|
|
DATES: Seminar: Jan. 29, 1998 (afternoon)
Workshop: Jan. 30, 1998 (morning & afternoon)
LOCATION: USC Davidson Conference Center
HOST: Ned Field (field@usc.edu)
DESCRIPTION:
It has been understood for more than 100 years that sediments can increase the level of earthquake ground motion relative to bedrock. However, there has been a long-standing and often contentious debate between seismologists and engineers on whether the response of sediments to strong-ground motion is similar to that of relatively well-studied weak motion. The prevailing view in the engineering community, based almost exclusively on laboratory studies, is that sediment response is nonlinear. That is, amplification factors are generally reduced for stronger ground motion because the finite strength of sediments causes a breakdown of Hooke's law. This perspective has been applied in engineering practice. Seismologist, on the other hand, have traditionally been skeptical because of a lack of evidence and a skepticism that laboratory studies represent in situ behavior. They've generally concluded that either sediment nonlinearity is insignificant, or that it is buried among the myriad of other complicating factors (i.e. uncertainties) in the data. Recent progress in several disciplines makes the time ripe for a seminar & workshop on this problem.
At the seminar (afternoon of Jan. 29, 1998), representatives from each discipline will give general overview talks on the following topics:
The purpose of this seminar is to both educate the general SCEC community, and to bring the members from the various disciplines up to speed with respect to the other disciplines.
With this introduction, the workshop on the next day will focus on specific technical issues that remain unresolved. Participants will be invited to present and discuss any results that pertain to these particular issues. Given the unprecedented diversity of disciplines that will be in attendance, it is hoped that this workshop will establish points of agreement and disagreement, stimulate crossbreeding, and identify priorities for future research.
If you are interested in participating in the workshop, please contact Ned Field (213-740-7088; field@usc.edu) to let him know. Also, please include your thoughts on what issues should be addressed at the workshop.
Seminar Agenda:
---------
1:00 "Introduction - Site Response
101"
by Ned Field, USC
1:20 "Soil Behavior Under Dynamic &
Cyclic Loads"
by Mladen Vucetic, UCLA
1:55 "Geotechnical Modeling of Site
Response - An Historical Perspective"
by Geoff Martin, USC
2:30 "Seismological Evidence For and
Against Nonlinear Site Effects"
by John Anderson, UNR
3:05 Break
3:25 "Rock Physics and Dynamic Nonlinear
Elastic Response"
by Paul Johnson, LANL
4:00 "Site-Amplification Factors as
Incorporated in Building Codes"
by Roger Borcherdt, USGS
4:25 "Perspectives from Japan"
by Susumu Iai, Ministry of Transportation, Japan
Workshop titles and abstracts:
"Description of the nonlinear behavior of geomaterials for site response analysis"
by J.P. Bardet of the University of Southern
California
From the point of view of mechanics, the characterization and description of material properties are the key issues of site response analysis in engineering seismology. Many stress-strain models in these studies (e.g., linear equivalent) are simplistic one-dimensional formulations, and as such do not describe accurately shear strength, nor can they be generalized to two- or three-dimensional problems without drastic assumptions. A few sophisticated constitutive models have been proposed to generalize the nonlinear and cyclic behavior of soils and rocks, based on the stress-strain curves measured from laboratory experiments (e.g., multiple yield plasticity and bounding surface plasticity). However, these generalized models have serious shortcomings for site response analysis applications. Some produce undesirable effects in cyclic responses (e.g., artificial ratchetting), while others have a complicated mathematical formulation, require numerous model constants, and are difficult to implement in computer codes. These models unfortunately have limited applicability and reliability for site response analysis.
Since 1994, we have developed a realistic and practical description of material behavior for site response analysis, and have introduced a constitutive model based on the novel concept of scaled memory (SM). SM transforms the nonlinear plastic modulus into a piecewise-linear distribution, and then uses this simplified distribution for generating the hysteretic stress-strain loops during cyclic loadings. SM generalizes the models of Ramberg-Osgood and Hardin-Drnevich, and is simpler than, but as capable as, multiple yield surface plasticity. SM also generalizes bounding surface theory, and corrects its ratchetting problems.
