and
multi-support excitation to capture variable input motions over the depth of
embedment. Baseline models incorporate vertical translation, horizontal
translation, and rocking degrees of freedom. A series of springs and dashpots are
developed based on site-specific soil properties. Seismic input includes base
translation, as well as kinematic loading of basement walls (simulated by
displacement histories applied to the ends of horizontal springs attached to
basement walls). Baseline models are calibrated to match building response
quantities interpreted from available recorded data.
Once baseline models have been successfully calibrated to match recorded data, the
soil-foundation interface is idealized using one of the following simplified modeling
approaches from Chapter 6:
 Model 1. In Model 1, only the above-ground portion of the structure is modeled,
the base is fixed at the ground surface, and the free-field ground motion, ug, is
applied at the base of the model.
 Model 2. In Model 2, the above-ground and subterranean portions of the
structure are modeled, the based is fixed at the bottom of the embedded portion
of the structure, the soil surrounding the embedded portion is ignored (i.e., no
horizontal foundation springs are used), and the free-field ground motion, ug, is
applied at the base of the model.
structure are modeled, horizontal and vertical soil springs are included, the far
end of each spring is fixed against translation, and the free-field ground motion,
ug, is applied at the base slab while horizontal springs at other levels remain
fixed. In the example applications, Model 3 is investigated using response
history analysis, although in practice, it is typically used in nonlinear static
(pushover) analysis applications.
structure are modeled, horizontal and vertical soil springs are included, the far
end of each spring is fixed against a rigid “bathtub” element, and the horizontal
foundation input motion, uFIM, or free-field motion, ug, is applied to the rigid
element. The key difference between Model 4 and Model MB is the manner in
which the seismic demand is applied. In Model 4, the effect of kinematic loading
on basement walls associated with depth-variable displacement histories is
neglected. Further, the use of free-field motions in lieu of foundation input
motions neglects kinematic interaction altogether by replacing the recorded

motions at the base of the building with equivalent free-field motions applied
uniformly at the ends of all horizontal foundation springs.
Table 7-1 summarizes the properties of the modeling approaches considered in the
example applications. Variations in Model MB (denoted MB.1 and MB.2) are
considered as follows:
 In Model MB.1, the embedded portion of the building is assumed to be rigid.
The specification of seismic demand is not modified. The objective of this model
is to investigate the effects of flexibility in the subterranean structural elements.
 In Model MB.2, there is a change in the way that soil flexibility is modeled. In
this model, springs are not allowed to develop tension to investigate the effects of
nonlinearities in the foundation springs.
Table 7-1 Summary of Modeling Approaches Considered in Example Applications
Model Variations
Parameters No. 1 No. 2 No. 3 No. 4 MB MB.1 MB.2
Structural Foundation
N/A Flexible Flexible Flexible Flexible Rigid Flexible
Elements
4a: uFIM
Input Motion ug ug ug uFIM uFIM uFIM
4b: ug
Depth-variable Ground
No No No No Yes Yes Yes
Motion
Not
Spring Tension n/a n/a Permitted Permitted Permitted Permitted
Permitted
All modeling schemes, except Model MB.2, permit tension to develop in the soil
springs. The rationale for this approach is that, prior to an earthquake, actual
basement wall-to-soil contact pressures range between ‘at rest’ and ‘active’ earth
pressures. Springs have an initial condition of zero force. Since earthquake shaking
will impose alternating cycles of increased and decreased pressures relative to the
initial state, wall-to-soil contact can be represented by a spring that develops tension,
provided that the level of deformation does not lead to gap formation. Given the
modest levels of demand imposed on the Sherman Oaks and Walnut Creek buildings,
this was expected to be a reasonable assumption, and this assumption was tested
using Model MB.2 on the Sherman Oaks building.
7.2 Sherman Oaks Building
The Sherman Oaks building is a 13-story structure above grade with two basement
levels below grade. The seismic force-resisting system consists of reinforced
concrete moment-resisting frames that extend from the roof to the foundation,
supplemented by perimeter concrete shear walls in the subterranean levels. The
gravity system consists of cast-in-place concrete slabs supported on concrete beams

and columns. The building was originally designed in 1964, and later seismically
upgraded with friction dampers following the 1994 Northridge earthquake.
Construction drawings of the Sherman Oaks building were made available for
inspection through the auspices of the California Strong Motion Instrumentation
Program. The building measures 50 m (164 ft) tall from the ground surface to the
roof. The plan dimensions of the superstructure are 21.9 m (72 ft) wide by 57.6 m
(189 ft) long, although it widens at the first floor to match the foundation dimensions.
The height of the first story is 7.0 m (23 ft), and other above-grade stories are 3.6 m
(11.75 ft). The basement levels are embedded approximately 6.2 m (20.5 ft) below
the ground surface. The height of the first basement level is 3.5 m (11.5 ft), and the
height of the second basement level is 2.7 m (9 ft). The characteristics of the
foundation are described in Section 7.2.3 below.
7.2.1 Site Characterization
A plan view of the Sherman Oaks site is shown in Figure 7-3. The ground surface
elevation is approximately 216 m (709 ft) above mean sea level. Boring information
and geophysical logs were obtained from geotechnical investigations conducted in
the vicinity of the site (LeRoy, Crandall and Associates, 1978 and 1982).
Figure 7-3 Plan view of the Sherman Oaks site showing locations of borehole
and geophysical logs used for site characterization (adapted from
LeRoy, Crandall and Associates, 1978 and 1982).
Information from a total of 47 borings ranging in depth from 15 m (50 ft) to 38 m

(125 ft) was obtained. Borings indicated relatively consistent soil conditions,
consisting of alluvial deposits composed of silt, with sand and clay overlying shale

bedrock at depths ranging from 21 m (70 ft) to 27 m (90 ft). The water table was
measured at an approximate depth of 12 m (40 ft) below the ground surface.
Geophysical data was developed from downhole seismic surveys. Based on available
geotechnical information, an interpretation of the soil layering at the site, in terms of
shear wave velocity profile and material profile, is shown in Figure 7-4.
Figure 7-4 Subsurface characteristics of the Sherman Oaks site: (a) shear wave velocity
profile; and (b) material profile (adapted from LeRoy, Crandall and Associates,
1978 and 1982).
The average moist unit weight was taken to be 20 kN/m3 (124 lb/ft3). To estimate
strength properties of the foundation soils, direct shear test results from samples were
utilized. In the shallow soils, where most of the soil-foundation load transfer will
occur, the soil is unsaturated, so drained shear strengths were used. For the range of
surcharge pressures over the soil profile extending to the bottom of piles at a depth of
parameters inferred from available data are: (1) c′ = 20.1 kN/m2, and ′ 30°, from
approximately 16.2 m (53 ft) below grade, the soil Mohr Coulomb strength
0.0 m to 6.1 m; and (2) c′ = 12.9 kN/m2, and ′ = 38°, from 6.1 m to 16.2 m.
7.2.2 Ground Motion Recordings
The Sherman Oaks building was instrumented in 1977 by the California Strong
Motion Instrumentation Program and designated CSMIP Station No. 24322.

Instrumentation includes 15 accelerometers at the locations shown in Figure 7-5,
including the second basement (foundation level), ground floor, second floor, eighth
floor, and roof levels. There is only one vertical sensor located at the foundation
level, so base rocking effects cannot be measured, and there are no free-field
instruments in the vicinity of the site.
Figure 7-5 Sherman Oaks building (CSMIP Station No. 24322) instrument
locations (CESMD, 2011, with permission).
Since 1977, six earthquake events have been recorded and processed by CSMIP at
this station. The events, along with peak accelerations for the second basement level,
ground floor, and roof are presented in Table 7-2. Note that in the 1987 Whittier and
1994 Northridge events, the peak accelerations followed an unusual pattern in which
the largest recorded motions occurred at the ground floor (i.e., the recorded motions
at the foundation and roof levels were smaller). As a result, subsequent calculations
are based on peak accelerations measured at the foundation level.
Horizontal translations recorded at the second basement (foundation) level were used
as foundation input motions, uFIM. The foundation input motion, uFIM, is the modified
free-field, ug, response due to base-slab averaging and embedment effects. Typically,
ug is known, and uFIM must be calculated based on transfer functions. In this case,
uFIM was measured and ug was inferred using transfer functions to remove the base-

slab averaging and embedment effects. The recorded uFIM and inferred ug response
spectra for the 1994 Northridge earthquake are shown in Figure 7-6.
Table 7-2 Summary of Earthquake Events Recorded at the Sherman Oaks Building
Second Basement Ground Floor Roof
Earthquake (Foundation), (g) (g) (g)
N-S E-W V N-S E-W V N-S E-W V
1987 Whittier 0.100 0.148 0.038 0.250 0.169 ---- 0.140 0.140 ----
1992 Landers 0.039 0.031 0.012 0.045 0.035 ---- 0.085 0.095 ----
1994 Northridge 0.453 0.227 0.181 0.886 0.392 ---- 0.467 0.257 ----
2007 Chatsworth 0.015 0.022 NR 0.045 0.058 ---- 0.025 0.036 ----
2008 Chino Hills 0.049 0.037 0.010 0.073 0.045 ---- 0.082 0.067 ----
2010 Calexico 0.004 0.004 0.003 0.004 0.005 ---- 0.049 0.022 ----
Symbols: N-S=north-south; E-W=east-west; V=vertical; NR=no record available.
Figure 7-6 Response spectra for recorded foundation input motion, uFIM, and inferred free-
field motion, ug, at the Sherman Oaks building, 1994 Northridge earthquake.
An equivalent linear ground response analysis was performed using DEEPSOIL v4.0
(University of Illinois, 2011) to estimate the depth-variable ground response adjacent
to the embedded portion of the structure, taking the recording of uFIM as input at
6.7 m with an elastic half-space. Input motions in ground response analyses can be
specified as “outcropping” or “within,” the former requiring an elastic half-space, and
the latter requiring a rigid base (Kwok et al., 2007). For the Sherman Oaks building,
the motion was recorded “within” the profile, but on a large foundation slab that
could be interpreted as representing an outcropping condition. Analyses were
performed for both conditions, with “outcropping” results appearing to be more
realistic.

Dynamic soil properties consisted of the Vs profile shown in Figure 7-4 and modulus
reduction and damping curves appropriate for the overburden pressures and soil types
present at the site, as given by Darendeli (2001). DEEPSOIL does not update the
shear modulus reduction of the elastic half-space during the analysis. Therefore, the
elastic half-space parameter was updated through iterative runs until convergence
was reached with the Vs of the deepest layer in the soil column. The computed
variation of peak ground acceleration (PGA) and peak ground velocity (PGV) with
depth is shown in Figure 7-7. Note that PGA changes significantly with depth
(approximately 50%) whereas the change in PGV is comparatively minor.
Figure 7-7 Computed variation of peak ground acceleration (PGA) and peak
ground velocity (PGV) with depth at the Sherman Oaks site, 1994
Northridge earthquake.
7.2.3 Foundation Conditions
The foundation is rectangular in plan, measuring 36.3 m (119 ft) wide by 57.6 m
(189 ft) long. The foundation consists of bored pile groups of varying configurations,
interconnected by pile caps and grade beams. The cast-in-place (bored) concrete
piles measure 51 cm (20 in) in diameter, and extend to a depth of approximately
9.9 m (32.5 ft) below the lowest basement level (i.e., approximately 16.2 m below
grade). Typical pile spacing varies between 0.9 m (3 ft) to 1.2 m (4 ft), center to
center. Concrete basement walls surround all embedded levels below grade. A
foundation plan developed based on information contained in construction drawings
provided by CSMIP is shown in Figure 7-8.

Figure 7-8 Sherman Oaks building foundation plan, based on construction drawings
provided by CSMIP.
7.2.4 Development of Foundation Springs and Dashpots
For the Sherman Oaks building, resistance is provided by both shallow and deep
foundation elements. It is assumed that the shallow foundation elements (principally
grade beams) remain in contact with the soil, which is justified by the sandy nature of
the soils at the site and the presence of friction piles rather than end-bearing piles.
The methodologies described in Chapter 2 were applied to the development of

foundation springs and dashpots for the Sherman Oaks building. Critical input
parameters included the foundation dimensions, fundamental vibration periods of the
structure, soil properties, and the amplitude of the input motion. Modeling of the
Sherman Oaks foundation and site conditions included the following:
 Determination of average effective profile velocity for shallow foundation
elements, considering foundation dimensions, overburden pressures from the
structure, and nonlinear effects.
 Determination of the structure-to-soil stiffness ratio for estimating the
significance of soil-structure interaction effects.
 Calculation of shallow foundation horizontal stiffness and damping coefficients,
and their separate contributions arising from base slab friction and passive
pressure resistance against the basement walls.

 Calculation of vertical stiffness and damping coefficients for springs under the
base slab, emphasizing the adjustment for rotational impedance for shallow
foundation elements.
 Contribution of pile groups to the vertical and lateral stiffness of the foundation
system.
 Distribution of shallow foundation springs and dashpots to the nodes in the
analytical model.
 Calculation of limiting spring forces (i.e., capacities).
Static stiffnesses, dynamic stiffness modifiers, and embedment correction factors,
were calculated for the Sherman Oaks building based on gross foundation dimensions
(21.9 m by 57.6 m). The use of gross foundation dimensions assumes that stress
bulbs below the pile caps and grade beams overlap at modest depth, so that the
shallow foundation components effectively perform as a unified foundation system.
This assumption does not require the relatively thin foundation slab between pile
caps and grade beams to resist bearing pressures.
Average Effective Profile Velocity. For soil profiles that vary with depth,
foundation stiffness and damping coefficients are based on an average effective
profile velocity calculated over an effective profile depth, zp. Table 7-3 summarizes
the effective profile depth, zp, the depth range, and the average effective profile
velocity calculated using Equation 2-18 for each foundation vibration mode over the
depth range considered.
Table 7-3 Summary of Effective Profile Depths and Average Effective Profile Velocities
for the Sherman Oaks Building
zp Depth Range Vs,avg
Vibration Mode Basis (1) (m) (m) (m/s)
Horizontal translation (x and y), overall BL 17.8 0 to 24.0 254.2
Horizontal translation (x and y), base spring BL 17.8 6.2 to 24.0 296.2
Vertical translation (z) BL 17.8 0 to 24.0 254.2
Rocking along x-axis (xx) 4

B3 L 14.0 0 to 20.2 243.0
Rocking along y-axis (yy) 4

BL3 22.6 0 to 28.8 277.7
Notes: (1) Calculated using overall foundation half-width, B=11 m; half-length, L=28.8 m.
In Table 7-3, the maximum depth considered is greater than zp by the amount of the
foundation embedment, D (i.e., the depth range extends to D + zp). For the specific
case of the horizontal base spring, the depth interval begins at the base of the
foundation.

Before averaging, measured free-field velocities, Vs, should be increased to account
for the presence of overburden pressures caused by the added weight of the structure.
For the Sherman Oaks building, the weight of the soil that was excavated for the
basement levels was slightly greater than the estimated weight of the structure, so no
net overburden pressure exists below the foundation, and no correction was applied.
Structure-to-Soil Stiffness Ratio. Taking the transverse (N-S) direction as critical,

the structure-to-soil stiffness ratio, h/(VsT), for the Sherman Oaks building is
approximately 0.06, calculated using two-thirds of the modeled building height from
foundation to roof, h = 2/3(56.2 m) = 37.5 m, the fundamental period for the fixed
base condition in the transverse direction, T = 2.71 sec, and the average effective
profile velocity for rocking about the xx-axis, Vs, avg = 243.0 m/s. Because this ratio is
less than 0.1, strong inertial SSI effects would not be expected to occur (i.e., period
lengthening near unity, foundation damping near zero). The building was further
analyzed, however, to study potential kinematic interaction effects on higher mode
responses. This study was used to determine if a building for which traditional first-
mode SSI metrics indicated no significant effect, could, in fact, exhibit potentially
significant impacts on the vertical distribution of structural response quantities used
in design (e.g., drift ratios and story shears).
Horizontal Stiffness and Damping. Calculations for horizontal stiffness and

damping ratios are shown in Table 7-4. For horizontal stiffness, contributions from
both base friction and passive pressure resistance were considered. Because base
friction mobilizes the soil below the embedment depth, D, the average effective
profile velocity for the base spring is taken between depths D and D+zp. The base
spring stiffness was subtracted from the overall horizontal stiffness to determine the
portion of horizontal stiffness attributed to passive pressure resistance against the
basement walls.
density, s, of 20 kN/m3/g, and values of average effective profile velocity, Vs, avg,
In Table 7-4, shear modulus, G, was evaluated from Equation 2-9, using a soil mass
determined using overburden-corrected shear wave velocities below the foundation,

shown in Table 7-3. Values of shear modulus should be reduced to account for large
strain effects associated with nonlinear behavior. Using Table 2-1, assuming Site
Class D, and peak acceleration (at the foundation level) of about 0.45 g from the
1994 Northridge earthquake, the shear modulus reduction factor, G/G0 = 0.5. A
modulus reduction factor of 1.0 was used for the other smaller events. Other
parameters were determined using the equations referenced in the table.

Table 7-4 Calculation of Shallow Foundation Stiffness and Damping Parameters for the Sherman Oaks Building
G(1) Ksur (2)  a0(3)  sur emb
Eq. 2-9 Table 2-2a Table Eq. Table Table Table Dynamic Stiffness Dashpot Coefficient
Spring/Dashpot (MPa) (kN/m; kN-m/rad) 2-2b 2-15 2-3a 2-3a 2-3b Eq. 2-14a Eq. 2-13a
   s 
Horizontal, overall
k x ,total   x K x , sur x cx,total  2k x ,total  emb 
k , cx ,total    
x-direction
65.2 6.48e6 1.45 0.13 1 0.076 0.114
x ,total
k y ,total   y K y , sur y    s 
Horizontal, overall
cy ,total  2k y ,total  emb
k , c y ,total   
 
y-direction 65.2 7.04e6 1.45 0.13 1 0.070 0.106
y ,total
   s 
Horizontal, base spring
k x ,base   x K x , sur cx,base  2k x ,base  sur 
k , cx ,base    
x-direction 88.6 8.81e6 --- 0.11 1 0.065 ---
x ,base
k y ,base   y K y , sur    s 
cy ,base  2k y ,base  sur
k , c y ,base 

  
y-direction 88.6 9.56e6 --- 0.11 1 0.060 ---
y ,base
   s 
k z   z K z , sur z cz  2kz  emb
 kz , cz  
Vertical, z
  
65.2 8.55e6 1.22 0.13 1 --- 0.132
   s 
k xx   xx K xx , sur xx cxx  2kxx  emb
 kxx , cxx  
  
58.0 1.05e9 1.74 0.14 1 --- 0.010
   s 
k , c yy 
k yy   yy K yy , sur yy cyy  2k yy  emb 
  
73.9 5.55e9 1.58 0.12 0.99 --- 0.003
yy
using overall foundation half-width, B=11 m; half-length, L=28.8 m; and =0.33.

Notes: (1) Calculated using a shear modulus reduction factor, G/G0=0.5 for the Northridge earthquake; for other events G/G0=1.0, and values would be doubled.
(2) Calculated
(3) Calculated at a frequency corresponding to the first-mode period of the flexible-base structure.

Dashpot coefficients were determined using Equation 2-13a and the total foundation
damping ratio. The total foundation damping ratio is equal to the sum of the
hysteretic damping ratio, s. For the Sherman Oaks site, the soil hysteretic damping
radiation damping ratio (determined using equations in Chapter 2) and the soil
ratio was evaluated using Darendeli (2001), and determined to be 0.088 for the
Northridge earthquake, and 0.009 for the other events.
Vertical Stiffness and Damping. Calculations for overall vertical and rotational
stiffness and damping ratios are shown in Table 7-4, using values of average
effective profile velocity from Table 7-3. Vertical springs were distributed over the
footprint of the foundation, as shown in Figure 7-9, using a vertical stiffness intensity
that is normalized by area (Equation 2-20). Stiffness intensities were adjusted near
the edges of the foundation to match the overall rocking stiffness values given in
Table 7-4 (using Equation 2-21 and an end length ratio of Re = 0.4). The stiffness of
any individual spring in the model was then computed as the product of stiffness
intensity and the tributary area for the spring.
k zi  kN/m3  czi  kN-sec/m3 

8,202 715
29,361 2,454
36,524 3,209
43,687 3,964
Figure 7-9 Vertical spring and dashpot intensities distributed over the footprint
of the Sherman Oaks building. Solid lines represent tributary area
boundaries and solid circles represent column nodes.
Vertical stiffness intensities in the central zone of the foundation were computed
using Equation 2-20a. Edge intensities were increased by factors of Rk = 5.33 for the
xx-direction and Rk = 3.58 for the yy-direction to correct for underestimation of
rotational stiffness (Equations 2-21a, 2-21b). Corner intensities were evaluated as the
average of the intensities in the xx- and yy-directions. Dashpot intensities were
calculated based on stiffness intensities using Equation 2-20b, and reduced by Rc
(Equations 2-21c, 2-21d), to correct for overestimation of rotational damping.

Contribution of Pile Groups. Pile group horizontal and vertical stiffness and
damping ratios were evaluated in accordance with Section 2.3. Active pile length, La,
was evaluated separately for the two vibration modes. It is approximately equal to
the pile length, Lp, for vertical modes, and is less than Lp for horizontal modes. For
the Sherman Oaks building, the active pile length for horizontal loading was
evaluated using the results from Syngros (2004). Static pile stiffnesses, K jp , were
calculated using the equations in Table 2-4a, modified for dynamic effects to
calculate the single pile stiffness, k jp , and then modified for group effects.
Calculations were based on average effective shear wave velocity calculated over an
effective profile depth from D to D+La, and shear modulus reduced for large strain
effects (G/G0=0.5 for the Northridge earthquake; 1.0 for other events). Approximate
dynamic pile group efficiency factors for the Sherman Oaks building were evaluated
using Figure 2-11.
Results are shown in Table 7-5. Results are provided for 3x3 and 4x4 pile group
configurations and applied to individual pile groups shown in Figure 7-8. Results for
3x3 pile groups were applied to groups with 12 piles, and results for 4x4 pile groups
were applied to groups with more than 13 piles. The modeled stiffness for each pile
group is the product of the individual pile stiffness, k jp , the pile group efficiency
factor, k Gj , and the number of piles, Npile, in each group.
Table 7-5 Calculation of Pile Stiffness and Damping Parameters for the Sherman Oaks
Building
Static Pile Pile Group
Active Pile Stiffness, Efficiency
Pile Group
Length, La Vs,avg G(1) K jp (2) Factor,
j (3)
Spring/ Damping Ratio,
Dashpot (m) (m/s) (MPa) (MN/m) k Gj (3)
Horizontal, x, y
4.1 180.1 63.4 237.3 0.47 0.2
(3x3 group)
Vertical, z
9.9 183.7 67.6 859.3 0.43 0.25
(3x3 group)
Horizontal, x, y
4.1 180.1 63.4 237.3 0.36 0.05
(4x4 group)
Vertical, z
9.9 183.7 67.6 859.3 0.30 0.29
(4x4 group)
Notes: (1) Calculatedusing a shear modulus reduction factor, G/G0=0.5 for the Northridge earthquake.
(2) Calculatedfor Ep=474.8 MPa, and d=0.6 m.
(3) Approximated using Figure 2-11.
Distribution of Springs to Foundation Nodes. A total of 68 nodes were included in

the Sherman Oaks foundation model. Horizontal springs and dashpots were
distributed across nodes on the basement walls (representing passive pressures) and
base slab (representing base shear). Vertical springs and dashpots were distributed

across nodes on the base slab. A plan view, transverse section, and longitudinal
section of foundation nodes are shown in Figure 7-10.
Pile Group Contribution k y , piles  k yG k x , piles  k xG
k y ,total  k y ,base
k1  k2  k1  
Translation in the k y ,base k y , piles
y-direction 48 16 16
k x ,total  k x ,base
k4  k5  k 4  
Translation in the k x ,base k x , piles
x-direction 27 6 6
k3  k zG  k z
Translation in the
z-direction Note: k3 is not a constant value; is a function of Npiles, and kz is a function
of spring intensities shown in Figure 7-9.
Figure 7-10 Plan view, transverse section, and longitudinal section of foundation nodes
for the Sherman Oaks building, and calculation of associated spring
stiffnesses.
The overall horizontal stiffness, including contributions from base shear, passive
pressure resistance against the basement walls, and pile groups, was distributed
around the foundation perimeter (i.e., interior nodes have no horizontal springs).

