Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 34
24
CHAPTER FIVE
CONCLUSIONS AND SUGGESTIONS FOR FURTHER RESEARCH
CONCLUSIONS lowed by FLAC. FLAC and ABAQUS are the most widely
used for nonlinear (i.e., bi-linear) 2-D and three-dimensional
Since their introduction in the 1970s, tools and techniques (3-D) site response analyses.
for performing site response analysis have continued to
evolve. Advances in both computer hardware and software The survey respondents recognize that analytical proce-
have played a role in improving analyses. Software devel- dures have developed much more quickly than the practical
opments have resulted in wide and almost exclusive use of procedures for developing the input parameters those ana-
site response analysis in practice, as revealed by the survey lytical procedures require. The literature search performed
performed as a part of this study. Moreover, it appears that as a part of this study reveals a lack of practical guidance
practitioners can now afford and do perform sensitivity documents for multidimensional 2-D and especially 3-D site
analyses and use larger suites of input ground motions. Nev- response analysis.
ertheless, significant confusion remains about how to select
appropriate input ground motions and the number of ground Dynamic soil properties appear to be most commonly
motion sets needed. Automated graphics capabilities allow determined by field testing and empirical correlation.
error-checking of input data and evaluation of the reason- Respondents cited uncertainties in input motions as the great-
ableness of analytical results. Ground response animation, est influence on site response analyses. Finally, sensitivity
available in some software, can provide useful insight into analyses appeared to be the most common tool for evaluating
site response. the effects of uncertainties on computed site response.
Other improvements to site response analysis practice It appears that 1-D equivalent-linear analyses have
have resulted from the development of more advanced ana- become the de facto standard for site response analysis of
lytical models, particularly for nonlinear, effective-stress highway facilities. It also appears that users have concerns
modeling of site response. Multidimensional nonlinear anal- about the applicability of equivalent-linear analyses in cases
ysis software employing advanced, plasticity-based consti- where site-specific response analyses are most useful: soft
tutive models are now available. The survey indicates that sites, liquefiable sites, and sites subjected to very strong
one-dimensional (1-D), equivalent-linear analyses are by far shaking. The DOT adoption of nonlinear analyses, however,
the most commonly used by departments of transportation appears to be restrained by uncertainty in how to develop
DOTs surveyed, but this is not the case among their consul- the input parameters required for available nonlinear mod-
tants and researchers. The most commonly used program els, and by the lack of well-documented validation studies
for equivalent-linear analysis is SHAKE2000, followed by for those models. Despite these limitations, more than half of
SHAKE and DEEPSOIL. The survey further reveals that a the respondents use nonlinear site response analyses. Never-
variety of modulus reduction and damping curves are used theless, consistent with findings by Kwok et al. (2007), these
to represent equivalent-linear and nonlinear dynamic prop- limitations of nonlinear codes must be overcome for non-
erties of soil. linear site response analysis to become more widespread in
geotechnical earthquake engineering practice.
It appears that nonlinear models for 1-D analyses are
becoming more common in DOT practice, especially among
their consultants. The most commonly used program for SUGGESTIONS FOR FURTHER RESEARCH
both nonlinear and nonlinear effective-stress analysis is
D-MOD2000, followed by DEEPSOIL and FLAC. Multidi- The results documented here reveal a mismatch between
mensional nonlinear analyses and multidimensional effec- current practice, as applied for highway facilities in T-3+
tive-stress analyses are not used often. Survey participants states, and the state-of-the-art, as implied from the literature
cite the engineering time required to develop multidimen- search. This mismatch between state-of-the-practice (SOP)
sional models as the main reason for their sparse use. The and state-of-the-art (SOA) is common in all engineering
most commonly used program for two-dimensional (2-D) disciplines. This gap might be narrowed by comparing the
equivalent-linear site response analysis is QUAD4M, fol- survey results to findings of the literature search to iden-
OCR for page 35
25
tify needs and opportunities. Those that appeared the most they can represent the soil constitutive behavior such that the
urgent and the most valuable to survey respondents are out- implied strength is reasonable and corresponds to that of the
lined below in a relative order of importance. soil profile during shaking.