In this SCEC presentation, we present the basic concepts of the SM theory, and illustrate its usefulness by simulating several cyclic stress-strain responses for clays and sands. The model is presently implemented in a computer program for one-dimensional site response analysis, which has a principle similar to that of SHAKE.
----------------------
"Seismic Response of Stiff Soil Sites During Recent California Earthquakes"
by Jonathan D. Bray.
The talk presents a summary of recently completed studies of site-dependent seismic site response incorporating the wealth of strong motion data provided by the 1989 Loma Prieta and 1994 Northridge Earthquakes. The empirical data, results of back-analyses of various strong motion recording sites, and analyses of the response of the sites to various design levels of shaking are combined with the results of previous similar analyses to develop recommendations for site classification, prediction of site-dependent amplification, and site-dependent design spectra. The adequacy of current U.S. building codes in addressing site-dependent site response is assessed in light of the strong motion data from these recent earthquakes. It is found that the current codes underpredict the intensity of shaking at stiff soil sites. In fact, in terms of PGA, the data suggest that stiff soil sites do not exhibit has much nonlinearity at high levels of shaking that the current UBC site factors indicate.
------------------------
"The Turkey Flat (near Parkfield) and Coalinga Experiences"
by Chris Cramer
The geotechnical site characterization experience at Turkey Flat and the effect of finite sources on the 1983 Coalinga aftershock strong motion data will be presented and discussed. Modeling is most sensitive to geotechnical site characterization variability in Vs and damping (Field and Jacob, 1993; Cramer, 1995). There are significant differences between laboratory and in situ damping measurements at Turkey Flat. Laboratory based measurements suggest that soil damping at Turkey Flat should be about 1.5% for weak motions while in situ measurements suggest a soil damping of ~7%. Modeling of weak-motion response indicates soil damping is 7% +/- 2% which implies, for strong motion data, little effect is to be expected because damping is already high. Additionally, concerning Vs measurement resolution, modeling suggests Vs resolution within 5% is needed to match response, but observed downhole Vs resolution is only 10-20%. Vs in situ measurements may not have sufficient resolution for modeling. Finally, observed strong motion effects due to directivity affecting both stations vs only one station in an observing station pair at Coalinga suggest that a finite source effect can look like a non-linear effect.
---------------------------
"Nonlinear effects in vibrator-induced ground-surface motion"
by Petros Dimitriu
Small-size structural vibrators are ideal sources to study sediment nonlinearity. They are cheap, controlable and capable of exciting the ground at various levels and at frequencies of engineering interest (typically from several Hz to several tens of Hz). By using two such vertical vibrators, not only was I able to reproduce two basic nonlinear effects (i.e. harmonic distortion of a sine signal and generation of combinational components -sum and difference-during interaction of two wavefields), but also to observe a qualitatively new class of nonlinear phenomena: modulational instability (self-modulation) and recurrence. Modulational instability of uniform sinusoidal wavetrains (chiefly Rayleigh waves, as it is known that during vertical excitation of the ground surface about 70% of the energy goes into them) occurred at sufficiently LOW excitation amplitudes (strains typically less than 10^-5) and distances from the vibrator of the order of several wavelengths. The phenomenon was confirmed in subsequent tests using a large (~10-ton) vertical prospecting vibrator, driven at a small fraction (~10%) of the full capacity, and at somewhat larger distances (~10 wavelengths). Qualitatively similar behaviour was earlier observed experimentally in water wavetrains and is known in nonlinear physics as the Fermi-Pasta- Ulam recurrence phenomenon. According to nonlinear theory, recurrent evolution can occur in nonlinear dispersive wavetrains whose envelope is subjected to perturbation (modulational instability) and involves the generation of SOLITONS, both in the carrier and the envelope.
--------------------------
"Identification of site response using downhole array seismic records"
by Ahmed-W. Elgamal
The dynamic shear properties of soils at a number of instrumented sites is identified from recorded downhole seismic accelerations. Properties from the Lotung and Hualien (Taiwan) sites, the Kobe Port Island (Japan) site, and the Treasure Island (California) site are identified from small tremor as well as large shaking events. Histories of the actual seismic shear stress-strain histories will be presented and discussed.