Half of the horizontal stiffness was distributed to each edge of the foundation in each
direction. Horizontal stiffness was partitioned between a base stiffness and a wall
stiffness in each direction. Wall stiffnesses were distributed equally to each
basement wall node, and base stiffnesses were distributed equally along each side of
the foundation.
Vertical stiffness, including contributions from shallow foundation elements and

piles, was modeled at all 24 foundation nodes. Shallow foundation stiffness
contributions were based on stiffness intensity, and pile stiffness contributions were
based on the vertical pile group stiffness calculations. Dashpots were distributed to
the foundation nodes in an identical manner.
Limiting Spring Capacities. Capacities of individual springs were evaluated using

conventional vertical and lateral pile group capacity analysis (e.g., Salgado, 2008).
Pile group capacities were assigned to foundation nodes as described above for
spring stiffness. Capacities are compared to spring force demands in the following
section.
7.2.5 Analysis Results
The Sherman Oaks building was analyzed in OpenSees (McKenna, 1997),

incorporating foundation springs and dashpots described above. Details for the
development and calibration of the Baseline Model (MB) are described in
Appendix A. Adjustments to the shear modulus for cracked concrete, Rayleigh
damping, and structural mass and stiffness were used to produce a reasonably close
match to recordings from the 1994 Northridge earthquake. Based on calibration
studies, Model MB was considered a reasonable engineering approximation for the
Sherman Oaks building. Using this model as a basis, alternative modeling
configurations for the Sherman Oaks building were developed and studied, and the
resulting response quantities were compared for the following variants:
 Model MB.1, which examined the importance of flexibility in subterranean
structural elements (walls and slabs);
 Model MB.2, which examined the importance of nonlinearity in foundation
springs by not allowing tension (allowing gap formation);
 Model 4, which removed the effects of depth-variable ground motions and
considered the use of free-field motions (versus foundation input motions);
 Model 3, which fixed the far ends of foundation springs against displacement and
applied input motions at the base slab level;
 Model 2, which ignored the effects of embedment by omitting the surrounding
soil and assumed a fixed base at the foundation level; and

 Model 1, which ignored the response of the subterranean levels by assuming a
rigid base at the ground level.
Fundamental periods of vibration for each of the model variants are shown in Table
7-6. The resulting modeled periods were only modestly affected by different
idealizations of the soil-foundation interface.
Modeling Configurations for the Sherman Oaks Building
Fundamental Period
(sec)
Model Longitudinal Transverse
MB (Baseline Model) 2.67 2.72
MB.1 (rigid subterranean structure) 2.35 2.68
MB.2 (no tension in foundation springs) 2.65 2.73
Model 4 (bathtub) 2.67 2.72
Model 3 (fixed horizontal springs) 2.34 2.65
Model 2 (fixed at foundation) 2.67 2.71
Model 1 (fixed at grade) 2.34 2.67
Comparisons of computed displacement histories, maximum displacement, drift

ratios, story shear coefficients, and peak floor accelerations between Model MB and
the other models are shown in Figures 7-11 through 7-15. A summary of peak
response quantities from all foundation modeling configurations is shown in Figure
7-16.
Model MB.1 and Model MB.2. Results for Model MB.1 in Figure 7-11 showed that
rigid subterranean structural elements caused an increase structural response,
particularly in the NS direction, likely due to the change in period. In the case of
Model MB.2, results in Figure 7-11 showed that allowing geometric nonlinearities
(i.e., gap formation) had no discernible impact on response. This suggests that gaps
would not be expected to form in the foundation springs between the soil and the
basement walls, and supports the equivalent-linear soil-foundation modeling
assumptions suggested in Section 7.1.3.
Model 4. Results for Model 4 in Figure 7-12 showed that the bathtub model
introduces negligible changes in displacement response over the height of the
structure (i.e., less than 3.6% different at the roof level). Changes in story drift and
story shear force profiles were also negligible. Peak floor accelerations were most
sensitive to the change in modeling configuration. Values above the ground level
were relatively unaffected, but values in the subterranean levels were sensitive to the

use of free field motions (results marked ‘4b’) versus foundation input motions
(results marked ‘4a’).
Model 3. Results for Model 3 in Figure 7-13 showed the least agreement with Model
MB. Fixing the upper-level foundation springs against displacement, and applying
input motions to the base slab, caused large differences in all response quantities
including building vibration periods, displacement histories, drift ratios, and story
shears. Given the significant discrepancies observed, use of this modeling
configuration is not recommended.
Model 2. Results for Model 2 in Figure 7-14 showed that modeling the subterranean
levels, even while ignoring the effects of the surrounding soil, can provide good
results for some response quantities. Model 2 exhibited good agreement for building
vibration periods and displacement histories. Reasonable agreement was observed
for maximum displacement and drift ratios, but story shears and peak floor
accelerations differed more significantly, particularly in the subterranean levels.
Model 1. Results for Model 1 indicated that ignoring the subterranean levels
significantly alters the period of vibration. As a result, displacement histories were
more out-of-phase than most other modeling configurations, as shown in Figure 7-15.
Differences in story drifts, story shears, and peak floor accelerations were relatively
large (up to 50% different) in some cases.
In Figure 7-16, peak displacement, drift, and story shear response quantities from all
modeling configurations were synthesized and plotted in a single figure. Results for
Model 3 are clear outliers for each of the parameters considered. Results for Model 4
are closest to Model MB, followed by Model 2, and then Model 1. Differences in
response quantities, when they occurred, were generally greater in the subterranean
levels than in the levels above grade.

between Model MB, MB.1 and MB.2 for the Sherman Oaks building.

between Model MB and Model 4 for the Sherman Oaks building.




Figure 7-16 Comparison of peak displacements, drift ratios, and story shears
from all model configurations, in the transverse direction.
7.3 Sherman Oaks Building Parametric Studies
Analysis of the Sherman Oaks building showed that key building response quantities
can be affected by the foundation modeling assumptions, even when the structure-to-
soil stiffness ratio, h/(VsT), does not indicate a strong dependence on SSI effects.
Additional parametric studies were performed to investigate how different structural
system characteristics might influence the magnitude of SSI effects on modeled
response. A series of simplified stick models of the Sherman Oaks building were
developed, and the resulting response quantities for each model were compared.

The key parameter varied between the models was structural stiffness. The moment
frame system of the Sherman Oaks building was idealized as a stick model, and the
stiffness scaled to different factors representing the stiffness of a series of core shear
wall and perimeter shear wall variations in the structural system. The parametric
stick models of the Sherman Oaks building are summarized in Table 7-7.
Table 7-7 Summary of Sherman Oaks Building Parametric Stick Models
Y-Direction X-Direction
(N-S) (2) (E-W) (3)
Stiffness Modeled
Scale Fdn. Model Height Ty Tx
System Factor Model ID (m) (1) (sec) h/(VsT) (sec) h/(VsT)
Moment
1 1 1-MF 50 2.83 0.05 2.61 0.05
Frame
2 2-MF 56.25 2.86 0.05 2.62 0.05
4a 4a-MF 56.25 2.87 0.05 2.63 0.05
MB MB-MF 56.25 2.87 0.05 2.62 0.05
Shear Wall
5 1 1-SW 50 1.32 0.10 1.17 0.10
Core
2 2-SW 56.25 1.39 0.11 1.20 0.11
4a 4a-SW 56.25 1.40 0.11 1.20 0.11
MB MB-SW 56.25 1.40 0.11 1.20 0.11
Perimeter
25 1 1-SW1 50 0.73 0.19 0.56 0.21
Shear Wall
2 2-SW1 56.25 0.84 0.18 0.63 0.21
4a 4a-SW1 56.25 0.88 0.18 0.62 0.22
MB MB-SW1 56.25 0.87 0.18 0.62 0.22
Perimeter
50 1 1-SW2 50 0.63 0.22 0.42 0.29
Shear Wall
2 2-SW2 56.25 0.75 0.21 0.52 0.26
4a 4a-SW2 56.25 0.79 0.20 0.51 0.26
MB MB-SW2 56.25 0.78 0.20 0.51 0.26
Perimeter
125 1 1-SW3 50 0.55 0.25 0.32 0.38
Shear Wall
2 2-SW3 56.25 0.69 0.22 0.44 0.31
4a 4a-SW3 56.25 0.73 0.21 0.43 0.31
MB MB-SW3 56.25 0.72 0.21 0.43 0.31
Notes: (1) h= 2/3 x Modeled Height
(2) Vs =243.0 m/s in the N-S direction (average effective profile velocity for xx-rotation)
(3) Vs =277.7 m/s in the E-W direction (average effective profile velocity for yy-rotation)

Each system was investigated considering the foundation modeling configurations
associated with Model 1, Model 2, Model 4, and Model MB. Based on results from
the Sherman Oaks building analyses, Model 3 was not considered further. As shown
in Table 7-7, the change in structural stiffness affects the modeled period in each
direction, and correspondingly changes the resulting structure-to-soil stiffness ratio,
h/(VsT). As the stiffness increases, the ratio eventually exceeds 0.1, and the expected
impact of SSI effects becomes more significant.
An idealized stick model for the Sherman Oaks building is shown in Figure 7-17.
Although the superstructure has been simplified into a single equivalent stick, each
model includes detailed modeling at the foundation level, essentially maintaining
foundation geometry, and spring and dashpot configurations. Details on stick model
development and calibration with the full-building model are provided in
Appendix B.
Figure 7-17 Elevation an idealized stick model of the Sherman Oaks building.
7.3.1 Parametric Study Results
Detailed analytical results for all parametric stick models are provided in
Appendix B. Results are compared between the moment frame stick model and the

full-building model to demonstrate the level of consistency in model predictions
using the idealized stick configuration. Results for key response quantities including
maximum displacement, story drift, story shear, and peak floor acceleration are
compared for each structural system variant across building all foundation modeling
configurations to document the change in relative SSI effects observed as the
structural system parameters were changed.
Table 7-8 summarizes relative comparisons between the results for each response
quantity for the moment frame (MF), core shear wall (SW), and perimeter shear wall
(SW3) models. Information is based on plots provided in Appendix B. An
explanation for the nomenclature used in the table is provided in the footnotes.
Table 7-8 Comparison of Results for Moment Frame, Core Shear Wall, and Perimeter Shear Wall
Stick Models and Alternative Foundation Modeling Configurations
Moment Core Shear Perimeter Shear Wall
Frame (MF) Wall (SW) (SW3)
Response X-Direct. Y-Direct. X-Direct. Y-Direct. X-Direct. Y-Direct.
Quantity Location (E-W) (N-S) (E-W) (N-S) (E-W) (N-S)
Period --- 1 < 2, 4a, MB Similar 1 < 2, 4a, MB 1 < 2, 4a, MB 1 < 2 < 4a, MB 1 < 2, 4a, MB
Displacement Superstructure 4a, MB <1, 2 4a, MB <2<1 1<< 4a, MB <2 1<<2<4a, MB 1< 4a, MB <2 1<<2, 4a < MB
Basement 4a, MB <1, 2 Similar 4a,MB < 2 4a,MB < 2 4a,MB < 2 4a,MB < 2
Story Drift Superstructure 2, 4a, MB <1 4a<MB <2<1 Varies 1 < 2, 4a, MB Similar 1, 2< 4a < MB
Basement 2,4a < MB 4a<2<< MB 4a<2< MB 4a<2<< MB 4a<2, MB 4a<2<< MB
Story Shear Superstructure 2, 4a, MB <1 2, 4a, MB <1 4a,MB <2<1 1< 4a, MB <2 4a< MB< 2<<1 2< 4a< MB<<1
Basement 4a<<2 < MB 4a<<2 < MB 4a<<2<< MB 4a<<2<< MB 4a<2<< MB 4a<<2<< MB
Peak Floor
Superstructure Similar 4a< 1< MB< 2 4a < MB <1, 2 4a < MB <1, 2 4a < MB <1< 2 2, 4a, MB <<1
Acceleration
Basement 4a, MB<<2 4a, MB<<2 4a, MB < 2 4a, MB < 2 4a, MB<<2 4a, MB<<2
Notes: (1) 1, 2, 4a, and MB refer to Model 1, Model 2, Model 4a, and Model MB foundation modeling configurations.
(2) “1<<2, 4a < MB” indicates, for example, that the response of Model 1 is less than the response of Models 2 and 4a; the
response of Models 2 and 4a are similar; and the response of all models is less than the response of Model MB.
(3) “<<” indicates a comparatively larger difference in response than “<”.
(4) “Similar” indicates that the results do not vary significantly across the model types.
(5) “Varies” indicates that the trend in the results is not clear; some models have larger values in some cases and smaller
values in other cases.
Figures 7-18 through 7-21 summarize maximum displacement envelopes

comparing results across different foundation modeling configurations for all five
building types (MF, SW, SW1, SW2, and SW3). Full scale plots are provided in
Appendix B. Compared in this way, the trends are clearly discernible in which
SSI effects are larger as the building stiffness increases. Response quantities in
the basement levels were consistently the most influenced by alternative
foundation modeling conditions.

Northridge, X-Direction Northridge, Y-Direction
14 14
12 12
10 10
Floor Number
Floor Number
8 8
6 6
4 4
2 1-MF-Stick 2 1-MF-Stick
2-MF-Stick 2-MF-Stick
0 4a-MF-Stick 0 4a-MF-Stick
MB-MF-Stick MB-MF-Stick
-2 -2
-40 -30 -20 -10 0 10 20 30 40 -40 -30 -20 -10 0 10 20 30 40
Max. Displacement [cm] Max. Displacement [cm]
14 14
12 12
10 10
Floor Number
Floor Number
8 8
6 6
4 4
2 1-SW -Stick 2 1-SW -Stick

2-SW -Stick 2-SW -Stick
0 4a-SW -Stick 0 4a-SW -Stick
MB-SW -Stick MB-SW -Stick
-2 -2
-40 -30 -20 -10 0 10 20 30 40 -40 -30 -20 -10 0 10 20 30 40
14 14
12 12
10 10
Floor Number
Floor Number
8 8
6 6
4 4
2 1-SW 1-Stick 2 1-SW 1-Stick

2-SW 1-Stick 2-SW 1-Stick
0 4a-SW 1-Stick 0 4a-SW 1-Stick
MB-SW 1-Stick MB-SW 1-Stick
-2 -2
-40 -30 -20 -10 0 10 20 30 40 -40 -30 -20 -10 0 10 20 30 40
14 14
12 12
10 10
Floor Number
Floor Number
8 8
6 6
4 4

-2 -2
-40 -30 -20 -10 0 10 20 30 40 -40 -30 -20 -10 0 10 20 30 40
14 14
12 12
10 10
Floor Number
Floor Number
8 8
6 6
4 4

-2 -2
-40 -30 -20 -10 0 10 20 30 40 -40 -30 -20 -10 0 10 20 30 40
Figure 7-18 Comparison of maximum displacement envelopes in each direction for foundation Models
1, 2, 4a, and MB, for each structural system variant (MF, SW, SW1, SW2, and SW3).

Northridge, Interstory Drifts, X-Direction Northridge, Interstory Drifts, Y-Direction
14 14
1-MF-Stick
12 12
2-MF-Stick
10 4a-MF-Stick 10
MB-MF-Stick
Story Number
Story Number
8 8
6 6
4 4
2 2 1-MF-Stick
2-MF-Stick
0 0 4a-MF-Stick
MB-MF-Stick
-2 -2
-0.015 -0.01 -0.005 0 0.005 0.01 0.015 -0.015 -0.01 -0.005 0 0.005 0.01 0.015
Max. Interstory Drifts [cm/cm] Max. Interstory Drifts [cm/cm]
14 14
12 12
10 10
Story Number
Story Number
8 8
6 6
4 4

-2 -2
-0.015 -0.01 -0.005 0 0.005 0.01 0.015 -0.015 -0.01 -0.005 0 0.005 0.01 0.015
14 14
12 12
10 10
Story Number
Story Number
8 8
6 6
4 4

-2 -2
-0.015 -0.01 -0.005 0 0.005 0.01 0.015 -0.015 -0.01 -0.005 0 0.005 0.01 0.015
14 14
12 12
10 10
Story Number
Story Number
8 8
6 6
4 4

-2 -2
-0.015 -0.01 -0.005 0 0.005 0.01 0.015 -0.015 -0.01 -0.005 0 0.005 0.01 0.015
14 14
12 12
10 10
Story Number
Story Number
8 8
6 6
4 4

-2 -2
-0.015 -0.01 -0.005 0 0.005 0.01 0.015 -0.015 -0.01 -0.005 0 0.005 0.01 0.015
Figure 7-19 Comparison of maximum story drift envelopes in each direction for foundation Models 1, 2,
4a, and MB, for each structural system variant (MF, SW, SW1, SW2, and SW3).

14 14
1-MF-Stick
12 12
2-MF-Stick
10 10 4a-MF-Stick
Floor Number
MB-MF-Stick
Floor Number
8 8
6
6
4
4
1-MF-Stick 2
2 2-MF-Stick
0
0 4a-MF-Stick
MB-MF-Stick -2
-2 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Normalized Story Shear
Normalized Story Shear
14 14
1-SW -Stick
12 12
2-SW -Stick
10 10 4a-SW -Stick
Floor Number
Floor Number
MB-SW -Stick
8 8
6 6
4 4
1-SW -Stick
2 2-SW -Stick 2
4a-SW -Stick
0 0
MB-SW -Stick
-2 -2
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Normalized Story Shear Normalized Story Shear
14 14
12 12 1-SW 1-Stick
2-SW 1-Stick
10 10
4a-SW 1-Stick
Floor Number
Floor Number
8 8 MB-SW 1-Stick
6 6
4 4
1-SW 1-Stick
2 2-SW 1-Stick 2
4a-SW 1-Stick
0 0
MB-SW 1-Stick
-2 -2
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
14 14
1-SW 2-Stick
12 12 2-SW 2-Stick
10 10 4a-SW 2-Stick
Floor Number
Floor Number
MB-SW 2-Stick
8 8
6 6
4 4
1-SW 2-Stick
2 2-SW 2-Stick 2
4a-SW 2-Stick
0 0
MB-SW 2-Stick
-2 -2
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
14 14
1-SW 3-Stick
12 12 2-SW 3-Stick
10 10 4a-SW 3-Stick
Floor Number
Floor Number
MB-SW 3-Stick
8 8
6 6
4 1-SW 3-Stick 4
2 2-SW 3-Stick 2
4a-SW 3-Stick
0 0
MB-SW 3-Stick
-2 -2
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Figure 7-20 Comparison of maximum story shear envelopes in each direction for foundation Models

14 14
12 12
10 10
Floor Number
Floor Number
8 8
6 6
4 4
2 1-MF-Stick 2 1-MF-Stick
2-MF-Stick 2-MF-Stick
0 4a-MF-Stick 0 4a-MF-Stick
MB-MF-Stick MB-MF-Stick
-2 -2
-1.5 -1 -0.5 0 0.5 1 1.5 -1.5 -1 -0.5 0 0.5 1 1.5
Peak Floor Accel. [g] Peak Floor Accel. [g]
14 14
12 12
10 10
Floor Number
Floor Number
8 8
6 6
4 4

-2 -2
-1.5 -1 -0.5 0 0.5 1 1.5 -1.5 -1 -0.5 0 0.5 1 1.5
14 14
12 12
10 10
Floor Number
Floor Number
8 8
6 6
4 4

-2 -2
-1.5 -1 -0.5 0 0.5 1 1.5 -1.5 -1 -0.5 0 0.5 1 1.5
14 14
12 12
10 10
Floor Number
Floor Number
8 8
6 6
4 4

-2 -2
-1.5 -1 -0.5 0 0.5 1 1.5 -1.5 -1 -0.5 0 0.5 1 1.5
14 14
12 12
10 10
Floor Number
Floor Number
8 8
6 6
4 4

-2 -2
-1.5 -1 -0.5 0 0.5 1 1.5 -1.5 -1 -0.5 0 0.5 1 1.5
Figure 7-21 Comparison of peak floor acceleration envelopes in each direction for foundation Models

7.3.2 Parametric Study Observations
The following general observations from parametric studies of the Sherman Oaks
building were made:
 Observed SSI effects became larger as the building stiffness increased. They
were relatively minor for the moment frame (MF) building, more discernible for
the core shear wall (SW) building, and readily apparent for the stiffest perimeter
shear wall building (SW3).
 Observed SSI effects correlated well with the structure-to-soil stiffness ratio,
h/(VsT); the larger the ratio, the greater the observed SSI effect.
The following additional observations were made, relative to the specific case of
these building variants and these input motions:
 For all building variants in this study, fundamental periods are past the peak
spectral response and located on the descending (i.e., velocity-controlled) branch
of the response spectrum for the Northridge earthquake. As such, any stiffening
in the structure could be assumed to lead to an increase in response.
 Observed differences in modeled periods between the different foundation
models were relatively small for a given structure. A possible exception
occurred in the case of shear wall buildings, in which Model 1 periods were
noticeably less than the other models.
 Large differences were observed between model results in the basement levels.
Different trends were observed in comparisons between superstructure results
and comparisons between basement results.
 As the superstructure stiffened, model-to-model variations in superstructure
response increased such that they approached the magnitude of observed
variations in basement level response. This was particularly noticeable in the
case of peak floor acceleration.
 In the superstructure, Model 4a overall results were the most similar to Model
MB results for all building types. Model 1 and Model 2 were usually
conservative for story shear and peak floor acceleration for all building types.
Model 1 was observed to result in conservative displacements and drifts in the
moment frame building, but unconservative displacements and drifts in the shear
wall buildings.
 In the basement levels, Model 2 results were higher than Models 4a results across
all response quantities. In the case of displacements and peak floor accelerations,
Model 2 results exceeded Model MB results, but in the case of story drift and
story shear, Model MB often had larger values than Model 2.

 In the case of moment frame structures, Model 1 and Model 2 led to comparable
or conservative results relative to Model MB. Therefore, Model 1 and Model 2
represent reasonable alternatives for practical foundation modeling of moment
frame structures.
 In the case of shear wall structures, Model 4a results were most consistent with
Model MB, and Model 2 results conservatively bounded Model MB results.
Therefore, Model 2 and Model 4a represent reasonable alternatives for practical
foundation modeling of shear wall structures.
 In the case of the SW3 shear wall building, the peak story shear in Model 1 was
1.2 times that of Model MB, and the peak floor acceleration of Model 1 was 1.5
times that of Model MB. Shear wall building designs based on Model 1 would
be “overdesigned” for shears and accelerations relative to response characterized
by Model MB.
7.4 Walnut Creek Building
The Walnut Creek building is a10-story structure with no subterranean levels. The
seismic force-resisting system consists of a reinforced concrete shear wall core with
perimeter precast and cast-in-place concrete frames. Construction drawings of the
Walnut Creek building were made available for inspection through the auspices of
the California Strong Motion Instrumentation Program. The building was designed
in 1970. The plan dimensions of the building are 31.8 m (104 ft) wide by 45.2 m
(148 ft) long. A typical floor plan is shown in Figure 7-22.
Figure 7-22 Walnut Creek building typical floor plan, based on construction
drawings provided by CSMIP.

The gravity system consists of 7 cm (2.75 in) lightweight concrete topping slabs over
7 cm (2.75 in) precast panel slabs supported on precast, prestressed reinforced
concrete beams. A longitudinal elevation of the building is shown in Figure 7-23.
The building measures 39.2 m (128.5 ft) tall from the ground surface to the roof. The
height of the first story is 4.9 m (16 ft), and all other stories are 3.8 m (12.5 ft). Note
that the ground surface is not level at the site, although for simplicity, this detail was
not considered in the structural model. The characteristics of the foundation are
described in Section 7.4.3, below.
Figure 7-23 Walnut Creek building longitudinal elevation, based on construction

drawings provided by CSMIP (Stewart and Stewart, 1997, with
permission).
7.4.1 Site Characterization
A plan view of the Walnut Creek site is shown in Figure 7-24. The ground surface
elevation is approximately 51.8 m (170 ft) above mean sea level. Soil conditions for
the site were obtained from portions of a geotechnical report, prepared by Harding,
Miller, Lawson & Associates (1970), and from a geotechnical report with seismic
refraction investigation of the subsurface soils, prepared by Raney Geotechnical
(1983), in the vicinity of the site. Information from a total of 19 borings ranging in
depth from 1.5 m (5 ft) to 17.4 m (57 ft), and four seismic refraction surveys, were
obtained from within and around the footprint of the Walnut Creek building.
Borings indicated predominantly west-dipping contacts of sandy clays and silts with
variable thicknesses of 0.6 m (2 ft) to 5.5 m (18 ft) overlaying siltstone and sandstone
of the Orinda Formation. The depth of the water table was measured between 6.1 m
(20 ft) and 12.2 m (40 ft) below the ground surface. The average moist unit weight

was taken to be approximately 18.1 kN/m3 (115 lb/ft3). Geophysical data was
developed from the seismic refraction surveys. Shear wave velocities in the soil
layers were estimated based on correlations by Fumal and Tinsley (1985), while the
shear wave velocity in the rock was based on the seismic refraction data presented by
Raney Geotechnical.
Figure 7-24 Plan view of the Walnut Creek site showing borehole and refraction
survey locations used for site characterization (adapted from
Harding, Miller, Lawson & Associates, 1970; Raney Geotechnical,
1983).
Based on available geotechnical information, an interpretation of the soil layering at

the site, in terms of a shear wave velocity profile and material profile, is shown in
Figure 7-25.
7.4.2 Ground Motion Recordings
The Walnut Creek building was instrumented in 1979 by the California Strong
Motion Instrumentation Program and designated CSMIP Station No. 58364.
Instrumentation includes 16 accelerometers at the locations including the ground
floor, third floor, eighth floor, and roof levels. There are two vertical sensors at the
ground level, allowing the base rocking effects to be measured. There are no free-
field instruments in the vicinity of the site.

Figure 7-25 Subsurface characteristics of the Walnut Creek site: (a) shear wave
velocity profile; and (b) material profile (adapted from Harding, Miller,
Lawson & Associates, 1970; Raney Geotechnical, 1983).
Since 1979, five earthquake events have been recorded and processed by CSMIP at
this station. The events, along with peak accelerations for the ground floor and roof
are presented in Table 7-9.
Table 7-9 Summary of Earthquake Events Recorded at the Walnut Creek Building
Ground Floor Roof
Earthquake (g) (g)
N-S E-W V1 V2 N-S E-W V
Livermore 80A 0.030 0.033 0.022 0.023 0.116 0.108 ----
Livermore 80B 0.061 0.048 0.024 0.018 0.137 0.192 ----
1989 Loma Prieta 0.102 0.046 0.053 0.047 0.208 0.164 ----
2007 AlumRock 0.005 0.003 NR NR 0.018 0.015 ----
2008 Alamo 0.029 0.018 NR NR 0.057 0.034 ----
Symbols: N-S=north-south; E-W=east-west; V=vertical; V1=Chan.12 (west); V2=Chan.13 (East); and
NR=no record available.
Instrumentation locations are shown in Figure 7-26. Horizontal translations recorded

at the ground level of the building were assumed to incorporate base-slab averaging
effects due to kinematic interaction. Because the bottom of the foundation is 4.2 m

(14 ft) below the ground level, it is reasonable to assume that embedment effects
were not captured in the recorded motion. Transfer functions derived using the
embedment procedures in Section 3.2 were used to develop foundation input
motions, uFIM, from the recorded motions.
Figure 7-26 CSMIP Station No. 58364: Walnut Creek 10-Story commercial building,
sensor location sketch (CESMD, 2011).
Since there were no free-field instruments in the vicinity of the site, free-field ground
motions, ug, were obtained from the recorded motions using transfer functions to
remove the base-slab averaging effects. The uFIM and ug response spectra for the
1989 Loma Prieta earthquake are shown in Figure 7-27.

(a): E-W Direction (b): N-S Direction
Figure 7-27 Response spectra for foundation input motion, uFIM, and free-field motion, ug, at the
Sherman Oaks building, 1989 Loma Prieta earthquake.
7.4.3 Foundation Conditions
The foundation is rectangular in plan, measuring 31.8 m (104 ft) wide by 45.2 m
(148 ft) long. The foundation consists of concrete spread footings, drilled shafts, and
mat elements. A foundation plan developed based on information contained in
construction drawings provided by CSMIP is shown in Figure 7-28.
Figure 7-28 Walnut Creek building foundation plan, based on construction

drawings provided by CSMIP.

The core walls extend 3 m (10 feet) below the ground level, supported on mat
foundation that is 1.2 m (4 feet) thick. The plan dimensions of the mat are 11.6 m
(38 ft) by 25.6 m (84 ft), and the bottom of the mat is a total of 4.3 m (14 feet) below
the ground level surface. Soil above the mat foundation was backfilled with concrete
and sand. Perimeter frame elements terminate at the ground level, where they are
supported on spread footings or belled caissons bearing on rock, and interconnected
with grade beams.
7.4.4 Development of Foundation Springs and Dashpots
The methodologies described in Chapter 2 were applied to the development of

foundation springs and dashpots for the Walnut Creek building. Critical input
parameters included the foundation dimensions, fundamental vibration periods of the
structure, soil properties, and the amplitude of the input motion. Modeling of the
Walnut Creek foundation and site conditions included the following:
 Determination of average effective profile velocity, considering foundation
dimensions, overburden pressures from the structure, and nonlinear effects.
 Determination of the structure-to-soil stiffness ratio for estimating the
significance of soil-structure interaction effects.
 Evaluation of the effects of mat foundation flexibility on foundation impedance.
 Calculation of mat foundation horizontal stiffness and damping coefficients,
considering base slab friction and neglecting passive pressure resistance against
the embedded core walls.
 Calculation of vertical stiffness and damping coefficients for springs under the
mat foundation, considering adjustment for rotational impedance for shallow
foundation elements.
 Calculation of springs and dashpots for the footings and drilled shafts supporting
the perimeter frames.
Average Effective Profile Velocity. Foundation stiffness and damping coefficients

were based on an average effective profile velocity calculated over an effective
profile depth, zp. Table 7-10 summarizes the effective profile depth, zp, the depth
range, and the average effective profile velocity calculated using Equation 2-18 for
each foundation vibration mode over the depth range considered. The maximum
depth considered is greater than zp by the amount of the foundation embedment, D
(i.e., the depth range extends to D + zp). For the specific case of the horizontal base
spring, the depth interval begins at the base of the foundation.
Values Vs, avg in Table 7-10 have not been corrected for overburden pressures due to
the added weight of the structure. Before averaging, measured free-field velocities,
Vs, should be increased to account for the presence of structural overburden.

Table 7-10 Summary of Effective Profile Depths and Average Effective Profile Velocities
for the Walnut Creek Building
zp Depth Range Vs,avg
Vibration Mode Basis (1) (m) (m) (m/s)
Horizontal translation (x and y), base spring BL 8.6 4.3 to 12.9 434.3
Vertical Translation (z) BL 8.6 0 to 12.9 415.9
Rocking along x-axis (xx) 4

B3 L 7.1 0 to 11.3 413.5
Rocking along y-axis (yy) 4

BL3 10.5 0 to 14.8 418.2
Notes: (1) Calculated using mat foundation half-width, B=5.8 m; half-length, L=12.8 m; depth D=4.3 m.
structure, v, was evaluated at two-foot intervals below the foundation assuming a
For the Walnut Creek building, the increase in vertical stress due to the weight of the
2V:1H distribution (Holtz et al., 2010). Overburden-corrected shear wave velocities

below the foundation, Vs, F, were calculated using Equation 2-17, and averaged using
Equation 2-18. Values of average effective profile velocity for both the overburden-
corrected and non-overburden-corrected cases are compared in Table 7-11.
Table 7-11 Comparison of Average Effective Profile Velocities with and without
Correction for Structural Overburden Weight
Vs,avg(1) Vs,F,avg(2)
Z (m/s) (m/s)
Horizontal Translation (x and y) 434.3 517.8
Vertical Translation (z) 415.9 463.7
Rocking along x-axis (xx) 413.5 460.8
Rocking along y-axis (yy) 418.2 465.5
Notes: (1) Vs,avg is the non-overburden-corrected average effective profile velocity.
(2) Vs,F,avg is the overburden-corrected average effective profile velocity.
Structure-to-soil stiffness ratio. Taking the transverse (E-W) direction as critical,

the structure-to-soil stiffness ratio, h/(VsT), for the Walnut Creek building is 0.1,
calculated using two-thirds of the modeled building height from foundation to roof,
h = 2/3(43.5 m) = 29.0 m, the average effective profile velocity for rocking about the
xx-axis, corrected for structural overburden, Vs,F,avg = 460.8 m/s (from Table 7-11),
and the fundamental period for the fixed base condition in the transverse direction,
T = 0.66 sec (from system identification, Stewart and Stewart, 1997). Because this
ratio is on the order of 0.1, inertial SSI effects (i.e., period lengthening, foundation
damping) would be expected to be significant.
Mat Foundation Flexibility. The Walnut Creek building has a central core
consisting of shear walls supported by an underlying reinforced concrete mat
foundation that is larger in plan and extends beyond the dimensions of the core.

Extension of the mat foundation beyond the core shear walls introduces flexibility
into the structural foundation elements. As discussed in Section 2.2.3, flexibility in
structural foundation elements can reduce stiffness and radiation damping in the
rocking mode (Iguchi and Luco, 1982).
To evaluate the effects of this flexibility on the resulting foundation impedance, plots
similar to Figure 2-7 (from Iguchi and Luco, 1982) were utilized with a ratio of core
to foundation radius equal to 0.75. Results are shown in Table 7-12, which indicate
that the foundation stiffness ratios are close to one, meaning there is little deviation
from the rigid foundation case. Therefore, the full mat dimensions of 11.6 m (38 ft)
by 25.6 m (84 ft) were used to calculate foundation impedance below the core walls
in the Walnut Creek building.
Table 7-12 Evaluation of Soil-to-Foundation Stiffness Ratios for Flexible Mat Foundation
below Shear Wall Core in the Walnut Creek Building
 (1)
Ratio
Basis for Core Mat rc k (2)
Equivalent Radius Radius
Vibration Mode Radius, rf rc (m) rf (m) rf (Eq. 2-19) k  0
Vertical Translation (z) Af /  7.0 9.7 0.72 100 0.9
Rocking along x-axis (xx) 4 4I f /  9.3 12.0 0.78 188 0.8
Rocking along y-axis (yy) 4 4I f /  5.5 8.1 0.68 59 0.8
Assessed based on Iguchi and Luco (1982), using rc/rf =0.75 and  =100.
Notes: (1) Calculated using Ef =22,894.7 MPa; f =0.2, tf =1.22 m, a0 =0.13, and G based on Vs,F,avg from Table 7-11.
(2)
Mat Foundation Horizontal Stiffness and Damping. Calculations for horizontal

stiffness and damping ratios are shown in Table 7-13. Horizontal stiffness
calculations were based on contributions from base friction only, and passive
pressure resistance along the embedded portion of the core walls was neglected.
Because the base shear reaction mobilizes only soil below the embedment depth D,
the effective profile velocity for the horizontal base spring was taken from D to D+zp
(shown in Table 7-10).
density, s, of 18.1 kN/m3/g, and values of average effective profile velocity
In Table 7-13, shear modulus, G, was evaluated from Equation 2-9, using a soil mass
determined using overburden-corrected shear wave velocities below the foundation

(shown in Table 7-11). Values of shear modulus should be reduced to account for
large strain effects associated with nonlinear behavior. Using Table 2-1, assuming
Site Class C, and peak acceleration (at the foundation level) of about 0.1 g from the
1989 Loma Prieta earthquake, the shear modulus reduction factor, G/G0 = 0.95.

Table 7-13 Calculation of Mat Foundation Stiffness and Damping Parameters for the Walnut Creek Building
G(1) Ksur (2)  a0(3)  sur emb
Eq. 2-9 Table 2-2a Table Eq. Table Table Table Dynamic Stiffness Dashpot Coefficient
Spring/Dashpot (MPa) (kN/m; kN-m/rad) 2-2b 2-15 2-3a 2-3a 2-3b Eq. 2-14a Eq. 2-13a
   s 
k x ,base   x K x , sur cx,base  2k x ,base  sur 
k , cx ,base    
x-direction 469.4 2.24e7 --- 0.11 1 0.057 ---
x ,base
   s 
k y ,base   y K y , sur cy ,base  2k y ,base  sur
k , c y ,base 

  
y-direction 469.4 2.40e7 --- 0.11 1 0.053 ---
y ,base
   s 
k z   z K z , sur z cz  2kz  emb
 kz , cz  
Vertical, z
  
376.4 2.35e7 1.22 0.12 1 --- 0.111
   s 
k xx   xx K xx , sur xx cxx  2kxx  emb
 kxx , cxx  
  
371.6 8.49e8 1.70 0.12 1 --- 0.005
   s 
k , c yy 
k yy   yy K yy , sur yy cyy  2k yy  emb 
  
379.3 2.78e9 1.54 0.12 0.99 --- 0.002
yy
using overall foundation half-width, B=5.8 m; half-length, L=12.8 m; and =0.33.

Notes: (1) Calculated using a shear modulus reduction factor, G/G0=0.95 for the Loma Prieta earthquake; for other events G/G0=1.0.
(2) Calculated
(3) Calculated at a frequency corresponding to the first-mode period of the flexible-base structure (from system identification, Stewart and Stewart, 1997).

A modulus reduction factor of 1.0 was used for the other smaller events. Other
parameters were determined using the equations referenced in the table.
Dashpot coefficients were determined using Equation 2-13a and the total foundation
damping ratio. The total foundation damping ratio is equal to the sum of the
hysteretic damping ratio, s. For the Walnut Creek site, the soil hysteretic damping
radiation damping ratio (determined using equations in Chapter 2) and the soil
ratio was evaluated using Menq (2003), and determined to be 0.01.
Mat Foundation Vertical Stiffness and Damping. Calculation of overall vertical

and rotational stiffness and damping ratios is shown in Table 7-13, using values of
average effective profile velocity based on overburden-corrected shear wave
velocities (shown in Table 7-11). Vertical springs were distributed over the footprint
of the foundation, as shown in Figure 7-29, using a vertical stiffness intensity that is
normalized by area (Equation 2-20). Stiffness intensities were adjusted near the
edges of the foundation to match the overall rocking stiffness values given in Table
7-13 (using Equation 2-21 and an end length ratio of Re = 0.4). The stiffness of any
individual spring in the model was then computed as the product of stiffness intensity
and the tributary area for the spring.
k zi  kN/m3  czi  kN-sec/m3 

96 275
308 778
418 1,235
526 1,697
Figure 7-29 Vertical spring and dashpot intensities distributed over the mat
foundation of the Walnut Creek building. Solid lines represent
tributary area boundaries.
Vertical stiffness intensities in the central zone of the mat foundation were computed
using Equation 2-20a. Edge intensities were increased by factors of Rk = 5.48 for the
xx-direction and Rk = 3.21 for the yy-direction to correct for underestimation of
rotational stiffness (Equations 2-21a, 2-21b). Corner intensities were evaluated as the

average of the intensities in the xx- and yy-directions. Dashpot intensities were
calculated based on stiffness intensities using Equation 2-20b, and reduced by Rc
(Equations 2-21c, 2-21d), to correct for overestimation of rotational damping.
Footing and Drilled Shaft Stiffness and Damping. Individual springs and dashpots
were calculated for horizontal and vertical translation modes for the spread footings
and drilled shafts supporting the columns of the perimeter frames. For simplicity,
spread footings and drilled shafts were treated as individual shallow square
foundations with dimensions of 3.05 m (10 ft) by 3.05 m (10 ft). Calculations were
similar to those presented for the mat foundation, using the same soil parameters.
For reference, calculated values were 3.91 x 106 kN/m for horizontal springs (kx and
ky); 4.97 x 106 kN/m for vertical springs (kz); 1.68 x 103 kN-s/m for horizontal
dashpots (cx and cy); and 2.75 x 104 kN-s/m for vertical dashpots (cz).
7.4.5 Analysis Results
The Walnut Creek building was analyzed in OpenSees (McKenna, 1997),

incorporating foundation springs and dashpots described above. Details for the
development and validation of the model are described in Appendix C. To expedite
analyses and post-processing of results, the three-dimensional model was collapsed
into a simplified two-dimensional model. Adjustments to the shear and flexural
stiffness for cracked concrete were used to produce a reasonably close match to the
periods and displacement histories in recordings from the 1989 Loma Prieta
earthquake. Based on validation studies, the model was considered a reasonable
engineering approximation for the Walnut Creek building.
Because the Walnut Creek core wall foundation has limited embedment, and there
are no basement levels, multi-support excitation along embedded portion of the
building was not considered. As a result, Model MB converges to Model 4 (bathtub
model). Using this model as a basis, alternative modeling configurations were
developed and studied, and the resulting response quantities were compared for the
following variants:
 Model 2, which included explicit modeling of the subterranean foundation
elements, assuming a fixed base at the foundation level and omitting the
surrounding soil; and
 Model 1, which ignored the response of the subterranean foundation elements by
assuming a rigid base at the ground level.
Implementation of Model 4 (bathtub model), Model 2, and Model 1 for the Walnut
Creek building is illustrated in Figure 7-30. Because of the minimal embedment of
the Walnut Creek building structure, Model 1 and Model 2 are nearly identical.

Figure 7-30 Schematic illustration of Model 4 (bathtub); Model 2 (fixed at
foundation level); and Model 1 (fixed at ground surface) for the
Walnut Creek building.
Fundamental periods of vibration for each of the model variants are shown in Table
7-14. The resulting modeled periods were only modestly affected by different
idealizations of the soil-foundation interface (i.e., Model 1 and Model 2 have the
same fundamental period; Model 4 is slightly more flexible with a longer
fundamental period).
Modeling Configurations for the Walnut Creek Building
Fundamental Period
(sec)
Model E-W Direction
Model 4 (bathtub) 0.83
Model 2 (fixed at foundation) 0.78
Model 1 (fixed at grade) 0.78

Comparisons of computed acceleration and displacement histories, maximum
displacement, drift ratios, story shear coefficients, and peak floor accelerations
between Model 4 (taken as Baseline Model MB), Model 2, and Model 1 are shown in
Figure 7-31 and Figure 7-32.
In Figure 7-31, the fixed-base models (Model 1 and Model 2) capture roof
acceleration and displacement histories reasonably well, but the phasing was quite
different. Observed differences in phasing are likely due to differences in higher
mode response.
Figure 7-31 Comparison of roof acceleration and displacement histories for

Model 4 (taken as Baseline Model MB), Model 1, and Model 2 for the
Walnut Creek building.
In Figure 7-32, there is a general trend towards over-prediction of response by the

fixed-base models (Model 1 and Model 2) relative to Model 4 (taken as Baseline
Model MB). Observed differences were largest in peak floor acceleration response,
and are likely the result of differences in higher-mode effects.

Figure 7-32 Comparison of maximum relative displacement, peak floor acceleration,
maximum story drift, and story shear coefficients for Model 4 (taken as Baseline
Model MB), Model 1, and Model 2 for the Walnut Creek building.
The presence of two vertical instruments at the ground level allowed for the
evaluation of base rotation, and computation of lateral roof displacement due to base
rocking. High-pass filtering of the base rocking data, which was needed to produce a
physically meaningful result, is described in Appendix C. Comparison of modeled
versus recorded roof displacement histories due to base rotation are shown in Figure
7-33.
Figure 7-33 Comparison of modeled versus recorded roof displacement histories

due to base rotation of the Walnut Creek building.

In the figure, the numerical model captured the maximum amplitude and approximate
frequency content of the observed rocking data. The phasing was reasonably well
predicted, although there were observed differences in amplitude over the length of
the record that can be attributed to discrepancies between predicted and observed
structural responses in upper levels due to standard modeling uncertainty. These
results showed that, in general, the foundation impedance functions for the Walnut
Creek building had about the correct level of flexibility.
7.5 Example Applications Summary and Conclusions
Based upon previous findings from Naeim et al. (2008), Tileylioglu et al. (2010), and
observations from example applications and parametric studies herein, the following
conclusions can be made:
 MB-type models have an encouraging ability to match observed building
responses from recorded motions with modest tuning of structural parameters
(e.g., damping ratios, building masses, or element stiffnesses). These results
suggest that the relatively simple equivalent-linear spring and dashpot approach
for SSI modeling presented in Chapter 2 can provide a satisfactory representation
of foundation impedance.
 Of the model types studied, Model 4, including foundation springs and dashpots
within a rigid bathtub element, provided the best match to MB-type models, and
is a reasonable and practical simplification to variable support excitation.
Model 2, which omits foundation springs, but explicitly includes modeling of the
subterranean structure, was the next best alternative. Model 1 results were
observed to vary significantly from MB-type model results, but generally
provided a conservative estimate of force-based demands for design use.
Model 3 results were highly variable, and use of Model 3 approaches to
foundation modeling is not recommended.
 Above-ground building responses, as measured by envelopes of peak response
parameters such as displacement, drift, story shear, and floor acceleration, were
only modestly affected by soil-structure interaction effects in the buildings
studied. The effects became more significant as the stiffness of the
superstructure increased. In contrast, below-ground responses were very
sensitive to soil-structure interaction effects and foundation modeling
assumptions (i.e., kinematic ground motion descriptions and spring/dashpot
configurations). Sensitivity of below-ground response was observed for all
structures, across the full parametric range of stiffnesses studied.

Chapter 8
Conclusions and Recommendations
In this chapter, conclusions and recommendations from prior chapters are distilled
into specific, step-by-step procedures to guide the implementation of soil-structure
interaction modeling in a design setting. Section 8.1 answers the three questions
posed in Chapter 1, which are intended to help guide decisions regarding when SSI is
likely to be important and what geotechnical information is needed to model SSI
effects. Section 8.2 provides guidance on implementation of inertial and kinematic
interaction effects. Section 8.3 describes further study and future research needed to
improve the state of knowledge for soil-structure interaction.
8.1 When is Consideration of Soil-Structure Interaction Important?
In Chapter 1, three questions were posed regarding implementation of SSI. In this

section, those questions are answered, along with references to more detailed
information.
Question 1: When is the use of foundation springs and dashpots important, and
which structural response parameters are affected?
The structure-to-soil stiffness ratio, h/(VsT), can be used as a relative measure for
determining when SSI effects will become significant. In this expression, h is the
structure height, Vs is the soil shear wave velocity; and T is the fixed-base building
period. In applying the structure-to-soil stiffness ratio, values of h, Vs, and T for a
given soil-foundation-structure system are not unique. The following values should
be used:
 Height. Height, h, is the effective height to the center of mass for the first mode
shape, taken as approximately two-thirds of the modeled building height.
 Shear wave velocity. Shear wave velocity should be taken as the average
effective profile velocity, Vs, avg, calculated based on overburden-corrected shear
wave velocities below the foundation, Vs, F (z). Guidance on calculating average
effective profile velocities is provided in Section 2.2.2 and illustrated in
Chapter 7.
 Period. Period should be taken as the best estimate value of the fixed-based
building period in the direction under consideration. The structure-to-soil
stiffness ratio should be evaluated separately in each direction.
GCR 12-917-21 8: Conclusions and Recommendations 8-1

When h/(VsT) > 0.1, SSI can significantly lengthen the building period and modify
(i.e., generally increase) damping in the system. This will modify the design base
shear (up or down, depending on spectral shape) and the distribution of force and
deformation demands within the structure, relative to a fixed-base analysis. The use
of springs and dashpots to represent the flexibility and damping at the soil-foundation
interface will be most significant for stiff structural systems such as shear walls and
braced frames.
When using the structure-to-soil stiffness ratio, it is important to recognize that the
ratio is an approximate relative measure, and not an absolute criterion. Even when
h/(VsT) < 0.1, relative distributions of moments and shear forces in a building can be
modified relative to the fixed-base condition, particularly in dual systems, structures
with significant higher-mode responses, and subterranean levels of structures.
Additional information is provided in Chapter 2, and SSI effects on other building
response quantities are investigated in Chapter 7.
Question 2: Under what conditions is consideration of the differences between

foundation input motions and free-field ground motions important?
Foundation motions differ from free-field ground motions because of kinematic

interaction effects resulting from the presence of stiff foundation elements on or in
the soil. Kinematic interaction includes base-slab averaging and embedment effects
in which the stiffness and strength of the foundation elements cause averaging or
reductions in localized maxima of spatially variable motions that would have
otherwise occurred within the footprint of the foundation.
Base-slab averaging effects become important within the period range of engineering
interest for foundation sizes (measured as an equivalent foundation half width) of
about 20 m (66 ft) or larger for typical soil or weathered rock sites in California.
Embedment effects are sensitive to the depth of embedment, and typically become
important when a structure has two or more subterranean levels. Both base-slab
averaging and embedment effects principally impact short period spectral ordinates
(at periods less than approximately 1.0 sec). Base-slab averaging and embedment
effects are introduced in Chapter 3. Implementation in standards and guidelines is
described in Chapter 4.
Inertial interaction effects are not a significant contributor to differences between

foundation motions and free-field motions, except in a narrow frequency range
centered on the fundamental frequency of the SSI system.
8-2 8: Conclusions and Recommendations GCR 12-917-21

Question 3: What field and laboratory investigations are necessary to develop
foundation springs and dashpots for SSI analysis?
When foundation springs and dashpots are to be implemented in an analysis, it is

necessary to have an engineering characterization of the soil stratigraphy (i.e., soil
types, layer thicknesses, depth to groundwater, depth to rock), seismic shear wave
velocities, and appropriate shear strength parameters for soil materials in the vicinity
of the foundation. In general, this type of information is routinely developed in
geotechnical investigations, so the scope of site investigations to support SSI analysis
is not significantly different than most typical investigations. The interpretation of
soil profile data for the development of foundation springs and dashpots is described
in Chapter 2. Checklists guiding the collection of geotechnical information for SSI
analyses are provided in Chapter 6, and examples of the use of this data are provided
in Chapter 7.
8.2 Summary of Soil-Structure Interaction Analysis Procedures
In this section, procedures for developing springs and dashpots, modifying ground
motions due kinematic interaction, and including soil-structure interaction in
response history analyses are summarized, along with references to more detailed
information.
8.2.1 Developing Springs and Dashpots
The steps for developing springs and dashpots in SSI analyses are summarized as
follows:
1. Develop the required input parameters for analysis:
a. Geotechnical and shear wave velocity profiles. Examples of the type of
information that is required are given in Figure 7-4 and Figure 7-25.
Uncertainty in the shear wave velocity profile should be considered on the
basis of scatter in data from measurements at the site.
b. Shear strength parameters and their variation with depth. Below the ground
water table, undrained strength parameters are required. Drained strength
parameters are generally acceptable above the ground water table.
Variability in shear strengths should be considered based on the range
observed in material-specific testing.
c. Poisson’s ratio, (. Can generally be taken as 0.3 for sands and 0.45 for
clays.
d. Soil hysteretic damping ratio, (s). Strain-dependent soil damping can be
measured using site-specific dynamic material testing, but for most projects
can be taken from existing empirical relationships. Several such

relationships are summarized in Kramer (1996). More recent relationships
are presented by Darendeli (2001) and Menq (2003). These relations are for
damping as a function of shear strain. For most design applications in the
Western United States, strain levels between about 0.1% to 0.5% would be a
reasonable, first-order estimate.
e. Foundation information. Required information includes gross dimensions
(e.g., length, width, depth), the foundation types envisioned for the project
(e.g., mats, inter-connected footings, piles), and the degree of connectivity
between foundation elements (e.g., tie beams, floor slabs). This information
should be developed in cooperation with the structural engineer.
f. Building period and mass. Fundamental mode periods are needed for the
two primary axes of the building. The building mass should include dead
load and the live load expected to be present during earthquake shaking.
Approximate estimates of period and mass are generally sufficient and
should be obtained from the structural engineer.
g. NEHRP site class and ground motion level. Site class can be developed from
the velocity profile using the average shear wave velocity in the upper 30 m
(Vs30). The primary ground motion parameter for the development of soil
springs is peak ground acceleration (PGA) in the free-field.
2. Evaluate the average effective profile velocity for calculation of foundation
stiffness values. For shallow foundations, proceed as follows:
a. For heavily loaded foundations, the overburden pressure due to the weight of
the structure will modify shear wave velocities relative to values measured in
the free-field. A correction for overburden effects can be made using
Equation 2-17.
b. Calculate the effective foundation sizes B eA and B eI using Equations 2-18.
c. Calculate the average effective profile velocity, Vs, avg, over profile depths
zp = B eA for horizontal springs and zp = B eI for rotational springs. Details are
provided in Section 2.2.2. Example applications are provided in Section
7.2.4 and Section 7.4.4.
d. Apply an appropriate shear wave velocity (or shear modulus) reduction to
account for large strain effects using Table 2-1.
e. Repeat as needed to develop alternative values of effective velocity that
account for uncertainty in the soil profile.
3. Calculate the static stiffness of the foundation for each vibration mode, j.
Vibration modes are designated as: horizontal (x and y), vertical (z), and
rotational (xx and yy).

a. Calculate the stiffness of a shallow foundation for a surface condition (i.e.,
not including embedment effects), Kj,sur, using the equations in Table 2-2a.
b. If applicable, calculate the stiffness modifier for embedment effects, j, using
equations in Table 2-2b.
c. The modified stiffness for embedment is taken as the product of Kj,sur and j.
d. Repeat the calculations as needed using the range of effective velocities that
account for uncertainty in the soil profile.
4. Apply dynamic stiffness modifiers, j, and calculate damping ratios, j:
structure (T) as  = 2/T. For three-dimensional analyses, the period can be

a. Calculate the angular frequency corresponding to the first mode period of the
taken as the average of the first mode periods in each orthogonal direction.
b. Calculate the dimensionless frequency, a0, using Equation 2-15.
c. Calculate the dynamic stiffness modifier, j, using equations in Table 2-3a.
d. Calculate radiation damping ratios, j, using equations in Table 2-3a for
surface foundations, or Table 2-3b for embedded foundations.
5. Evaluate frequency-dependent spring and dashpot (impedance) coefficients:
modifier for embedment, j, and the dynamic stiffness modifier, j, using
a. Stiffness, kj, can be taken as the product of the static stiffness, Kj,sur, the
Equation 2-14a.
b. Dashpot coefficient, cj, can be evaluated from radiation damping ratios, j,
using Equation 2-13a.
c. Repeat as needed for the range of static stiffnesses that account for
uncertainty in the soil profile.
6. Evaluate limiting spring forces (i.e., capacities), which depend on the strength of
the foundation soils, as described in Section 2.2.4:
a. For vertical springs, the capacity is the unfactored foundation bearing
capacity.
b. For horizontal springs at the base-slab level, the capacity is derived from
shear-sliding resistance at the soil-foundation interface.
c. For basement walls, the capacity is derived using passive earth pressure
theory.
d. Repeat as needed to account for uncertainty in shear strength parameters.
7. Distribute vertical springs and dashpots around the foundation:

a. Calculate stiffness and damping intensities kzi and czi using Equations 2-20a
and 2-20b.
b. Select a value for the end length ratio, Re, (typically 0.3 to 0.5).
c. Calculate the spring and dashpot amplifier, Rk, and dashpot reduction factor,
Rc, using Equations 2-21a through 2-21d.
d. Apply stiffness intensity Rk kzi and dashpot intensity RkRc czi within the
end regions.
e. Evaluate individual springs and dashpots based on the appropriate intensity
multiplied by the tributary area (at each location within the foundation
footprint).
f. Repeat as needed for the range of stiffness and dashpot values that account
for uncertainty in the soil profile.
8. Distribute horizontal springs and dashpots around the footprint of the foundation:
a. The horizontal spring at the base-slab level is evaluated as kj,sur = kj/j and is
applied at the base-slab level. A similar procedure is applied for damping.
b. For embedded foundations, the remaining lateral stiffness is distributed over
stiffness is taken as kj(1–1/j), as specified in Section 2.2.3 (similar for

the depth of the embedment (e.g., height of basement walls). The remaining
damping).
Procedures for deep (e.g., pile) foundations, described in Section 2.3, are similar to
those presented above for shallow foundations. Differences include the depth range
over which dynamic properties are evaluated; the equations used to calculate static
stiffness, dynamic modifiers, and damping ratios; and the need for group modifiers in
the case of pile groups. Computer programs (e.g., DYNA6, Western Engineering,
2011; SASSI, Lysmer et al., 1999) can perform Step 2 through Step 5 for finite
element analyses, utilizing procedures similar to those discussed in this report. The
DYNA6 program can also be used for analysis of pile impedance as well.
An important consideration when deep foundation elements (e.g., piles) are combined
with shallow foundation elements (e.g., spread footings or mats) is whether or not
resistance from both shallow and deep foundation elements can be combined. If the
soil is expected to settle away from the shallow foundation elements (e.g., the case of
consolidating soils and end-bearing piles), then lateral load resistance should be
derived on the basis of the piles, pile caps, and basement walls only, and the
resistance provided by shallow foundation elements should be ignored.

8.2.2 Modifying Ground Motions due to Kinematic Interaction
The steps for modifying ground motions due to kinematic interaction are summarized
as follows:
1. Collect the required input for the analysis:
a. Specification of seismic demand. Determine if seismic demands for design
are to be specified in the form of an acceleration response spectrum a set of
ground motion acceleration time histories.
b. Foundation Dimensions. Determine the area of foundation, as represented
by B eA , and the embedment depth, D.
c. Shear wave velocity profile. Determine shear wave velocities to depth, D.
2. Calculate the transfer function for base-slab averaging:
a. Transfer function, Hu(), is calculated using Equation 3-3 and Equation 3-4.
Input parameters include B eA and a.
b. Parameter, a, is evaluated using Equation 3-5. This base-slab averaging
model is calibrated for B eA in the range of 15 m to 40 m.
Hu(), for embedment is calculated using Equations 3-6a through 3-6d. Input
3. Calculate the transfer function for embedment (if D > 0). The transfer function,
parameters include depth, D, and the average effective profile velocity, Vs, avg,
over that depth.
4. For each frequency, , the combined transfer function ordinate for base-slab
averaging and embedment is taken as the product of the results from (2) and (3).
5. Modify free-field ground motions to foundation input motions:
a. Calculate response spectrum modifiers using Equation 3-7a or Equation
3-7b, which relate the ratio of foundation to free-field response spectral
ordinates to the transfer function at the corresponding frequency.
b. Modify acceleration time histories, as needed, using the procedures in
Section 3.4. Note the limitations described in Section 3.4 are for ground
motions dominated by long-period energy (e.g., soft soils and near-fault
effects).
8.2.3 Incorporating Soil-Structure Interaction in Response History

Analyses
Summary guidance for implementing soil-structure interaction in response history

analyses includes the following:
1. As a minimum, the subterranean levels of a structural system (if present) should
be included in the structural model that is fixed at the base. Seismic excitation

can be specified using the free-field motion or a foundation input motion derived
considering kinematic interaction effects. This corresponds to Model 2
illustrated in Figure 6-2. This approach neglects inertial SSI effects.
2. Based on current software capabilities, inertial SSI effects can be incorporated
into the structural model in a practical manner using a “bathtub” configuration.
In this approach, springs and dashpots are developed for the foundation system,
each of which is connected to a rigid bathtub surrounding the foundation.
Seismic excitation is applied to the bathtub using the free-field motion or a
foundation input motion derived considering kinematic interaction effects. This
corresponds to Model 4 illustrated in Figure 6-2.
3. The most complete modeling that can be implemented, within the context of
substructure methods of analysis, extends the bathtub model to include the effects
of multi-support excitation along the basement walls. This corresponds to the
Baseline Model (Model MB) illustrated in Figure 6-2. Multi-support excitation
requires an analysis of wave propagation in the free-field over the depth of
embedment. Depth-dependent motions from wave propagation analyses are
applied as displacements at the ends of the horizontal foundation springs. Depth-
dependent motions inherently include the effects of embedment described in
Section 3.2, and should be further adjusted for base-slab averaging.
8.3 Future Research Needs
Future research needs in soil-structure interaction have been organized into two
general thematic areas. The first theme involves relatively short-term
recommendations expanding on current studies to: (1) provide tangible insights into
the benefits of SSI analysis for owners and practicing engineers; and (2) further
explore the benefits and limitations of SSI response history analysis procedures,
possibly leading to improved procedures. The second theme involves relatively long-
term recommendations intended to address fundamental limitations in the state of SSI
knowledge, which limit the accuracy and reliability of SSI models available for use
in engineering practice.
8.3.1 Theme 1: Expansion of Current Studies
Example applications presented in Chapter 7 presented case studies on buildings with

available earthquake recordings. This effectively limited range of possible structural
configurations that could be investigated to the types of structural systems for which
data were available. In particular, case-study buildings had modest SSI effects and
lacked strongly nonlinear responses. Additional studies should be performed on a
common building configuration that is specifically selected to produce large SSI
effects.

One such example might include a low-rise dual system building (i.e., shear wall plus
moment frame system) with a basement. The building could be designed using a
fixed base model (i.e., Model 2 from Figure 6-2) using a typical design spectrum and
then redesigned using a bathtub model (i.e., Model 4 from Figure 6-2) for the same
design spectrum.
Both large-amplitude and modest-amplitude ground motions could be applied to the

model to evaluate the effects of nonlinearity in the structural response on the impact
of SSI effects. Alternative SSI element configurations that better account for uplift
and nonlinear soil behavior, as described in Section 2.4, could also be applied to
assess their relative impact on elastic-plastic procedures. Analysis using a direct
approach (as opposed to a substructure approach) could be performed, especially in
the case of strongly nonlinear response.
Expanded studies on such a building could be used to provide quantitative

comparisons of the effects of different foundation modeling approaches on structural
member sizes and, ultimately, construction costs.
8.3.2 Theme 2: Research to Address Knowledge Gaps
Research needs for expanding the state of knowledge for soil-structure interaction are
as follows:
 The foundation damping model of Veletsos et al. (various) produces different
results than similar models by others. A critical examination of the derivation of
that model is needed, followed by the development of equations for foundation
damping that properly consider hysteretic damping from soil response, radiation
damping from rotational and translational vibration modes, and the sensitivity of
radiation damping to different soil stiffness profiles. This problem is discussed
in Section 2.1.
 The rotational stiffness of shallow foundation systems with non-rigid structural
foundation elements is poorly understood. In particular, the effects of coupled
versus uncoupled rotations at the base of lateral-load bearing elements on
radiation damping and overall system impedance need to be investigated. This
problem is discussed in Section 2.2.3.
 The impedance of pile-supported foundations is poorly understood for realistic
pile and soil conditions, especially for pile groups. Elasto-dynamic solutions for
piles in idealized soil profiles exist in the literature, but are not used in practice,
in part because they only apply at low displacement levels. The discrete element
models that are used (e.g., LPILE, APILE) are poorly constrained for stiffness,
and are intended for non-seismic problems. Next-generation element models for
dynamic loading of piles are needed that accurately capture the stiffness from

elasto-dynamic analyses and the capacity from discrete element models. This
problem is discussed in Section 2.3.
 The kinematic interaction problem for pile-supported mat foundations subjected
to incoherent wave fields has not been explored to a sufficient degree. Data from
Japan suggest that existing models (for base-slab averaging and kinematic
response of pile foundations) are unable to capture observed foundation/free-field
transfer functions, which decay rapidly with frequency. This problem is
discussed in Section 3.3.
 The conversion of transfer functions, derived from kinematic interaction
analyses, to ratios of response spectra currently relies on guidelines provided in
FEMA 440, Improvement of Nonlinear Static Seismic Analysis Procedures
(FEMA, 2005). A relatively robust statistical model is needed for the
relationships between these ordinates as a function of frequency and ground
motion characteristics. This problem is discussed in Section 3.4.
 Values of the R-factor used in force-based methods for seismic design are based
on engineering judgment, considering observations of building performance in
past earthquakes and anticipated performance of similarly designed buildings in
future earthquakes. Nonlinear response (without collapse) of good performing
systems is used to justify high R-values. However, for some buildings, good
performance may result, in part, from soil-structure interaction effects that also
serve to reduce seismic demands. SSI effects may be partially reflected in
current values of the R-factor, resulting in a potential for double-counting. There
is a need to revisit the definition of R factors with respect to SSI effects, to make
sure that the specified values represent the effects of structural ductility alone.
This issue should be considered in the implementation of NIST GCR 12-917-20,
Tentative Framework for Development of Advanced Seismic Design Criteria for
New Buildings (NIST, 2012). This problem is discussed in Section 4.1.
 Available experimental data described in Chapter 5 should be distilled into a
common format of impedance ordinates, and compared to a consistent set of
predictions utilizing the procedures and parameter selection protocols given in
this report.
 Case studies of buildings with recordings from seismic instrumentation are
extremely valuable. Unfortunately, current protocols for structural
instrumentation seldom provide the information needed for SSI studies, so
additional data are needed. At a minimum, sensors are needed to record
structural translations at the foundation and roof, at least two vertical sensors on
the foundation to record rocking, and a ground instrument near the building.

Appendix A
Sherman Oaks Building Model
Development
This appendix provides detailed information on the structural modeling and
calibration for the Sherman Oaks building example application presented in
Chapter 7.
A.1 Baseline Model Development
Response history analysis of a three-dimensional model of the soil-foundation-

structure system of the Sherman Oaks building was performed using OpenSees, Open
System for Earthquake Engineering Simulation (McKenna, 1997; OpenSees, 2011).
This model was based on a previous OpenSees structural model of the building
provided by the California Strong Motion Instrumentation Program (CSMIP), and
United States Geological Survey (USGS). Construction drawings of the Sherman
Oaks building were made available for inspection through the auspices of CSMIP.
The structural system and foundation configuration of the Sherman Oaks building are
Renderings of the Sherman Oaks building configuration are provided in Figure A-1.
The renderings include cut-away views showing details in the first story, basement,
and foundation regions. The structural configuration in the first story and basement
levels were modeled in a simplified manner to keep the structural modeling aspects
tractable, and to focus on comparisons between variations in the idealization of the
soil-foundation interface.
In the first story of the building, the one-story extension of the structure along the
south longitudinal face of the building (shown in Figure A-1) was not included in the
structural model (i.e., both the mass and the stiffness of the one-story extension were
excluded). In the basement region, the geometry was simplified such that the shear
walls were placed in-line with the perimeter frames of the superstructure. This kept
the footprint of the model at 21.9 m wide by 57.6 m long (72 ft by 189 ft) from
foundation to roof. Although this geometric simplification was made, the flexibility
inherent in the real configuration was captured through elastic connecting springs
modeled with the stiffness properties of the horizontal slab components that
interconnect the frame lines of the superstructure with the basement wall lines. This
is shown schematically in Figure A-2.
GCR 12-917-21 A: Sherman Oaks Building Model Development A-1

Figure A-1 Renderings of the Sherman Oaks building, including cut-away views showing
structural details in the first story, basement, and foundation regions.
Figure A-2 Schematic illustration of elastic springs connecting the framing lines of the
superstructure with the wall lines in the basement levels.
A-2 A: Sherman Oaks Building Model Development GCR 12-917-21

The superstructure of the Sherman Oaks building model is a three-dimensional frame
of fiber elements (nonlinearBeamColumn in OpenSees). The model is a centerline
model that does not include finite joint elements. The two basement levels include
fiber element beams and columns, basement slabs, shear walls, and basement grade
beams. A rigid diaphragm constraint is used at each floor level.
Fiber elements are composed of both core and cover concrete materials (Concrete04
in OpenSees) and reinforcing steel materials (Steel02 in OpenSees). Fiber elements
model flexural behavior using nonlinear concrete and steel material models. Shear
and torsional flexibilities are modeled to be linear-elastic, and are combined together
into a single component model using the SectionAggregator approach in OpenSees.
An expected yield strength of 462 MPa (67 ksi) was used for the grade 60 steel
(Melchers, 1999). This value is slightly lower than, but still comparable to, an
expected yield strength of 517 MPa (75 ksi), which is suggested in ASCE/SEI 41-06
(ASCE, 2007). An expected initial stiffness of 20,000 GPa (29,000 ksi) was used,
along with a post-yield hardening stiffness of 2% of the initial stiffness. A nominal
concrete strength of 35 MPa (5.0 ksi) was included in the model, without provision
for expected strength. Use of nominal concrete strength in lieu of expected strength
was judged to have minimal impact on the structural response predictions because of
the mild nonlinearity experienced in the Northridge earthquake.
The calculated building mass included the mass of all structural elements (beams,
columns, and slabs); 0.5 kPa (10 psf) for partitions; 0.6 kPa (12 psf) for mechanical,
electrical, and plumbing components; and 25% of the design live load taken as 0.6
kPa (12 psf) for a design live load of 2.5 kPa (50 psf). These masses are accounted
for both above and below grade.
Damping was modeled as 4.5% Rayleigh damping, anchored to the first and second
mode periods of the building (2.9 sec and 1.0 sec, respectively). This level of
damping is a calibrated value, which is discussed below. In developing the Rayleigh
damping matrix, degrees-of-freedom associated with foundation springs were
excluded. This was necessary to avoid double counting of foundation damping
because the soil-foundation model included dashpots at these degrees-of-freedom.
Basement shear walls are 30 cm (1 ft) thick, modeled in the simplified manner
described above (and shown in Figure A-2). Figure A-2 shows how the moment
resisting frame, simplified shear wall, and slab elements connect to the framing nodes
in the two basement levels of the building. At each connection location, the nodes
are placed at the same coordinates in the model (but are shown offset in Figure A-2
to illustrate the connectivity). Vertical and rotational degrees of freedom are
constrained together for all nodes at each connection location. Figure A-2 also
depicts the rigid beams needed to support the simplified shear wall models and the
soil springs at the base of the building (soil dashpots are not shown). Stiffness in the

plane of the shear walls was computed based on shear behavior using G=0.4Ec, as
recommended in PEER/ATC-72-1, Modeling and Acceptance Criteria for Seismic
Design and Analysis of Tall Buildings (ATC, 2010), Ec=23,200 MPa (3,370 ksi), and
a factor to account for the cracked properties of the wall (another calibrated value
discussed below). The out-of-plane stiffness of the walls was not included in the
model.
The stiffness and damping properties of the soil are modeled using vertical and
horizontal springs and dashpots, as described in Chapter 7. Soil springs are linear-
elastic (with model variant MB.2 considering the no-tension gap springs). Dashpots
are linear in all model variations.
The depth-variable nature of input ground motions over the height of the basement
walls were specified using the MultipleSupportExcitation approach in OpenSees. In
this approach, the acceleration, velocity, and displacement acceleration-histories are
all specified at each subterranean level. Specifying each ground motion history
removes the need to integrate motions within OpenSees.
A.2 Baseline Model Calibration
The Sherman Oaks Baseline Model (MB) was calibrated against recordings from the
1994 Northridge earthquake. The calibrated model was then used to predict the
response of the building for the 1992 Landers and 1987 Whittier earthquakes to
assess the stability of the calibration.
The first stage of calibration targeted the shear stiffness of the reinforced concrete
walls and slabs in the subterranean levels. A multiplier on the theoretical shear
modulus of the uncracked concrete, Gc, was used as the calibration parameter to
match near-ground response. A definitive ratio of cracked to uncracked stiffness in
shear walls is not available in the literature. For example, ASCE/SEI 41-06 suggests
that the full (unreduced) value of Gc be used for both cracked and uncracked walls.
PEER/ATC-72-1 states that the cracked shear stiffness should be “substantially
lower” than the uncracked shear stiffness, but there are limited test data available for
use in quantifying this ratio. It was found that, over the range of 0.25 to 0.40
considered, near-ground response was not highly sensitive to this parameter, and
acceptable results were achieved with a stiffness multiplier of 0.25.
The second stage of calibration targeted the building period. Adjustments to the
structural mass and stiffness were used as the calibration parameters for the modeled
building period. The structure mass was modified through the application of a scale
factor. In fiber element models, stiffness is not an input parameter, but is computed
from component dimensions and material properties. Component stiffness was
adjusted through modification of the stiffness of reinforcing steel within the fiber
element model. This approach is not meant to suggest that the rebar stiffness is

highly variable; rather it is an indirect approach to modifying the stiffness of
reinforced concrete elements, which is known to be highly variable (Haselton et al.,
2011)
The initial modeled building period was found to be too low. To increase the period,
final calibration factors included multipliers of 0.7 on rebar stiffness and 1.0 on mass.
Table A-1 compares the first mode periods of the building in each direction with the
periods predicted by the Baseline Model (MB). Measured building periods were
computed using non-parametric system identification through the calculation of
transfer functions between the base-level and roof acceleration histories. Although
not an exact match in all cases, Table A-1 shows that modeled periods agree
reasonably well with the measured periods in the three events studied.
Table A-1 Comparison of Measured and Modeled Periods for the Sherman Oaks Building
Identified Period Identified Period MB Period MB Period
Earthquake Mode 1 (sec) Mode 2 (sec) Mode 1 (sec) Mode 2 (sec)
Long Trans Long Trans Long Trans Long Trans
1994 Northridge 2.92 3.28 0.84 0.94 2.67 2.72 0.84 0.94
1992 Landers 2.56 2.72 0.73 0.80 2.14 2.68 0.74 0.80
1987 Whittier 2.33 2.29 0.82 0.83 2.33 2.29 0.82 0.83
The third stage of calibration targeted Rayleigh damping. Adjustments to the level of
damping were used to match the amplitude of displacements in the superstructure.
The OpenSees model used full Rayleigh damping with two matching frequencies,
taken as the first- and second-mode vibration frequencies. A target level of 4.5% of
critical damping was found to provide reasonable results.
Soil spring capacities were compared to demands computed for the 1994 Northridge
using the calibrated Model MB to investigate the potential for failure of the
foundation during earthquake shaking. Limiting spring capacities for the Sherman
Oaks building (presented in Chapter 7) exceeded spring force demands by a factor of
three or more. This margin of safety indicates that foundation failure is unlikely, and
limiting spring capacities were not needed in the present simulations.
Displacement histories obtained from the calibrated Model MB are compared to

recordings from the 1994 Northridge earthquake in Figure A-3. The match in both
horizontal directions at the foundation and ground floor levels is excellent. Over the
height of the building, the quality of the match is generally better in the longitudinal
(EW) direction than in the transverse (NS) direction. Base rocking results in the
bottom frames show that rocking does not significantly contribute to expected roof
displacements. Because there was only one recording instrument at the ground level,
there are no base rocking data available for comparison to modeled results.

Model MB in the 1994 Northridge earthquake: roof level (a) and (b);
(h); and roof displacement due to base rocking (i) and (j).

Profiles of peak floor displacements and peak floor accelerations obtained from the
calibrated Model MB are compared to recordings from the 1994 Northridge
earthquake in Figure A-4. The fit for displacements is better than for accelerations.
Peak accelerations are surprisingly high at ground level and are slightly larger than
expected over the height of the structure. In general, the quality of the fit is
acceptable, and meets the intent of the calibration criteria.

accelerations for Model MB in the 1994 Northridge earthquake.
The calibrated Model MB was then used to predict the response of the building for
the 1992 Landers and 1987 Whittier earthquakes to assess the stability of the
calibration. Results for displacement histories and peak displacements are shown in
Figure A-5 through Figure A-8. The quality of the match for Whittier is better than
for Northridge, likely due to a lack of nonlinearity in the response. In the case of
Landers, the predicted response is consistently weaker than the recorded response.