Benchmarking of One-Dimensional Site Response Benchmarking of Multidimensional Total and Effective-
Analysis with Pore Water Pressure Generation Stress Site Response Software
A landmark benchmarking project was performed in The survey and literature review show increasing reliance
2006/2007 by PEER (Kwok et al. 2007). The project on 2-D analysis software such as FLAC, OpenSees, and
involved benchmarking available total stress nonlinear PLAXIS to evaluate site response and liquefaction. There
site response analysis computer software against theoreti- is a pressing need for a rigorous benchmarking study of the
cal problems, but also included conducting back analysis of software being used and usage protocols. This research topic
relevant case histories. The results of that study, described could be considered after 1-D effective-stress site response
in a series of presentations and technical papers, speeded analysis has been benchmarked.
up practical application of nonlinear analysis to the lev-
els identified in this survey. (The respondents used non- Benchmarking of Vs Correlations and Evaluation
linear methods to conduct approximately half of 1-D site
response analyses.) This survey revealed that most respondents use correlations
between standard penetration test (SPT) blow counts and Vs
Given the importance of use on nonlinear effective- to develop the Vs profile for site response analysis. Although
stress analysis for site class E (soft soils) and site class F this is generally poor practice, it cannot be entirely avoided
(liquefiable soils and very soft clays in the profile), and its given the large legacy SPT data sets available to many DOTs
gradual increase in use, a rigorous benchmarking study and their (erroneous) belief that direct Vs measurements are
of 1-D nonlinear software with pwp generation should cost prohibitive.
be conducted.
Given the importance of the Vs profile in site response
Threshold for Equivalent-Linear Versus Nonlinear Site analyses, a systematic study of Vs SPT and other correla-
Response Analysis tions (e.g., Vs qc and Vs Su) could be undertaken. The
work would involve a comparison of shear wave velocity
Nonlinear site response analyses generally demand more profiles established by correlative expressions to the results
resources and technical expertise than their equivalent- of actual measurements. It is important that the findings
linear counterparts. Therefore, practicing engineers have of such a study clearly identify which correlations may
a keen interest in the development of criteria for deciding be recommended for given site conditions (e.g., should an
when an equivalent-linear site response analysis is accept- SPT-based correlation be used when site-specific results of
able and when nonlinear site response is necessary. The Su measurements are available?) and the possible range of
available guidance is not sufficient and a systematic study is error. The study would also identify available Vs measure-
needed to develop such guidance. ment tools and the relative merits of various Vs measure-
ments, including downhole, sCPT, OYO suspension logging,
Input Ground Motion Selection ReMi, SASW, and MASW.
The selection of input ground motion time histories for site Evaluation of Liquefaction from Site Response Analysis
response analysis remains a challenging task. A detailed
study should develop guidelines for the selection of seed The survey revealed significant confusion on the part of
ground motions, scaling, and spectral matching specific to the users with regard to the use of site response analysis
site response analysis. Such a study should consider issues, for liquefaction evaluation either directly from models that
including the appropriateness of the input and propagated generate pwp or indirectly through the use of the simplified
motions for a number of engineering applications, including method for liquefaction analysis. Guidance on this issue
liquefaction assessment, bridge design, slop stability, and would be greatly welcomed by the profession.
soil-structure interaction analysis. This research can ben-
efit from a number of recent studies on the ground motions Site Response in Deep Deposits
selection for structural applications.
A challenging issue in site response analysis in deep soil
Implied Strength in Modulus Reduction Curves deposits is how to select the depth of the soil column to be
considered because the significant impedance contrast can
In site response analyses that mobilize large strains, widely be several kilometers deep. Guidance is needed on conduct-
used modulus reduction curves need to be improved so that ing site response analysis of deep deposits.
OCR for page 36
26
Vertical Site Response Evaluation aftershocks. This data set provides a unique opportunity to
validate and improve site response analysis models. A study
In dynamic analysis of structures and soil structure inter- that focuses on the use of this data set will be very useful for
action, 3-D motions are required as input. This survey increasing the reliability of site response analysis procedures,
found that the overwhelming majority of work done on site especially for long duration earthquakes, and the proper rep-
response analysis is related to horizontal motion. The lit- resentation of cyclic soil behavior under repeated cycles of
erature is insufficient on how to handle local site effects on loading, which has been studied only in the laboratory.
vertical ground motion propagation. Currently, PEER has an
effort focused on vertical site response as part of an update Verification and Validation of Software
of NGA-West. A detailed study on vertical site response
would be timely and fill a major gap in the body of knowl- As DOTs and/or their consultants adopt specific software,
edge in site response. or new and improved software becomes available, a need
emerges for software use guidance and for software verifi-
Calibration of Nonlinear Site Response Analysis from cation and validation procedures. The experience from the
Recent Japan Earthquake recent PEER-sponsored 2G02 Project (Stewart et al. 2008)
indicates that these procedures should be developed by a
The March 2011 earthquake in Japan has provided the team of software developers under the guidance of indepen-
research community with an extensive data set from mul- dent third parties. The survey and literature review reveal
tiple large events. This data set includes a significant num- the absence of such software use guidance and software ver-
ber of downhole arrays (KiK-net) that have been excited by ification and validation procedures, which causes significant
the main shock as well as by several large foreshocks and confusion on the part of users.