------------------------
"The Effects of Spectral Content on Non-linear Site Response"
by Eric Jones
I will discuss our work on modeling non-linearity for the Northridge event and for a hypothetical rupture of the Elysian Park Fault. We have done simulations with elastic-plastic material models and with the Masing Rule and get qualitatively similar results. The most striking feature of the results is that nonlinearity is important at the frequencies at which strain reversals occur and not at lower frequencies. That is, for a hypothetical seismic signal with two spectral peaks, such that the high frequency amplitude is great enough to produce strain reversals, the effects of non-linearity will be seen at the higher frequency but not at the lower. If this phenomenon occurs in nature, then conclusions about the importance of non-linearity will depend, at least in part, on the spectral content of the incoming signal and on the frequencies at which observations are made.
------------------------
"Needed Modifications in Methods for Equivalent Linear Soil Response Calculations"
by William B. Joyner
------------------------
"Evidence for Nonlinear Soil Response at the Van Norman Complex Following the 1994 Northridge, California, Earthquake"
talk by William Joyner, study by Giovanna Cultrera, David M. Boore, and William B. Joyner.
------------------------
"Stress-Induced Anisotropy and Nonlinear Waves in Damaged Rocks"
by V.Lyakhovsky (USC + Hebrew University of Jerusalem) & Y.Hamiel (Hebrew University of Jerusalem)
The macroscopic effects of distributed cracking and other types of damage requires constitutive models which exhibit nonlinear elasticity caused by opening and closure of cracks. To evaluate this effect we derive the elastic energy function that includes a non-analytical, second order term with third elastic modulus responsible for coupling between shear strain and dilatancy. The suggested theoretical model allows the differing properties of the solid under compression and tension, and variation of the effective elastic moduli with type of loading. These type of behavior has been directly observed during deformational four point beam test.
Theoretical stress-strain relations for 3-D elastic solid are used for simulation of small amplitude acoustic waves and of finite amplitude waves.
In the first case the model predicts the relation between acoustic wave velocity and state of stress. The calculated stress-strains and stress-induced velocities fit well the experimentally measured deformation and velocities in Barea sandstone.
In the second case, the model predicts a generation of additional harmonics away from the source as well as shift of a resonance frequency as observed in different laboratory and field experiments.
------------------------
"Seismic response of weak rock"
by Neven Matasovic
Briefly, will discuss some problems associated with the modeling
of *weak rock* in two-dimensional seismic site response analyses.
The *weak rock* is a term commonly used by seismologists for various
*stiff* geologic formations in southern California. While these
*stiff* formations are relatively well defined in seismological
terms, they are not well defined in the terms of the geotechnical
(earthquake) engineering. In particular, little is known about
the nonlinear and hysteretic characteristics of these materials
(modulus reduction and damping). This in turn poses some problems
associated with numerical modeling of earthern structures embedded
in weak rock. I will discuss these problems, offer some solutions
for discussion. My presentation will be supported with at least
two examples from my local practice where various alternatives
of modeling of the weak rock were found to significantly influence
the results.
------------------------
"Observed 'Rock' Nonlinearity at the Lucerne Valley Site"
by Bob Nigbor
I'll discuss a PG&E funded study of site response at the Lucerne Valley strong motion station, which included microtremor measurements, weak motion measurements, geophysics, and analysis. A frequency shift is observed between weak and strong motion which corresponds to a 50% loss of stiffness at this rock site with shallow (10m) weathering.