Model MB in the 1992 Landers earthquake: roof level (a) and (b); 8th
floor (c) and (d); ground level (e) and (f); foundation level (g) and (h);
and roof displacement due to base rocking (i) and (j).

1992 Landers Earthquake

accelerations for Model MB in the 1992 Landers earthquake.

1987 Whittier Earthquake

Model MB in the 1987 Whittier earthquake: roof level (a) and (b); 8th
floor (c) and (d); ground level (e) and (f); foundation level (g) and (h);
and roof displacement due to base rocking (i) and (j).

1987 Whittier Earthquake

accelerations for Model MB in the 1987 Whittier earthquake.

Appendix B
Sherman Oaks Building Stick Model
Development
calibration for the Sherman Oaks building stick model variations used in parametric
studies presented in Chapter 7. This appendix also includes detailed results
comparing key response quantities from each model variant.
B.1 Stick Model Development
Stick models were used to condense, or simplify, the seismic-force-resisting system

in the superstructure for parametric studies on structural system properties. The key
parameter varied between the models was structural stiffness. The moment frame
system of the Sherman Oaks building was first idealized as a stick model and
calibrated to the results of the full-building Baseline Model (MB). Once comparable
results were achieved for the moment frame stick model, the stiffness was scaled to
different factors representing the stiffness of a series of core shear wall and perimeter
shear wall variations.
B.1.1 Moment Frame Stick Model Development
The Sherman Oaks building moment frame is a 13-story structure above grade, with
a total of seven bays in the east-west (E-W) direction (X-direction) and a total of two
bays in the north-south (N-S) direction (Y-direction). The overall configuration of
the full-building Baseline Model (MB) is shown in Figures B-1 through B-3. These
figures also show the location of the master nodes used in the stick model. The stick
model configuration is shown in Figure B-4.
In the stick model, the mass of each floor, the effective lateral, axial, and rotational
stiffness, and the effective yield strength, are all lumped at the master node for each
level. Geometric and material definitions from the full-building OpenSees model
code were extracted and the rotational and axial stiffness at each floor level were
calculated. A static pushover analysis of the full-building Model MB was used to
determine the lateral stiffness and yield strength at each level.
First, a series of static pushover analyses were performed involving the application of
a single point load at the master node of the floor under consideration, while
constraining all other master nodes of the structure. The pushover analysis was
displacement-controlled, and the recorded force-displacement data at the master node
GCR 12-917-21 B: Sherman Oaks Building Stick Model Development B-1

Figure B-1 Plan view of the full-building model of the Sherman Oaks building.
Figure B-2 Longitudinal elevation (X-direction) of the full-building model of the Sherman
Oaks building.
B-2 B: Sherman Oaks Building Stick Model Development GCR 12-917-21

Figure B-3 Transverse elevation (Y-direction) of the full-building model of the
Sherman Oaks building.
under consideration was used to determine the effective lateral stiffness at each floor
level. Adjustments to the effective lateral stiffness were made as needed to account
for the stiffness contribution of adjacent floors. Once all contributions to stiffness
had been accounted for, force-displacement curves were created. These force-
displacement curves were then used to create idealized force-displacement curves per
ASCE/SEI 41-06 (ASCE, 2007). An example force-displacement curve is shown in
Figure B-5.
This methodology, however, resulted in overall stiffness and yield strength values for
the stick model that exceeded the full-building Model MB values. Another pushover
analyses was performed involving the application of a single point load at the master
node of the roof, pushed to a defined displacement of 20 cm (8 in). This

Figure B-4 Elevation the idealized stick model of the Sherman Oaks building.
Floor 11 Pushover - X Direction

7000
Structure
Stick
6000
5000
Force (kips)
4000
3000
2000
1000
0
0 1 2 3 4 5 6 7
Displacement (in)
Figure B-5 Idealized force-displacement curve for the Sherman Oaks building.

displacement level provided adequate story drift over the height of the structure,
without yielding of the structural elements. The force-displacement data recorded at
the master nodes of all levels was then used to calculate the story drifts at each floor
level for a given force. This method produced elastic stiffness values consistent with
the full-building Model MB. Because nonlinear behavior was not observed in the
full-building model for pushover displacements in the range of 20 cm to 50 cm (8 in
to 20 in), the stick model was idealized as an elastic element.
A basement structure node connectivity scheme was developed to properly associate

the idealized stick element with the ground level diaphragm. Node definitions and
connectivity at Floor 1 and below are shown in Figure B-6. All of the nodes in a
given floor level exist at the same location. The nodes labeled 301010, 201010, and
101010 are nodes that define the diaphragms of the substructure. The actual
basement walls occur outside of the lateral frame lines. Fixed support nodes provide
connectivity of the soil-to-wall spring that define interaction between the soil and the
basement wall. Multi-support constraint nodes provide nodal definitions that
simulate the stiffness of the basement walls, both in-plane and out-of-plane, and
account for the offset of the basement wall to the frame line. Rigid body constraints
simulate the vertical stiffness of the basement wall.
Figure B-6 Shear wall and basement level node definitions, constraints, and
connectivity for the Sherman Oaks stick model.

Using these nodal definitions, the base connectivity of the stick element at Floor 1
was defined using rigid elements linking the base of the stick to each of the
unconstrained nodes defining the diaphragm at Floor 1. This produced rigid body
constraints at the ground level that distributed the stick element demands to the walls
of the subterranean structure.
B.1.2 Moment Frame Stick Model Calibration
A limited calibration process was undertaken, with a focus on achieving closer

convergence between floor displacements and peak floor accelerations in the stick
model and full-building model. It was recognized, however, that the process of
idealizing a complex superstructure into a simplified stick representation is an
approximation.
The first stage of calibration targeted the hysteretic behavior observed at each floor
during the response histories. Hysteresis curves were used to compare the predefined
lateral stiffness of the stick model to the full-building Model MB. Softening of the
lateral stiffness of the stick model was needed in both lateral directions at all floor
levels. Plots of absolute displacement and peak acceleration response showed greater
convergence after this calibration.
The second stage of calibration targeted the mild bilinear behavior observed in story
levels above the first story. This stage consisted of defining the yield force at each
level where bilinear behavior occurred in an attempt to match maximum floor
displacements and peak floor accelerations. Each iteration caused divergence of
results with no clear trend toward better calibration, so further adjustment of this
parameter was abandoned.
Final Stick Model Properties. The final stick model was created using
twoNodeLink element objects. This allowed the stick elements spanning from floor
to floor to be modeled as axial, shear, and rotational springs. Final element
properties for the moment frame stick model are given in Tables B-1 through B-3.
The axial and rotational spring stiffness values in Table B-1 are based on length
(story height), area (total column area per story level), and elastic modulus
information taken directly from the full-building Model MB. Rotational stiffness
about global X- and Y-directions were determined by solving a system of linear
equations for a unit rotation. The lateral force-deformation properties in Table B-2
are based on lateral pushover analyses described above. Masses (weights) used in the
stick model are provided in Table B-3.
Each element property was defined using uniaxial stress-strain (force-deformation)

relationships. Axial and rotational element properties were constructed with an
elastic material object. Properties in both lateral directions were constructed with
uniaxial bilinear material objects with kinematic hardening.

Table B-1 Axial and Rotational Properties of the Moment Frame Stick Model
Property Story 1 Story 2 to Story 13
Length 276 in 141 in
Area 25056 in2 24480 in2
Elastic Modulus 4287 ksi 3713 ksi
Rotational Stiffness about Global x-axis 4508x107 kip-in / rad 7359x107 kip-in / rad
(E-W)
Rotational Stiffness about Global y-axis 2243x108 kip-in / rad 3608x108 kip-in / rad
(N-S)
Torsional Stiffness about Global z-axis Fixed Fixed
Table B-2 Lateral Force-Deformation Properties of the Moment Frame Stick Model
Story Yield Strength (kips) Initial Stiffness (kips/in) Strain-Hardening Ratio
E-W N-S E-W N-S E-W N-S
1 5500 6500 2566.8 2521.9 0.0371 0.0142
2 12000 13050 5342.1 4415.0 -0.0541 -0.0176
3 10171 11605 4893.8 3755.8 -0.0348 -0.0478
4 9600 10200 4724.2 3521.5 -0.0318 0.0000
5 8000 8214 4584.9 3429.5 -0.0643 -0.0234
6 6700 7300 4452.4 3389.5 -0.0460 -0.0255
7 5286 6000 4375.8 3351.1 -0.0345 -0.0312
8 5000 5733 4261.2 3293.0 -0.0210 -0.0370
9 4343 5086 4169.9 3242.0 0.0183 -0.0225
10 3771 4571 4141.1 3205.5 0.0040 -0.0090
11 3000 3743 4142.6 3215.3 0.0245 0.0339
12 3000 3600 4235.6 3346.7 0.0263 0.0115
13 5500 5000 4920.4 3864.7 0.0820 0.2642
Table B-4 compares the first-, second-, and third-mode periods of the moment frame
(MF) stick model and the full-building Model MB. Development and calibration of
the stick model was performed using a relatively early version of the Sherman Oaks
OpenSees Model MB, which changed over time as model components and ground
motion inputs were refined. For this reason, Model MB results presented in this
appendix are not an exact match to those presented in Section 7.2.

Table B-3 Masses in the Moment Frame Stick Model
Level Node Weight (kips)
Floor -1 Distributed 3824
Floor 0 Distributed 3824
Floor 1 Distributed 3010
Floor 2 4000000 2522
Floor 3 5000000 2522
Floor 4 6000000 2522
Floor 5 7000000 2522
Floor 6 8000000 2522
Floor 7 9000000 2522
Floor 8 10000000 2522
Floor 9 11000000 2522
Floor 10 12000000 2522
Floor 11 13000000 2522
Floor 12 14000000 2522
Floor 13 15000000 2522
Roof 16000000 3314
Model and the Full-Building Model MB
Identified Period Identified Period Identified Period
Mode 1 (sec) Mode 2 (sec) Mode 3 (sec)
Translational Torsional Translational
Model (Global Y-axis) (about Global Z-axis) (Global X-axis)
MB-MF-Full 2.45 2.31 2.21
MB-MF-Stick 2.87 N.A. 2.62
Figure B-7 compares the roof displacement response history for the moment frame
stick model to the center of the full-building Model MB in the X-direction and the
Y-direction. Responses in the Y-direction are more in phase than in the X-direction.
Displacement amplitudes match fairly well. In some cases, larger amplitudes are
observed in the stick model, but in other cases, larger amplitudes are observed in the
full-building model. Similar results were obtained for nodes located at the mid-
height of the building.

Northridge Record, Channel 1
30
MB-MF-Stick
20 MB-MF-Full
Lateral Displacement (cm)
10
-10
-20
-30
-40
0 10 20 30 40 50 60
Time (sec)
Northridge Record, Channel 2

30
MB-MF-Stick
20 MB-MF-Full
Lateral Displacement (cm)
10
-10
-20
-30
0 10 20 30 40 50 60
Time (sec)
Figure B-7 Comparison of roof displacement histories for the moment frame
stick model and the full-building Model MB.
Figures B-8 through B-11 compare response envelopes for maximum displacement,
story drift, story shear, and peak floor acceleration for each direction. Results for the
moment frame stick model are typically within about 10% of those for the full-
building Model MB. Maxima in the positive and negative directions are reported. In
some cases, results for the moment frame stick model are larger in one direction, but
smaller in the perpendicular direction. It was found that calibration attempts often
improved results in one direction at the expense of the other direction.
Based on the following comparisons, the calibration process for the moment frame
stick model was judged to be sufficient for the purposes of the parametric study.

Northridge, X-Direction
14
12
10
Floor Number
8
0 MB-MF-Stick
MB-MF-Full
-2
-40 -30 -20 -10 0 10 20 30 40
Max. Displacement [cm]
Northridge, Y-Direction
14
12
10
Floor Number
0 MB-MF-Stick
MB-MF-Full
-2
-40 -30 -20 -10 0 10 20 30 40

frame stick Model MB and the full-building Model MB.

Northridge, Interstory Drifts, X-Direction
14
MB-MF-Stick
12
MB-MF-Full
10
Story Number
-2
-0.015 -0.01 -0.005 0 0.005 0.01 0.015
Max. Interstory Drifts [cm/cm]
Northridge, Interstory Drifts, Y-Direction

14
12
10
Story Number
0 MB-MF-Stick
MB-MF-Full
-2
-0.015 -0.01 -0.005 0 0.005 0.01 0.015
Figure B-9 Comparison of maximum story drift envelopes for the moment frame
stick Model MB and the full-building Model MB.

14
12
10
Floor Number
8
0 MB-MF-Stick
MB-MF-Full
-2
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
14
12 MB-MF-Stick
MB-MF-Full
10
Floor Number
-2
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
Figure B-10 Comparison of maximum story shear envelopes for the moment
As mentioned earlier, development and calibration of the stick model was performed
using a relatively early version of the Sherman Oaks OpenSees Model MB, which
changed over time. In Figure B-10, a spike in story shear force can be observed in
the basement level for the stick model. This is because early versions of the full-
building Model MB included an artificial phase shift in the ground motions used for
multi-support excitation, which caused a similar spike in shear demand. This phase
shift was removed in subsequent versions of the full-building Model MB, and the
spike no longer appears in the results for the final full-building model depicted in the
figure.

14
12
10
Floor Number
0 MB-MF-Stick
MB-MF-Full
-2
-1.5 -1 -0.5 0 0.5 1 1.5
Peak Floor Accel. [g]
14
12
10
Floor Number
0 MB-MF-Stick
MB-MF-Full
-2
-1.5 -1 -0.5 0 0.5 1 1.5

B.2 Additional Moment Frame Stick Model Comparisons to Full-
Building Model Results
Calibrations were applied to Baseline Model (MB) results only. The calibrated
moment frame stick model was imported into models capturing the foundation
modeling configurations associated with Model 1, Model 2, and Model 4a. Table
B-5 compares the first-, second-, and third-mode periods of the moment frame stick
model and the full-building model for each foundation configuration.
Model and the Full-Building Model for each Alternative
Foundation Configuration
1-MF-Full 2.26 2.19 2.09

1-MF-Stick 2.83 N.A. 2.61
2-MF-Full 2.46 2.31 2.22
2-MF-Stick 2.86 N.A. 2.62
4a-MF-Full 2.47 2.32 2.21
4a-MF-Stick 2.87 N.A. 2.63
Periods in each simplified stick model are longer than the periods observed in the
corresponding full-building model. This was attributed to the stick elements being
modeled as linear elements and calibrated with an elastic stiffness that was softer
than the full-building model.
Figures B-12 through B-23 compare response envelopes for maximum displacement,
story drift, story shear, and peak floor acceleration for each alternative foundation
model in each direction. In general, comparisons are similar to trends observed for
Model MB.

14
12
10
Floor Number
0 1-MF-Stick
1-MF-Full
-2
-40 -30 -20 -10 0 10 20 30 40
14
12
10
Floor Number
0 1-MF-Stick
1-MF-Full
-2
-40 -30 -20 -10 0 10 20 30 40
frame stick Model 1 and the full-building Model 1.

14
12 1-MF-Stick
1-MF-Full
10
Story Number
8
-2
-0.015 -0.01 -0.005 0 0.005 0.01 0.015
14
12
10
Story Number
0 1-MF-Stick
1-MF-Full
-2
-0.015 -0.01 -0.005 0 0.005 0.01 0.015
stick Model 1 and the full-building Model 1.

14
12
10
Floor Number
0 1-MF-Stick
1-MF-Full
-2
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
14
12
10
Floor Number
0 1-MF-Stick
1-MF-Full
-2
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4

14
12
10
Floor Number
8
0 1-MF-Stick
1-MF-Full
-2
-1.5 -1 -0.5 0 0.5 1 1.5
14
12
10
Floor Number
0 1-MF-Stick
1-MF-Full
-2
-1.5 -1 -0.5 0 0.5 1 1.5

14
12
10
Floor Number
0 2-MF-Stick
2-MF-Full
-2
-40 -30 -20 -10 0 10 20 30 40
14
12
10
Floor Number
0 2-MF-Stick
2-MF-Full
-2
-40 -30 -20 -10 0 10 20 30 40

14
12 2-MF-Stick
2-MF-Full
10
Story Number
8
-2
-0.015 -0.01 -0.005 0 0.005 0.01 0.015
14
12
10
Story Number
0 2-MF-Stick
2-MF-Full
-2
-0.015 -0.01 -0.005 0 0.005 0.01 0.015
stick Model 2 and the full-building Model 2.

14
12
10
Floor Number
0 2-MF-Stick
2-MF-Full
-2
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
14
12
10
Floor Number
0 2-MF-Stick
2-MF-Full
-2
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4

14
12
10
Floor Number
8
0 2-MF-Stick
2-MF-Full
-2
-1.5 -1 -0.5 0 0.5 1 1.5
14
12
10
Floor Number
0 2-MF-Stick
2-MF-Full
-2
-1.5 -1 -0.5 0 0.5 1 1.5

14
12
10
Floor Number
0 4a-MF-Stick
4a-MF-Full
-2
-40 -30 -20 -10 0 10 20 30 40
14
12
10
Floor Number
0 4a-MF-Stick
4a-MF-Full
-2
-40 -30 -20 -10 0 10 20 30 40
frame stick Model 4a and the full-building Model 4a.

14
4a-MF-Stick
12
4a-MF-Full
10
Story Number
8
-2
-0.015 -0.01 -0.005 0 0.005 0.01 0.015
14
12
10
Story Number
0 4a-MF-Stick
4a-MF-Full
-2
-0.015 -0.01 -0.005 0 0.005 0.01 0.015
stick Model 4a and the full-building Model 4a.

14
12
10
Floor Number
0 4a-MF-Stick
4a-MF-Full
-2
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
14
12
10
Floor Number
0 4a-MF-Stick
4a-MF-Full
-2
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4

14
12
10
Floor Number
8
0 4a-MF-Stick
4a-MF-Full
-2
-1.5 -1 -0.5 0 0.5 1 1.5
14
12
10
Floor Number
0 4a-MF-Stick
4a-MF-Full
-2
-1.5 -1 -0.5 0 0.5 1 1.5

B.3 Moment Frame Stick Model Comparisons for Different
Foundation Configurations
Figures B-24 through B-27 overlay response envelopes for maximum displacement,
configuration of the moment frame models in each direction. The following trends
were observed:
 Displacements for Model 1 are typically the largest, followed by Model 2.
Displacements for Model 4a are very similar to Model MB.
 Drifts trends are similar to displacement trends, except that Model MB has larger
drifts than the others models in the basement levels.
 Story shear results for Model 1 are the largest. Story shears for Model 2, Model
4a and Model MB are similar.
 Peak floor acceleration results for all models are similar in the upper floors.
Model 2 produces significantly larger results in the ground story and basement
levels.

14
12
10
Floor Number
8
2 1-MF-Stick
2-MF-Stick
0 4a-MF-Stick
MB-MF-Stick
-2
-40 -30 -20 -10 0 10 20 30 40
14
12
10
Floor Number
2 1-MF-Stick
2-MF-Stick
0 4a-MF-Stick
MB-MF-Stick
-2
-40 -30 -20 -10 0 10 20 30 40

frame stick Models 1, 2, 4a, and MB.


14
1-MF-Stick
12
2-MF-Stick
10 4a-MF-Stick
MB-MF-Stick
Story Number
-2
-0.015 -0.01 -0.005 0 0.005 0.01 0.015

14
12
10
Story Number
2 1-MF-Stick
2-MF-Stick
0 4a-MF-Stick
MB-MF-Stick
-2
-0.015 -0.01 -0.005 0 0.005 0.01 0.015
stick Models 1, 2, 4a, and MB.

14
12
10
Floor Number
8
2 1-MF-Stick
2-MF-Stick
0 4a-MF-Stick
MB-MF-Stick
-2
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
14
12 1-MF-Stick
2-MF-Stick
10 4a-MF-Stick
MB-MF-Stick
Floor Number
-2
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4

14
12
10
Floor Number
2 1-MF-Stick
2-MF-Stick
0 4a-MF-Stick
MB-MF-Stick
-2
-1.5 -1 -0.5 0 0.5 1 1.5
14
12
10
Floor Number
2 1-MF-Stick
2-MF-Stick
0 4a-MF-Stick
MB-MF-Stick
-2
-1.5 -1 -0.5 0 0.5 1 1.5

B.4 Shear Wall Stick Model Development
To investigate the effects of structural stiffness on SSI, moment frame stick model
parameters were scaled to represent the stiffness associated with shear wall systems.
ASCE/SEI 7-10 (ASCE, 2010) and ACI 318-08 (ACI, 2008) were used as references
for determining appropriate scale factors for stiffness. A shear wall stick model was
created by converting the base shear of the full-building Model MB into a design
base shear for an equivalent shear wall using the ratio of response modification
coefficients (R factors) for the special reinforced concrete moment frame and special
reinforced concrete shear wall systems, respectively. Assuming a bearing wall
system, this ratio is 1.6. Based on an amplified base shear, and material properties
predefined in the full-building Model MB, a required shear wall area was determined.
In the case of the core wall configuration, a geometry using two separate interior core
walls, 16 inches thick, was assumed. The cores were configured to match the size
and dimensions of typical elevator, stair, and restroom core areas in typical buildings.
The stiffness was then determined by summing the flexural and shear deflections of a
slender cantilever wall using the following equation:
  b  b  
Vh3 1.2Vh
(B-1)
12 EI Gc A
Where the shear modulus of concrete, Gc, was assumed to be 40% of the elastic
modulus of concrete. Total deflection was primarily controlled by the flexural term,
and the stiffness of the shear wall system was approximately three times the stiffness
of the moment frame system. The stiffness contribution from the gravity system was
also considered. When combined with the stiffness of the shear wall system, a total
stiffness scale factor of 5 was obtained. The core shear wall building model, with a
scale factor of 5, was designated SW. Table B-6 compares the first-, second-, and
third-mode periods of the shear wall stick models for each foundation configuration.
Table B-6 Comparison of Modeled Periods for the Shear Wall Stick Models
for each Alternative Foundation Configuration
1 Stick (SW) 1.32 N.A. 1.17
2 Stick (SW) 1.38 N.A. 1.19
4a Stick (SW) 1.40 N.A. 1.20
MB Stick (SW) 1.40 N.A. 1.20

Additional shear wall systems with scale factors of 25, 50, and 125, representing
perimeter shear wall configurations, were also considered. These models were
designated SW1, SW2, and SW3, respectively.
B.5 Shear Wall Stick Model Comparisons
Figures B-28 through B-43 overlay response envelopes for maximum displacement,
configuration, for each shear wall model, in each direction. The following trends
were observed:
 Differences between the models are much more significant in the stiffer shear
wall systems than the moment frame system, demonstrating that SSI effects have
a larger influence on response in stiffer structural systems.
 Displacements in Model 1 are typically the smallest. Displacements in Model 4a
are very similar to Model MB.
 Drift results are similar to displacement results, except that the Model MB has
larger drifts in the basement levels.
 Story shear results vary between the X-direction and Y-direction for all models.
 Peak floor accelerations in the superstructure are larger in Model 1 and Model 2.
Model 2 produces significantly larger floor accelerations in the ground story and
basement levels.

14
12
10
Floor Number
8
2 1-SW -Stick
2-SW -Stick
0 4a-SW -Stick
MB-SW -Stick
-2
-40 -30 -20 -10 0 10 20 30 40
14
12
10
Floor Number
2 1-SW -Stick
2-SW -Stick
0 4a-SW -Stick
MB-SW -Stick
-2
-40 -30 -20 -10 0 10 20 30 40
Figure B-28 Comparison of maximum displacement envelopes for the SW stick

Models 1, 2, 4a, and MB.

14
12
10
Story Number
2 1-SW -Stick
2-SW -Stick
0 4a-SW -Stick
MB-SW -Stick
-2
-0.015 -0.01 -0.005 0 0.005 0.01 0.015

14
12
10
Story Number
2 1-SW -Stick
2-SW -Stick
0 4a-SW -Stick
MB-SW -Stick
-2
-0.015 -0.01 -0.005 0 0.005 0.01 0.015
Figure B-29 Comparison of maximum story drift envelopes for the SW stick

14
12
10
Floor Number
8
2 1-SW -Stick
2-SW -Stick
0 4a-SW -Stick
MB-SW -Stick
-2
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
14
12 1-SW -Stick
2-SW -Stick
10 4a-SW -Stick
MB-SW -Stick
Floor Number
-2
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
Figure B-30 Comparison of normalized story shear envelopes for the SW stick

14
12
10
Floor Number
2 1-SW -Stick
2-SW -Stick
0 4a-SW -Stick
MB-SW -Stick
-2
-1.5 -1 -0.5 0 0.5 1 1.5
14
12
10
Floor Number
2 1-SW -Stick
2-SW -Stick
0 4a-SW -Stick
MB-SW -Stick
-2
-1.5 -1 -0.5 0 0.5 1 1.5
Figure B-31 Comparison of peak floor acceleration envelopes for the SW stick

14
12
10
Floor Number
8
2 1-SW 1-Stick
2-SW 1-Stick
0 4a-SW 1-Stick
MB-SW 1-Stick
-2
-40 -30 -20 -10 0 10 20 30 40
14
12
10
Floor Number
2 1-SW 1-Stick
2-SW 1-Stick
0 4a-SW 1-Stick
MB-SW 1-Stick
-2
-40 -30 -20 -10 0 10 20 30 40

14
12
10
Story Number
2 1-SW 1-Stick
2-SW 1-Stick
0 4a-SW 1-Stick
MB-SW 1-Stick
-2
-0.015 -0.01 -0.005 0 0.005 0.01 0.015

14
12
10
Story Number
2 1-SW 1-Stick
2-SW 1-Stick
0 4a-SW 1-Stick
MB-SW 1-Stick
-2
-0.015 -0.01 -0.005 0 0.005 0.01 0.015

14
12
10
Floor Number
8
6
4 1-SW 1-Stick
2 2-SW 1-Stick
4a-SW 1-Stick
0
MB-SW 1-Stick
-2
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
14
12 1-SW 1-Stick
2-SW 1-Stick
10
4a-SW 1-Stick
Floor Number
8 MB-SW 1-Stick
6
4
2
0
-2
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

14
12
10
Floor Number
2 1-SW 1-Stick
2-SW 1-Stick
0 4a-SW 1-Stick
MB-SW 1-Stick
-2
-1.5 -1 -0.5 0 0.5 1 1.5
14
12
10
Floor Number
2 1-SW 1-Stick
2-SW 1-Stick
0 4a-SW 1-Stick
MB-SW 1-Stick
-2
-1.5 -1 -0.5 0 0.5 1 1.5

14
12
10
Floor Number
8
2 1-SW 2-Stick
2-SW 2-Stick
0 4a-SW 2-Stick
MB-SW 2-Stick
-2
-40 -30 -20 -10 0 10 20 30 40
14
12
10
Floor Number
2 1-SW 2-Stick
2-SW 2-Stick
0 4a-SW 2-Stick
MB-SW 2-Stick
-2
-40 -30 -20 -10 0 10 20 30 40

14
12
10
Story Number
2 1-SW 2-Stick
2-SW 2-Stick
0 4a-SW 2-Stick
MB-SW 2-Stick
-2
-0.015 -0.01 -0.005 0 0.005 0.01 0.015

14
12
10
Story Number
2 1-SW 2-Stick
2-SW 2-Stick
0 4a-SW 2-Stick
MB-SW 2-Stick
-2
-0.015 -0.01 -0.005 0 0.005 0.01 0.015

14
1-SW 2-Stick
12
2-SW 2-Stick
10 4a-SW 2-Stick
Floor Number
8 MB-SW 2-Stick
6
4
2
0
-2
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
14
1-SW 2-Stick
12 2-SW 2-Stick
10 4a-SW 2-Stick
Floor Number
MB-SW 2-Stick
8
6
4
2
0
-2
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

14
12
10
Floor Number
2 1-SW 2-Stick
2-SW 2-Stick
0 4a-SW 2-Stick
MB-SW 2-Stick
-2
-1.5 -1 -0.5 0 0.5 1 1.5
14
12
10
Floor Number
2 1-SW 2-Stick
2-SW 2-Stick
0 4a-SW 2-Stick
MB-SW 2-Stick
-2
-1.5 -1 -0.5 0 0.5 1 1.5

14
12
10
Floor Number
8
2 1-SW 3-Stick
2-SW 3-Stick
0 4a-SW 3-Stick
MB-SW 3-Stick
-2
-40 -30 -20 -10 0 10 20 30 40
14
12
10
Floor Number
2 1-SW 3-Stick
2-SW 3-Stick
0 4a-SW 3-Stick
MB-SW 3-Stick
-2
-40 -30 -20 -10 0 10 20 30 40

14
12
10
Story Number
2 1-SW 3-Stick
2-SW 3-Stick
0 4a-SW 3-Stick
MB-SW 3-Stick
-2
-0.015 -0.01 -0.005 0 0.005 0.01 0.015

14
12
10
Story Number
2 1-SW 3-Stick
2-SW 3-Stick
0 4a-SW 3-Stick
MB-SW 3-Stick
-2
-0.015 -0.01 -0.005 0 0.005 0.01 0.015

14
12
10
Floor Number
8
6
4 1-SW 3-Stick
2 2-SW 3-Stick
4a-SW 3-Stick
0
MB-SW 3-Stick
-2
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
14
1-SW 3-Stick
12 2-SW 3-Stick
10 4a-SW 3-Stick
MB-SW 3-Stick
Floor Number
8
6
4
2
0
-2
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

14
12
10
Floor Number
2 1-SW 3-Stick
2-SW 3-Stick
0 4a-SW 3-Stick
MB-SW 3-Stick
-2
-1.5 -1 -0.5 0 0.5 1 1.5
Northridge,X-Direction
14
12
10
Floor Number
2 1-SW 3-Stick
2-SW 3-Stick
0 4a-SW 3-Stick
MB-SW 3-Stick
-2
-1.5 -1 -0.5 0 0.5 1 1.5

B.6 Moment Frame Stick Model and Shear Wall Stick Model
Comparisons
Figures B-44 through B-47 show comparisons between moment frame stick Model
MB and shear wall stick Model MB results for maximum displacement, story drift,
story shear, and peak floor acceleration in each direction. Results clearly indicate a
significant difference in building response between stiffer and more flexible
structural systems.
14
12
10
Floor Number
0 MB-MF-Stick
MB-SW -Stick
-2
-40 -30 -20 -10 0 10 20 30 40
14
12
10
Floor Number
0 MB-MF-Stick
MB-SW -Stick
-2
-40 -30 -20 -10 0 10 20 30 40

frame stick Model MB and the shear wall stick Model MB.

14
MB-MF-Stick
12
MB-SW -Stick
10
Story Number
-2
-0.015 -0.01 -0.005 0 0.005 0.01 0.015

14
12
10
Story Number
0 MB-MF-Stick
MB-SW -Stick
-2
-0.015 -0.01 -0.005 0 0.005 0.01 0.015
stick Model MB and the shear wall stick Model MB.

14
12
10
Floor Number
8
0 MB-MF-Stick
MB-SW -Stick
-2
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
14
12 MB-MF-Stick
MB-SW -Stick
10
Floor Number
-2
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4

14
12
10
Floor Number
0 MB-MF-Stick
MB-SW -Stick
-2
-1.5 -1 -0.5 0 0.5 1 1.5
14
12
10
Floor Number
0 MB-MF-Stick
MB-SW -Stick
-2
-1.5 -1 -0.5 0 0.5 1 1.5

Appendix C
Walnut Creek Building Model
Development
validation for the Walnut Creek building example application presented in Chapter 7.
C.1 Walnut Creek Building Model Development
Response history analysis of a three-dimensional model of the soil-foundation-

structure system of the Walnut Creek building was performed using OpenSees, Open
System for Earthquake Engineering Simulation (McKenna, 1997; OpenSees, 2011).
Construction drawings of the building were made available for inspection through the
auspices of CSMIP. The structural system and foundation configuration are
Renderings of the Walnut Creek building configuration are provided in Figure C-1.
The renderings include cut-away views showing details of the core wall and
perimeter frame configurations, as well as the interior layout of a typical floor level.
Figure C-2 provides plan views of the Walnut Creek building shear wall core at the
foundation, first floor, and typical floor levels.
The building model is composed of the core walls and the perimeter frame. The
framing elements were modeled with force-based beam-column elements using fiber
sections (nonlinearBeamColumn elements in OpenSees). Two methods of modeling
the core wall were considered: (1) use of four-node quadrilateral elements (Quad or
Shell elements in OpenSees); and (2) use of horizontal and vertical boundary frame
elements in combination with diagonal strut elements.
The core walls are 30.5 cm (12 in) thick. To account for the cracked section
properties of the concrete walls, a multiplier on the theoretical uncracked stiffness
must be applied. This multiplier should be dependent on the level of in-situ cracking
of the shear wall and on the intensity of ground shaking, which would dictate the
level of cracking that the wall will experience during an earthquake. A definitive
ratio of cracked to uncracked stiffness in shear walls is not available in the literature.
In the case of flexural stiffness, ASCE/SEI 41-06 (ASCE, 2007) suggests a multiplier
of 0.8 for nominally uncracked walls, and a multiplier of 0.5 for cracked walls.
GCR 12-917-21 C: Walnut Creek Building Model Development C-1

Figure C-1 Renderings of the Walnut Creek building, including cut-away views showing: (a) the
exterior facade; (b) structural framing; (c) core wall elevations; (d) core wall layout;
(e) details of the precast cladding system; and (f) interior layout of a typical floor level.
C-2 C: Walnut Creek Building Model Development GCR 12-917-21

Figure C-2 Plan views of the Walnut Creek building shear wall core: (a)
foundation level; (b) first floor; and (c) typical floor level.

In the case of shear stiffness, ASCE/SEI 41-06 suggests that the full (unreduced)
value of Gc be used for both cracked and uncracked walls. The PEER/ATC-72-1
report (ATC, 2010) states that the cracked shear stiffness should be “substantially
lower” than the uncracked shear stiffness, but there are limited test data available for
use in quantifying this ratio. Considering that the level of shaking during the Loma
Prieta earthquake was modest at the Walnut Creek site, a factor of 0.3 was used for
the approximate ratio between cracked and uncracked concrete shear stiffness.
When using plane-stress Quad elements, the stiffness in the plane of each wall
Poisson’s ratio,  (taken as 0.25). The material chosen to represent the concrete was
element is based on the input values of wall dimensions, material behavior, and
an elastic isotropic material (nDMaterial type called ElasticIsotropic in OpenSees).

The out-of-plane stiffness of each wall element is not represented in this type of
model. In the current version of OpenSees, Quad elements with an elastic isotropic
material can only capture the elastic behavior of the wall elements. Also, the
implementation of quadrilateral elements into a three-dimensional model can be
complicated because of the difference in the number of degrees of freedom at the
nodes.
Because of the above limitations, a combination of boundary frame elements and

diagonal struts was also considered for modeling the core walls. This type of model
permits modeling of inelastic behavior and out-of-plane behavior. In this type of
model, the boundary elements are intended to represent the flexural characteristics of
the wall, and diagonal compression-only truss elements are used to represent
diagonal compression struts in the concrete, which carry a significant portion of the
shear in the concrete walls. The beam and column elements were assigned high
elastic axial stiffness. The beams were also given high flexural stiffness, but the
columns were assigned finite stiffness based on a square section with dimensions
equal to the wall thickness and an elastic modulus reduced to 30% of the gross value.
The dimensions of the diagonal truss elements were also set equal to the wall
thickness, and the elastic modulus was reduced to 30% of the gross value.
Reductions in section stiffness were estimated to account for cracking as well as the
reduced efficiency of the diagonal compression strut. Stiffness reductions were
validated through comparisons between the modeled response and the measured
response of the building to the 1989 Loma Prieta earthquake input motions.
Frame elements were composed of nonlinear beam-column elements with fiber

sections representing inelastic behavior of both core and cover concrete materials
(Concrete02 in OpenSees) and reinforcing steel materials (Steel02 in OpenSees).
Shear and torsional flexibilities can be modeled as linear-elastic, and combined
together into a single component model using the SectionAggregator approach in
OpenSees. For expediency, however, the current model did not consider these modes
of deformation. An expected yield strength of 462 MPa (67 ksi) was used for the

grade 60 steel (Melchers, 1999), with an initial stiffness of 20,000 GPa (29,000 ksi),
and a post-yield hardening stiffness of 2% of the initial stiffness. A nominal concrete
strength of 35 MPa (5.0 ksi) was modeled, and the expected strength was taken as 1.5
times the nominal strength (52 MPa, 7.5 ksi), in accordance with ASCE/SEI 41-06.
The three-dimensional OpenSees model is depicted in Figure C-3. To expedite the

analyses and post-processing of results, the model was simplified to a two-
dimensional model, also shown in Figure C-3. The two-dimensional model includes
the framing elevations spanning the E-W direction, which was generated by
collapsing the coordinates of the three-dimensional model into two dimensions, and
constraining overlapping nodes to have the same horizontal displacement. The
stiffness and damping properties of the soil are modeled using vertical and horizontal
springs and dashpots, as described in Chapter 7. The soil springs are linear-elastic,
and the dashpots are linear in all model versions. Foundation spring and dashpot
locations are shown in Figure C-4. At the base of the mat, a grid of rigid beams was
created to model the larger footprint of the mat foundation (since it is larger than the
core wall in plan).
The calculated building mass included the mass of all structural elements (beams,
columns, and slabs); 0.5 kPa (10 psf) for partitions; 0.6 kPa (12 psf) for mechanical,
electrical, and plumbing components; and 25% of the design live load taken as 0.6
kPa (12 psf) for a design live load of 2.5 kPa (50 psf). The base of the core wall
extends 4.3 m (14 feet) below the ground level, which includes a 3 m (10 feet)
extension of the core wall, 1.2 m (4 feet) of mat foundation, and sand backfill. The
mass of the core wall (including the sand) is also included in the calculated building
mass. In the model, the mass of the slab and framing elements is assigned to the
nodes, and the mass of each shear wall is distributed to the lower boundary beam.
Damping was modeled as 5% of critical Rayleigh damping, anchored to the first and
third mode periods of the building. The damping was uniformly distributed to all
structural elements, but degrees-of-freedom associated with foundation springs were
excluded. This was necessary to avoid double counting of foundation damping
because the soil-foundation model included dashpots at these degrees-of-freedom.
A preliminary evaluation of the relative accuracy of two-dimensional and three-

dimensional models was performed. A comparison between modeled and recorded
roof displacement histories is shown in Figure C-5. The data indicate that a
significant improvement in accuracy is obtained using the three-dimensional
representation of the numerical model. Although the three-dimensional model was
more accurate in representing the amplitude of the recorded displacements, the two-
dimensional model was a good match to the phasing, and was judged to be a
reasonable approximation of the physical structure for the purposes of this study.

(a) Three-dimensional model
(b) Simplified two-dimensional model

Figure C-3 Illustrations of the Walnut Creek building OpenSees model:
(a) three-dimensional model; and (b) simplified two-dimensional
model.

Figure C-4 Horizontal and vertical spring and dashpot locations for the Walnut
Creek building model.
Figure C-5 Comparison of relative roof displacement histories from recorded

data, two-dimensional (2D) model results, and three-dimensional
(3D) model results for the Walnut Creek building.
C.2 Model Validation
Because of limitations in available site-characterization data, it was decided to pursue

model validation, with a focus on matching overall stiffness and deformation
characteristics, rather than full model calibration. As a result, only the effective
stiffness of the wall elements (columns and struts) was adjusted. In addition, only the
response in the E-W direction was evaluated.
Recorded motions from the 1989 Loma Prieta earthquake were used to validate the
numerical model. The first step in the validation was a system identification to

compare the fundamental period of the system. The 5%-damped elastic response
spectra were calculated for the ground-floor and roof level recorded accelerations, as
shown in Figure C-6.
Figure C-6 5%-damped elastic response spectra of recorded motions for the
Walnut Creek building in the 1989 Loma Prieta earthquake.
The first two periods identified for the E-W direction were 0.81sec and 0.18sec,
respectively. The effective stiffness of the wall columns and diagonal struts were
then modified so that the numerical model gave an accurate representation of the first
two periods as well as the relative displacement history between the ground floor and
the roof. An effective stiffness equal to 30% of the gross section properties resulted
in the first two modeled periods equal to 0.83sec and 0.23sec, respectively.
Comparisons of modeled versus recorded acceleration and relative displacement

histories for the Walnut Creek building in the 1989 Loma Prieta earthquake are
shown in Figure C-7. The plots indicate that the numerical model is a reasonable
representation of the dynamic response of the physical structure.

Figure C-7 Comparison of modeled versus recorded acceleration and relative displacement
histories for the Walnut Creek building in the 1989 Loma Prieta earthquake: roof level
(a) and (b); 8th floor (c) and (d); 3rd floor (e) and (f); and ground floor (g) and (h).
Comparisons of the modeled versus recorded maximum relative displacements and

peak accelerations are shown in Figure C-8. Maximum relative displacements are
somewhat over-predicted relative to observed maxima. The maximum story drift
profile of the model is shown in Figure C-9.
Because two vertical instruments were present at the ground floor level, it is possible
to compute base rocking (i.e., the difference between the two vertical recordings
divided by the horizontal distance between them). As shown in Figure C-10,
however, low frequency noise causes base rocking contributions evaluated in this
manner to have unrealistic features. Accordingly, the vertical records were filtered
using high-pass Butterworth filters with a corner frequency of 0.4 Hz.

Figure C-8 Comparison of modeled versus recorded maximum relative
displacements and peak accelerations for the Walnut Creek building
in the 1989 Loma Prieta earthquake.
Figure C-9 Maximum story drift profile for the Walnut Creek building model
subjected to the 1989 Loma Prieta earthquake.
A corner frequency of 0.4 Hz was selected through examination of displacement

histories with different corner frequencies shown in Figure C-10, following the
recommendations of Bommer and Boore (2005), which results in the minimum filter
corner that removes the low frequency noise features in the rocking displacement
history. The resulting comparison between modeled and recorded roof displacement
due to base rotation of the Walnut Creek building is presented in Chapter 7.

Figure C-10 High-pass filtering of vertical motions at base level showing the effects of corner frequency on displacement and rocking
histories.

Symbols
A footing area, m2 (ft2)

Ac footing-soil contact area required to support vertical and shear loads, m2 (ft2)
a0 dimensionless frequency for footings
a0k dimensionless frequency related to base contact area
a 0p dimensionless frequency for piles
B foundation half-width (small plan dimension), m (ft)
BeA half-dimension of an equivalent square footing matching the area of the
actual footing, m (ft)
BeI half-dimension of an equivalent square footing matching the moment of
inertia of the actual footing, m (ft)
c′ soil shear strength, kN/m2 (kips/ft2)
cj dashpot coefficient for translation along the j-axis, kN-sec/m (kipf-sec/ft)
cjj dashpot coefficient for rotation about the j-axis, kN-m/rad (kipf-ft/rad)
c jp dashpot coefficient for single pile, kN-m/rad (kipf-ft/rad)
c zi dashpot intensity, kN-sec/m3 (kipf-sec/ft3)
Cr OpenSees parameter for BNWF models
Cs seismic coefficient
C s seismic coefficient at an elongated period T
d pile diameter, m (ft)
D embedment depth, m (ft)
DL internal node spacing, m (ft)
Dr relative density
dw height of effective side wall contact of embedded foundation, m (ft)
Ef foundation Young’s modulus, MPa (ksi)
Es Young’s modulus of the soil, MPa (ksi)
Ep Young’s modulus of a pile, MPa (ksi)
F force, kN (kips)
GCR 12-917-21 Symbols D-1

fE fundamental frequency (Hz)
fL limiting frequency (Hz)
FSv vertical factor of safety
G soil shear modulus, kPa (ksi)
G0 maximum soil shear modulus (at small strains), kPa (ksi)
Ga soil shear modulus at atmospheric pressure, kPa (ksi)
h height of the center of mass for the first-mode shape, m (ft)
Hu kinematic transfer function
area moment of inertia of soil-foundation contact, m4 (ft4); i denotes the axis

about which the surface is taken
j index denoting modes of translation displacement or rotation
Jt polar moment of inertia of soil-foundation contact surface, m4 (ft4)
k spring stiffness, translation, kN/m (kips/ft), and rotation, kN-m/rad
(kip-ft/rad)
kj spring stiffness, translation, along the j-axis, kN/m (kips/ft)
Kj foundation static stiffness for mode j, translation, kN/m (kips/ft), and
rotation, kN-m/rad (kip-ft/rad)
kj complex-valued impedance function for mode j, translation, N/m (lbf/ft), and
rotation, N-m/rad (lbf-ft/rad)
kjj spring stiffness, rotation, about the j-axis, kN-m/rad (kip-ft/rad)
k Gj efficiency factor for a group of piles
k jp spring stiffness for a single pile, kN/m (kips/ft)
K jp pile static stiffness for mode j, kN/m (kips/ft)
k zi stiffness intensity (coefficient of subgrade reaction), kN/m3 (kips/ft3)
L foundation half-length (large plan dimension), m (ft)
La active pile length (typically in lateral mode, < Lp), m (ft)
le spring spacing, m (ft)
Lend end region length of foundation with increase stiffness, m (ft)
Lp pile length, m (ft)
m mass, Mg (lbm, kipm)
M first mode participating mass, Mg (lbm, kipm)
D-2 Symbols GCR 12-917-21

n, ns, nx, nyy exponents related with foundation damping
p-x horizontal passive load-displacement behavior against side of footing
q spring force, kN (kips)
q0 upper limit of the linear elastic region, kPa (ksi)
qult ultimate load, kPa (ksi)
q-z vertical load-displacement curve
Q vertical load on footing, kN (kips)
QULT ultimate vertical load, kN (kips)
Rc dashpot intensities scalar
Re end length ratio
Rk spring stiffness intensities scalar
Rv rebound ratio
r, rf foundation radius, m (ft)
s spring deflection, m (ft)
s0 displacement at load q0, m (ft)
s50 displacement at which 50% of the ultimate load is mobilized, m (ft)
S spring spacing, m (ft)
Sa fixed-base spectral acceleration, m/s2, (ft/s2), g
Sa spectral acceleration for a flexible-base structure, m/sec2, (ft/sec2), g
S ae equivalent spectral acceleration, m/sec2 (ft/sec2), g
Sa-FIM fixed-base spectral acceleration with FIM effects, m/s2, (ft/s2), g
Sd spectral displacement, m (ft)
S de equivalent spectral displacement, m (ft)
T undamped natural vibration period, sec
T flexible-base period, sec
tf foundation thickness, m (ft)
Tp predominant period of ground motion, sec
Tx, Tyy fictitious vibration periods, calculated as if the only source of the vibration
was translation or rotation, sec
T T period lengthening ratio due to soil-structure interaction

t-x horizontal shear-sliding curve at the base of a footing
uf horizontal deflection amount in a SDOF structure due to base shear force F,
cm (in)
uFIM foundation input motion, acceleration, m/sec2; velocity, m/sec; or
displacement, m
ug free-field ground motion, acceleration, m/sec2; velocity, m/sec; or
displacement, m
V base shear force, kN (kips)
Vapp apparent wave propagation velocity, m/sec (ft/sec)
Vs shear wave velocity in a homogeneous isotropic half-space; measured free-
field shear wave velocity, m/sec (ft/sec)
Vs (z) measured free-field shear wave velocity at depth z, m/sec (ft/sec)
Vs, avg average effective profile velocity over depth zp, m/sec (ft/sec)
Vs, F overburden-corrected shear wave velocity below foundation, m/sec (ft/sec)
VULT ultimate vertical load, kN (kips)
W effective weight of structure, kN (kips)
wpj, wsj, wbj weight factors, for the pile, soil, and pile tip stiffness contributions for
mode j
z depth, m (ft)
zp effective profile depth, m (ft)
zw depth to centroid of effective sidewall contact, m (ft)
j dynamic (frequency-dependent) stiffness modifier for surface foundation for
mode j
 jp dynamic (frequency-dependent) stiffness modifier for single pile for mode j
v vertical angle of incident waves
0 flexible-base system damping ratio
f foundation damping ratio
i fixed-base structural damping ratio
j dynamic damping ratio for mode j
 jp pile dynamic damping ratio for mode j
p pile damping ratio
rj radiation damping ratio for mode j

s soil hysteretic damping ratio
 deflection, m (ft)
 total deflection with respect to free-field at the top of a SDOF structure, cm
(in)
max maximum earthquake-induced displacement in a structural member, cm (in)
 max maximum earthquake-induced displacement at the top of a structure, cm (in)
y yield displacement of a structural member, cm (in)
 y yield displacement at the top of a structure, cm (in)
v increment of vertical stress at depth z from the structural weight, kPa (ksi)
j dimensionless modulus of subgrade reaction
j, jj embedment correction factor for rigid footing spring constants for mode j
 rotation deflection amount in a SDOF structure due to base moment (rad)
elastic elastic rotation limit (rad)
a dimensionless parameter accounting for effects of incoherence from wave
passage and stochastic processes, as well as Fourier amplitude variability
 dimensionless parameter for active pile length
L dimensionless pile length parameter
 dimensionless constant exponent for active pile length
s
ductility demand, respectively
global structural system ductility demand, individual structural member
 soil Poisson’s ratio

f foundation Poisson’s ratio
p pile mass density, Mg/m3 (lbm/ft3)
s soil mass density, Mg/m3 (lbm/ft3)
m effective confining stress, kPa (ksi)
 v 0 effective stress from the self-weight of the soil, kPa (ksi)
′ angle of internal friction, rad (deg)
j loss angle due to damping, rad (deg)
j dimensionless constant for vibration for mode j
 soil-to-foundation stiffness ratio, representing the flexibility of the
foundation

 dimensionless pile tip stiffness
 undamped natural vibration frequency (rad/sec)

References
Abrahamson, N.A., 1985, Estimation of Seismic Wave Coherency and Rupture

Velocity Using the SMART 1 Strong-Motion Array Recordings, Report No.
UCB/EERC-85/02, Earthquake Engineering Research Center, University of
California, Berkeley, California.
Abrahamson, N.A., 1992, Spatial Variation of Earthquake Ground Motion for
Application to Soil-Structure Interaction, Report. No. EPRI TR-100463,
Electrical Power Research Institute, Palo Alto, California.
Abrahamson, N.A., Schneider, J.F., and Stepp, J.C., 1991, “Empirical spatial
coherency functions for application to soil-structure interaction analyses,”
Earthquake Spectra, Vol. 7, No. 1, pp. 1-27.
ACI, 2008, Building Code Requirements for Structural Concrete, ACI 318-08,
American Concrete Institute, Farmington Hills, Michigan.
Allotey, N., and Naggar, M.H.E., 2003, “Analytical moment–rotation curves for rigid
foundations based on a Winkler model,” Soil Dynamics and Earthquake
Engineering, Vol. 23, pp. 367-381.
Allotey, N., and Naggar, M.H.E., 2007, “An investigation into the Winkler modeling
of the cyclic response of rigid footings,” Soil Dynamics and Earthquake
Anoyatis, G., and Mylonakis, G., 2012, “Dynamic winkler modulus for axially
loaded piles,” Geotechnique, Vol. 62, pp. 521-536.
Ancheta, T.D., 2010, Engineering Characterization of Spatially Variable Ground
Motions, Ph.D. Dissertation, University of California, Los Angeles,
California.
Ancheta, T.D., Stewart, J.P., and Abrahamson, N.A., 2011, “Engineering
characterization of earthquake ground motion coherency and amplitude
variability,” Proceedings, 4th International Symposium on Effects of Surface
Geology on Seismic Motion, IASPEI / IAEE, University of California Santa
Barbara, California.
Andrus, R.D., Stokoe, K.H. II, and Juang, C.H., 2004. “Guide for shear-wave-based
liquefaction potential evaluation,” Earthquake Spectra, Vol. 20, No. 2,
pp. 285-308.
GCR 12-917-21 References E-1

Apsel, R.J., and Luco, J.E., 1987, “Impedance functions for foundations embedded in
a layered medium: an integral equation approach,” Earthquake Engineering
and Structural Dynamics, Vol. 15, No. 2, pp. 213-231.
Aviles, J., and Perez-Rocha, L.E., 1996, “Evaluation of interaction effects on system
period and the system damping due to foundation embedment and layer
depth,” Soil Dynamics and Earthquake Engineering, Vol. 15, pp. 11-27.
ASCE, 2000, Seismic Analysis of Safety-Related Nuclear Structures and
Commentary, ASCE 4-98, American Society of Civil Engineers, Reston,
Virginia.
ASCE, 2005, Seismic Design Criteria for Structures, Systems, and Components in
Nuclear Facilities, ASCE/SEI 43-05, American Society of Civil Engineers,
Reston, Virginia.
ASCE, 2007, Seismic Rehabilitation of Existing Buildings, ASCE/SEI 41-06,
American Society of Civil Engineers, Reston, Virginia.
ASCE, 2010, Minimum Design Loads for Buildings and Other Structures, ASCE/SEI
7-10, American Society of Civil Engineers, Reston, Virginia.
ASCE, 2013, Seismic Evaluation and Retrofit of Existing Buildings, ASCE/SEI
41-13, American Society of Civil Engineers, Reston, Virginia.
ASCE, (in preparation), Seismic Analysis of Safety-Related Nuclear Structures and
Commentary, ASCE 4, Reston, Virginia.
ATC, 1996, Seismic Evaluation and Retrofit of Concrete Buildings, ATC-40 Report,
Applied Technology Council, Redwood City, California.
ATC, 2010, Modeling and Acceptance Criteria for Seismic Design and Analysis of
Tall Buildings, PEER/ATC-72-1 Report, prepared by Applied Technology
Council in cooperation with Pacific Earthquake Engineering Research
Center, Redwood City, California.
Barghouthi, A.F., 1984, Pile Response to Seismic Waves, Ph.D. Dissertation,
University of Wisconsin, Madison, Wisconsin.
Barkan, D.D., 1962, Dynamics of bases and foundations, McGraw-Hill, New York,
New York.
Bielak, J., 1975, “Dynamic behavior of structures with embedded foundations,”
Earthquake Engineering and Structural Dynamics, Vol. 3, pp. 259-274.
Bielak, J., 1976, “Modal analysis for building-soil interaction,” Journal of
Engineering Mechanics, Vol. 102, pp. 771-786.
E-2 References GCR 12-917-21

Blaney, G.W., Kausel, E., and Roesset, J.M., 1975, “Dynamic stiffness of piles,”
Proceedings, 2nd International Conference on Numerical Methods in
Geomechanics, Virginia Polytechnical Institute and State University,
Blacksburg, Virginia, pp. 1010-1012.
Boore, D.M., and Boomer, J.J., 2005, “Processing of strong-motion accelerograms:
needs, options and consequences,” Soil Dynamics and Earthquake
Borja, W.I., and Wu, W.H., 1994, “Vibration of foundations on incompressible soils
with no elastic region,” Journal of Geotechnical Engineering, Vol. 120,
No. 9, pp. 1570-1592.
Boulanger, R.W., Curras, C.J., Kutter, B.L., Wilson, D.W., and Abghari, A., 1999,
“Seismic soil-pile-structure interaction experiments and analyses,” Journal
Geotechnical and Geoenvironmental Engineering, Vol. 125, pp. 750-759.
CESMD, 2011, “CESMD – A Cooperative Effort,” Center for Engineering Strong
Motion Data (CESMD), accessed in 2011, http://www.strongmotioncenter
.org.
Chang, B.J., Thomas, J.M., Raychowdhury, P., Gajan, S., Kutter, B.L., and
Hutchinson, T.C., 2007, Soil-Foundation-Structure Interaction: Shallow
Foundations, SSRP 07/24, University of California, San Diego, California.
Chatzigogos, C.T., Pecker, A., and Salencon, J., 2009, “Macroelement modeling of
shallow foundations,” Soil Dynamics and Earthquake Engineering, Vol. 29,
pp. 765-781.
Chopra, A.K., and Goel, R.K., 1999, “Capacity-demand-diagram methods based on
inelastic design spectrum,” Earthquake Spectra, Vol. 15, No. 4, pp. 637-656.
Chopra, A., and Yim S.C., 1985, “Simplified earthquake analysis of structures with
foundation uplift,” Journal of Structural Engineering, Vol. 111, No. 4,
pp. 906-930.
Ciampoli, M., and Pinto, P.E., 1995, “Effects of soil-structure interaction on inelastic
seismic response of bridge piers,” Journal of Structural Engineering,
Vol. 121, No. 5, pp. 806-814.
Clough, R.W., and Penzien, J., 1993, Dynamics of Structures, McGraw Hill, New
York.
Coduto, D.P., 1994, Geotechnical Engineering: Principles and Practices, 2nd Edition,
Prentice-Hall, Upper Saddle River, New Jersey.
Cox, W., Reese, L., and Grubbs, B., 1974, “Field testing of laterally loaded piles in
sand,” Proceedings, Offshore Technology Conference, Houston, Texas,
Paper 2079.

Cremer, C., Pecker, A., and Davenne, L., 2001, “Cyclic macro-element of soil
structure interaction: Material and geometrical nonlinearities,” International
Journal of Numerical and Analytical Methods in Geomechanics, Vol. 25,
pp. 1257-1284.
Crouse, C.B., 2001, “Commentary on soil-structure interaction in U.S. seismic
provisions,” Proceedings, 2nd UJNR Workshop on Soil-Structure Interaction,
Tsukuba, Japan.
Crouse, C.B., Hushmand, B., Luco, J.E., and Wong, H.L., 1990, “Foundation
impedance functions: Theory versus experiment,” Journal of Geotechnical
Engineering, Vol. 116, No. 3, pp. 432-449.
Darendeli, 2001, Development of a New Family of Normalized Modulus Reduction
and Material Damping Curves, Ph.D. Dissertation, University of Texas at
Austin, Texas.
Day, S.M., 1978, “Seismic response of embedded foundations,” Preprints of
Conference Proceedings of ASCE Convention and Exposition, Chicago, IL,
Preprint No. 3450, American Society of Civil Engineers, New York, New
York.
De Barros, F.C.P., and Luco, J.E., 1995, “Identification of foundation impedance
functions and soil properties from vibration tests of the Hualien containment
model,” Soil Dynamics and Earthquake Engineering, Vol. 14, pp. 229-248.
Dobry, R., and Gazetas, G., 1988, “Simple method for dynamic stiffness and
damping of floating pile groups,” Geotechnique, Vol. 38, No. 4, pp. 557-574.
Dobry, R., Gazetas, G., and Stokoe, K.H. II, 1986, “Dynamic response of arbitrarily
shaped foundations: Experimental verification,” Journal of Geotechnical
Dobry, R., Vicente, E., O'Rourke, M.J., and Roesset, J.M., 1982, “Horizontal
stiffness and damping of single piles,” Journal of the Geotechnical
Engineering Division, Vol. 108, No. 3, pp. 439-459.
El-Naggar, M.H., and Novak, M., 1994, “Non-linear model for dynamic axial pile
response,” Journal of Geotechnical Engineering, Vol. 120, No. 2,
pp. 308-329.
El-Naggar, M.H., and Novak M., 1995, “Nonlinear analysis for dynamic lateral pile
response,” Soil Dynamics and Earthquake Engineering, Vol. 15, No. 4,
pp. 233-244.
EPRI, 2007, Hard-Rock Coherency Functions Based on the Pinyon Flat Array Data,
ADAMS Accession No. ML071980104, Electric Power Research Institute,
Palo Alto, California.

Erden, S.M., 1974, Influence of Shape and Embedment on Dynamic Foundation
Response, Ph.D. Dissertation, University of Massachusetts, Amherst,
Massachusetts.
Fadum, R.E., 1948, “Influence values for estimating stresses in elastic foundations,”
Proceedings, 2nd International Conference on Soil Mechanics and Foundation
Engineering, Rotterdam, Netherlands, Vol. 3, pp. 77-84.
Fan, K., Gazetas, G., Kaynia, A., and Kausal, E., 1991, “Kinematic seismic response
of single piles and pile groups,” Journal of Geotechnical Engineering,
Vol. 117, No. 12, pp. 1860-1879.
FEMA, 1997, NEHRP Guidelines for the Seismic Rehabilitation of Buildings, FEMA
273, prepared by the Applied Technology Council and the Building Seismic
Safety Council for the Federal Emergency Management Agency,
Washington, D.C.
FEMA, 2000, Prestandard and Commentary for the Seismic Rehabilitation of
Buildings, FEMA 356, prepared by the American Society of Civil Engineers
for Federal Emergency Management Agency, Washington, D.C.
FEMA, 2004a, NEHRP Recommended Provisions for Seismic Regulations for New
Buildings and Other Structures, Part 1: Provisions, FEMA 450-1/2003
Edition, prepared by the Building Seismic Safety Council for the Federal
Emergency Management Agency, Washington, D.C.
FEMA, 2004b, NEHRP Recommended Provisions for Seismic Regulations for New
Buildings and Other Structures, Part 2: Commentary, FEMA 450-2/2003
Edition, prepared by the Building Seismic Safety Council for the Federal
Emergency Management Agency, Washington, D.C.
FEMA, 2005, Improvement of Nonlinear Static Seismic Analysis Procedures, FEMA
440, prepared by the Applied Technology Council for Federal Emergency
Management Agency, Washington, D.C.
FEMA, 2009, NEHRP Recommended Seismic Provisions for New Buildings and
Other Structures, FEMA P-750/2009 Edition, prepared by the Building
Seismic Safety Council of the National Institute of Building Sciences for the
Federal Emergency Management Agency, Washington, D.C.
Fleming, W.G.K., Weltman, A.J., Randolph, M.F., and Elson, W.K., 1992, Piling
Engineering, 2nd Edition, John Wiley & Sons, New York, New York.
Flores-Berrones, R., and Whitman, R.V., 1982, “Seismic response of end-bearing
piles,” Journal of the Geotechnical Engineering Division, Vol. 108, No. 4,
pp. 554-569.

Fumal, T.E., and Tinsley, J.C., 1985, “Mapping shear wave velocities of near surface
geologic materials, in evaluating earthquake hazards in the Los Angeles
region-an earth-science perspective,” USGS Survey Professional Paper 1360,
pp. 101-126.
Gadre, A., and Dobry, R., 1998, “Lateral cyclic loading centrifuge tests on square
embedded footings,” Journal Geotechnical and Geoenvironmental
Gajan S., 2006, Physical and Numerical Modeling of Nonlinear Cyclic Load-
Deformation Behavior of Shallow Foundations Supporting Rocking Shear
Walls, Ph.D. Dissertation, University of California, Davis, California.
Gajan, S., Hutchinson, T.C., Kutter, B.L., Raychowdhury, P., Ugalde, J.A., and
Stewart, J.P., 2008, Numerical models for analysis and performance-based
design of shallow foundations subject to seismic loading, Report No. PEER-
2007/04, Pacific Earthquake Engineering Research Center, University of
Gajan, S., and Kutter, B.L., 2008, “Capacity, settlement, and energy dissipation of
shallow footings subjected to rocking,” Journal Geotechnical and
Geoenvironmental Engineering, Vol. 134, No. 8, pp. 1129-1141.
Gajan, S., and Kutter, B.L., 2009, “A contact interface model for shallow foundations
subjected to combined cyclic loading,” Journal Geotechnical and
Geoenvironmental Engineering, Vol. 135, pp. 407-419.
Gajan, S., Phalen, J.D., and Kutter, B.L., 2003, Soil-Foundation-Structure
Interaction: Shallow Foundations, Report No. UCD/CGMDR-03/01, Center
for Geotechnical Modeling, University of California, Davis, California.
Gajan, S., Raychowdhury, P., Hutchinson, T.C., Kutter, B.L., and Stewart, J.P., 2010,
“Application and validation of practical tools for nonlinear soil-foundation
interaction analysis,” Earthquake Spectra, Vol. 26, No. 1, pp. 111-129.
Gazetas, G., 1991, “Foundation vibrations,” Foundation Engineering Handbook, 2nd
Edition, Chapter 15, H.-Y. Fang, ed., Chapman and Hall, New York, New
York.
Gazetas, G., 2006, “Seismic design of foundations and soil-structure
interaction,”Proceedings, First European Conference on Earthquake
Engineering and Seismology, Geneva, Switzerland.
Gazetas, G., and Dobry, R., 1984a, “Horizontal response of piles in layered soil,”
Journal of the Geotechnical Engineering Division, Vol. 110, No. 1,
pp. 20-40.

Gazetas, G., and Dobry, R., 1984b, “Simple radiation damping model for piles and
footings,” Journal of the Geotechnical Engineering Division, Vol. 110,
No. 6, pp. 937-956.
Gazetas, G., and Stokoe, K.H. II, 1991, “Free vibration of embedded foundations:
Theory versus experiment,” Journal of Geotechnical Engineering, Vol. 117,
No. 9, pp. 1382-1401.
Ghosh, B., and Madabhushi, S.P.G., 2007. “Centrifuge modelling of seismic soil
structure interaction effects.” Nuclear Engineering and Design, Vol. 237,
No. 8, pp. 887-896.
Givens, M.J., 2013, Dynamic Soil-Structure Interaction of Instrumented Buildings
and Test Structures, Ph.D. Dissertation, Department of Civil and
Environmental Engineering, University of California, Los Angeles,
California.
Guzina, B.B., and Pak, R.Y.S., 1998, “Vertical vibration of a circular footing on a
linear-wave-velocity half space,” Geotechnique, Vol. 48, No. 2, pp. 159-168.
Harden, C.W., and Hutchinson, T.C., 2009, “Beam-on-nonlinear-Winkler-foundation
modeling of shallow, rocking-dominated footings,” Earthquake Spectra,
Vol. 25, pp. 277-300.
Hardin, B.O., and Black, W.L., 1968, “Vibration modulus of normally consolidated
clay,” Journal of the Soil Mechanics and Foundation Division, Vol. 94,
No. 2, pp. 353-369.
Harding, Miller, Lawson, and Associates, 1970, Foundation Investigation Walnut
Creek Office Building for Dillingham Corporation, Walnut Creek,
California, Report 3008,002.04, San Francisco, California.
Haselton, C.B., Leil, A.B., Deierlein, G.G., and Chou, J.H., 2011, “Seismic Collapse
Safety of Reinforced Concrete Buildings: I, Assessment of Ductile Moment
Frames,” Journal of Structural Engineering, Vol. 137, pp. 481-491.
Holtz, R.D., Kovacs, W.D., and Sheahan, T.C., 2010, An Introduction to
Geotechnical Engineering, 2nd Edition, Prentice Hall, Englewood Cliffs, New
Jersey.
Houlsby, G.T., and Cassidy, M.J., 2002, “A plasticity model for the behavior of
footings on sand under combined loading,” Geotechnique, Vol. 52, No. 2,
pp. 117-129.
Iguchi, M., and Luco, J.E., 1982, “Vibration of flexible plate on viscoelastic
medium,” Journal of Engineering Mechanics, Vol. 108, No. 6,
pp. 1103-1120.

Jeremic, B., Jie, G.Z., Preisig, M., Tafazzoli, N., 2009, “Time domain simulation of
soil-foundation-structure interaction in non-uniform soils,” Journal of
Earthquake Engineering and Structural Dynamics, Vol. 38, No. 5,
pp. 699-718.
Kagawa, T., and Kraft, L.M., 1981, “Lateral pile response during earthquakes,”
Journal of Geotechnical Engineering, Vol. 107, pp. 1713-1731.
Karatzia X., and Mylonakis, G., 2012, “Horizontal response of piles in layered soil:
Simple analysis,” 2nd International Conference on Performance-Based
Design in Geotechnical Engineering, Taormina, Italy.
Kausel, E., 2010, “Early history of soil-structure interaction,” Soil Dynamics and
Earthquake Engineering, Vol. 30, No. 9, pp. 822-832.
Kausel E., Whitman A., Murray J., Elsabee F., 1978, “The spring method for
embedded foundations,” Nuclear Engineering and Design, Vol. 48,
pp. 377-392.
Kaynia, A.M., and Kausel, E., 1982, Dynamic Stiffness and Seismic Response of Pile
Groups, Research Report R82-03, Massachusetts Institute of Technology,
Cambridge, Massachusetts.
Kaynia, A.M., and Novak, M., 1992, “Response of pile foundations to Rayleigh
waves and to obliquely incident body waves,” Earthquake Engineering and
Structural Dynamics, Vol. 21, No. 4, pp. 303-318.
Kim, S., and Stewart, J.P., 2003, “Kinematic soil-structure interaction from strong
motion recordings,” Journal Geotechnical and Geoenvironmental
Kramer, S.L., 1996, Geotechnical Earthquake Engineering, Prentice Hall, Upper
Saddle River, New Jersey.
Kwok, A.O., Stewart, J.P., Hashash, Y.M.A., Matasovic, N., Pyke, R., Wang, Z., and
Yang, Z., 2007, “Use of exact solutions of wave propagation problems to
guide implementation of nonlinear seismic ground response analysis
procedures,” Journal Geotechnical and Geoenvironmental Engineering,
Vol. 133, No. 11, pp. 1385-1398.
LeRoy, Crandall and Associates, 1978, Report of Supplementary Studies Proposed
Retaining Walls Proposed Sherman Oaks Galleria, Ventura and Sepulveda
Boulevards, Los Angeles, California, Report No. ADE-78044, prepared for
Kendall International, provided by MACTEC, Los Angeles, California.

LeRoy, Crandall and Associates, 1982, Completion of Exploration Program
Proposed Office Building and Parking Structure, Ventura and Sepulveda
Boulevards, Los Angeles, California, Report No. ADE-81384, prepared for
McNeill Enterprises, provided by MACTEC, Los Angeles, California.
Lin, A.N., and Jennings, P.C., 1984, “Effect of embedment on foundation-soil
impedances,” Journal of Engineering Mechanics, Vol. 110, No. 7,
pp. 1060-1075.
Liou, G.S., and Huang, P.H., 1994, “Effect of flexibility on impedance functions for
circular foundations,” Journal of Engineering Mechanics, Vol. 120, No. 7,
pp. 1429-1446.
Luco, J.E., Trifunac, M.D., and Wong, H.L., 1988, “Isolation of soil-structure
interaction effects by fullscale forced vibration tests,” Earthquake
Engineering and Structural Dynamics, Vol. 16, No. 1, pp. 1-21.
Luco, J.E., and Westmann, R.A., 1971, “Dynamic response of circular footings,”
Journal of Engineering Mechanics, Vol. 97, No. 5, pp. 1381-1395.
Lysmer, J., Ostadan, F., and Chen, C., 1999, Computer Program SASSI2000,
Geotechnical Division, University of California, Berkeley, California.
Mamoon, S.M., and Banerjee, P.K., 1990, “Response of piles and pile groups to
traveling SH waves,” Earthquake Engineering and Structural Dynamics,
Vol. 19, No. 4, pp. 597-610.
Maravas, A., Mylonakis, G., and Karabalis, D.L., 2007, “Dynamic characteristics of
simple structures on piles and footings,” Proceedings, 4th International
Conference on Earthquake Geotechnical Engineering, Thessaloniki, Greece,
Paper No. 1672.
Marcuson, W.F., and Wahls, H.E., 1972, “Time effects on dynamics shear modulus
of clays,” Journal of Soil Mechanics and Foundations Division, Vol. 98,
No. SM 12, pp. 1359-1373.
Matlock, H., 1970, “Correlations for design of laterally loaded piles in soft clay,”
Proceedings, 2nd Offshore Technology Conference, Houston, Texas, Vol. 1,
pp. 577-594, Paper No. 1204.
McClelland, B., and Focht, J.A., Jr., 1958, “Soil modulus for laterally loaded piles,”
Transactions, Vol. 123, pp. 1049-1063.
McKenna, F., 1997, Object Oriented Finite Element Programming Frameworks for
Analysis, Algorithms and Parallel Computing, Ph.D. Dissertation, University
of California, Berkeley, California.
Melchers, R.E., 1999, Structural Reliability Analysis and Prediction, 2nd Edition,
John Wiley & Sons Inc., New York, New York.

Menq, F.-Y., 2003, Dynamic Properties of Sandy and Gravelly Soils, Ph.D.
Dissertation, University of Texas at Austin, Texas.
Michaelides O., Gazetas, G., Bouckovalas, G., and Chrysikou, E., 1998,
“Approximate nonlinear analysis of piles,” Geotechnique, Vol. 48, No. 1,
pp. 33-54.
Mikami, A., Stewart, J.P., Ostadan, F., and Crouse, C.B., 2006, “Representation of
ground motion incoherence for the analysis of kinematic soil-structure
interaction,” Proceedings, 8th U.S. National Conference on Earthquake
Engineering, April 18-22, San Francisco, California, Paper No. 1071.
Mikami, A., Stewart, J.P., and Kamiyama, M., 2008, “Effects of time series analysis
protocols on transfer functions calculated from earthquake accelerograms,”
Soil Dynamics and Earthquake Engineering, Vol. 28, No. 9, pp. 695-706.
Mylonakis, G., 1995, Contribution to Static and Seismic Analysis of Piles and Pile-
Supported Bridge Piers, Ph.D. Dissertation, Department of Civil
Engineering, University at Buffalo, State University of New York, Buffalo,
New York.
Mylonakis, G., 2011, Personal Communication (unpublished study), University of
Patras, Greece.
Mylonakis, G., and Gazetas, G., 1998, “Settlement and additional internal forces of
grouped piles in layered soil,” Geotechnique, Vol. 48, No. 1, pp. 55-72.
Mylonakis, G., and Gazetas, G., 1999, “Lateral vibration and internal forces of
grouped piles in layered soil,” Journal Geotechnical and Geoenvironmental
Mylonakis, G., and Gazetas, G., 2000, “Seismic soil-structure interaction: Beneficial
or detrimental,” Journal of Earthquake Engineering, Vol. 4, pp. 377-401.
Mylonakis, G., Nikolaou, S., and Gazetas, G., 2006, “Footings under seismic loading:
Analysis and design issues with emphasis on bridge foundations,” Soil
Dynamics and Earthquake Engineering, Vol. 26, pp. 824-853.
Mylonakis, G., and Roumbas, D., 2001, “Dynamic stiffness and damping of piles in
inhomogeneous soil media,” Proceedings, 4th International Conference on
Recent Advances in Geotechnical Earthquake Engineering and Soil
Dynamics, San Diego, California, Paper No. 6.27.
Naeim, F., Tileylioglu, S., Alimoradi, A., and Stewart, J.P., 2008, “Impact of
foundation modeling on the accuracy of response history analysis of a tall
building,” Proceedings, SMIP2008 Seminar on Utilization of Strong Motion
Data, California Strong Motion Instrumentation Program, Sacramento,
California, pp. 19-55.

Newmark, N.M., 1969, “Torsion of symmetrical buildings,” Proceedings, 4th World
Conference on Earthquake Engineering, Santiago, Chile.
Nie, J., Xu, J., Hofmayer, C., and Ali, S., 2008, “An approach for assessing structural
uplifting using blast motions,” Proceedings, ASME Pressure Vessel and
Piping Division Conference, Chicago, Illinois, Paper No. PVP2008-61278.
Nii, Y., 1987, “Experimental half-space dynamic stiffness,” Journal of Geotechnical
Nikolaou, A., Mylonakis, G., Gazetas, G., and Tazoh, T. 2001, “Kinematic pile
bending during earthquakes: analysis and field measurements,”
Geotechnique, Vol. 51, No. 5, pp. 425-440.
NIST, 2012, Tentative Framework for Development of Advanced Seismic Design
Criteria for New Buildings, GCR 12-917-20, prepared by the NEHRP
Consultants Joint Venture, a partnership of the Applied Technology Council
and the Consortium of Universities for Research in Earthquake Engineering,
for the National Institute of Standards and Technology, Gaithersburg,
Maryland.
Nogami, T., 1983, “Dynamic group effect in axial response of grouped piles,”
Journal of Geotechnical Engineering, Vol. 109, No. 2, pp. 228-243.
Nogami, T., Otani, J., Konagai, K., and Chen, H.-L., 1992, “Nonlinear soil-pile
interaction model for dynamic lateral motion,” Journal of Geotechnical
Nova, R., and Montrasio, L., 1991, “Settlements of shallow foundations on sand,”
Novak, M., 1987, “Discussion of dynamic response of arbitrarily shaped foundations:
experimental verification,” Journal of Geotechnical Engineering, Vol. 113,
No. 11, pp. 1410-1412.
Novak, M., 1991, “Piles under dynamic loads: state of the art,” Proceedings, 2nd
International Conference on Recent Advances in Geotechnical Earthquake
Engineering and Soil Dynamics, St. Louis, Missouri, pp. 233-245.
OpenSees, 2011, Open System for Earthquake Engineering Simulation: OpenSees,
University of California, Berkeley, http://opensees.berkeley.edu.
Ostadan F., 2005, “Seismic soil pressure for building walls – an updated approach,”
Soil Dynamics and Earthquake Engineering, Vol. 25, pp. 785-793.
Ostadan, F., Deng, N., and Kennedy, R., 2005, “Soil-structure interaction analysis
including ground motion incoherency effects,” Proceedings, 18th
International Conference on Structural Mechanics in Reactor Technology
(SMiRT 18), Beijing, China.

Ostadan, F., Hadjian, A.H., Tseng, W.S., Tang, Y.K., Tang, H.T., 1991, “Parametric
evaluation of intermediate SSI solutions on final response,” Proceedings, 11th
International Conference on Structural Mechanics in Reactor Technology
(SMiRT 11), Tokyo, Japan, Paper K04/1, pp. 63-68.
Ostadan F., and White, W., 1998, “Lateral seismic soil pressure – an updated
approach,” Proceedings, U.S.-Japan SSI Workshop, U.S. Geological Survey,
Menlo Park, California.
Pais, A., and Kausel, E., 1988, “Approximate formulas for dynamic stiffnesses of
rigid foundations,” Soil Dynamics and Earthquake Engineering, Vol. 7,
No. 4, pp. 213-227.
Paulay, T., and Priestley, M.J.N., 1992, Seismic Design of Reinforced Concrete and
Masonry Buildings, John Wiley & Sons, New York, New York.
Pecker, A., and Chatzigogos, C.T., 2010, “Non-linear soil-structure interaction:
Impact on the seismic response of structures,” Earthquake Engineering in
Europe, M. Garevski and A. Ansal (editors), Springer, New York,
pp. 79-103.
PEER, 2010, Guidelines for Performance-Based Seismic Design of Tall Buildings,
Report No. 2010/05, developed by the Pacific Earthquake Engineering
Research Center as part of the Tall Buildings Initiative, University of
Pender, M., 1993, “Aseismic pile foundation design analysis,” Bulletin of New
Zealand National Society of Earthquake Engineering, Vol. 26, No. 1,
pp. 49-160.
Penzien, J., 1970, “Soil-pile-foundation interaction,” Earthquake Engineering, R.L.
Wiegel (editor), Prentice Hall, New York, New York.
Perez-Rocha, L.E., and Aviles, J., 2003, “The evaluation of interaction effects in
inflexible resistances,” Revista de Ingenieria Sismica, No. 69 (in Spanish).
Powell, G.H., 2006, “Static pushover methods – explanation, comparison and
implementation,” Proceedings, 8th U.S. National Conference on Earthquake
Engineering, San Francisco, California.
Priestley, M.J.N., and Park, R., 1987, “Strength and ductility of concrete bridge
columns under seismic loading,” ACI Structural Journal, Vol. 84, No. 1,
pp. 61-76.
Randolph, M.F., 1981, “The response of flexible piles to lateral loading,”

Randolph, M.F., and Wroth, C.P., 1978, “Analysis of deformation of vertically
loaded piles,” Journal of Geotechnical Engineering, Vol. 104, No. 12,
pp. 1465-1488.
Raney Geotechnical, 1983, Foundation Investigation North Main Centre, North Main
St., and Principle Ave., Walnut Creek, CA, Report 029-005, West
Sacramento, California.
Rathje, E.M., Faraj, F., Russell, S., and Bray, J.D., 2004, “Empirical relationships for
frequency content parameters of earthquake ground motions,” Earthquake
Spectra, Vol. 20, No. 1, pp. 119-144.
Raychowdhury, P., 2008, Nonlinear Winkler-Based Shallow Foundation Model for
Performance Assessment of Seismically Loaded Structures, PhD Dissertation,
University of California, San Diego, California.
Raychowdhury, P., and Hutchinson, T.C., 2008, “Nonlinear material models for
Winkler-based shallow foundation response evaluation,” GeoCongress 2008:
Characterization, Monitoring, and Modeling of GeoSystems, American
Society of Civil Engineers, Reston, Virginia.
Raychowdhury, P., and Hutchinson, T.C., 2009, “Performance evaluation of a
nonlinear Winkler-based shallow foundation model using centrifuge test
results,” Earthquake Engineering and Structural Dynamics, Vol. 38, No. 5,
pp. 679-698.
Richart, F.E., Jr., and Whitman, E.V., 1967, “Comparison of footing vibration tests
with theory,” Journal of Soil Mechanics and Foundations Division, Vol. 93,
No. 6, pp. 143-168.
Riggs, H.R., and Waas, G., 1985, “Influence of foundation flexibility on soil-
structure interaction,” Earthquake Engineering and Structural Dynamics,
Vol. 13, No. 5, pp. 597-615.
Roesset, J.M., 1980, “Stiffness and damping coefficients of foundations,”
Proceedings, ASCE Geotechnical Engineering Division National
Convention, pp. 1-30.
Rosebrook, K.R., and Kutter, B.L., 2001a, Soil-Foundation Structure Interaction:
Shallow Foundations, UCD/CGMDR-01, Center for Geotechnical Modeling,
University of California, Davis, California.
Rosebrook, K.R., and Kutter, B.L., 2001b, Soil-Foundation structure interaction:

Rosebrook, K.R., and Kutter, B.L., 2001c, Soil-Foundation Structure Interaction:
Salgado, R., 2008, The Engineering of Foundations, McGraw-Hill, New York, New
York.
SCEC, 2002, Recommended Procedures for Implementation of DMG Special
Publication 117 Guidelines for Analyzing and Mitigating Landslide Hazards
in California, Southern California Earthquake Center, June 2002,
http://www.scec.org/resources/catalog/hazardmitigation.html.
Scott, R., 1981, Foundation Analysis, Prentice Hall, Upper Saddle River, New Jersey.
Soubra, A.-H., 1999, “Upper-bound solutions for bearing capacity of foundations,”
Journal Geotechnical and Geoenvironmental Engineering, Vol. 125, No. 1,
pp. 59-68.
Sozen, M.A., and Moehle, J.P., 1993, Stiffness of Reinforced Concrete Walls
Resisting In-Plane Shear, Electric Power Research Institute, Research
Project 3094-01, Palo Alto, California.
Stewart, J.P., and Fenves, G.L., 1998, “System identification for evaluating soil-
structure interaction effects in buildings from strong motion recordings,”
Earthquake Engineering and Structural Dynamics, Vol. 27, pp. 869-885.
Stewart, J.P., Fenves, G.L., and Seed, R.B., 1999b, “Seismic soil-structure interaction
in buildings II: Empirical findings,” Journal Geotechnical and
Stewart, J.P., Kim, S., Bielak, J., Dobry, R., and Power, M., 2003, “Revisions to soil
structure interaction procedures in NEHRP design provisions,” Earthquake
Spectra, Vol. 19, No. 3, pp. 677-696.
Stewart, J.P., Seed, R.B., and Fenves, G.L., 1999a, “Seismic soil-structure interaction
in buildings I: Analytical aspects,” Journal Geotechnical and
Stewart, J.P., and Stewart, A.F., 1997, Analysis Of Soil-Structure Interaction Effects
On Building Response From Earthquake Strong Motion Recordings At 58
Sites, Report UCB/EERC-97/01, University of California, Berkeley,
California.
Stewart, J.P., and Tileylioglu, S., 2007, “Input ground motions for tall buildings with
subterranean levels,” Structural Design of Tall and Special Buildings,
Vol. 16, pp. 543-557.

Stokoe, K.H., and Erden, S.M., 1985, Influence of Base Shape on Dynamic Response
of Surface Foundations, Geotechnical Engineering Report GP85-1,
University of Texas at Austin, Texas.
Syngros, K., 2004, Contributions to the Static and Seismic Analysis of Piles and Pile
Supported Bridge Piers Evaluated through Case Histories, Ph.D.
Dissertation, City University of New York, New York.
Thomas, J.M., Gajan S., and Kutter B.L., 2005, Soil-Foundation-Structure
Interaction: Shallow Foundations, UCD/CGMDR-05/02, Center for
Geotechnical Modeling, University of California, Davis, California.
Tileylioglu, S., 2008, Evaluation of Soil-Structure Interaction Effects from Field
Performance Data, Ph.D. Dissertation, University of California, Los
Angeles, California.
Tileylioglu, S., Naeim, F., Alimoradi, A., and Stewart, J.P., 2010, “Impact of
foundation modeling on the accuracy of response analysis for a tall
building,” Proceedings, 9th U.S. National & 10th Canadian Conference on
Earthquake Engineering, EERI and Canadian Association for Earthquake
Engineering, Paper No. 1666.
Tileylioglu, S., Stewart, J.P., and Nigbor, R.L., 2011, “Dynamic stiffness and
damping of a shallow foundation from forced vibration of a field test
structure,” Journal Geotechnical and Geoenvironmental Engineering,
Vol. 137, No. 4, pp. 344-353.
Ugalde, J., 2007, Centrifuge Tests on Bridge Columns on Shallow Square Footings,
Master’s Thesis, University of California, Davis, California.
University of Illinois, 2011, DEEPSOIL V4.0, University of Illinois at Urbana-
Champaign, accessed in 2011, http://deepsoil.cee.illinois.edu/.
Veletsos, A.S., and Meek, J.W., 1974, “Dynamic behavior of building-foundation
systems,” Earthquake Engineering and Structural Dynamics, Vol. 3,
pp. 121-138.
Veletsos, A.S., and Nair, V.V., 1975, “Seismic interaction of structures on hysteretic
foundations,” Journal of Structural Engineering, Vol. 101, pp. 109-129.
Veletsos, A.S., and Prasad, A. M., 1989. “Seismic interaction of structures and soils:
Stochastic approach,” Journal of Structural Engineering, Vol. 115, No. 4,
pp. 935-956.
Veletsos, A.S., Prasad, A.M., and Wu, W.H., 1997, “Transfer functions for rigid
rectangular foundations,” Earthquake Engineering and Structural Dynamics,
Vol. 26, No. 1, pp. 5-17.

Veletsos, A.S., and Wei, Y.T., 1971, “Lateral and rocking vibrations of footings,”
Journal of Soil Mechanics and Foundations Division, Vol. 97, No. 9,
pp. 1227-1248.
Vrettos, C., 1999, “Vertical and rocking impedances for rigid rectangular foundations
on soils with bounded non-homogeneity,” Earthquake Engineering and
Structural Dynamics, Vol. 28, pp. 1525-1540.
Vucetic, M., and Dobry, R., 1991, “Effect of soil plasticity on cyclic response,”
Journal of Geotechnical Engineering, Vol. 117, pp. 89-107.
Watson, G.N., 1995, A Treatise on the Theory of Bessel Functions, Cambridge
University Press, United Kingdom.
Western Engineering, 2011, “DYNA6 Software,” Western Geotechnical Research
Centre, Western University, accessed December 2011, http://www.eng.uwo
.ca/research/grc/dyna6.htm.
Wolf, J.P., 1985, Dynamic Soil-Structure Interaction, Prentice-Hall, Upper Saddle
River, New Jersey.
Wolf, J.P., and Von Arx, G.A., 1978, “Impedance function of a group of vertical
piles,” Proceedings, Specialty Conference on Earthquake Engineering and
Structural Dynamics, Pasadena, California, Vol. 2, pp. 1024-1041.
Wong, H.L., 1979, Soil-Structure Interaction: A Linear Continuum Mechanics
Approach (CLASSI Users Manual), Report No. 79-04, Department of Civil
Engineering, University of Southern California, Los Angeles, California.
Wong, H.L., and Luco, J.E., 1985, “Dynamic interaction between rigid foundations
in a layered half space,” Earthquake Engineering and Structural Dynamics,
Vol. 5, No. 3, pp. 149-158.
Wong, H.L., Trifunac, M.D., and Luco, J.E., 1988, “A comparison of soil-structure
interaction calculations with results of full-scale forced vibration test,” Soil
Dynamics and Earthquake Engineering, Vol. 7, pp. 22-31.
Xu, J., Costantino, C., and Hofmayer, C., 2003, Collaborative Study of NUPEC
Seismic Field Test Data for NPP Structures, Report BNL-NUREG-71355-
2003, Brookhaven National Laboratory, Upton, New York.
Yamada, S., Hyodo, M., Orense, R.P., Dinesh, S.V., and Hyodo, T., 2008, “Strain-
dependent dynamic properties of remolded sand-clay mixtures,” Journal
Geotechnical and Geoenvironmental Engineering, Vol. 134, No. 7,
pp. 972-981.

Yu, E., Whang, D.H., Conte, J.P., Stewart, J.P., and Wallace, J.W., 2005, “Forced-
vibration testing of buildings using the linear shaker seismic simulation
(LSSS) testing method,” Earthquake Engineering and Structural Dynamics,
Vol. 34, pp. 737-761.

Project Participants

John (Jack) R. Hayes, Jr. Steven L. McCabe
Engineering Laboratory (MS8604) Engineering Laboratory (MS8604)
National Institute of Standards and Technology National Institute of Standards and Technology
100 Bureau Drive 100 Bureau Drive
Gaithersburg, Maryland 20899 Gaithersburg, Maryland 20899
www.NEHRP.gov www.NEHRP.gov
John (Jay) L. Harris III

Engineering Laboratory (MS8604)
100 Bureau Drive
Gaithersburg, Maryland 20899
www.NEHRP.gov

APPLIED TECHNOLOGY COUNCIL CONSORTIUM OF UNIVERSITIES FOR
201 Redwood Shores Parkway, Suite 240 RESEARCH IN EARTHQUAKE ENGINEERING
Redwood City, California 94065 1301 S. 46th Street, Building 420
www.ATCouncil.org Richmond, California 94804
www.CUREE.org
Joint Venture Management Committee

James R. Harris Christopher Rojahn
J.R. Harris & Company Applied Technology Council
1775 Sherman Street, Suite 1525 201 Redwood Shores Parkway, Suite 240
Denver, Colorado 80203 Redwood City, California 94065
Robert Reitherman Andrew Whittaker

Consortium of Universities for Research in Dept. of Civil, Structural, and Environ. Engin.
Earthquake Engineering 230 Ketter Hall
1301 S. 46th Street, Building 420 University at Buffalo
Richmond, California 94804 Buffalo, New York 14260
Joint Venture Program Committee

Jon A. Heintz (Program Manager) Michael Constantinou
Applied Technology Council Dept. of Civil, Structural, and Environ. Engin.
201 Redwood Shores Parkway, Suite 240 University at Buffalo
Redwood City, California 94065 132 Ketter Hall
Buffalo, New York 14260
GCR 12-917-21 Project Participants F-1

C.B. Crouse Jack P. Moehle
URS Corporation University of California Berkeley
1501 4th Avenue, Suite 1400 325 Davis Hall – MC 1792
Seattle, Washington 98101 Berkeley, California 94720
William T. Holmes James R. Harris (ex-officio)

Rutherford & Chekene Andrew Whittaker (ex-officio)
55 Second Street, Suite 600
San Francisco, California 94105
Project Manager
David A. Hutchinson
Buehler & Buehler, Structural Engineers
600 Q Street, Suite 200
Sacramento, California 95811
Project Technical Committee

Jonathan P. Stewart (Project Director) Bret Lizundia
University of California, Los Angeles Rutherford & Chekene
Dept. of Civil and Environmental Engineering 55 Second Street, Suite 600
5731 Boelter Hall San Francisco, California 94105
Los Angeles, California 90095
Farzad Naeim
C.B. Crouse John A. Martin & Associates, Inc.
URS Corporation 950 S. Grand Avenue, 4th Floor
1501 4th Avenue, Suite 1400 Los Angeles, California 90015
Seattle, Washington 98101
Farhang Ostadan
Tara C. Hutchinson Bechtel Corporation
University of California, San Diego 50 Beale Street
Department of Structural Engineering San Francisco, California 94105
9500 Gilman Drive
La Jolla, California 92093
Working Group Members

Fortunato Enriquez Silvia Mazzoni
California State University, Chico Degenkolb Engineers
475 East 10th Avenue 235 Montgomery Street, Suite 500
Chico, California 95926 San Francisco, California 94104
Michael J. Givens Erik Okstad

University of California, Los Angeles Rutherford & Chekene
5731 Boelter Hall 55 Second Street, Suite 600
Los Angeles, California 90095 San Francisco, California 94105
Curtis B. Haselton Andreas Schellenberg

California State University, Chico Rutherford & Chekene
Department of Civil Engineering 55 Second Street, Suite 600
475 East 10th Avenue San Francisco, California 94105
Chico, California 95926
F-2 Project Participants GCR 12-917-21

Project Review Panel
Craig D. Comartin Gyimah Kasali
CDComartin Inc. Rutherford & Chekene
7683 Andrea Avenue 55 Second Street, Suite 600
Stockton, California 95207 San Francisco, California 94105
Youssef Hashash George Mylonakis

University of Illinois, Urbana-Champaign University of Patras, Greece
Dept. of Civil and Environmental Engineering 26500 Patras, Greece
205 North Mathews Avenue
Urbana, Illinois 61801 Graham Powell
Graham H. Powell, Inc.
Annie M. Kammerer 14710 Wolfgang Road
U.S. Nuclear Regulatory Commission Truckee, California 96161
Office of Nuclear Regulatory Research
Washington, D.C. 20555
Workshop Participants
Peter Behnam Robert D. Hanson
KPFF Consulting Engineers 2926 Saklan Indian Drive
6080 Center Drive, Suite 300 Walnut Creek, California 94595
Doug Hohbach
Lauren Carpenter Hohbach-Lewin
WHL Consulting Engineers, Inc. 260 Sheridan Avenue, Suite 150
350 South Figueroa Street, Suite 400 Palo Alto, California 94306
Ben Hushmand
Michael Cochrane Hushmand Associates, Inc.
Weidlinger Associates, Inc. 250 Goddard Road
4551 Glencoe Avenue, Suite 350 Irvine, California 92618
Marina del Rey, California 90292
Peter Lee
John Gavan Skidmore, Owings & Merrill LLP
KPFF Consulting Engineers One Front Street, Suite 2500
6080 Center Drive, Suite 300 San Francisco, California 94111
Anne Lemnitzer
Tony Ghodsi University of California, Los Angeles
Englekirk Structural Engineers Dept. of Civil and Environmental Engineering
925 Fort Stockton Drive, Suite 202 5731 Boelter Hall
San Diego, California 92103 Los Angeles, California 90095
Christine Goulet Marshall Lew

Pacific Earthquake Engineering Research Center MACTEC Engineering & Consulting, Inc.
325 Davis Hall, University of California 5628 E. Slauson Avenue
Berkeley, California 94720 Los Angeles, California 90040
GCR 12-917-21 Project Participants F-3

Weian Liu Michael Mehrain
University of California, San Diego URS Corporation
Jacobs School of Engineering 915 Wilshire Boulevard, Suite 700
9500 Gilman Drive, SERF 252 Los Angeles, California 90017
La Jolla, California 92093-0403
Navid Nastar
Michael Mahoney University of Southern California
Federal Emergency Management Agency Civil and Environmental Engineering
500 C Street, SW Kaprielian Hall, Room 225
Washington, D.C. 20472 3620 S. Vermont Avenue
Los Angeles, California 90089-2531
David Mar
Tipping Mar Patrick Smith
1906 Shattuck Avenue Fugro West, Inc.
Berkeley, California 94704 4820 McGrath Street, Suite 100
Ventura, California 93003
Neven Matasovic
Geosyntec Consultants
2100 Main Street, Suite 150
Huntington Beach, California 92648
F-4 Project Participants GCR 12-917-21

Soil Structure Interaction For Building

Uploaded by

Copyright:

Available Formats

Soil Structure Interaction For Building

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Soil Structure Interaction For Building

Uploaded by

Copyright:

Available Formats

What is the purpose of this report?

What is the purpose of this report?

What organizations were involved in producing this report?

What organizations were involved in producing this report?

NIST GCR 12-917-21

NEHRP Consultants Joint Venture

Soil-Structure Interaction for

U.S. Department of Commerce

National Institute of Standards and Technology

NEHRP Consultants Joint Venture

Joint Venture Management Joint Venture Program Committee

Project Technical Committee Project Review Panel

Working Group Members Project Manager

The NEHRP Consultants Joint Venture is a partnership between the Applied

Soil-structure interaction (SSI) analysis evaluates the collective response of three

GCR 12-917-21 Preface iii

NEHRP Consultants Joint Venture also gratefully acknowledges Jack Hayes

iv Preface GCR 12-917-21

Preface ..................................................................................... iii

List of Figures ............................................................................. ix

List of Tables............................................................................. xix

1. Introduction ..................................................................... 1-1

2. Inertial Interaction ............................................................ 2-1

3. Kinematic Interaction ......................................................... 3-1

4. Implementation in Standards and Design Guidelines ................... 4-1

5. Experiments and Case Studies .............................................. 5-1

GCR 12-917-21 Table of Contents v

6. State of Practice ............................................................... 6-1

7. Example Applications ......................................................... 7-1

vi Table of Contents GCR 12-917-21

Appendix A: Sherman Oaks Building Model Development ...................... A-1

Appendix B: Sherman Oaks Building Stick Model Development ............... B-1

Appendix C: Walnut Creek Building Model Development ....................... C-1

Symbols ................................................................................... D-1

References ............................................................................... E-1

Project Participants .................................................................... F-1

GCR 12-917-21 Table of Contents vii

Figure 1-1 Schematic illustration of a direct analysis of soil-structure interaction

Plot of period lengthening ratio ( T T ) and foundation damping (f )

rotational radiation damping ratio (rr) for circular foundations

supporting a rigid core and flexible perimeter walls............................ 2-20

GCR 12-917-21 List of Figures ix

loading; and (c) vertical loading .......................................................... 2-26

x List of Figures GCR 12-917-21

GCR 12-917-21 List of Figures xi

xii List of Figures GCR 12-917-21

GCR 12-917-21 List of Figures xiii

xiv List of Figures GCR 12-917-21

GCR 12-917-21 List of Figures xv

xvi List of Figures GCR 12-917-21

GCR 12-917-21 List of Figures xvii

GCR 12-917-21 List of Tables xix

xx List of Tables GCR 12-917-21

The response of a structure to earthquake shaking is affected by interactions between

1.1 Background and Objectives

GCR 12-917-21 1: Introduction 1-1

This report is intended to provide background for addressing three fundamental

2. Under what conditions is consideration of the differences between foundation

3. What field and laboratory investigations are necessary to develop foundation