------------------------
"Two Vertical Arrays in the Los Angeles Region: An Experiment to Predict
High-Strain Site Response"
by Albert M. Rogers
This project was established with the following goals: (1) to complete borehole drilling and installation of two vertical strong-motion arrays in the Los Angeles region; (2) to obtain borehole geotechnical and lithologic data at the sites; (3) to provide strain-dependent site response estimates, based on these data, using linear, equivalent linear, and fully non-linear one-dimensional codes; (4) to compare the predicted site response estimates with observations using future strong-motion recordings. The vertical arrays are located at Cerritos College in Norwalk and at the central fire station near downtown San Bernardino, California. The geologic conditions at these sites span a range of Los Angeles basin deposits from gravelly alluvial fan deposits to overbank deposits of sands and silts.. The first two phases of this project, concerning the site geotechnical properties and instrumentation, are the subject of this talk. The strong-motion stations are instrumented with Quanterra 6160 data loggers, Kinemetrics FBA 13 surface accelerometers, and FBA 23DH borehole instruments. The data are recorded on internal hard-disk drives in both continuous and event trigger modes and simultaneously transmitted by spread spectrum radio to the digital recording facility at California Institute of Technology. The data loggers provide 160 Hz sample rates and dynamic range sufficient to record ground motions from about the city noise level to 2g. The borehole instruments are at 100 meters below the surface. Two additional surface accelerometers form a triangular array on the college campus to study wave direction of approach, and a linear array of equally spaced accelerometers extends north-northeast to Whittier, from the deepest part of Los Angeles basin to the basin edge. Shear- and P-wave velocities and lithologic logs have been obtained for the boreholes. The San Bernardino site is underlain by thin fill, gravelly sand, sandy gravel, and increasing percentage of gravel with depth. Shear-wave velocities vary from about 287 m/s at the surface to 481 m/s at about 60 meters and below. The Cerritos College site is underlain by loose fine sand, soft silt with clay, dense silt and sand, sand and gravel, and compact gravels with clay and silt binder interbedded with tens of feet of silty clays and clayey silts. Shear-wave velocities vary from about 200 m/s near the surface to 445 m/s at 80 meters and below.
------------------------
"Issues in 1- D Site Response"
by Walt Silva
The following issues will be discussed:
1. Observations of Soil Nonlinearity from Strong Ground Motions,
2. Development of G/Gmax and Hysteretic Damping curves using Strong Ground Motions,
3. Validation of Conventional Vertical Propagating Shear-wave Model, equivalent-linear, and nonlinear approaches,
4. Validation of laboratory based nonlinear properties for a range in site conditions,
5. Nonlinearity in Rock Verses Soil,
6. Nonlinearity, P/SV coupling, and approaches to modeling Vertical Motions at soil sites,
7. Issues in developing design ground motions at soil sites; NRC and DOE perspective,
8. Objectives, purpose, and scope of ROSRINE 1, 2, 3, and 4.
------------------------
"Soft Soil to Stiff Rock: Laboratory Evalaution of Behavior in the Linear and Nonlinear Ranges."
by Kenneth H. Stokoe, II
------------------------
"Nonlinear site resposne at Tarzana derived from the Northridge earthquakes"
by Yuehua Zeng
Since the Tarzana station has recorded the largest peak acceleration so far from the Northridge earthquake, it has attracted much attention among engineers and seismologists. Spudich et al. (1996) have studied the site effect of the hill using array measurement across the area. While obtaining the largest topographic amplification perpendicular to the strike of the hill, they were puzzled by the fact that the largest mainshock ground motion occurred along the east-west direction. We added the weak motion site amplification function derived from the aftershock recordings (Su et al., 1995, 1996) into the synthetic simulation using a composite source model. We found the seismic motions along the east-west direction are about comparable to the observation. However, the synthetic seismogram on the north-south component after site response correction has a peak ground acceleration several times higher than the observation. We resolve the problem to nonlinear site response. We assumed a nonlinear shear modulus reduction relation with strain, which we believe is suitable to nonlinear rock behavior (Stokoe et al., 1997), for the shallow surface layer beneath the station, and computed the nonlinear response of the model to the input ground motion. We found that a model for nonlinear resposne reduces the acceleration of the north-south to a level comparable to the observation. The same nonlinear effect has much smaller influence on the east-west component because this direction is nearly nodal to the source radiation and it has relatively lower level of ground motion. Thus a model invoking nonlinear rock behavior in concert with otherwise high site effects provides a simple explanation of ground accelerations that is consistent with the well-documented source mechanism.
Phone 213/740-5843
Fax 213/740-0011
e-mail: SCECinfo@usc.edu
Last Updated: