Geotechnical Earthquake Engineering and Soil Dynamics IV

One-dimensional seismic ground response analyses are often performed using equivalent-linear procedures, which require few, generally well-known parameters. Nonlinear analyses have the potential to more accurately simulate soil behavior, but their implementation in practice has been limited because of poorly documented and unclear parameter selection and code usage protocols as well as inadequate documentation of the benefits of nonlinear modeling relative to equivalent linear modeling. Regarding code usage/parameter selection protocols, the following are described: (1) when input motions are from ground surface recordings, we show that the full outcropping motion should be used without converting to a "within" condition; (2) Rayleigh damping should be specified using at least two matching frequencies with a target level equal to the small strain soil damping; (3) the "target" soil backbone curves used in analysis can be parameterized to capture either the soil's dynamic shear strength when large-strain soil response is expected (strains approaching 1%), relatively small-strain response (i.e., γ < 0.3%) as inferred from cyclic laboratory tests, or a hybrid of the two; (4) models used in nonlinear codes inevitably represent a compromise between the optimal fitting of the shapes of backbone and hysteretic damping curves, and we present two alternatives for model parameterization. The parameter selection and code usage protocols are tested by comparing predictions to data from vertical arrays. We find site amplification to be generally underpredicted at high frequencies and overpredicted at the elastic site period, where a strong local resonance occurs that is not seen in the data. We speculate that this bias results from over-damping.

The current state-of-the-art ground motion models for application to the United States is summarized. The recent NGA ground motion models for shallow crustal earthquakes are in much more advanced than the currently available models for subduction earthquakes and for stable continental regions. The models for the subduction and stable continental regions will likely change significantly in the next several years as major studies are being started to improve these models. There is a move to using the larger horizontal component as compared to the average horiontal component in US building codes to define the design ground motion. A key issue for all ground motion models is the standard deviation. In particular, the standard use of the ergodic assumption tends to overestimate the standard deviation by about 40%, but removing the ergodic assumption leads to increased complexity in the models for the median ground motion as increased epistemic uncertainty in the median for a specific source/site combination is needed. In the future, there will be a greater reliance of finite-fault numerical simulations to estimate the ground motion, but there is a large range of results from different models that needs to be resolved before the can be wide-spread application of these approaches.

As non-linear response history analyses are becoming more prevalent in practice, there is a need to better understand how the selection and modification (e.g., amplitude scaling or spectrum matching) of records will influence the resulting structural response predictions. There are currently many methods of ground motion selection and modification available, but little guidance is available to engineers on which methods are appropriate for their specific application. The Ground Motion Selection and Modification Program was formed within the Pacific Earthquake Engineering Research (PEER) Center to address this issue. This paper presents the current methodology developed by the Program as well as sample results from the first pilot study completed in 2006. Preliminary results show that for a first-mode-dominated structure, one can improve the prediction of its response by taking into account record properties that are important to the non-linear response of the building when selecting and scaling ground motion records.

Pulse-like near-fault ground motions resulting from directivity effects are a special class of ground motions that are challenging to characterize for seismic performance assessment. These motions contain a pulse in the velocity time history of the motion, often occurring in the direction perpendicular to the fault rupture at locations near the fault where the earthquake rupture has propagated towards the site. A recently proposed wavelet-based signal processing approach is used on a large ground motion library to empirically identify these pulses in ground motions. Example results are presented to demonstrate that the identified motions are often observed at sites where directivity effects are expected (although no claim is made that all observed pulses are due to directivity). The response spectra of these records are then studied using this approach, and it is seen that their spectra can be described using an existing ground motion prediction (attenuation) model coupled with a narrow-band amplification function in the region of the pulse period. The modified prediction can be incorporated in probabilistic seismic hazard analysis, providing a direct and transparent method of accounting for directivity effects.

A model for horizontal peak ground strain (PGS) is developed in consideration of three fundamental contributions to spatially variable ground motion (SVGM): (1) spatial incoherence effects, which contribute to phase variability in a stochastic sense; (2) wave passage effects, which contribute to phase variability in a deterministic sense; and (3) amplitude variability. Previous models for each of these effects are reviewed and compared to array data from Borrego Valley, California. Published empirical models for coherency and amplitude variability are found to represent reasonably well the Borrego data. We extend previous work by considering correlations of amplitude and phase variability (generally found to be small) and characterizing the coherency-dependent probabilistic distribution of phase variability. Using the aforementioned amplitude and phase variability models, a procedure is developed to generate simulated acceleration records from a seed record. The procedure is applied to a suite of Northridge earthquake recordings to predict ground strains, which are found to be strongly dependent on the peak ground velocity (PGV) of the seed motion and the separation distance between the seed and simulated motions. The dependence of PGS on PGV saturates for large PGV (> 50 cm/sec).

When earthquake engineering is concerned with both the amplitude and durational aspects of ground motion, it commonly uses seismic loading in the form of a time history. The translation of traditional probabilistic seismic hazard analysis (PSHA), i.e., based on peak acceleration/response, into a time history can be complicated by deaggregation sensitivity to choice of motion parameter. The complication is exacerbated when peak motion only partially addresses damage potential. Energy-based parameters, such as Arias Intensity, address both amplitude and durational aspects of seismic shaking. To the extent that energy content correlates with damage potential, PSHA using AI (or some other energy measure) yields hazard curves that facilitate the translation of probabilistic load into a time series and, through deaggregation, facilitates the development of design earthquake scenarios and time histories. Potential applications range from assessing the conservatism of a time history's loading, to developing a suite of time histories with specific probabilistic loads for damage sensitivity analysis. Examples of AI-based and peak acceleration-based PSHA for exceedance return periods suggested for dams (thousands of years) show that AI-based deaggregation focuses on larger earthquakes where peak acceleration-based deaggregation indicates competing scenarios.

Ground motions are derived for application to Project-specific design criteria. Both probabilistic and deterministic seismic design ground motions are developed satisfying design requirements, and using seismic source and ground motion attenuation models similar to those employed for earlier extensions to the BART system. Specifically, the Silicon Valley Rapid Transit Project (SVRT) alignment seismic design criteria for permanent structures require determination of the greater of ground motions from a site-specific 10% in 50 years [approximately, 500-year return period] probabilistic analysis or the median deterministic ground motions from the maximum magnitude events on the major active faults in the immediate vicinity: the San Andreas, Hayward, and Calaveras faults. For convenience, the continuum of design ground motion response spectra within this range were segregated into smooth enveloping horizontal and vertical surface outcrop spectra over the frequency range of 100 to 0.1 Hz and for National Earthquake Hazards Reduction Program (NEHRP) Site Class D for "North," "Central," and "South" reaches of the alignment. Based on these analyses this report defines seismic design ground motion response spectra for permanent and temporary structures for these three reaches.

The evaluation of the seismic hazard for a particular site involves the estimation of ground motions associated with potentially damaging earthquakes. Typically the seismic hazard is first computed at the bedrock level. To obtain the seismic hazard at the soil surface, the mean rock hazard is usually multiplied either by a mean amplification function or by a set of average site amplification factors. In this type of approach the variability in the site response is often lost or is not coupled appropriately with the variability associated with the rock hazard. Recent studies have proposed methods to account for the site response in probabilistic seismic hazard analyses using a probabilistic approach. Although these more refined procedures take into account the uncertainty in the entire earthquake process, the most accurate approach to develop site-specific ground motions is the use of site-specific soil attenuation relationships in the probabilistic seismic hazard calculations. Ground motion attenuation relationships were developed for soil sites in the Upper Mississippi Embayment that incorporate uncertainties in the earthquake source, path, and site processes. These attenuation relationships have been used for deterministic and probabilistic seismic hazard analyses for the region. Suites of probabilistic ground motions have been generated for selected cities within the Upper Mississippi Embayment.

A persistent problem with the more widely used non-linear soil models employed in site response analysis is their inability to match measured modulus reduction and damping curves simultaneously. One possible approach for selecting a model's material properties is to match the modulus reduction curve obtained from soil tests while disregarding the mismatch with the damping curve. Although, this approach might work well when the soil is subjected to small strains, for large strains the damping is generally overestimated. Alternatively model parameters can be selected to approximately match both modulus reduction and damping curves over a desired strain range. Various authors proposed possible solutions to this mismatch. One approach uses a reduction factor for the hysteretic damping to better match the damping ratio at larger strains. Other approaches use advanced material constitutive models. This paper presents a new damping reduction factor, which modifies the hysteresis loops from the extended Masing rules, and a fitting procedure for both modulus reduction and damping curves. The damping reduction factor is implemented in conjunction with the hyperbolic model in the site response analysis program DEEPSOIL. Two examples of 1-D site response analyses are presented to illustrate the performance of the new formulation.

Site effects play a very important role in characterizing earthquake ground surface motions. At the edges of sedimentary basins, the large velocity contrast between soil and rock generates basin-induced surface waves that propagate in the horizontal direction inside the basin resulting in ground motions with long duration and high low-frequency energy. The 2-D non-linear seismic response of the Upper Mississippi Embayment has been computed to evaluate these basin effects in the region. The analyses have been performed using an indirect boundary integral — discrete wave number method. The non-linear soil behavior has been incorporated by using an equivalent linear approach. For the range of periods implemented in the analyses, 2 to 10 seconds, no significant difference was observed between the soil amplification of 1-D and 2-D models of the Upper Mississippi Embayment, except for sites located in a narrow region along the basin edge and periods longer than 4 seconds, where higher ground amplification was observed due to basin-edge effects. Possible differences in ground motion duration due to basin effects are not included in these analyses and must be accounted for separately if necessary.

The prediction of soil response at a site is an important aspect of estimating ground motion from future earthquakes. The soil response at the Turkey Flat Site Effects Test Area from the 2004 Parkfield M6.0 earthquake (0.3g) is compared via spectral ratios to the response from the 1995 weak-motion based predictions of Cramer (1995) and from 0.1g and less strong-motion recordings from seven M3-5 Parkfield and the M6.5 San Simeon earthquakes. At Turkey Flat four recording sites are spaced across the two-kilometer wide, shallow, stiff-soil valley: Rock South, Valley Center, Valley North, and Rock North. Rock South is located 5 km from the San Andreas Fault and the 2004 Parkfield rupture. Near-field and azimuthal effects in the 0-5 Hz frequency band are present in the 2004 Rock South records, but not in the records at the other sites 6–7 km from the San Andreas Fault. Thus the 2004 Rock North mainshock record provides better predictions of the 2004 strong ground motions at both valley sites. Comparisons of the mainshock spectral ratios to both sets of weaker-motion mean spectral ratios suggests nonlinear soil behavior in the shallow soils at the Valley Center site and linear soil behavior elsewhere.

Surface wave tests are based on the solution of an inverse problem for soil profile identification from the experimentally measured dispersion curve. Most criticism for this testing methodology arises from the non-uniqueness of the solution, i.e. several sets of model parameters can give very similar dispersion curves. In this paper some implications of solution non-uniqueness in seismic response studies are investigated using both synthetic numerical simulations and experimental data. Adopting a Monte Carlo approach for the inverse problem, possible solutions of the inversion process are selected such that they can be considered equally good with respect to the experimental data. This set of soil profiles is subsequently used to evaluate the seismic response of the site. It is shown that equivalent profiles with respect to surface wave testing are equivalent also with respect to site amplification and V_S,30 determination.

This paper presents a practical approach to develop a site-specific design acceleration response spectral (ARS) curve for a soil profile Type F site due to liquefaction. A site development consisted of four main structures up to 12-stories high over a two- to three-level parking structure in Marina Del Ray, California. Based on the 2001 California Building Code (CBC) Table 16-J, the site was classified as Type F, requiring a site-specific evaluation. A site-specific probabilistic seismic hazard analysis (PSHA) was performed to develop the target uniform hazard spectra (UHS). An equivalent-linear one-dimensional seismic response analysis using the computer program SHAKE91 was performed and compared with the nonlinear effective-stress seismic response analyses using the computer code FLAC and D-MOD2000. Differences and deficiencies of the response spectra between using the equivalent-linear and nonlinear effective-stress seismic response analyses are discussed and the recommendations for structures having a different fundamental period of vibration are made.

This paper presents a micro-mechanical study of the response of unsaturated non-collapsible and collapsible soil deposits when subjected to a dynamic base excitation. The discrete element method was employed to idealize the soil skeleton. The effects of interparticle water bridges in a pendular state were modeled using suction forces. Periodic boundaries in the two lateral directions (El Shamy, 2004) and a high gravitational field were used to reduce the number of particles, required to achieve a realistic simulation of a soil deposit, to a computationally manageable value. Numerical simulations were conducted to assess the impact of moisture content on the dynamically induced settlements of unsaturated non-collapsible and collapsible soil deposits. The outcome of these simulations provided a valuable insight on the associated response mechanisms.

The move of the offshore industry into ever greater water depths presents significant challenges for geotechnical and seismic engineering. Soil conditions at deepwater sites are well outside the range of typical onshore experience. Given the very low strength soils, earthquake loads often control foundation design. In these conditions the evaluation of seismic design criteria is critical to project success. This paper discusses the site response analysis for a field in 650 m water depth offshore the Nile Delta. Soil amplification predicted using the SHAKE type equivalent linear analyses is compared to a non-linear approach. The equivalent linear approach is seen to predict significantly larger spectral acceleration than the non-linear analysis for these soil conditions. Implications for seismic design criteria are discussed.

Both of equivalent linear and non-linear site response analyses utilize soil dynamic properties expressed by strain amplitude dependent modulus reduction and damping ratio relations. Most of the non-linear site response methods use Masing rule to define the unloading and re-loading paths that results in overestimated hysteretic damping at large shear strain amplitude. In this paper, site response analyses of a soft clay site under very strong input motions at the rock base are performed using the equivalent linear approach (SHAKE analysis) and a non-linear approach (i.e. program FLAC with a user-defined soil model). A `generalized Masing rule' was implemented in the user defined soil model to correct the hysteretic damping for the unloading and re-loading stress-strain loops to be consistent with the experimental damping. The non-linear site response analyses, we performed, indicate that, with the hysteretic damping correction, the computed surface acceleration and peak shear stress along the depth do not change significantly compared with those without hysteretic damping correction. But the induced peak shear strains in the soft layers are noticeably different between the case with hysteretic damping correction through the generalized Masing rule and that using the original Masing rule.

Laboratory test results indicate that cyclic cohesive soil behavior is influenced by the rate of loading. However, such behavior is most often ignored in the site response analysis. This paper develops three models of rate-dependent dynamic soil behavior based on available laboratory data and implemented in a one-dimensional equivalent linear site response analysis program. The site response analysis results show that rate-dependent shear modulus and damping can have a pronounced influence on propagated weak ground motion, but a secondary influence on propagated strong motion. Rate dependence of damping ratio has a greater impact on the computed response than rate-dependence of the shear modulus.

We investigate the variability in ground motion predictions that results from the methodology selected for the realization of strong motion site response analyses of typical soil profiles in the Los Angeles (LA) Basin. For this purpose, site-specific regional velocity and attenuation structures are compiled using geotechnical data at three downhole arrays, low-strain velocity and attenuation profiles obtained via waveform inversion of weak motion records, and crustal velocity structures as described by the Southern California Earthquake Centre Community Velocity Model (SCEC CVM IV). Successively, broadband ground motions are simulated by means of a finite source model for rupture scenaria of weak, medium and large magnitude events (M=3.5–7.5) at stations located on a regular 2D grid within distances 5–75km from the fault. Observed estimates of site response at the stations of interest are first compared to simulated small magnitude rupture scenaria. Parametric studies are next conducted for each fixed magnitude (fault geometry) scenario by varying the source-to-site distance and source parameters for the ensemble of site conditions. Elastic, equivalent linear and nonlinear simulations are implemented for the base-model velocity and attenuation structures and nonlinear soil properties, to examine the variability in ground motion predictions as a function of ground motion amplitude and frequency content, and nonlinear site response methodology. Results of this study are currently being used to develop supplementary criteria for the NEHRP site classification system to describe the nonlinearity susceptibility of soft sediments, and thus establish a cost-effective and computationally-efficient framework for site parameterization and response simulation defined as a function of source-path-site-effects and design sophistication level. The reflection of nonlinear site response modeling uncertainty is finally depicted in the estimation of synthetic ground motion amplification factors, which are compared to the currently employed NEHRP factors.

Linear or equivalent linear wave propagation models, such as SHAKE, assume that the shear wave velocity is constant, hence the wave form is not distorted during propagation. If the tangent shear modulus is not constant, then perturbations in stress travel at different speed depending on the tangent modulus at that point and time. In cyclic loading, the shear stress-strain curve can be concave down, which we call strain-softening (near failure, for example) or concave up, which we call strain-stiffening (if there are negative pore pressures being generated due to dilatancy). In strain stiffening soil, the peak of a stress wave travels faster than the front, the wave front sharpens, and there is a possibility to form a shock wave. If the soil is strain softening, the peak travels slower than the front and the wave front elongates to form a dispersed wave. In this study, we describe evidence of nonlinear wave propagation including formation of shock waves that has been observed using a vertical array of accelerometers in centrifuge model tests and in real earthquake data. Implications of non-linear wave propagation in strain-stiffening and strain-softening soil are discussed. We also look at the reflection phenomena of nonlinear waves at rigid and free boundaries. One result is that for a soil that is undergoing cyclic mobility (negative pore pressure development during loading, and positive pore pressure development during unloading), the wave sharpens as it approaches the ground surface and the reflected wave is almost negligible.

Since the 2D modeling of topographic features, through numerical analysis, has been done up to now, all by symmetrical geometries, there is a problem that, whether the seismic response of asymmetrical models are so different or not. As the natural features are not symmetric, it is so important for engineers to know whether they can analyze them based on symmetrical models and how much is their risk. Therefore, in this paper we have evaluated the 2D site response based on parametric study of several asymmetric geometry models of hill (asymmetric semi-sine, triangular and trapezoidal shapes) and have compared them with the response of their corresponded symmetrical models. The studies have been done in time domain and soil material is considered linear elastic. Also, the assumed input motion was a Ricker type wavelet which was considered as SV and P plane wave and applied vertically. All calculations are executed in time-domain using the direct boundary element method. Clear perspectives of the amplification patterns of the hill are presented by investigation of the frequency-domain responses. In general the results shown that the increase of asymmetry rate between two side of features model haven't has a considerable effect on their seismic response and the response are mainly affected by geometric properties of the part of models which includes larger surface of model plane. Also the opposite component of incident motion wave response at the crest has been more amplified due to asymmetry affects.

Whereas seismic response analyses are commonly performed to evaluate the effects of deep soil sites on earthquake ground motions, relatively limited experience is available with the response of sites underlain by deep deposits of geologically young clays under strong earthquake shaking. This paper illustrates the use of nonlinear and equivalent-linear analysis techniques for evaluation of the seismic response of the San Francisco International Airport airfield, which is underlain by Young San Francisco Bay Mud and fill deposits. Site response analyses for rock motions representative of the 1989 Loma Prieta earthquake were able to simulate the recorded surface motions at the airport during the earthquake, and indicate that peak accelerations on the airfield were lower than at the recording station.

The Tito Scalo site, located near the city of Potenza in the Basilicata region in Southern Italy, is a high seismic risk area. Borings and dynamic in situ tests have been performed. Among them Down Hole (D-H), Noise Analysis Surface Waves (NASW) and Seismic Dilatometer Marchetti Tests (SDMT) have been carried out, with the aim to evaluate the of shear-wave velocity (V_s) profile of the soil. The results show a very detailed and stable shear-wave profile. After evaluating the synthetic accelerograms at the bedrock, the ground response analysis at the surface, in terms of time histories and response spectra, has been obtained by two non-linear models GEODIN and EERA.

Evaluation of ground response is one of the fundamental problems in earthquake geotechnical engineering. Current practice has tended to focus on horizontal earthquake motion and ignore the effect of vertical motion. Because of strong vertical ground motions recorded in recent earthquakes, great concern has arisen over the behavior of vertical motion and its effect on engineering structures. This paper presents a practical procedure for the analysis of site response to vertical earthquake motion. The procedure involves the use of dynamic stiffness matrix method and is built in the modern MATLAB environment to take full advantages of the matrix operations in MATLAB. The nonlinear soil behavior is accounted for through the well-accepted equivalent linear approach. Comparisons of the predictions from the procedure against downhole array records show reasonably good agreement.

Damage distribution surveys have long been used as a means for estimating relative ground shaking intensity. In the absence of dense instrumental arrays, these surveys can be used to identify and delineate geotechnical earthquake hazard zones. The traditional means of analyzing spatially distributed damage survey data is qualitative or semi-quantitative in nature. While this may be a satisfactory approach in situations where there is a marked contrast in building performance, its effectiveness is limited when damage contrasts are subtle, and spatial patterns are less obvious. In this paper we present a quantitative approach [Getis-Ord (Gi*) statistic] for interpreting and analyzing earthquake damage surveys. By implementing neighborhood-scale spatial analysis within a Geographical Information Systems (GIS) framework, statistically significant concentrations of elevated (and/or low) damage concentrations can be identified and quantified. An example application of this method is applied to a damage survey from the 2001 Southern Peru Earthquake (M_w8.4). In contrast to preliminary studies that suggested damage concentrations were a consequence of topographic amplification, the spatial analyses demonstrated that hotspots of damage are generally correlated with flatter ground in the study area.

Strain-dependent shear moduli and damping ratios are estimated for the deposit of the Service Hall at the Kashiwazaki-Kariwa nuclear power plant based on shear stress-strain time histories that can be directly determined from its downhole array recordings during the 2007 Niigata-ken Chuetsu-oki earthquake. The Holocene dune sand with V_s = 310–350 m/s at depths less than about 70 m shows strong nonlinearity with a shear modulus ratio of about 0.3 and a damping ratio of about 30 % at a maximum shear strain of about 3 x 10^–1%. The reduction in shear modulus ratio is consistent with previous laboratory test results for sand consolidated under similar confining pressures but the corresponding damping ratio seems to be slightly higher than that of the previous studies. The mudstone with V_s =500 m/s located at depths greater than about 100 m does not show any strong nonlinearity with a shear modulus ratio of about 0.6 and a damping ratio of about 10 % at a maximum shear strain of about 1.5 x 10^–1%. The strong nonlinearity of the Holocene dune sand during the main shock might have lowered the amplitude ratios between the near-surface layer and the bedrock in the short period range significantly compared with those of other smaller events.

A blind prediction experiment was conducted for the Turkey Flat test area data after the September 28, 2004 Parkfield earthquake. The results of the test were reviewed at the Turkey Flat Blind Prediction workshop in 2006. The predicted ground motions at the center of the valley significantly exceeded the observed ground motions. In this study, records of Parkfield aftershocks and two other events, including the 2003 San Simeon earthquake, are analyzed and compared with the 2004 Parkfield mainshock records. Comparison of the observed ground motions at the surface and at depth at the rock outcrop station shows little amplification from a depth of 24 m to the surface. This observation may be expected as the instruments at both the surface and at depth are located in the same rock formation. Average rock motions beneath the valley center site are considerably weaker than those recorded 800 m away at a similar depth below the rock outcrop. This observation implies that the recorded motion at depth under the rock outcrop should not be applied directly as the input motion at depth under the sedimentary valley for the purpose of predicting ground motion at the surface. The peak ground accelerations at the surface at the center of the valley are largest for nearly all events. The ratio of the peak acceleration from the surface to depth varies from about 1.5 to more than 6. The smallest amplification is for the distant M6.5 San Simeon earthquake of 2003, which has a lower dominant frequency content than the frequencies amplified at the valley center site.

A blind test has been conducted for site response of the Turkey Flat Site Effects Test Area, near Parkfield, California to the September 28, 2004 M6.0 Parkfield Earthquake. Acceleration time histories recorded on bedrock near one valley edge were provided along with detailed geotechnical properties at all recording sites to interested participants. Participants were asked to make predictions of the ground motions at the five other recording locations (3 surface and 2 downhole), while records were withheld by CSMIP until all predictions were received and officially logged. A second series of predictions were made based on ground motions recorded in bedrock beneath the valley floor. In all there were 15 participating teams from 4 countries that submitted a total of 92 sets of blind predictions, 45 for Part 1 and 47 for Part 2 of the test. Of the 92 sets of predictions, 18 are from industry, 54 from academia, and 20 from government sectors. Site response models tested fall into three categories, with 55 prediction sets based on equivalent linear methods, 33 based on non-linear, and 4 based on other approaches. Predictions generally overestimate valley response, with little improvement using independently interpreted verses standard geotechnical models. Results indicate a marked improvement for site-response predictions based on input of actual rock motions recorded immediately beneath the valley sediments. The strong-motion test demonstrates that surface motions at a valley edge may not always be reliable input motion for estimating site response of valley sediments. This paper provides an overview of the Turkey Flat strong-motion blind test, and summarizes the various soil models and computational models submitted for the test. A companion paper in these proceedings discusses quantitative results of the blind test and comparison with other data.

Ground response analysis in two dimensional modeling is one of the problems considered in geotechnical earthquake engineering. In many cases, a site might be located on an alluvial valley and two dimensional analysis should be performed instead of one dimensional analysis in order to obtain reliable surface accelerations. In this research rectangular valleys with different ratios of width to thickness are modeled in two dimensions and the acceleration response of soil on the middle point of surface is evaluated. Results indicate that the amplification of the motions in two dimensional models with different width to thickness ratios of the valley differs from those calculated by one dimensional technique. The ratio of the width to thickness for which the results of one and two dimensional analysis are similar, is also determined.

Ground rupture caused by an earthquake is a major consideration for the siting of engineered structures and mitigation of existing lifelines that cross active faults. Ready availability of new digital data sources, including digitized mapped fault traces and high-resolution topographic datasets, allows for rapid incorporation of existing information towards evaluation of fault rupture hazard. However, available digital and non-digital fault data generally are not detailed enough for site-specific evaluation and design. Precise information from targeted fault (paleoseismic) trenching provides additional information required for siting and mitigation of critical facilities and lifelines. In addition to defining the locations of active fault strands that can be used to develop set backs for building footprints and mitigation plans for existing structures, paleoseismic trenching can provide vital information on the style, amount, and probability of surface faulting. Trenching of `secondary' faults that are structurally related to underlying, larger faults can provide indirect evidence for the determination of timing and probable fault offset on deeper, often inaccessible (`blind') faults. This paper provides several current applications of paleoseismic trenching performed to address unique project-specific surface-rupture issues and obtain seismic parameters for engineering design.

Many investigators have applied the liquefaction potential index (LPI) to map regional liquefaction hazard. LPI, which integrates the liquefaction potential of susceptible soil elements at a specific location into a single value, has been used to assess both (1) spatial variability of liquefaction potential, and (2) liquefaction potential of surficial geologic units. A promising application to mapping has been the establishment of a median threshold LPI value at which liquefaction effects occur. This threshold when applied to complementary cumulative frequency distributions of LPI for spatially homogeneous surficial geologic units yield liquefaction probability curves for surficial geologic units that can be used for hazard mapping. Both probabilistic liquefaction scenario maps and PSHA-based liquefaction hazard maps have been produced with this approach. The scenario maps compare favorably with historical liquefaction in the mapped areas.

A deformation-based approach for mapping earthquake-induced liquefaction hazard has been developed that utilizes recently generated predictive relationships for shear-strain potential. These new relationships are used to calculate "Displacement Potential Index" given induced cyclic stress ratio and normalized fines-corrected standard penetration resistance. Conventional liquefaction hazard mapping is based on qualitative assessments of surface deposits, their geomorphology, environment and recency of deposition, and past performance from post-earthquake investigations. This new method is a more quantitative, deformation-based approach that emphasizes the severity of ground failure rather than mere likelihood of liquefaction. These new methods are being investigated for use in improving delineation of California's regulatory liquefaction hazard zones. Over 650 boring logs have been analyzed for 18 Quaternary geologic map units in a pilot study covering a portion of the southern San Francisco Bay Area in Santa Clara County. A sufficient number of boring logs were available for nine of the Quaternary units to permit reclassification into four liquefaction susceptibility units, for which parametric modeling was performed and ground deformation hazard mapped over the test area. Pilot study results show good agreement with historical occurrences of liquefaction.

The recent five earthquakes in Iran offered many good quality strong motion data at 242 Iranian free-field strong motion stations. These data provide a unique opportunity for site effect analysis and the study of a practical site classification system. The horizontal to vertical spectral ratio (HVSR) of S-wave was used to estimate the soil amplification characteristics at the sites. The strong motion station sites were classified using identified site dominant period and geological data according to Japanese code and 2003 IBC provision. Furthermore, the response spectral shape (RSS) method was also employed in site classification mostly for checking purpose. A choice of site class for each strong motion station was made by comparison among different classifications. The station sites were classified into I, II, III and IV categories (from hard to soft), which are comparable to B, C, D, and E in the 2003 IBC provision. A successful comparison between HVSR and RSS results gives a good support to the classification results. Furthermore, the mean of HVSR for all the records from the earthquakes were calculated according to the site classes, and compared with those of pervious studies in Taiwan and Japan. A good conformity among different outcomes used to validate the classification results. Finally, a new classification system for the site categorization in Iran is introduced, and compared with the one proposed in Iranian building code.

Since 1976, the General Directorate of Disaster Affairs of Turkey has deployed several strong-motion accelerographs at selected sites, particularly along the North Anatolian Fault zone. Within the framework of the project entitled Compilation of National Strong Ground Motion Database in Accordance with International Standards, initiated in 2006, site conditions at a total of 153 strong-motion sites were investigated within the upper 30 m depth through boreholes including Standard Penetration Testing (SPT) and surface seismics. A preliminary assessment of the site characterization was held by making use of the mean shear-wave velocity of the uppermost 30 m depth with reference to the NEHRP provisions. The recorded accelerograms and corresponding catalogue information were compiled for these strong-motion sites using the available information from national and international databases. The 5% damped normalized elastic acceleration spectra of the accelerograms recorded for events with M_w ⩾ 4.0 were computed and categorized according to the site class and M_w. Mean normalized spectra computed for different M_w and site class bins were compared in order to investigate the significance of site class on the spectral response. NEHRP site classification is evaluated within the limitations of available data and findings.

A software package, KoeriLoss2, is developed to perform earthquake damage estimation for urban areas based on detailed analyses of local site effects. The first stage of the approach involves generation of microzonation maps with respect to ground motion parameters for the investigated area using separately calculated regional seismic hazard input data. The effects of local geological and geotechnical site conditions are taken into account using representative soil profiles with shear wave velocities extending down to the engineering bedrock (V_s ⩾ 750m/s). 1D site response analyses are conducted using Shake91 to calculate an average site specific PGA and elastic acceleration response spectrum on the ground surface. In the second stage of the approach, vulnerability of the building and pipeline inventories are estimated using ground shaking parameters and empirical relationships. A pilot study is conducted for Zeytinburnu and Bakirkoy districts in Istanbul, Turkey to demonstrate the applicability of KoeriLoss2.

Heavy damage was observed in Muzaffarabad, Pakistan due to the 2005 Kashmir earthquake, especially on terraces with thick well compacted gravel deposits and along the terrace rims. We evaluated the damage distribution visually from Ikonos satellite images. In an effort to contribute to the restoration of Muzaffarabad city and its future urban and infrastructure development, we investigated the area's geomorphology and measured microtremors at some 20 locations. The four terraces in the city consist of well compacted gravel deposits and our microtremor results indicate slightly higher amplification values and lower fundamental frequencies on the thicker deposits in the South-East of the city, where damage was very severe. We hope our study will be a good complement to JICA's (Japan International Cooperation Agency) Master plan for the city's reconstruction.

The effects of local geology, including uncertainty, have been incorporated into several central and eastern U.S. seismic hazard products. The approach is to generate site amplification distributions on a grid for a given period using three dimensional geological/geophysical/geotechnical models (urban hazard maps) or reference profiles keyed to surface geology and depth-to-bedrock (regional hazard maps). A derived shear-wave velocity profile for a grid point is randomized along with dynamic soil properties and the choice of input ground motion to represent model uncertainties. A Monte Carlo approach is used to estimate the lognormal mean and standard deviation of the site amplification distribution. Probabilistic seismic hazard maps use the full site amplification distribution and scenario (deterministic) maps use the appropriate fractile site amplification (median or 84^th percentile). Computational efficiency and soil response program calibration are important aspects of this procedure.

This paper aims to study the soil dynamics characteristics of central part of Shiraz city near the monument of Persepolis, southwest of Iran. The parameter investigated is natural period of soil sediment. The technique used to obtain this parameter is microtremor measurements. Microtremor measurements were obtained at 120 stations in the city. The data was recorded for two periods of fifteen minutes. Segmental Cross Spectra were applied to calculate spectra. Nakamura's method (H/V) was used for analyzing the measured data. At each site, natural period values were obtained. Because of low accuracy of amplification factors that achieved by microtremor measurements, this factor is neglected in this study. Iso-period map of this part of the city was prepared. The calculated natural periods were compared to those obtained through one dimensional analysis for some existing geotechnical borehole information. It showed that the predicted fundamental period by microtremor method compares well comparable to that obtained by one dimensional analysis.

The paper reports a series of centrifuge model tests designed to study the effect of foundation position on the interaction of reverse, dip-slip faults with shallow foundations. All tests are conducted with a 15 m deep, dry sand layer and a reverse fault with dip angle 60°. Results of a `free-field' test are first presented which allows the fault rupture propagation through the soil to be investigated when no foundation is present. Additional tests are then reported in which shallow foundations are placed at different positions relative to the fault. It is seen that interaction between the foundation and the fault often causes deviation of the fault from the free-field condition, and this may protect the foundation from damage. This has also been observed in recent case histories. In addition, the response of the foundation as characterized by its rotation is very sensitive to the exact position of the foundation — the interaction between the foundation and the fault is subtle. Design should consider carefully the position of the fault emergence with respect to shallow foundations.

The present study aims to develop urban area on an existing landfill. To achieve this goal, it is important to understand the mechanical properties of municipal solid waste. Since particle size of waste is large, a large triaxial shear device was employed for mechanical waste tests. Among results, this paper addresses triaxial compression, cyclic behavior, and mitigation of creep by means of preloading. The second topic concerns modulus and damping ratio changing with strain amplitude.

This study investigates the density effect on the aging behavior of sands and explore whether the aging-induced stiffness gain is isotropic by nature. The results of resonant column tests on Ottawa sand samples reveal that the medium-dense sample had the highest aging rate, as determined by its shear-modulus increase, followed by the dense and loose sand samples. This result might be attributed to two competing processes, the tendency toward force-chain homogenization and the susceptibility to local structure collapse. This attribution does not apply to Toyoura sand samples, which suggests that the effect of packing density on the aging behavior of sand indeed varies with the type of sand. Bender element tests in the true triaxial apparatus demonstrate that aging behavior is not isotropic by nature because the increase of shear wave velocities in different polarization planes varies under isotropic stress states. Increasing the stress in one direction leads to more significant increases in the velocities that are either polarized or propagating in that direction.

A new system, capable of generating and receiving P-waves in the laboratory specimens, is developed. P-waves were generated using a multi-layered piezoelectric actuator made of ceramics placed above the top cap and were received at two different levels at the side of specimen using two piezoelectric accelerometers. Triggering to generate P-waves was made by employing single pulse, single sinusoidal or continuous sinusoidal waves. Dynamic vertical Young's moduli (E_d) were computed from the travel time of the waves between two accelerometers and the corresponding travel distance between them. E_d obtained from this system were compared with the statically measured vertical Young's moduli (E_s) evaluated by conducting small unloading/reloading cycles on the same specimens. The stress level was kept constant to make both measurements just one after the other. For this study, four tests were conducted using medium and large scale triaxial apparatuses. The testing material included Toyoura sand (Uniformly-graded sand from Japan) and Chiba gravel (Well-graded sandy gravel from Japan). As a result of this study single pulse or single sinusoidal triggering was found better an continuous sinusoidal triggering. Moreover, no significant effect of specimen size was observed in E_s and E_d for Toyoura sand specimens. Next, E_d obtained was larger than E_s in both types of materials. Lastly, the difference between E_s and E_d increased with the increase in particle size of specimens. Such difference was about 15–25% in Toyoura sand specimens and it was about 100% in loose Chiba gravel specimens.

The use of vertical drains, along with pre-loading, is the most popular method of soil improvement in coastal areas to meet growing demand for diverse uses of limited land space. Piezometers are the key instrument for continuous monitoring throughout the consolidation period. Basically, piezometers have several disadvantages such as incapability of providing any mechanical properties, meticulous installation procedures, misleading results caused by unexpected hydrogeologic boundary condition, and relatively high damage rate. As an alternative instrument for monitoring soil improvement, a piezo-electric sensor was elected to measure stiffness changes during consolidation. The piezo-electric sensor is composed of two bender elements mounted on a frame. Bender elements, composed of piezo-materials and metal shims, have favorable features suitable to field instrumentation such as simple principle of energy conversion, excellent control capability and small physical size. The prototype sensors have been developed through extensive in-house program to optimize in terms of data quality and installation convenience. The first prototype sensor was implemented in coastal mud using SPT(standard penetration test) rods pushed with a routine boring machine and excellent shear wave signals were measured. The sensor is on the further collective refining process.

Over the past 40 years, approximately 15 mine tailings dams have failed worldwide as a result of earthquake shaking. Methods to analyze the dynamic response and liquefaction resistance exist for soil, but fine coal refuse, including slurry that is hydraulically placed behind the dike, is inherently different than soil. Therefore, a research study is underway to understand and predict the dynamic behavior of fine coal refuse. To date, field testing has been performed at two coal refuse impoundments in eastern Kentucky. Laboratory resonant column testing was used to measure modulus reduction curves (i.e. the relationship between shear modulus, material damping, and shear strain) of undisturbed specimens. Field standard penetration testing, seismic cone penetrometer testing, and geophysical surface wave testing was performed to measure blow count (N₆₀), cone tip resistance (q_c), and shear wave velocity (v_s). Field vane shear data were acquired to correlate peak and residual undrained shear strength to q_c. Preliminary results of this study include development of modulus reduction curves for fine refuse and development of methods to estimate peak and undrained shear strength based on in situ CPT testing. Future results will include correlation of overburden-corrected field data to cyclic resistance ratio (CRR) to predict liquefaction resistance.

A fiber reinforced soil behaves as a composite material in which fibers of relatively high tensile strength are embedded in a matrix of relatively plastic soil. Shear stresses in the soil mobilize tensile resistance in the fibers, which in turn imparts greater strength to the soil. In this paper a study on the influence of synthetic fibrous materials in improving the dynamic strength characteristics of a fine sandy soil is reported. The project is aimed at converting fibrous carpet waste into a value-added product for soil reinforcement. A series of strain controlled large cyclic triaxial tests were carried out on Toyoura sand specimens reinforced with randomly distributed carpet strips. The dynamic strength characteristics of the reinforced sand are defined in terms of deformation modulus and damping ratio. The effects of parameters such as fiber content and fiber aspect ratio, have been studied. The results clearly indicate the effectiveness of the reinforcement in improving the dynamic properties of fine sand.

Failure of tailings dams due to earthquakes occurred several times in the past, causing significant loss of properties and human lives as well as severe environmental problems. In the US, government regulatory agencies such as the Office of Surface Mining (OSM) and the Mine Safety and Health Administration (MSHA) require comprehensive seismic analyses in the design of tailings dams. This paper presents the results of an extensive laboratory testing program on coarse and fine coal waste refuses. Representative samples were taken from a tailings dam in the Appalachian coalfield. Shear modulus and damping ratio versus shear strain curves for the two types of refuse are established. In addition, the response of fine refuse to cyclic loading including the generation of excess pore pressure was investigated using cyclic triaxial tests. Implications for design are discussed.

Dynamic properties of saturated compacted soils are very important parameters used for analyzing seismic responses of compacted earth dams. The complete test results of them are rare since that saturating the compacted soil specimens is time-consuming and laborious by using traditional saturation methods. In this study, two vacuum-saturation methods were proposed to improve saturating the compacted SM, ML and CL specimens. Both methods can reliably make their Skempton coefficient B greater than 0.95. Their complete dynamic properties were then measured by resonant column test (RCT). The test results showed that their G/G_max versus strain relationships and damping ratio curves are much different from those of ordinary soils. The empirical equations correlating the maximum shear modulus, void ratio and effective confining pressure were also suggested for different compacted soils.

The main objective of this work was to experimentally investigate the influence of key environmental factors, including compaction-induced matric suction and confining pressure, on small-strain stiffness properties of partially saturated sandy and clayey soils using the pressure plate, resonant column, and bender element testing techniques. To achieve this goal, a comprehensive series of resonant column (ASTM D 2325-68), bender element (ASTM C 778), pressure plate (ASTM D 4015-92), and filter paper (ASTM D 5298) tests were conducted on several identically prepared specimens of loose, poorly graded sand (SP) and compacted, high plasticity clay (CH) subject to different compaction-induced (post-compaction) suction and confining stress states prior to resonant column and bender element testing. Compaction-induced suctions in all specimens were assessed prior to testing via a full set of previously calibrated soil-water characteristic curves (SWCC). Compaction-induced matric suction was found to exert a significant effect on the small-strain stiffness response of partially saturated SP and CH soils both from bender element and resonant column testing techniques.

This paper presents an investigation of the value of soil density that controls the velocity of small strain shear waves in saturated soil. An "effective soil density" is defined that is related to the fraction of pore water that moves with the solid skeleton during shear wave propagation. Biot theory indicates that the ratio of effective soil density to saturated soil density is always ⩽ 1 and is a function of the specific gravity of solids, porosity, hydraulic conductivity, and shear wave frequency. Except for extreme cases, the effective density ratio will range from 0.75 to 1.0. For many geotechnical applications, effective soil density will be equal to saturated soil density for low hydraulic conductivity materials (clays and silts). On the other hand, consideration of effective soil density may be important for high hydraulic conductivity materials (clean sands and gravels) in some cases. The findings are relevant to applications involving the propagation of small strain shear waves through saturated soil, and in particular for laboratory and field tests in which shear modulus is back-calculated from measured shear wave velocity.

Four series of drained tests on sands were performed with a hollow cylinder apparatus in order to investigate the fundamental deformation behavior of sands and the effects of the intermediate principal stress parameter b during cyclic rotation of principal stress axes. The test condition that principal stress axes rotated cyclically from 0° to 180° relative to vertical under constant effective principal stresses was maintained. The experimental results indicated that strains were accumulated steadily with the increasing number of cycles, while b had considerable influences on their changes during the process of cyclic rotation. The increase rate of the maximum contractive volumetric strain within each cycle was lowered with its progressive accumulation. It was found from four test series that the magnitude of the volumetric strain at the same number of cycles increased with the increase of b. The hysteretic loop in shear stress-strain curves was observed to be unclosed for the first several cycles but tend to become closed with the increase of the number of cycles. The test results in this paper also demonstrated the significant effects of the density, the effective mean normal stress and deviatoric stress on the deformation behavior of sands under cyclic rotation of principal stress axes.

Understanding of the effect of cementation on dynamic behaviour of cemented soil in an earthquake prone area could be crucial for earthquake resistance design. The major section of the city of Tehran has been developed on cemented coarse-grained alluvium. In order to understand the dynamic behaviour of this soil, a series of undrained cyclic triaxial tests were performed on uncemented and artificially calcite-cemented samples. In this paper, two dynamic parameters including dynamic shear modulus and damping ratio based on different definitions are investigated. In this regard the effects of cement content, confining pressure and cyclic deviatoric stress were studied as well. In general by increase in the amount of cementation, maximum shear modulus increases and damping ratio show no significant variation. Also, increase in deviatoric strain results in the increase in damping ratio and decrease in shear modulus. In addition the higher confining pressures lead to higher shear modulus and lower damping ratio. Furthermore the shear modulus decreases with increase in number of cycles.

This paper presents a theoretical study of pore fluid induced damping of saturated soil in resonant column tests. The solid skeleton is treated as equivalent linear. Pore fluid induced damping due to Biot flow is evaluated based on the spectral response of the soil column under harmonic excitations. Closed-form analytical solution indicates that pore fluid induced damping is frequency and hydraulic conductivity dependant. At the first resonance, pore fluid induced damping is dependent on a dimensionless hydraulic conductivity parameter K. For K within the range of 10^–3 to 10, corresponding to coarse sands and gravels, pore fluid induced damping may have an important contribution to total soil damping, especially at small strain levels.

Aging effects in sand, such as increases in penetration resistance with time after deposition, densification, and/or liquefaction, are known to occur in situ, but the causes of these effects are not fully understood. Nonetheless, these effects have important ramifications in earthquake engineering. First, the lack of understanding of the phenomenon is an impediment to quality assurance/quality control (QA/QC) for ground densification projects aimed at mitigating the damaging effects of liquefaction. This can be understood by considering that most liquefaction evaluation procedures correlate liquefaction susceptibility to in situ indices, such as penetration resistance (SPT and CPT) and small strain shear wave velocity (V_s), all of which are influenced by aging. Consequently, it is unclear as to how long after ground densification QA/QC in situ tests should be performed to ensure that the densification was sufficient to mitigate liquefaction susceptibility. Presented herein is an overview of an ongoing sand aging field study where liquefaction is being induced by explosives, vibrocompaction (using a vibroflot), and a NEES vibroseis in a heavily instrumented sand deposit. The state and properties of the sand are being monitored as a function of time after the disruption of the soil structure.

Determination of strain associated with shear modulus and damping from hysteresis loops generated on torque-rotation plane during torsional test is complicated. This is due to nonuniform stress-strain variation occurring linearly with radius in a soil specimen in torsion. A conventional equivalent radius approach proposed by Chen and Stokoe (1979) is adequate for evaluating strain associated with shear modulus at low to intermediate strain levels. The approach is less accurate for damping measurement, as well as measurements at high strain. Stress integration developed by Sasanakul and Bay (2008) is used to account for the nonuniform stress-strain effect more precisely. The approach involving integration of an assumed soil stress-strain model is used to generate a plot of equivalent radius ratio versus strains based on shear modulus and damping separately. This paper presents the development of equivalent radius ratio curves used for damping measurement. The stress integration approach is applied for hyperbolic, modified hyperbolic, and Ramberg-Osgood models. Variation of equivalent radius ratio with normalized rotation for each model is presented. The results illustrate the usefulness of the stress integration approach and suggest that using a single value of equivalent radius ratio to calculate strains is not appropriate.

The author proposes an alternative to the traditional representation of soil damping. Rather than using damping ratio, this author advocates using viscosity as the specific soil property, especially for saturated permeable soils. Thus represented, the imaginary part of the complex shear modulus will vary directly with frequency. The point is particularly relevant in cases where the water table may change, thus affecting the dynamic design of foundations or structures composed of soils.

Large deposits of naturally-cemented soils occur in the semi-arid regions of the Southwest, United States. Some of these soils have relatively large void ratios and often exhibit unique characteristics under static and dynamic loads. The dynamic properties of these cemented soils with large void ratios have received little attention in technical literature, mainly due to the difficulty associated with recovering undisturbed specimens of this type of material. This paper presents the results of a cyclic triaxial test program performed on relatively undisturbed samples of cemented silts obtained from block samples. Tests were carried out both under natural moisture contents and after saturation. The results of this study show that the Young's modulus of the soil increases with the degree of cementation, while the damping ratio is relatively insensitive to the degree of cementation.

Three surface wave techniques, SASW, active MASW, and passive MASW, were conducted at a well-characterized test site. Crosshole shear wave velocity, SPT N-value, and geotechnical boring logs were also available for the test site. For the multi-sensor active test data, it was observed that the cylindrical beamformer transform provided the best resolution in dispersion data over a wide range of frequencies. The dispersion data obtained from all three surface wave techniques was generally in good agreement, and the inverted shear wave profiles were consistent with the crosshole, SPT N-value, and material log results.

Surface wave velocity measurements have recently been performed in the Salt Lake Valley region in northern Utah and the Mississippi Embayment region of the Central United States using equipment from the Network for Earthquake Engineering Simulation (NEES) program. A unique, low-frequency servo-hydraulic vibrator (termed Liquidator), developed at the University of Texas at Austin as part of the NEES program, was used to excite the requisite low-frequency energy for these studies. These measurements are the first application of this unique equipment for deep (200 m and greater) shear wave velocity (V_S) profiling of soil sites. This profiling depth was achieved because of the field performance of Liquidator in the 1- to 4-Hz frequency range. Profiles from these studies have contributed to a better understanding of the V_S structure and variability of the deep sediments in these seismically vulnerable regions. The field performance of Liquidator, the shear wave velocity structures, and the importance of such profiles in developing community velocity models for use in earthquake engineering are discussed.

The dynamic stiffness of the ground is one of the key parameters in earthquake-resistant designs involving local ground response and in assessment of geotechnical structures subjected to various man-made vibrations. Its use is being extended even to static settlement and deformation problems. Its importance and extended usage require a more manageable method than conventional seismic techniques such as crosshole, downhole test, and suspension logging. In general, conventional borehole seismic methods are expensive and technically complex. This paper presents a new method which has been developed to reduce testing cost and to improve the easiness in its use. An in-hole probe has been developed to be fit in three-inch boreholes with functional improvements including a pair of mechanical couplings driven with two servomotors, and a servomotor triggering device. The in-hole seismic method was adopted at various sites such as compaction fills of railroad subgrade, bedrocks of tunnel construction sites, tunnel-face rock mass, and abutment of rock fill dam. In-hole seismic test results were verified by comparing with reliable crosshole results. Presently, the method is very cost effective and easy to use in residual soils, compaction fills and bedrock where uncased boreholes can stay open.

Mine spoils are crushed rock fragments resulting from mountaintop removal mining. Mine spoils are typically placed in valleys as fill, but are susceptible to large amounts of differential settlement. Settlement mechanisms in mine spoil are different than those in soil, and include 1) stress-induced settlement due to particle crushing under increased confinement, and 2) hydrocompression due to wetting of the material. Herein, a method has been developed to quantitatively predict settlement in mine spoils by measuring the effect of confining stress and wetting on void ratio in reconstituted laboratory specimens. Settlement potential was found to be much more severe for dry specimens that had not been previously wetted. Upon wetting, the dry specimens experienced almost instant volume change corresponding to vertical strains of around 8%. Resonant column testing revealed that wetted specimens had lower shear wave velocities (v_s) than specimens that had not been wetted. Thus, v_s can be used to indicate whether or not spoils have been previously wetted, and settlement potential can be estimated based on deformation observations made in the laboratory specimens. With this method, mine spoil sites can be screened, and sites less susceptible to excessive amounts of settlement can be identified.

The downhole seismic test is a widely used method to evaluate in-situ seismic wave velocity profiles. In this test, elastic waves are propagated from a stationary source point that is placed at the ground surface to the receivers, which are placed down in a borehole. The wave velocities are usually determined by evaluating the travel times either between the source and each of the receivers, called the direct method, or between receivers, called the true interval or pseudo-interval method. These conventional approaches of determining wave velocities make simplifying assumptions including a straight raypath between the source and the receivers and velocity isotropy. In a layered and anisotropic media, these assumptions can result in errors in the velocity profile. An inversion algorithm is presented to evaluate the velocity profile from downhole seismic test data. It models the soil profile as a stack of horizontal layers with elliptical velocity anisotropy. The raypaths are calculated using a nonlinear iterative algorithm based on the bending method, fulfilling Fermat's principle of minimum travel time and Snell's Law. The inversion algorithm models wave propagation in real soil media more accurately and thus enables a more robust estimate of the seismic wave velocity profile.

The occurrence of liquefaction in soils is often evaluated using the simplified procedure originally proposed by Seed and Idriss based on in situ indices. Although numerous studies have been conducted to improve and extend this procedure, the effect of age on liquefaction resistance is still poorly understood and correction factors have not been generally accepted. Nine published studies on the effect of age are reviewed in this paper. A regression line representing the average variation in liquefaction strength grain with time was developed from cases where strength was expressed based on recommended liquefaction resistance curves. This regression line is considered an update of the previously proposed relationship by Arango et al. The results indicate that commonly used liquefaction evaluation procedures can be overly conservative in many natural soil deposits.

A number of newer liquefaction evaluation methods have been proposed since the consensus on liquefaction analysis procedures was reached by Youd et al. (2001). These methods are compared and applied to two candidate sites for construction of a power plant along the Ohio River. Comprehensive subsurface investigations were carried out and indicated primarily clean fluvial sands at both sites. Assessment of soil liquefaction susceptibility was made using field measurements consisting of SPT N-values, cone tip stress q_T, and shear wave velocity V_s. By comparing the results given by the various newly proposed liquefaction analysis methods, their relative conservatism is evaluated. At these two sites, it was concluded that the methods based on SPT N-value and cone tip resistance q_T are more conservative than the methods based on shear wave velocity V_s.

Saturated specimens of Ottawa sand prepared with 0%, 3% and 5% bentonite by dry mass of sand are tested under cyclic loading to investigate the effects of bentonite on the cyclic response. For the same skeleton relative density and cyclic stress ratio (CSR), the cyclic tests on the sand-bentonite mixtures show a significant increase of the number of cycles required for liquefaction compared to the clean sand. This is caused, as observed in resonant column tests, by an increase of the elastic threshold due to the presence of bentonite, which delays the generation of excess pore pressure. Such behavior can be explained by the rheological properties of the pore fluid. Oscillatory tests conducted with a rheometer on bentonite slurries show that for shear strains as large as 1% these materials exhibit elastic behavior with a constant shear modulus. Moreover, due to the thixotropic nature of the bentonite slurries, their storage modulus shows a marked increase with time. This observation is consistent with the increase in the liquefaction resistance of the sand-bentonite mixtures with time also observed in cyclic triaxial experiments.

This paper presents the results of static and cyclic undrained triaxial testing on underflow (coarse fraction) tailing materials for a very high proposed tailing storage facility in seismically active southern Peru. The results of these tests, ranging in initial effective confining pressures from about 0.1 to 5.8 MPa, showed contractive but generally ductile behavior at higher initial effective confining pressures likely caused by grain breakage. These test results formed an important part of the design of the tailing storage facility.

In California several geotechnical projects located in seismically active areas have encountered recent alluvial deposits with varying mica contents. Post-construction liquefaction susceptibility evaluations can be made, however case histories do not exist for soils with considerable mica contents, so it is difficult to evaluate if the results from the post-construction investigation are accurate. Cyclic triaxial tests on isotropically consolidated specimens were performed to evaluate the effect of mica content on the liquefaction of poorly-graded sand. The results show that fewer cycles are required to initiate liquefaction for specimens with higher mica contents. It should be noted that the impact of mica particles on the maximum and minimum void ratio suggests that the maximum compacted density of mica-sand mixtures decreases with an increase in mica content. Therefore, if relative density is used as an indicator of liquefaction potential for sands with varying mica contents, than a correction factor should be applied to the in-situ field tests, such as SPT blow counts or CPT tip resistance and sleeve friction values.

Two variables are essential for the assessment of liquefaction susceptibility of sandy soil layers; the seismic demand expressed in terms of cyclic stress ratio, CSR; and the capacity of the soil layers to resist liquefaction, expressed in terms of cyclic resistance ratio, CRR. The variation of the safety factors (CRR/CSR) with depth were determined for set of soil profiles where CSRs were calculated using stress reduction factors proposed in the literature and CRRs based on SPT blow counts. In addition, CSRs were also determined based on site response analyses. A total stress approach (Shake91) and an effective stress approach (Cyclic1D) were used for site response analysis. Safety factors for effective stress analysis were calculated using the CSRs and CRRs both based on the calculated excess pore water pressures for the analyzed soil profiles. A relationship between CSR and number of cycles with respect to excess pore water pressure ratio, r_u was derived based on previously conducted cyclic laboratory test results. In the second stage of the study, the effects of input motion on liquefaction susceptibility were evaluated. Site response analyses were conducted using 24 previously recorded acceleration time histories that are compatible with the earthquake hazard for the investigated site scaled to the same PGA. All the results are compared and the differences are discussed in terms of the final assessment of the liquefaction susceptibility.

This paper describes laboratory-based liquefaction testing of aged soil deposits to address the need for quantifying the influence of aging on liquefaction potential. The in situ aging of liquefiable soil deposits has been qualitatively shown to result in an increase in liquefaction resistance. For example, Pleistocene age sandy soils will generally have a higher resistance to seismic induced liquefaction than similar sandy soils of Holocene age. However, a quantitative measure of this liquefaction resistance increase for use in liquefaction analysis is still unrealized. Cyclic simple shear testing is used in this study to systematically quantify the influence of aging on liquefaction potential. The influence of aging is then measured as the difference between liquefaction triggering of an aged sample versus a similar reconstituted sample. Uncertainly in the field and lab testing results is explicitly defined in a probabilistic manner so that the age-based correction factor can be used in performance-based analysis. Preliminary results from this ongoing research project are presented.

The critical state is the state at which a mass of soil is sheared at a constant volume, constant effective stress, and constant shear stress. This state is reached at an undrained or constant volume condition and at large strains at which all potential net particle reorientation and crushing are complete. Triaxial compression tests are most commonly used to assess critical state conditions; however, as a result of the limited displacement of the device, triaxial testing may not reach a true critical state. In contrast, the ring shear test allows very large shear displacements where all net particle reorientation and crushing are complete. At the large displacements achieved in the ring shear, particle damage and grain crushing can lead to contractive behavior in initially dense materials. These effects result in critical state conditions that can differ significantly between the triaxial and ring shear devices.

The liquefaction resistance of soils has been noted to be influenced by fines content. At present piezocone measurements are not capable of providing accurate predictions of fines content and are therefore open to uncertainty. The work presented herein, extends previous analyses of liquefaction resistance and develops the framework for prediction of liquefaction resistance through the CPT recovered parameter of excess pore pressure (u₂) measured in flight. The main driving force behind liquefaction is the build-up of excess pore pressure in the soil structure; therefore the use of the piezocone measured excess pore pressure is a preferred variable to fines content in the analysis of liquefaction resistance.

This paper presents the results of shaking table model tests and numerical modelling studies which were carried out to gain insight to liquefaction mechanisms in layered sand deposits. The experiments were performed by means of a shaking table testing system on uniform sand columns, silt interlayered sand columns and two layered sand columns deposited at various relative densities and subjected to different input excitations. Excess pore water pressures generated during dynamic loading were measured by pore water pressure transducers installed at three different depths. The physical model tests are also numerically modelled to analyse the dynamic behavior of uniform and layered sand columns by using the Towhata-Iai liquefaction model. The experimental and numerical analysis results revealed that the presence of a less permeable silt interlayer within the sand deposit and existence of a loose sand layer underlying dense sand deposits can have significant effect on the liquefaction behavior.

Sand boils are a long-recognized phenomenon associated with soil liquefaction. Deep liquefied sand is acknowledged to be able to burst through overlying impermeable layers. However, the exact conditions required to produce sand boils are unclear, as is their relative importance to engineering structures. To investigate the phenomenon further, a series of simple model tests were performed. The models consisted of small sand samples containing one or more layers of consolidated silt, visible in profile, which are then vibrated. Visual monitoring was taken of structural settlement and the influence of multiple layering, and the ensuing water films and sand boils observed. It was found that relatively heavy foundations over relatively thin silt layers were susceptible to sand boils. If the silt layer was too thick to be unduly strained by the foundation then the occurrence of water films was suppressed. Buried silt layers were seen to experience a hitherto unobserved phenomenon whereby silt was carried up towards the soil surface with the pore fluid during shaking. It is concluded that the relative importance of sand boils and associated layered-soil phenomena is an area requiring further investigation.

Variation of pore structures of soils due to seismic activities affects density and fluid distribution in the pores, which in turn could affect the strength and liquefaction potential of the porous media. This paper, based on experimental investigation, studied the effect of vibrations, on pore sizes of porous media. A monolayer of glass beads of different sizes under full saturation condition was used to simulate porous media, and the glass beads were subjected to the vibrations provided by a small-scale shaking table. A microscopic camera, which was positioned above the glass beads and connected to a computer, captured the pore variations during the vibration at 1sec interval. Then, graphical software was employed to analyze the changes of pore size distributions before, during, and after the vibrations. The experimental study revealed that the pore size distributions of saturated and densely packed glass beads changed little before and after the vibration. During the vibration, however, the number of bigger pores decreased and the number of smaller pores increased. It may be concluded from this preliminary experimental study that although the pore structure of dense subsoil may remain relatively unchanged before and after seismic activity, the soil could experience significant change of pore structure during the vibration, which could affect the soil's density, strength, and liquefaction potential.

The Standard Penetration Test (SPT) has long been used by the geotechnical community to evaluate the factor of safety (FOS) against liquefaction and to estimate liquefaction induced soil strength loss, vertical settlement and lateral spread. There are other investigative methods of evaluating liquefaction risk, the most common being the cone penetration test (CPT). However, the SPT method is often the first choice for geotechnical professionals because the testing is economical, the equipment readily available, and it returns physical samples that can be tested in the laboratory. The public and governing jurisdictions trust that the geotechnical community is using field-testing methods that provide relatively accurate soil strength data, which is critical to liquefaction evaluation. Because most jurisdictions in the United States do not require the use of calibrated SPT hammers, the majority of SPT liquefaction analyses are based on uncalibrated SPT hammer data. There appears to be a perception by governing jurisdictions that it is acceptable for geotechnical engineers to rely on their engineering judgment to determine SPT hammer energy. However, numerous past studies demonstrate that SPT hammer energy varies widely depending upon the hammer type, manufacturer and lifting mechanism. The variance in energy can be as much as 200 to 250 percent, which is too large to reasonably estimate using engineering judgment alone. This paper focuses on the influence of hammer energy on the SPT liquefaction analyses results. We examine the variability within the results as a function of measured hammer energy. Finally, the paper encourages public agencies to require the use of calibrated SPT hammers in the evaluation of liquefaction triggering, strength loss, dynamic settlement and lateral spread hazards. It is our hope that this practice will be adopted in local and state building codes.

On the performance based design, it is very important to predict the deformation of structures. In order to predict the deformation of structures, we have to know how much the ground will move or how much earth pressure will work. Especially, a prediction of behavior of a liquefied ground beneath structures is very important because of its large deformation. Many studies on the behavior of liquefied soil were conducted to evaluate the characteristics of soil in the post liquefaction stage. There are two kinds of idea about the characteristics of post liquefied soil. Some papers demonstrate that the liquefied soil behaves like a solid which has very low shear strength. In the other idea the soil is considered as a liquid. At the first stage of deformation, the soil behaves like a very weak solid or a liquid. However, it recovers the shear strength after the large deformation due to dilatancy caused by the deformation. This study aims to clarify the effects of some factors on the rigidity recovery of post liquefied soil. This paper describes results of a series of torsional hollow cylinder shearing tests to evaluate the shear stress — strain relationship of liquefied soil. Effects of fine particles content, relative density, overburden pressure, and maximum shear strain during the shaking are discussed.

This paper present a K_D-CRR corrected relation for sandy soils potential liquefaction evaluation under cyclic loading, on the basis of previous tentative CRR-K_D curves proposed. The new tentative correlation for evaluating CRR from K_D, was formulated by combining previous CRR-K_D correlations with the vast experience incorporated in current methods based on CPT and SPT for the evaluation of liquefaction potential. The Seismic Dilatometer Marchetti Test (SDMT) has the advantage, in comparison with CPT and SPT tests, to measure independent parameters, such as the Horizontal Stress Index (K_D). A re-evaluation of CRR-K_D correlation has been also made taking into account the numerous recent SDMT. The results show that the Horizontal Stress Index (K_D) is a key-parameter and is very sensitive to potential liquefaction behaviour under cyclic loading. The plotted correlation with critical values of K_D is very simple to use for detecting liquefaction potential of soils. CPT, SPT and SDMT data have been then used in the evaluation of the potential liquefaction at San Giuseppe La Rena site in Catania (Sicily, Italy), which is recognized as a typical Mediterranean city at high seismic risk.

Damages attributed to earthquake-induced liquefaction have resulted in substantial loss of lives and property around the globe. This paper reports on a study that evaluates the effectiveness of bagged and boosted decision-tree machine learning methods to develop models for assessing liquefaction potential. Major advantages of decision tree learning include their ease of use and their production of easily inspected models with which engineers can better understand the mechanism of the liquefaction phenomenon and verify the trained models. This study demonstrates that the accuracies of the models learned by either bagged or boosted decision-tree learning are competitive with accuracies reported for other learning methods, such as neural network learning. Test-set accuracies of the models learned in this study were 96% on a dataset of 178 cone penetrometer soundings and 89% on a dataset of 225 shear wave velocity soundings, with combined training and testing accuracies of these models on the datasets of 98% and 96%, respectively.

An experimental study on the effect of seepage force on liquefaction of soil in embankments was performed using centrifuge modeling. A total of 5 model tests were conducted on the centrifuge at Case Western Reserve University. The model configuration was an embankment made of sand overlying a sand foundation. Three of the models were submerged while the other two had pore fluid only on one side of the embankment with seepage in the embankment. In the tests, the relative densities of the sand making the embankments were varied while the density of the soil in the foundation was kept the same. Model preparation techniques and test procedures are described. Pore pressure transducers, accelerometers and LVDTs were used to record the response during the earthquakes. The settlement of embankment was measured and profiles of the embankment before and after earthquakes were compared. In two of the tests, a viscous pore fluid was used to satisfy scaling laws for centrifuge modeling. The effect of the pore fluid viscosity on liquefaction behavior was investigated by comparing the results with that from model tests with water as pore fluid. The study found no conclusive evidence that seepage force affects the liquefaction of soil in embankments.

Although there exists some consensus regarding seismic soil liquefaction triggering assessment of free field soil sites, assessing liquefaction triggering potential beneath building foundations still stays as a controversial and a difficult issue. Liquefaction triggering potential under building foundations is affected by both the static and cyclic stress state of the soil medium. As part of these studies, conventionally used normalized cyclic demand term, cyclic stress ratio corrected for K_α and K_σ effects adjacent to and under building foundations are to be estimated. For this purpose series of 3-D numerical simulations have been performed for a number of generic soil-structure systems. A representative CSR value, CSR_SSEI,rep varying with depth was defined, which has potential use in assessing liquefaction-induced bearing capacity failures. Then a simplified, semi empirical procedure was recommended for the estimation of CSR_SSEI,rep values which was further calibrated through the results of 3-D numerical simulations. Good match between the predictions of the recommended semi-empirical procedure and the results of numerical simulations are concluded to be convincing and promising for the assessment of liquefaction triggering potential for foundation soils.

Centrifuge experiments were performed at the NEES@Davis Geotechnical Centrifuge Facility to develop insight into the seismic performance of structures at sites undergoing liquefaction. Three structural models, each representing simple structures founded on shallow rigid mats, were placed on each soil model. The soil profile contained a shallow layer of potentially liquefiable sand to investigate the effects of foundation width and contact pressure on building performance at sites with soil liquefaction. The base of the model container was excited by a sequence of three horizontal shakings, simulating increasing intensities of the Port Island down-hole recordings from the 1995 Kobe Earthquake. Two experiments are described wherein the liquefiable saturated, loose Nevada Sand layer varied in thickness from 3 m to 6 m (prototype scale). A summary of building and free-field movements observed in each experiment is presented with a comparison of the centrifuge results with actual building settlements observed in previous earthquakes. The results from these experiments provide valuable insight, because engineers generally apply methods developed for estimating free-field seismic settlement of clean sand deposits in evaluation of potential building movements at sites prone to liquefaction.

The dynamic lateral subgrade reaction generated on a pile foundation in saturated ground consists of an effective stress component and a pore water pressure component. These components were measured through a shaking table test using a centrifuge facility. The total stress and the pore water pressure acting on the pile surface were measured directly using earth pressure meters and pore pressure meters, respectively. These components were measured on opposing sides of the pile in the direction of shaking. The effective stress was evaluated as the difference between the total stress and the pore water pressure. It was found that the compressive effective stress dominates on the side of a pile where the adjacent soil is under compression. On the other hand, the tensile pore water pressure dominates on the side of a pile where the adjacent soil is under extension.

The paper presents a brief outline of a simplified analysis procedure to analyze the flexural response of axially and laterally loaded piles under liquefied soil condition. Parametric study has also been presented in this paper to bring out in to focus the influence of various factors like gradient of surface topography, depth of liquefied zone, pile boundary condition, presence of axial and lateral loads etc, controlling the pile behavior under such condition.

Finite difference (FD) solutions for static and dynamic Winkler's foundation models are applicable in modeling pile responses. In this paper, relevant solutions on liquefaction induced lateral spreading are presented. Direct and indirect earth-pressure approximations and implementations are introduced. Pile foundation failures of 1995 Kobe earthquake were examined using these models. It was found that both the static and dynamic analyses could provide rational pile displacements in agreement with field observations. The largest pile displacements were found at the pile head from static modeling and the dynamic one using indirect earth pressures. For dynamic solutions with direct earth pressure approximations, maximum pile displacements were found at pile tip. These solutions seem reasonable to model different types of lateral spreading. The mechanism of ground motion, strongly affected by geologic and geographic site conditions as well as the soil-foundation-structure interactions, needs to be carefully verified before applying these solutions.

This paper examines the settlement of instrumented 2 x 2 model pile groups in liquefiable soil based on the results of dynamic centrifuge tests. The piles are end-bearing in dense sand, and are instrumented such that base, shaft and total pile load components can be measured. The data suggest that the overall co-seismic group settlement is accrued from incremental settlements of the individual piles as the group rocks under the action of the kinematic and inertial lateral loads. A Newmarkian framework for describing this behaviour is presented in which permanent settlement is incremented whenever the load in any of the piles exceeds the capacity of the soil to support the pile. This bearing capacity of the piles in liquefied soil is estimated based on measured dynamic soil properties during shaking and observations of the changes in load carried by the piles. The contribution of the pile cap in reducing settlement is also discussed.

Investigation of the correlation between so-called engineering demand parameters (EDPs) and various intensity measures (IMs) has received substantial attention in structural earthquake engineering, as accurate prediction of seismic demand is desired in performance-based seismic evaluation of structures. Little work however has been performed quantifying the seismic response of pile foundations within a performance-based context. In this study the seismic demands of pile foundations are investigated in a performance-based approach. A simple soil-pile-structure model consisting of a two layer soil deposit with a single pile and a single-degree of freedom superstructure is used in a parametric study to determine the salient features in the seismic response of the soil-pile foundation system. A suite of 40 `ordinary' ground motion records were scaled to various ranges of intensity to probabilistically investigate the full range of pile behaviour, from initial elastic response to failure. Various IMs are used to inspect their correlation with the seismic demand, which is measured in terms of: peak lateral displacement of the pile head (relative to pile toe) and peak pile curvature. The lowest scatter in the prediction of pile response was observed when seismic intensity was measured in terms of velocity-based IMs.

Liquefaction induced soil pressures and lateral spreading demands are often the major cause of pile foundation failures. Review of case studies and data indicate that: (i) liquefaction-induced loads include soil pressures and potentially large displacements due to laterally spreading soil, both of which impose significant deformation demands on below-ground piles, (ii) these large deformation demands may result in plastic hinging of the piles, and (iii) the magnitude, location, and spread of the plastic hinging has not been comprehensively investigated. This numerical study evaluates the inelastic behavior of reinforced concrete pile foundations under liquefiable and lateral spreading soil conditions. The parametric study considers different structural and soil details, as well as different simplified analyses methods commonly used for applying liquefaction-induced demands. The application of liquefaction-induced demands is studied using either (i) applied pressures or (ii) applied lateral spread displacements. Of the available methods for estimating lateral spread displacements, the four parameter empirical model proposed by Bardet et al. (2002), is used and observed to be conservative in comparison with other empirical and analytical models. This investigation reveals that the most pronounced variation in inelastic demands, as characterized by maximum bending moment depth and plastic hinge length, is attributed to the strength and stiffness of the pile as well as the layering condition of the soil. In particular, the existence of a non-liquefiable layer (crust) at the ground surface significantly affects the demands on the pile shaft.

Two centrifuge model tests at 40g level were examined in order to evaluate the use of particle size reduction as an alternative method to construct centrifuge model tests of a pile foundation undergoing lateral spreading of liquefied soil. The main difference between both tests was the method used to match the prototype permeability coefficient: (1) viscous pore fluid instead of water and (2) reduction of the particle size while using water as pore fluid. The two soils were placed at the same relative density, which resulted in two very different soil compressibilities and different model responses. Other centrifuge tests reported elsewhere (Gonzalez M., 2007), indicate that when compressibility rather than relative density is matched, particle size reduction produces the correct soil and pile response that agrees. Therefore, both permeability and compressibility need to be matched when using the particle size reduction technique.

The 1999 Kocaeli earthquake caused severe damages, most notably in areas closer to the epicenter. Most of the major property damage resulted from liquefaction and cyclic failure. Many buildings settled and tilted due to the liquefaction and bearing capacity failures. Damages of loss of coastal lands and subsidence of the ground occurred at several sites along Izmit Bay due to liquefaction and cyclic failure. Several site surveys, soil investigations and static and dynamic laboratory soil tests were conducted covering most of the affected area. The results presented in this paper are part of an investigation including series of tests on the behavior of saturated natural fine-grained soil samples, taken from the area. The samples were tested using stress-controlled cyclic triaxial and cyclic torsional apparatuses. The results were analyzed for the assessment of dynamic instability of fine-grained soils by using the energy criteria, which is frequently used for liquefaction analysis. Analysis of cyclic test data suggests that portions of fine-grained soil deposits, i.e. silts and silty clays, likely developed relatively high excess pore pressures and significant shear strains during the earthquake that led to particle structure breakdown and strain softening, and thus, likely contributed to the observed overall stability problems.

The cyclic shear response of natural low-plastic Fraser River silt was investigated using constant volume direct simple shear tests, employing specimens prepared from undisturbed field samples and by reconstitution of the same silt material. The specimens of undisturbed silt, despite having a looser density under identical consolidation stress conditions, exhibited extensively more dilative response and larger shear resistance in comparison to those displayed by counterpart reconstituted specimens. In addition to consolidation stress conditions and resulting void ratios, it appears that other naturally inherited parameters such as soil fabric and aging effects would critically influence the shear response of natural silt. Since the shear response observed from reconstituted specimens seem to result in conservative design parameters, it would be prudent to use specimens from undisturbed samples to assess the expected field behavior of natural silt.

The shear force on a metal coupon dragged through liquefied sand was measured in a modified triaxial apparatus. This device allowed much larger strains and strain rates to be imposed than in a conventional triaxial liquefaction test. The effect of varying fines content on the post-liquefaction large-strain residual shear resistance was measured for specimens prepared at a range of different densities, and compared with residual strength values back-calculated from field liquefaction slope failures. Increasing fines content was found to decrease the residual strength compared to that of clean sand under similar placement conditions. Values measured in the laboratory under truly undrained conditions were comparable to lower-bound values obtained from field cases, suggesting that the laboratory values represent a conservative lower bound for fully-undrained conditions during sliding.

Equivalent granular void ratio and equivalent granular steady state line was used as the common basis for synthesizing test results. In isolating the influence of host sand, two sand-fines mixes with different host sand, but the same fines type and content, were tested. The influence of fines type was then investigated by testing Sydney sand mixed with three different types of fines. Scanning Electron Microscope (SEM) photography was also used to elucidate the role of angularity of fines. This led to the finding that the angularity of fines may need to be considered in determining equivalent granular void ratio.

A series of monotonic and cyclic triaxial tests were performed on a sand with fines sourced from Christchurch, New Zealand. The sand was sieved and then mixed to give three soils with different fines contents. The undrained tests were used to examine the effects of fines on the strain softening behaviour under monotonic loading and liquefaction resistance in cyclic loading. Two reference states were used as a basis for evaluation of the effects of fines: the relative density and the steady state line within the state-concept framework for sand characterization. The addition of fines to the sand base caused downward movement of the steady state line in the D_r-p' plane (e-p' plane), and this effectively increased the potential for strain softening or flow deformation. Samples prepared at an identical relative density showed decreasing cyclic strength with increasing fines content. Conversely, samples at an identical initial state relative to the steady state line showed increasing cyclic strength with the fines content.

Sands and silty sands with the same liquefaction resistance can have significantly different hydraulic conductivity and coefficient of consolidation. Numerical simulations of cone penetration resistance taking into account the effects of silt content and consolidation characteristics show that the penetration resistances are significantly affected by consolidation characteristics for sands and sand-silt mixes with similar liquefaction resistances. The influence of permeability and consolidation characteristics on cone penetration resistances are presented. Relationships between liquefaction resistance, cone penetration resistance, and a normalized penetration rate are presented and compared with current field-based CPT-liquefaction screening methods.

Civil infrastructure built on alluviums and recent deposits, such as dams, contain significant amount of silts. The static and dynamic behavior of these fine-grained soils has been investigated less than the clay-like or sand-like soils. Low plasticity silts (PI = 6) obtained east of St. Louis in Illinois are known as loess that has been re-deposited by water in the floodplains of the Mississippi River. These silts were reconstituted in the laboratory by slurry at water content above the liquid limit and then consolidated to an initial effective stress. The initial laboratory characterization under monotonic loading included a series of consolidated undrained triaxial compression tests at different effective confinement to determine the critical state parameters. A series of stress-controlled cyclic triaxial compression tests were run under normally and overconsolidated conditions. The liquefaction behavior of the silt at different over consolidation ratios and its relationship to the monotonic behavior is presented and discussed.

Results of undrained cyclic triaxial tests on sand-(nonplastic) silt mixtures (under conditions of constant global void ratio) reported by Xenaki and Athanasopoulos (2003) are further analyzed in the framework of recent advances on the subject of liquefaction behavior of fine-grained soils. The observed behavior, for all values of fines content (0% to 100%), is found to have the characteristics of the "sand-like" behavior described in the literature. The same characteristics are also exhibited by plastic sand-clay mixtures with plasticity index ranging from 2 to 11. The undrained initial modulus of the specimens and the amplitude of pre-liquefaction strains were found to decrease and increase, respectively, with increasing fines content, thus negating the beneficial effects of dilatancy in this type of soils. The existence of a critical value of fines content, FC_cr, at which liquefaction resistance is minimized, is confirmed and shown to be useful in describing the behavior of the mixtures in terms of their relative density.

This paper presents the results of a laboratory study addressing the post-liquefaction undrained shear behavior of a silty-sand characterizing the main liquefiable layer within Bonneville sediments at the Farmington Siding Landslide Complex (FSLC). FSLC is located in Davis County, Utah where liquefaction flow-slides associated with past earthquake activity in the region have been documented. Undrained cyclic triaxial tests to trigger liquefaction in water-pluviated samples of Bonnevile silty-sand consolidated isotropically under various effective confining stresses were followed by undrained monotonic triaxial compression tests on the liquefied specimens. The experimental results revealed a dilative behavior of the liquefied soil element, and the post-liquefaction undrained shear strength showed an increase with increasing initial effective consolidation stress. An infinite slope analysis addressing the onset of flow-slide in a liquefied Bonneville silty-sand deposit, using the post-liquefaction undrained shear strength data from laboratory testing, resulted in required minimum slope inclinations much greater than the actual slope gradient at the FSLC.

This paper presents a probabilistically-based semi-empirical model for the assessment of cyclically-induced shear and post-cyclic volumetric (reconsolidation) straining of saturated fine-grained soils. Consolidated-undrained, strain-controlled static, and stress-controlled cyclic triaxial tests have been performed on undisturbed silty and clayey samples for the purpose of compiling a database composed of induced maximum cyclic shear and post-cyclic volumetric strains along with Atterberg limits, natural moisture content, undrained static shear strength. The maximum likelihood methodology is used to develop limit-state models incorporating the selected descriptive variables for the estimation of cyclically-induced soil straining. Results are summarized in the form of a semi-empirical stochastic model which enables the estimation of cyclically-induced maximum shear and post-cyclic volumetric straining as a function of liquid limit, plasticity index, natural moisture content, undrained shear strength, cyclic shear stresses.

Based on the results of assessment of various preparation methods, the specimen of the Mailiao silty sand was prepared by the staged sedimentation method in the large biaxial laminar shear box of 1880 x 1880 x 1520 mm (L x W x H) at the National Center for Research on Earthquake Engineering (NCREE). It was found that there is no significant difference of the densities and fines contents of sand at various depths, except a very thin layer (< 1 mm) of soil with a higher fines content at the top of the specimen. The P-wave velocity measurement confirmed the saturation of the silty sand specimen. The relatively low permeability and high compressibility of the Mailiao silty sand resulted in a longer duration of dissipation of generated pore water pressure and a larger volumetric strain after shaking compared with those of the clean Vietnam sand. Sand boils were observed when water rushed out through the thin relatively impervious layer on the surface of the specimen.

The majority of large embankment dams in the United States in operation today were constructed over half a century ago — before major earthquakes in 1964 jump-started geotechnical earthquake engineering and initiated systematic evaluations of the seismic stability of these earth structures and the soil deposits on which they are built. Many dams have required strengthening owing to increased levels of anticipated shaking or inadequate resistance of the embankment or foundation materials. Soil stabilization and ground improvement methods have included buttress fills, removal and replacement of weak foundation soils, deep dynamic compaction, vibro-compaction, vibro-replacement, explosive compaction, deep soil mixing, jet grouting, compaction grouting, and permeation grouting. In most cases the design strategy is to strengthen the dam to prevent vertical crest deformations that would exceed the freeboard and transverse cracking to depths greater than the remaining freeboard after the earthquake. Special attention must also be given to assuring that seepage paths cannot form along conduits or other discontinuities. In many cases stabilization to prevent excessive upstream deformations is costly and difficult, so in recent years there have been several cases where ground improvement and construction of stabilizing trenches and berms has been confined to the downstream slope and foundation area.

In order to reduce potential settlements and liquefaction susceptibility, two passes of deep dynamic compaction (DDC) were adopted to density the loose, saturated granular backfill between steel sheet pile walls in peat bogs at the relocated U.S. Route 44 in Carver, Massachusetts. Piezocone penetration tests (CPTu) were performed before and after DDC to verify compaction effects. The comparison of pre- and post-compaction CPTu results indicated that the strength and stiffness of the backfill were significantly improved after each pass of dynamic compaction. Under the design load, around 40 percent of the estimated settlements in the backfill could be reduced after one pass of DDC, and up to 84 percent after two passes. At the same time, the backfill was densified to the level of resisting potential liquefaction under the design earthquake motions.

One potential method for soil improvement against soil liquefaction is accelerated drainage through in situ vertical drains. Vertical drains are utilized to expedite the dissipation of excess pore water pressures by reducing the length of the pore water drainage path. For more than thirty years, gravel drains have been employed to ensure the excess pore water pressure ratio (r_U = Δu/σ'_v) remains below a prescribed maximum value. Recently, interest in the use of prefabricated vertical drains has increased because these elements can be installed with less site disruption than with other soil improvement methods. This paper presents the results from initial centrifuge experiments intended to evaluate the effectiveness of prefabricated vertical drains for liquefaction remediation. The centrifuge model includes sloping ground towards a central channel. Prefabricated vertical drains were installed in the soil upslope of one side of the central channel, while the other side was left untreated. Based on the results of the experiment, the prefabricated vertical drains were effective at draining and dissipating the excess pore water pressures, as well as significantly reducing the horizontal and vertical deformations.

A comparative study of two different concentrations of colloidal silica gel and sand mixtures is presented. The dynamic properties, shear modulus and damping ratio, of No. 120 Nevada sand permeated with colloidal silica gel were determined using resonant column tests. The behaviour of the gelled samples is compared to untreated sand samples. The effects of the concentration of colloidal silica gel by percent weight, cyclic shear strain and aging on the dynamic properties of the gel-sand mixtures were studied. Results show there is a small difference in shear modulus with concentration of colloidal silica gel by percent weight and the gel-sand mixtures have a slightly higher shear modulus than untreated sand. The presence of the silica gel does not greatly affect the damping ratio. A 5% by weight gel-sand sample subjected to a long-term, small-strain cyclic shear resonant column test for a duration of 28 days after gelation shows an increase of 6 MPa in the shear modulus.

This paper presents a case study of compaction grouting used as a mitigation measure for soil liquefaction beneath a proposed two-story retail building with integral parking structure. The site is in an area that experienced significant seismic shaking, liquefaction, and structural damage during the 1994 "Northridge" earthquake. Analysis suggested potential for a lateral spreading hazard into the adjacent Los Angeles River. Compaction grouting was used to treat loose to medium dense saturated alluvium to a depth of 17.1 m. A test grout section was conducted prior to production to help determine optimum spacing of grout injection locations. Records of injected grout quantities were maintained in order to provide an indication of relative soil displacement with respect to treatment depth. A grout flow logger was used periodically to acquire real-time pressure and flow data during production grouting. Evaluation of these data provided a means of detecting soil response, providing a basis for establishing practical quantity limits when working near a buried storm drain pipe and a concrete-lined embankment. Comparison of pre-grouting and post-grouting soil test data showed significant improvement in liquefaction resistance, and estimated total and differential settlement was reduced from approximately 280 to 35 mm. Ground improvement was sufficient to allow the local building department to approve a previously infeasible project and to allow an alternate (more conventional) foundation system, representing significant project economy.

Blast liquefaction testing was used to evaluate the effectiveness of vertical drains in preventing liquefaction. As part of the study, settlements associated with drain installation and blasting were measured. In addition, CPT soundings were made before and after drain installation and at 1, 2, 4, and 8 weeks after blasting. Drains installed with high vibration produced 350 mm of settlement while drains installed with low vibration produced only 100 mm of settlement. Despite the settlement, CPT soundings 3 to 5 days after drain installation showed decreases of 66% to 50% in the cone resistance at the low and high vibration test areas, respectively. In contrast, 13 days after drain installation and 7 days after blasting, which produced another 300 to 400 mm of settlement, the cone tip resistance was 30% to 60% higher than prior to treatment. Although there was some evidence of increased penetration resistance at times greater than 7 days, the trends were inconsistent.

To investigate the consequences of liquefaction with regard to foundation performance, subsurface blasting was used to induce liquefaction around a test shaft for a major foundation load testing program in Charleston, South Carolina, USA. Cone penetration test data were recorded before blasting, immediately following two blast-induced liquefaction events while pore pressures were still elevated, one month after the blasting, and again, seven years after the blasting. All testing was located inside of the blast array (i.e., within the liquefied area). Tip resistance, sleeve friction, and u₂ porewater pressure measurements were all lower in the two soundings that were performed while the sands were still liquefied. The post-blast soundings performed one month later and again seven years later indicated that the upper portion of the liquefied deposit (immediately below an impermeable clay cap) had been loosened. There was little change between the data collected one month after blasting and those data collected 7 years after blasting. Densification was not apparent. However, the sand was Pleistocene Age and the destruction of aging effects may have countered blast-induced densification effects in the CPT data.

Stone-columns were used as part of Interstate 5 (I-5)/Interstate 805 (I-805) Highway Widening Project undertaken by the California Department of Transportation (Caltrans). During the stone-column construction, excessive slope movements were observed. Since no measurements were taken during construction, it was not known if the movements were caused by the vibration generated during the stone-column installation. A series of dynamic response analyses was conducted using the 2-D finite element computer program Plaxis to determine the probable causes of the observed slope movements by evaluating the possible correlations between the observed movements and the nearby stone-column installation activities. The overall study consisted of a test stone-column piling vibration measurement program and dynamic responses analyses. The piling measurement program was conducted to gather vibration data at various embankment points due to pilings. Two sets of dynamic response analysis were performed. The first set involved the embankment response analysis during the June 15, 2004, San Clemente Island earthquake. The results of the analysis were used to calibrate the dynamic soil parameters. The second set of analysis was carried out to assess the potential adverse impacts of the stone-column installation to the embankment stability. The second set of analysis included calibration of the dynamic soil parameters using the recorded vibrations from the test stone-column piling program. The results of the analyses indicate that the potential for excessive slope movements and failure due to stone-column piling exists.

The use of geogrid reinforced soil mats has become a viable method to mitigate differential settlements caused by deformations in the underlying soils, including those resulting from fault rupture and seismically induced settlements. In this paper a case study is presented where numerical modeling was used to evaluate the effectiveness of geogrid reinforced soil to mitigate as much as 100 mm of differential seismic settlement. The subsurface investigation, which made extensive use of the cone penetration test, is first presented. Using this information, numerical models were then developed. Descriptions of the modeling procedures are presented and important/sensitive aspects of the models are discussed. The results of the modeling are presented and observations are made regarding performance of the geogrid-reinforced soil system. Areas where additional studies are warranted are identified.

The Vibro-Replacement Method, also known as Stone Columns is one of the most prevalent liquefaction mitigation measure used in the U.S. However, difficulties have been encountered occasionally when applying this method in urban or residential environments because of the potential for negative impacts such as noise and vibration on project neighbors, and the potential for structural damage when it is applied too close to the exiting structures. The SAVE (Silent, Advanced, Vibration-Erasing) Compozer is a newly developed Non-Vibratory Replacement method that does not induce vibration and creates very little noise. The same degree of densification can be achieved when using the SAVE Compozer as can be achieved with conventional Vibro-replacement without the negative impacts of noise or vibration. This paper presents case histories of the first two applications of the SAVE Compozer in the U.S. including verification results of densification performance and vibration monitoring records, together with an overview of the SAVE Compozer method including densification concept, equipment, and execution.

Bender elements embedded in a centrifuge model were used to image a concrete column grouted in sand. The bender elements were oriented in a polar array with the center of the column near the center of the array, and travel-time based tomographic inversion clearly identified the presence of the high-shear wave velocity anomaly. Tomographic imaging was performed using a pixel-based representation, in which the domain is divided into pixels and the shear wave velocity of each pixel is solved by inversion, and using a parametric-based representation, in which the geometry of the domain is assumed and the domain parameters are varied to minimize inversion error. Elimination of time delay present in the system was required to reduce inversion errors. Though the inversion is sensitive to input parameters and data quality, the grouted column was evident in the center of the array for a range of inversion techniques.

Drilled displacement piles are installed by the displacement of soil and subsequent placement of fluid cement grout within the evacuated volume. Depending on the soil grain-size characteristics, soil behavior, in situ soil density, pile spacing, and pile diameter, the installation process can result in measurable densification and an increase in lateral stress. Thus, it follows that the drilled displacement process may be used to mitigate cyclic liquefaction in sandy soils. In an effort to quantify the increase in cyclic liquefaction resistance resulting from drilled displacement piles, the authors performed pre- and post-installation cone penetration testing at a site in North Myrtle Beach, South Carolina. The results support that the installation of drilled displacement piles generally increases the liquefaction resistance of the surrounding soil. For a specific area replacement ratio, the trend is that improvement decreases with increasing pre-installation (q_tlN)_cs and approaches unity at very high values of pre-installation (q_tlN)_cs. The improvement can be significant for values of pre-installation (q_tlN)_cs generally as high as 250. Overall, the degree of improvement is greater for higher area replacement ratios.

Mission Bay is a 1.21-square-kilometer (300-acre) reclaimed site in San Francisco, California, that is currently being redeveloped. The composition of the fill varies and includes loose sand, loose to medium dense silty and clayey sand, and loose to medium dense sandy gravel with clay. The fill is liquefiable throughout much of Mission Bay. The groundwater is typically 1 to 3 meters (3 to 10 feet) below the ground surface. This paper will present case studies of the mitigation of liquefaction potential at two sites using Rapid Impact Compaction (RIC). RIC is a method of soil improvement using a track-mounted machine that imparts energy by dropping an approximately 6,800 kilogram (7.5-ton) weight from a controlled height, about 1 meter, onto a patented foot. The energy is delivered at a rate of 40 to 60 blows per minute. Drop height, number of blows, and penetration per blow are monitored and/or controlled by an on-board data acquisition system. Case studies will include results of pre-treatment borings and/or CPTs, results of post-treatment CPTs, and a comparison of before and after liquefaction potential. In addition, the results of vibration monitoring performed during RIC will be discussed.

Computer aided installation of stone columns combined with insitu instrumented surcharging provided the necessary feedback to improve a difficult site for support of a new gymnasium and health center. The low lying site consists of an ancient landslide with loose sands over bedrock which is overlain by peaty soils. Both liquefaction of the loose saturated sands and settlement of the organic soils needed improvement. Vibro-displacement stone columns combined with surcharging was employed to improve both types of soil present, thus providing the long term performance needed. Computer monitored vibro-replacement provided the necessary real-time information to confirm the stratigraphy, and control the diameter of column installed within the softer surface soils. Final improvement of the surface organic soils was provided by surcharge. An in-situ vibrating wire water level system was installed after the vibro improvement, providing the necessary feedback to confirm completion of the primary consolidation.

Ground improvement using stiff columnar reinforcement, such as stone, jet-grout, and soil-mix columns, is commonly used for mitigation of seismic damage in weak ground. Seismic shear stress reduction in the reinforced soil mass is often counted on for reducing liquefaction potential. Current design methods assume composite behavior of the reinforced soil, where the shear stress reduction is based on the ratio of the columnar stiffness relative to the soil as well as the area, replacement ratio. This implicitly assumes the stiff columns will deform in pure shear along with the soil. We performed 3-D dynamic finite element modeling to better understand the column deformation and shear stress reduction behavior. We found that the columns deform in both shear and flexure, providing little seismic shear stress reduction and current methods may be unconservative.

The effectiveness of ground improvement by the sand compaction pile (SCP) method is generally evaluated through penetration resistances, such as cone resistance (q_c) and N-values of the standard penetration test (SPT), measured at the center point in-between piles before and after the pile installation. However, it is known that the penetration resistance of the ground adjacent to piles does not only increase, but the lateral stress increases as well, resulting in increase in the resistance to liquefaction. Therefore, in order to evaluate the effectiveness of the ground improvement on compacted soils, it is necessary to quantify the effect of K_c-values on the penetration resistance and the resistance to liquefaction. In this paper, based on the results of SPT and CPT performed in chamber box in the laboratory, the relationships between penetration resistance, resistance to liquefaction and relative density were re-examined, and the influence of K_c was investigated. Although the results indicated that generally the resistance to liquefaction increases with increasing K_c-value, little difference was noted when the density of the deposit was high. Based on the results, recommended charts incorporating the effect of K_c were proposed.

This study presents the seismic response of a deep mixed (DM) treated foundation under the fourteen-story Oriental Hotel building in Japan using a simplified method proposed by Siddharthan and Porbaha (2008). This hotel was subjected to intensive shaking during the 1995 Hyogoken-Nanbu (Kobe) Earthquake (M = 6.9) and extensive liquefaction and ground movement at locations near the building have been observed. The hotel that was retrofitted with grid type deep mixing method survived the earthquake with little or no damage. The proposed simplified approach showed clearly the effectiveness of treated columns in reducing the pore water pressure response at locations closer to DM treated zone. The effectiveness of treatment is significant, especially near the surface. Absence of liquefaction within the cells and at locations within up to 2.7 m from the edge column would have played a positive role relative to the accepted performance of the Oriental Hotel during the 1995 Hyogoken-Nanbu earthquake. Accordingly, the proposed simplified approach can be effectively used to analyse various DM configurations, site and excitation conditions. The soil responses are needed within and around the DM columns to ascertain the effectiveness of DM treatment and obtain an optimum design.

The installation of stone columns or piers in a loose, saturated sand deposit can potentially mitigate the risk of liquefaction by decreasing the seismic demand on the soil by redistributing the induced shear stresses from the sand to the columns or piers. In calculating the shear stress redistribution, it is commonly assumed that both the soil and the columns or piers respond as shear beams. Although less common, it has also been assumed that the soil responds as a shear beam and that the columns or piers respond as flexural beams, with the redistribution of the shear stresses computed accordingly. However, the results presented herein show that the columns or piers, and the soil immediately surrounding the column/pier, deform in a combination of shear and flexure. The percent contribution of shear versus flexural deformation of the column or pier varies with depth, with the column/pier deforming predominantly in flexure near the ground surface and predominantly in shear at depth. The percent contribution of each mode of deformation governs the redistribution of the shear stresses from the soil to the pier. The distribution of the shear stresses between the soil and columns or piers are quantified for "typical" properties of a loose, saturated sand profile reinforced with Impact^TM Rammed Aggregate Piers^TM.

Two series of pullout loads tests were conducted on anchors embedded in submerged sand in this study. First series of tests were in submerged sand and the second series of test were in submerged sand reinforced with single layer of geogrid. The geogrid was positioned directly on the anchor and its width ratio was varied (Br/B = 2, 3 and 4). Other parameters varied in both monotonic and cyclic mode of loading were embedment ratio of anchor (H/B=2, 3 and 4) and density of sand (loose, medium and dense conditions). All the cyclic load tests were conducted for the cyclic load ratio of 60%. From the results of study it was found that the load - displacement response of unreinforced and reinforced sand for a given density and depth of embedment are similar except higher peak and residual loads due to the inclusion of geogrid reinforcement. Both peak and residual pullout loads increased with width ratio of reinforcement for the densities and embedment ratios studied in this investigation. In the case of cyclic load, the displacement of anchor was increased with decreasing rate and reached almost a constant value after 350 cycles. The post cyclic peak pullout load was marginally higher than the peak pullout load of monotonic condition for all the comparable test conditions.

Recent changes and consolidation in building codes across the U.S. have put a larger emphasis on the need for cost-effective foundation support of both bearing and uplift foundation loads resulting from wind and seismic forces. The Rammed Aggregate Pier (RAP) system is a commonly-accepted soil reinforcement method that results in highly-densified aggregate columns for liquefaction mitigation, bearing capacity enhancement, and uniform settlement control for conventional shallow foundations. The RAP system provides a cost beneficial alternative to deep foundations such as pre-cast driven piles, auger-cast piles, or concrete piers and to mass excavation and replacement for shallow foundations. In high seismic and wind areas, uplift loads created by overturning forces on foundations are resisted with uplift RAPs that are equipped with a structurally designed steel anchor assembly. Typical uplift capacities in California range from 178 to 445 kN (40 to 100 kips) for 762 to 838 mm (30 to 33 inch), 3.6 to 7.0 m (12 to 23 feet) meter long RAP elements using ASD (allowable stress design) and incorporating a minimum factor of safety of 2. Full scale load tests on uplift RAPs show repeatable tangent stiffness after multiple overstress load cycles. This paper presents the first published results of psuedo-cyclic uplift load tests on uplift RAPs installed on projects in California.

The history of dynamic analysis in geotechnical earthquake engineering is traced from 1952 to 2008. The role of dynamic analysis is illustrated by case histories from several areas of engineering practice. Attempts to validate methods of analysis by blind prediction trials based on data from element tests, centrifuge tests and from instrumented sites in the field subjected to earthquake excitation are described. The factors controlling the reliability of non-linear effective stress analysis are examined and guidelines are suggested for conducting defensible and reliable analyses in practice.

This paper presents an attempt to model the seismic response of a spread footings system founded on liquefiable granular soil deposit using a three-dimensional transient fully-coupled hydromechanical model and taking into account the effects of soil-foundation interaction. The porous soil medium is modeled as a mixture of two interpenetrating phases, namely the fluid phase (water) and the particulate solid phase. The fluid is idealized as a continuum by using averaged Navier-Stokes equations that accounts for the presence of the solid particles. The Discrete Element Method (DEM) is employed to model the assemblage of these particles. The interphase momentum transfer is modeled using established relationship that accounts for the dynamic change in porosity. The spread footing is modeled as a rigid block and its motion is described by the resultant forces and moments acting upon it. A computational simulation is conducted to investigate the response of spread-footings on a saturated granular deposit when subjected to a dynamic excitation. Results of the conducted simulation showed that the foundation sustained excessive settlement as the ground shaking progressed and the underlying soil liquefied. The conducted simulation appears to capture essential dynamic response patterns typically observed in such systems.

A new numerical methodology is presented for the simulation of the non cohesive soil response under small, medium and large cyclic shear strains, with special interest given to liquefaction phenomena. The new methodology is based on a recently proposed elastoplastic bounding surface model, which has been implemented to the 2D finite difference code FLAC, via its User-Defined-Model capability. In particular, the emphasis in this paper is on comparing measurements to numerical results for the seismic response of a liquefiable sand layer that has a mild slope leading to the detrimental phenomenon of lateral spreading. The new methodology is proved capable of accurately predicting the time-histories of ground acceleration, excess pore pressure buildup and accumulation of displacements at all depths within the mildly-sloping liquefiable layer. Moreover, the new methodology provides insight to the mechanism of accumulation of displacement during lateral spreading.

A fully coupled nonlinear dynamic numerical approach and an advanced constitutive model have been implemented in a three dimensional finite element computer code to predict the response of saturated porous media under different types of loading conditions including earthquake loading. The formulation and implementation are already verified and validate, and were then applied to the problem of sand liquefaction and cyclic mobility phenomena for investigation on a specific configuration of layered soil column, where liquefaction of the deeper loose elements prevents transmission of earthquake induced shear stresses to the upper layers.

The dynamics of saturated poroelastic media has received attention of several investigators from the mid twentieth century but problems concerning unsaturated soils necessitate the theoretical developments for mechanics of unsaturated media. By the development of high speed digital computers numerical methods are used to solve complex engineering problems. One of the most common problems in geotechnical earthquake engineering is the seismic wave propagation from the seismic source to the structure's foundation, throughout the stratified sedimentations. The well known boundary element method (BEM) is an effective boundary type numerical method for dynamic analysis of linear elastic bounded and unbounded media. The BEM consists of two key ingredients: the governing boundary integral equation and the corresponding fundamental solutions. This paper presents the full dynamic boundary integral equation of unsaturated isotropic poro-elastic media for the first time. The introduced boundary integral equation can be solved by BEM once the corresponding fundamental solutions are determined.

Earthquake recordings from a model test structure located at Garner Valley California are used to evaluate inertial soil-structure interaction (SSI) effects. The test structure consists of steel columns and bracings 4.06 m in height supporting a 4m x 4m x 0.4m reinforced concrete roof slab. The structure rests on a 4m x 4m x 0.5m reinforced concrete foundation with no embedment. Surficial soil conditions consist of organic soil overlying silty sand alluvium with an average shear wave velocity in the upper 15 m of 200m/s. Seismic monitoring is performed with uni-axial (horizontal) accelerometers on the roof and foundation as well as vertical sensors at the corners of the foundation. Parametric system identification procedures are implemented to evaluate fundamental-mode vibration frequencies for a fixed- and flexible-base condition, from which SSI effects can be inferred. Data is available for the structure with and without bracing from two earthquakes. The lengthening of the fixed-base period due to SSI is a period lengthening of 12% and 74% for the braced and unbraced structure, respectively. The corresponding foundation damping levels for these two cases are about 1% and 5%. The different levels of period lengthening and foundation damping for the two structural configurations reflect the strong increase of inertial SSI effects with the ratio of structure-to-soil stiffness, which is naturally greater for the fixed- base configuration. The observed levels of SSI are reasonably well predicted by available theoretical models.

This paper investigates in depth the dynamic responses of soil-foundation-structure systems using dimensional analysis in order to evaluate the significance of soil-structure interaction effects when subjected to near-fault ground motions. A soil-foundation-structure interacting (SFSI) system is idealized into a simplified 2DOF system where both linear and elastic-plastic structures with flexible foundations are considered. Through rigorous dimensional analysis, the response of superstructure is presented in terms of the dimensionless II-products. This approach brings forward the self-similarity, the invariance with respect to changes in scale or size, which decisively describe the interactive behaviour of a foundation-structure system. Extensive simulations have been conducted on various foundation-structure systems subjected to pulse-type ground motions to numerically describe the structural response, i.e. dimensionless relative displacement, as function of distinctive properties of superstructure and foundation. Results show that soil-structure interaction is insignificant for a variety of structure-foundation systems in practice. However, there are conditions under which the response of foundation-structure system is significantly higher than that of the fixed-base structure.

This paper introduces a new way of estimating the inelastic response of yielding structures with behavior that can be approximated with the elastoplastic idealization. The proposed approach emerges from formal dimensional analysis and is liberated from the response of the elastic system. The application of the proposed method hinges upon the existence of a distinct time scale and a length scale that characterize the most energetic component of the ground shaking. The most decisive feature of this work is that the inelastic response curves that result with the proposed approach assume similar shapes for different values of the normalized yield displacement. Because of this similarity the paper proposes a single inelastic response curve which offers directly the maximum inelastic displacement of the structure given the energetic pulse period and pulse amplitude of the ground shaking.

Foundations for machines with impact loads are widespread powerful sources of industrial vibrations. These foundations mostly transmit vertical dynamic loads on the ground and generate ground vibrations which may harmfully affect surrounding buildings. Dynamic effects range from serious disturbances of working conditions for sensitive devices and people to visible structural damage. Natural frequency of vertical machine foundation vibrations and complete vibration records of ground and structure vibrations can be predicted prior to installation of machine foundations. Diverse measures can be used to mitigate dynamic effects of impact machine foundations.

A numerical model was developed to simulate the response of two instrumented, centrifuge model tests on soft clay. The centrifuge tests consist of a rigid containment to simulate the behavior of rectangular ten-storey buildings on 30 m uniform and layered soft soils. Each test model was subjected to several earthquake- like shaking events i.e. scaled versions of an artificial western Canada earthquake and the Port Island ground motion recorded during the 1995 Kobe Earthquake at a centrifugal acceleration level of 80g. The centrifuge model was simulated with the 3-D finite-difference-based Fast Lagrangian Analysis of Continua (FLAC) program. The results predicted with the use of nonlinear elastic—plastic models for the soil are shown to be in good agreement with measured acceleration, soil response, and structural behaviour. Numerical results were used to study the effect of soil layering and soil-structure interaction on foundation input motion.

This paper presents a framework for finite element response sensitivity analysis of structural and/or geotechnical systems based on the direct differentiation method (DDM). A frame-soil system subjected to both quasi-static pushover and dynamic loads is used as application example. Focus is on the sensitivity of the FE predicted response of this SSI system with respect to the material parameters used to represent the elastoplastic behavior of the various soil and structural materials. The sensitivity analysis results provide the relative importance of the various material parameters for selected global and local response parameters of the SSI system.

This paper presents the results of experimental and analytical studies of the seasonally frozen soil effects on a three-story reinforced-concrete parking garage. With a 16-channel portable accelerograph, vibration test data were collected during the winter of 2006–2007. The structural dynamic properties were identified from the recorded data. Meanwhile, a three-dimensional Finite Element (FE) model including both superstructure and soil has been developed in order to verify the frozen soil effects. The FE modeling results indicate that the fundamental frequency of the soil-structure system increases 50% as the frost penetrates from 0 to 1.4 m. It is found by push-over analysis that the system stiffness increases 114% due to soil freezing. Most importantly, the shear distribution between columns and shear walls is significantly changed due to soil freezing. When soil thaws, the shear force taken by columns increases 10∼15%, and that taken by shear walls decreases by 10∼15%. It also shows that more columns experience yielding at thawing condition. This indicates that crucial zones in the frame-shear wall structures might move from shear walls to columns.

Analytical and numerical solutions are presented for the rigid-plastic response of geo-structures to idealized ground acceleration pulses. These shock-like waveforms are typical of near-fault earthquake motions generated by forward fault-rupture directivity and may inflict large permanent displacements in the absence of substantial residual soil strength. The geo-structures are modeled as rigid blocks resting on inclined frictional planes. Although idealized, these models are widely accepted by geotechnical engineers, for simulating a variety of structures including retaining walls, embankments and slopes. Four basic simple pulse waveforms are examined: (1) rectangular; (2) sinusoidal; (3) triangular; (4) exponential. An analytical study is presented on the effect of frictional strength and number of excitation cycles on peak displacements. Results are presented in the form of dimensionless graphs and closed-form expressions that elucidate the salient features of the problem. It is shown that Newmark approaches based on conventional motions may under- or over-estimate peak displacements depending on the circumstances. It is also shown that all three time histories of ground motion (i.e., acceleration, velocity, and displacement) control peak response — contrary to the widespread view that ground velocity alone is of leading importance. Issues related to scaling laws of peak displacement are discussed.

Seismic soil-structure interaction performance evaluations of projects located in deep soft clay deposits warrant special attention, particularly when designing strategic infrastructure that must remain operating after a major earthquake. The development of numerical analytical platforms in recent decades, along with faster computing tools, have made possible to include in the state of practice quite sophisticated solution techniques aimed at better representing the physics of the problem at hand. This paper presents the application of a 3-D finite difference model for evaluating the static and dynamic response of a 118 by 100 m cellular-raft foundation to be built in soft clay. The raft foundation is a 2.5 m high box-type foundation embedded 1 m and supported by a grid of peripheral and internal walls, 2.5 m long and 0.40 m thick, which integrates a cellular structure. The model is used to obtain first the static behavior exhibited by the foundation for the construction stages, including long term consolidation, and then the design earthquake is considered and the equation of motion is solved in time domain.

Novel analytical solutions are presented for single-degree-of-freedom (SDOF) oscillators founded on footings and piles on compliant ground. First, exact formulas are derived for the fundamental viscoelastic natural period and effective damping of these structures. It is shown that certain approximations proposed in the literature may be inaccurate for highly-damped systems. Second, the influence of foundation mass on natural period and damping is investigated. Third, the effect of coupled swaying-rocking oscillations at the pile head is addressed by translating the reference system to the location down the pile, where the resultant soil reaction is applied. This ensures a diagonal foundation impedance matrix and simplifies the solution. Fourth, the amount of radiation damping generated from a single pile and a surface footing are compared. To this end, a new concept of statically and geometrically equivalent SSI systems is introduced. It is shown that a structure founded on a pile may generate twice the radiation damping of a similar structure on a spread footing.

The rocking of a relatively rigid footing on soil can be visualized as a moving contact problem, with a contact area moving from one side of the footing to the other as the footing rocks. Thus the partial separation of footing (uplift) and the resulting moment capacity can be more closely related to the critical contact area ratio, A/Ac (ratio between the actual footing area (A) and the area required to support the vertical and shear loads (Ac)). This paper compares the moment capacities calculated by using A/Ac with those obtained from centrifuge and 1-g experiments conducted on rocking shallow footings for a range of A/Ac values. It is shown that the moment capacity of a rocking footing is well defined and, unlike bearing capacity, moment capacity is less sensitive to the uncertainty in soil properties as A/Ac increases.

In material modeling, a number of physical properties such stiffness, strength and post-yield capacity can be directly determined through element-level tests. However, to actually generate a numerically robust representation of a materials response, additional parameters describing the backbone and/or cyclic response (e.g., the shape of transitions from linear to nonlinear regions and the full shape of the post-yield region) are needed. These backbone parameters are difficult to characterize directly from an element-level experiment, and thus are not well correlated with typical soil parameters. However, they may be more readily obtained from system-level tests. Therefore, this study focuses on the parameters needed to describe the nonlinear response of Winkler springs, used in the context of Winkler-based shallow foundation modeling. The foundation subgrade is assumed as a system of discrete, mechanistic, uncoupled springs having nonlinear uplift, sliding and damping capabilities. For this type of modeling problem, elastic-plastic spring representations are commonly used in practice. However, under cyclic rocking, physical changes to the underlying soil occur (mostly due to the densification and shearing at the edges), therefore an elastic-plastic representation will not provide a realistic description of the soil-footing behavior. In this paper, data sets from centrifuge and large-scale footing experiments, whereby the response is controlled in a single degree of freedom, are used to determine the parameters describing the nonlinear backbones of mechanistic springs. For this purpose, models of each test are constructed and through comparison of model results with the experimental response, and isolating one parameter at a time, best-fit parameters are determined. Finally, a sensitivity study is carried out to observe the effect of parameters that describe the spring backbone response.

A numerical model is developed using OpenSees to represent the dynamics of a soil model-model container-centrifuge shaking table system. The centrifuge shaker-system includes soil model, the flexible shear beam model container, shaking table and its reaction mass. All of these different components interact with the soil model during dynamic excitation, with some absorbing energy and others allowing undesired modes to affect the response observed in the experiment. This interaction of soil model and centrifuge/actuation system might attenuate or amplify the discrepancies in the responses of the numerical and physical models. The relative error between a numerical simulation and a physical simulation depends on how the boundary conditions and interaction among different components in the physical model are included in the numerical model. Assessment of the quality of a comparison between a numerical and physical simulation should account for the effects of the boundary conditions and dynamic interaction among different components in the dynamic system. This paper outlines the details of the simulation model, presents some representative results from simulations, discusses the effect of interaction among different components on the responses, and presents how the sensitivity of simulation outputs to uncertainties in the material properties depends on boundary conditions in the physical and numerical simulations.

The risk associated with structures subjected to dynamic loading needs a rigorous analysis that takes into account dynamic soil-structure interaction. A key step of this analysis consists of estimating the dynamic response of the foundations by calculating their impedance (i.e. dynamic stiffness K and damping C) or displacement functions (real part F₁ and imaginary part F₂). The purpose of this work is to evaluate the displacement functions (inverse of impedance functions) of strip footings on the surface of some homogeneous soil. Validation studies indicate the accuracy and versatility of the models performed with the software FLAC (Fast Lagrangian Analysis of Continua). It is known that the assumption of homogenous layer or half space with constant shear modulus G may not be realistic as the shear wave velocity increases as a function of the effective overburden stress. In this paper, three soil models are considered. For each case, the adopted mechanical characteristics correspond to a type of soil with constant or variable shear wave velocity. Calculations are performed over a practically sufficient range of oscillating frequency ratios a_o. Comparison of results obtained with varying and constant shear wave velocity shows the importance to consider this velocity increasing with depth. Additional calculations conducted on two-layer soil are also presented in order to recommend the thickness of soil that is required in the model to capture soil-structure interaction.

Recently developed equipment allows measurement of the shear modulus of soil in situ as a function of level of strain. In these field experiments, the excitation is applied on the ground surface using large scale shakers, and the response of the soil deposit is recorded through embedded receivers. The focus of this paper is on the simulation of signals which would be recorded at the receiver locations in idealized conditions to provide guidelines on the interpretation of field measurements. Discrete and finite element methods are employed to model one dimensional and three dimensional geometries, respectively, under various lateral boundary conditions. When the first times of arrival are detected by receivers under the vertical impulse, they coincide with the arrival of the P wave, related to the constrained modulus of the material, regardless of lateral boundary conditions. If one considers, on the other hand, phase differences between the motions at two receivers the picture is far more complicated and one would obtain propagation velocities, function of frequency and depth, which do not correspond to either the constrained modulus or Young's modulus. It is thus necessary to apply some care when interpreting the data from field tests based on vertical steady state vibrations. The use of inverse analysis can be considered as a way of extracting the shear modulus of soil from the field test measurements.

Dynamic behavior of a pile group is simulated by a two-dimensional finite element analysis. Results of the full-scale lateral-load tests of a 3 x 5 pile group conducted in the Salt Lake City International Airport site are compared with the simulation. The soil profile at the test site generally consisted of cohesive sandy silt and silty sand. Closed-end steel pipe piles were driven to a depth of 11.6 m. Pile head was rotation free. Dynamic lateral loads were applied to pile head with the statnamic device. To conduct numerical analysis, model parameters for soils are taken from the geotechnical investigation data at the test site. Response of five piles in the middle of the 3 x 5 pile group are simulated. Measured dynamic loads are applied at the pile head. Overall load-deflection behavior of the pile group agrees well with the measured ones. Some discrepancies are found on the bending moment profile. In the full-scale dynamic tests, peak moments for larger target deflections occurred at the same depths for all piles, while in computation, the peak moment occurs with greater depths as target deflection increases. Numerical analysis overestimates the group effects in dynamic tests. This calls up further improvement of a numerical modeling.

A lateral load test was performed on a full-scale steel pipe pile installed in a soil profile consisting of cohesive, fine-grained soils with occasional sand layers. The pile was loaded in nine increments from 0 to 89 mm of displacement, with 15 loading cycles applied for each increment. P-y curves representing static monotonic and cyclic loading conditions were derived from strain gauge data using a moving cubic polynomial technique. The results of push-over analyses using the derived p-y curves as input agreed well with the measured load-displacement response. It was observed that the cyclic p-y curves exhibited a sharp concave-up shape which contrasted with the broader concave-down shape of the static p-y curves. The cyclic p-y curves also exhibited lower peak resistances relative to the static p-y curves. The initial portion of the cyclic p-y curves is relatively flat with near-zero resistance which represents gapping of the soil. The displacement at which the soil reaction component of the cyclic p-y curves begins to dramatically increase is a function of the maximum displacement during previous loading cycles. It was also observed that the effects of cyclic loading were more pronounced in shallower soils.

Recently, a great attention has been devoted to performance based design in geotechnical engineering, and many design codes have initiated applying its concepts. As a result, estimation of deformation of structures subjected to intense loading, earthquake, is becoming more important. In this regard, physical modeling in geotechnics is a key measure to study the behavior of different structures including pile foundation and quay wall. 1-g shaking model test is one of the most common tools in the field of experimental geotechnical engineering. However, limitation of the size of model decreases its efficiency. To overcome the disadvantage of low confining stress, the world's largest shaking table, 20m x 15m, which is called E-Defense, was established in the National Research Institute for Earth Science and Disaster Prevention (NIED), Japan where full-scale tests on engineering structures including RC building, quay wall, pile foundation, and masonry house are run. This paper aims to illustrate E-Defense facility by giving an example of a full-scale test on pile group and sheet pile quay wall which was subjected to the liquefaction-induced large ground deformation. The quay wall was displaced laterally, and the pile group, which was placed behind it, was damaged extensively. Over 880 channel data of various parameters were recorded and some of the results are presented in this paper mainly on rate dependency behavior of liquefied sand.

A numerical method is presented for analyzing the nonlinear behavior of single piles subjected to harmonic vertical or torsional loading. The method combines the finite element method accounting for nonlinear soil behavior in the near field and a semi-analytical formulation for radiation damping in the far field. Nonlinear stress-strain characteristics of soil are considered with a shear modulus degradation model used in soil dynamics and earthquake engineering. Mohr-Coulomb strength criterion is used to approximate the contact property at the pile-soil interface. The method is simplified but versatile, and the soil parameters required are relatively few being all the standard geotechnical parameters. The comparisons between the predictions and the results from field tests and a boundary element method are made to verify the present method. A parametric study indicates that the linear analysis overestimates both the stiffness of piles and the interaction factor for the pile group effect.

Deepwater subsea petroleum developments present unique challenges for the geotechnical engineer. The combination of very soft soils and moderate to strong seismic ground motion requires innovative foundation solutions. This paper describes the seismic design of suction caisson foundations for subsea infrastructure in a deepwater field. Six-meter diameter, 12-m long suction piles were used to support underwater manifold assemblies. The seismic response of the structures was evaluated using a non-linear analysis in the time domain to predict foundation displacements under the design seismic event. The methodology used to define foundation response and assess structure displacements is presented. Results of the analysis showed that although there was local yielding of the foundation soils, the expected dynamic and permanent displacements were acceptable.

Piles are preferred type of foundation for supporting structures in seismic areas. Their performance is affected by soil, pile and earthquake related parameters. In non-liquefying soils, the shear modulus may degrade and material damping may increase with increasing strain/displacement. In liquefying soils, the lateral ground displacement may lead to increased lateral loads and moments on the pile leading to pile failure in bending. The loss of lateral support can also lead to buckling type failure. These issues for pile design in seismic areas are examined.

The response of pile foundations during past earthquakes have shown that piles in firm soils generally perform well, while the performance of piles in soft or liquefied ground can raise some questions. Pile foundations are very frequently used in soft clay to achieve the required bearing in stiff stratum. This paper presents the results of dynamic centrifuge model tests carried out at the National University of Singapore to physically model dynamic soil-pile response in soft clay. The experimental results strongly suggest that there are significant effects arising from stiffness degradation due to imposed inertial loading of surrounding soil on piles with repetitive shaking during earthquake, and corresponding changes in time period and amplification of ground motion. The study also examined the influence of different relative stiffness of the clay-pile system, and different superstructural inertial loads. The centrifuge tests were back-analyzed via 3-D finite element modeling using the general purpose code ABAQUS. The analyses show that the use of simple non-linear hypoelastic soil models can provide reasonable agreement with the experimental observations.

Lateral pile cap tests were performed on a pile cap with two backfills to evaluate its static and dynamic behavior. In one case the backfill was loose silty sand while in the other case, a 0.91-m wide zone of gravel was compacted between the pile, cap and the loose silty sand. The relatively thin dense gravel zone increased the lateral resistance by 75 to 150% relative to the loose silty sand backfill. The dense gravel zone also increase lateral resistance to 59% of the resistance that would be provided by a backfill entirely composed of dense gravel. The dynamic stiffness for the pile cap with the gravel zone decreased about 10% after 15 cycles of loading, while the damping ratio remained relatively constant with cycling. Stiffness increased by about 10% at higher deflections, while the damping ratio decreased from an initial value 0.30 to around 0.26 at higher deflections.

A method for assessment of the undrained drilled shaft tip load — tip displacement response in sands under earthquake or impact loading has not been available heretofore. Statnamic test results on two WSDOT drilled shaft with the shaft tips in sand below the water table have yielded field evidence with which to compare/calibrate such assessment. A method for tip load — tip displacement evaluation based on the assessed undrained stress-strain-strength of a sand as established from drained tests results and formulation is presented here. It is used to provide assessed response to compare with field recorded behavior.

This paper discusses the results of a dynamic numerical analysis in which a pair of propped retaining walls embedded in a dry coarse-grained soil are subjected to two different earthquakes. The retaining walls have a small embedded length, so that they undergo significant displacements under seismic conditions. Soil behaviour is described with a non-linear elastic-plastic model, with damping resulting from irreversible strains. The seismic input is made of two real acceleration time histories, having a similar Arias intensity but a different frequency content. The response of the soil and the retaining structures to these different inputs is presented, devoting a special attention to the description of soil-wall interaction that occurs during an earthquake.

A closed-form stress plasticity solution is presented for earthquake-induced earth pressures and distribution of these pressures on inflexible retaining walls. The solution is essentially an approximate yield line approach that over- and under-estimates active and passive pressures, respectively. Results are presented in the form of dimensionless graphs and charts that elucidate the salient features of the problem. Compared to Mononobe-Okabe equations, the proposed solution is simpler, more accurate and safe. In addition, it provides a rational means for determining the distribution of limit thrusts on the wall. It is shown that the pseudo-dynamic seismic problem does not differ fundamentally from the gravitational one, as the former can be derived from the latter by means of a revolution of the reference axes. In the second part of the paper, the solution is extended to determine the distribution of limit pressures on a gravity wall by means of simple wave equations. The proposed approach has advantages over earlier efforts by Steedman and Zeng, as it satisfies the stress boundary conditions of the problem.

A set of two dynamic centrifuge experiments on stiff and flexible U-shaped retaining structures with dry medium dense sand backfill was performed in order to evaluate the validity of the various assumptions and analysis procedures currently used for the evaluation of seismically induced forces on retaining structures. The experimental results show that dynamic moments and dynamic earth pressures tend to be overestimated using the current analysis methods. More importantly, the dynamic earth pressure increases monotonically with depth and the maximum measured moments are not necessarily in with maximum measured earth pressure. These results are consistent with results obtained independently for gravity retaining structures and call into question the validity of the basic assumptions in the currently used analysis and design procedures.

A series of dynamic response analyses of cantilever retaining walls were conducted in order to assess the effects of ground shaking characteristics (amplitude and frequency) and wall flexibility on the seismically induced lateral earth pressures occurring on the walls. The investigation was focused on the seismically induced thrust and its application point on the wall stem. Idealized backfill-wall systems were modeled by finite element method. Results were compared with solutions obtained by pseudo-static approach of Mononobe-Okabe and pseudo dynamic approach of Steedman-Zeng. The analyses indicate that the acceleration amplitude of the horizontal ground motion is the major effective parameter on magnitude of the dynamic thrust acting on the walls. The difference between the dynamic thrust predicted by finite element and analytical methods becomes more significant for large acceleration amplitudes. Magnitude of the dynamic thrust calculated by finite element method reaches its highest value when the excitation frequency overlaps with the natural frequency of the backfill-wall system Numerical modeling study indicates a decrease in dynamic earth pressures for more flexible walls.

Geofoam is a type of cellular geosynthetic that can be used to reduce lateral earth pressures acting on rigid earth-retaining structures (RERSs) under a variety of soil and load conditions. One of the geosynthetically-functional ways in which geofoams do this is as a compressible inclusion to allow shear-strength mobilization of retained soil through controlled yielding which involves the well-known soil mechanics mechanism of arching. This paper describes how a mathematical model developed by Veletsos and Younan for a seismic-load application not involving geofoam was modified and extended to create a novel solution involving the use of geofoam as a compressible inclusion. An illustration of how this Extended Veletsos-Younan model can be applied in practice is also included using a case-history and full-scale, 1-g physical tests.

This paper presents a simplified procedure for computing lateral seismic earth pressures on soldier pile retaining walls. This approach is to develop two curves of lateral load versus displacement behavior to consider the soil-pile-wall interaction in a "decoupled" way. A curve of lateral earth driving force versus displacement is first established by slope stability analysis for the sliding soil mass behind the wall based on the Newmark chart that relates the standardized maximum displacement of the soil mass to the limiting acceleration coefficient. A second curve of lateral load versus displacement for the soldier pile/wall is then obtained by lateral analysis of the pile/wall structure under the lateral earth driving force. The intersection between the two load-displacement curves is considered as the lateral earth driving force on the pile/wall. An example of analyzing one soldier pile retaining wall is presented in this paper to illustrate the proposed method. The simplified method of accounting for the soil-pile-wall interaction may be used to estimate the lateral soil seismic driving forces on piles/walls without the uncertainty in determining the seismic coefficient related to the use of the Mononobe-Okabe method and choosing the design wall displacement. However, the simplified method is not validated and the results from the method should be verified against finite element studies or field or model testing.

Among the six MSE earth walls tested on the National Geotechnical Centrifuge at the University of California, Davis, one wall (Wall 2) was provided with uneven reinforcement. There was no catastrophic failure observed in any of the six walls, even though the base excitation was as high as 0.9g. Wall 2 showed the lowest lateral displacement. A relatively simple rigid plastic model for predicting the seismic permanent displacement of MSE walls with uneven reinforcement is presented. It extends an existing model that was developed for MSE walls with uniform reinforcement. The variation of the acceleration in the backfill has been included since significant deamplification in acceleration was observed in the backfill in large excitations. The study reveals that the proposed analytical model captures many aspects of the characteristic behavior of MSE walls under seismic loading. Computed and measured lateral MSE wall displacements were compared and the agreement between them is very good.

The seismic earth pressures behind a retaining wall that bends, rotates about its base, and translates horizontally and vertically are evaluated numerically using an ideally elastic model. The following factors are analyzed: (1) flexural movements of the wall; (2) rotation of the wall at its base; and (3) horizontal and vertical movements of the wall. The following results are obtained: (1) the flexural flexibility of the wall, the horizontal translation, and the rotation at the base have significant effects on the earth pressures behind the wall; and (2) within the range of relative flexibilities of typical retaining walls (relative between wall and backfill), the effects of the horizontal translation are more significant than those of the vertical translation and rotation. Four extreme cases are also analyzed: rigid wall with stiff foundation, rigid wall with soft foundation, flexible wall with stiff foundation, and flexible wall with soft foundation. The seismic pressures behind a rigid wall with a stiff foundation are the largest, and they are the smallest for a flexible wall with a soft foundation. The pressures behind a rigid wall with a soft foundation are affected the most by the horizontal translation of the wall, while the pressures behind a flexible wall with a soft foundation are affected both by the flexural flexibility of the wall and by the horizontal movements.

Geofoam is a type of cellular geosynthetic that can be used to reduce lateral earth pressures acting on a wide range of earth-retaining structures (ERSs). There are two distinct geosynthetic-functional ways in which geofoams can provide this benefit: as lightweight fill and as a compressible inclusion. These functions can be implemented either in new construction or retroactively with existing ERSs. Depending on the specific design details chosen, it is possible to reduce lateral earth pressures acting on an ERS to almost zero under both gravity and seismic loading. Unfortunately, the existence and capabilities of this geotechnology, which has the potential to revolutionize the way ERSs are designed, are still not widely known and thus have been significantly underutilized to date. This paper summarizes the state of knowledge concerning the use of geofoams to reduce lateral earth pressures on ERSs and addresses both analytical methods (with an emphasis on seismic loading) as well as geofoam materials and products. It is hoped that this will encourage both usage and research of this intriguing geotechnology.

This paper describes the seismic modeling performed for the design of a 135-foot-tall mechanically stabilized earth (MSE) wall for the Third Runway Project at the Seattle-Tacoma International Airport. The FLAC finite difference program was used to model the MSE wall under construction and seismic loading conditions. This wall was constructed in 2005 and is the tallest MSE wall in North America. The wall was built of granular fill reinforced with steel strips. Wall fascia consists of concrete facing panels. The wall was designed in accordance with the 1996 AASHTO standard guidelines incorporating the interim revisions through 2000. The numerical modeling was performed to provide the design team with additional insight into the static and seismic performance of the MSE wall.

Horizontal Slice Method (HSM) is an approach employed to evaluate the seismic stability of soil slopes and walls in the framework of limit equilibrium method and pseudo-static analysis. HSM is also capable of applying variable accelerations at different levels within the horizontal slices of a slope or retaining structure, which is difficult when using other limit equilibrium procedures. In this paper the effect of amplification of seismic acceleration on the stability of slopes and soil structures has been evaluated. In combination with the HSM a pseudo-dynamic method is adopted in the present analysis, which also considers the effect of phase difference in the shear waves traveling through the backfill due to seismic excitation. The analysis is based on the observations of Nova-Roessig and Sitar relating to centrifuge tests performed on a number of scaled reinforced soil slopes. Two main sets of analyses are conducted. In the first one, seismic amplification is considered in the analysis by assigning amplified accelerations at different levels to the corresponding slices with different distributions. The second set of analyses is conducted for slopes having the same seismic acceleration for all the slices, which means there is no amplification. The main results of the analysis are presented in dimensionless form relating to the total force in the reinforcement layers to maintain the stability of the reinforced structure. This definition is equivalent to the coefficient of active lateral earth pressure. Therefore the effect of amplification on the active lateral earth pressure can be evaluated and compared with the case of no amplification. The failure surfaces for each set of analysis are also obtained and compared with the observed deformations in the physical model.

Earthquake-induced deformation of piled bridge abutments in approach embankments underlain by liquefied soils may be reduced by the restraining forces provided by the piles and bridge superstructure. Design methods accounting for these "pinning" effects have been applied in practice, but they lack significant comparison with actual physical data. Three dynamic centrifuge model tests were performed to investigate pile pinning effects, and the results are presented in this paper. The centrifuge models were composed of two identical embankments underlain by liquefiable soil, one with a pile group and another without. Data from the tests demonstrated that the embankment with a pile group had less lateral movement and settlement than the embankment without a pile group. The experimental results and back-calculated stress-strain responses are described, and the implications for design practice discussed.

The abutment-soil-structure-interaction plays a significant role in the global response of bridge structures. This paper presents nonlinear closed-form formulas to describe force-displacement backbone curves of bridge abutments with typical backfill soils. The original closed-form formulas are modified Hyperbolic Force-Displacement ("HFD") equations developed to simulate the results of two field abutment tests for an abutment wall height of 5.5 feet. To extend the formulas to be applicable to abutment walls with other heights, a series of analyses are performed using the limit-equilibrium method that implements mobilized Logarithmic-Spiral failure surfaces coupled with a modified Hyperbolic soil stress-strain behavior ("LSH" model). The limit-equilibrium method with LSH model, which is a relatively simple approach, has been verified through analyses using a complex finite element model and also has been validated by abutment and pile cap field tests conducted on various typical structure backfills. Based on the results of the analyses, two height adjustment factors are derived for typical sandy and clayey backfill soils. The recommended closed-form formulas are the products of the original closed-form formulas and the height adjustment factors. The formulas are simple enough to be used by the practicing engineers and are capable of capturing the major characteristics of the force-displacement backbone curves.

Preliminary results from a study to develop innovative bridge foundations to optimize bridge performance under earthquake events are presented. Numerical models of bridges with both flexible and stiff footings are compared to demonstrate the potential importance of soil-footing stiffness on the performance of a soil-footing-column-deck-abutment system. Detailed numerical models were developed to capture the inelastic behavior of soil-footing systems. This article summarizes two types of foundation models developed using the OpenSees computational platform. One model uses a contact interface model, soilFootingSection2D, to simulate the foundation and has been proven to be effective in analyzing the cyclic response and displacement of soil-foundation-bridge system. The other uses 2-D nonlinear Winkler foundation consisting of various constitutive springs. Although the former model is more detailed and fundamentally sound in modeling the footings, in its current state it is limited to 2-D use only. The latter model, once verified via comparison to experimental data, can be extended to 3-D. Future work will use the 3-D nonlinear footing models to more accurately simulate the 3-D response of a bridge system to 3-D ground motions.

Liquefaction-induced lateral spreading has caused major damage to bridge structures in past earthquakes. Lateral spreading may cause large lateral deformations at the abutments, which may induce large forces in the foundation, superstructure, and connections, leading to severe damage or even the collapse of bridge structures. The post-earthquake condition of a bridge structure is strongly related to the level of residual lateral displacement of the pile caps and abutments in the longitudinal direction, and several factors may affect this level of residual deformation. In this paper, the relative importance of several key factors is evaluated through the use of a realistic design example. Results suggest that the most influential parameters in this problem are the characteristics of the ground motion in terms of its spectral acceleration at the degraded period of the potential sliding mass, followed by the post-liquefaction residual undrained shear strength of the potentially liquefiable material, and the error term in the estimation of the residual lateral displacements induced by lateral spreading.

An AASHTO-sponsored project, NCHRP 12-70, was conducted to address current limitations and deficiencies in the methods being used within the AASHTO LRFD Bridge Design Specifications for the seismic design of retaining walls. The existing AASHTO Specifications use the Mononobe-Okabe (M-O) equations. Under certain combinations of ground motions and wall geometry, unrealistically high seismic active earth pressures are estimated. NCHRP 12-70 Project was conducted to address this and other limitations of the M-O equations for estimating seismic earth pressures. The project also considered seismic design requirements for slopes and embankments, and for buried structures. The products of the project include (1) recommendations on ground motions to use, (2) an updated method for estimating permanent ground movement, (3) charts for including the effects of soil cohesion in seismic earth pressure calculations, and (4) a generalized limit-equilibrium approach for evaluating seismic active earth pressures. This paper provides a summary of the NCHRP 12-70 work with a focus on seismic ground motions and earth pressure evaluations.

A comprehensive seismic investigation of the Brooklyn Bridge was completed to assess its potential retrofit needs. The Brooklyn Bridge, built in 1883, has become a national treasure and architectural and engineering marvel. To ensure that the seismic retrofit needs of the bridge were based on a rational framework, avoiding overconservatism that would potentially lead to unnecessary retrofit and impacting negatively on the architecture of the bridge, advanced engineering investigations of the condition of the bridge and its seismic response were made. Specifically, the seismic investigation of the main bridge was performed following two approaches, referred to as the global and local analyses. In the global analysis, the entire main bridge with its foundation caissons was modeled, and the effects of soil-foundation interaction were incorporate through the use of foundation impedances. In the local analysis, each bridge tower with its caisson and the surrounding soils was investigated with a model using solid finite difference and slip and gap interface elements. The local analyses of the towers were performed to confirm quality of the motions and foundation impedances used in the global analysis, and to ensure that the conclusions regarding the potential need for foundation retrofitting was realistic and essential. This paper presents the details of the two seismic evaluation approaches, and compares the bridge foundation responses from both analyses. It also demonstrates the benefits of local analysis in the seismic evaluation of long-span bridges.

For reliable estimation of bridge performance, especially when subjected to liquefaction and lateral spreading, appropriate modeling of soil-pile-structure interaction and understanding of the global bridge behavior are essential. In practice, soil and foundation systems are often approximated using very simple foundation springs or unrealistic lateral spreading mechanisms that may not represent all important aspects of the global system behavior. In this context, a detailed bridge model of a typical highway overpass bridge in liquefiable soils was developed with an emphasis on accurate characterization of the structural and soil behavior, including various soil-pile-structure (SPSI) modeling strategies for several components in the bridge system, such as piles, pile caps, and abutment structures. Using this model, the global behavior of the bridge system is more readily understood by capturing realistic force boundary conditions at the bridge pier bases and bridge deck ends. In addition, important bridge damage mechanisms are identified.

Static seismic analyses are used to characterize the vulnerability of an older-vintage bridge to liquefaction and lateral spreading. A sensitivity study is performed by varying key parameters and observing their influence on the predicted free-field lateral spreading displacement required to mobilize damage states ranging from minor damage to collapse. The thickness of the nonliquefiable crust layer was found to be the most influential parameter. Published fragility curves were reasonably consistent with the displacements required to cause major damage and collapse of the bridge.

A pushover analysis on the inelastic response of bridge columns supported on pile groups consisting of full moment connection cast-in-place-steel-shell (CISS) piles is presented in this paper. An original component of the work presented in this paper is that the analytical models considered the nonlinear actions that develop in the column, supporting piles and soil. Analysis confirmed that variations in the horizontal and vertical soil stiffness affect the pile cap lateral deflection and rotation, respectively. Results from this analysis also indicate that for columns with aspect ratios lower than six the contribution of soil-structure interaction is significant; however, for columns with higher aspect ratios and for stiffer soils the effects of soil-structure interaction are negligible.

The paper includes seven recent earthquakes investigated by the Earthquake Investigation Committee of ASCE Technical Council on Lifelines Earthquake Engineering: 1993 Hokkaido Nansei Oki, Japan, 1993 Guam, 1999 Kocaeli (Izmit), Turkey, 2001 El Salvador, 2001 Nisqually, Washington, 2003 Zemmouri, Algeria and 2004 Sumatra, Indonesia earthquakes. These events occurred over the past 15 years both inside the United States and overseas, and encompassed crustal, interplate and intraslab events. Following a brief outline of the characteristics and damages of these earthquakes, some of the geoscience and geotechnical aspects observed by the authors are discussed with lessons learned with respect to lifelines facilities outlined.

Bridge abutments are earth-retaining structures which support the superstructure at the ends of a bridge and provide resistance to deformation and earthquake induced inertial forces from the bridge deck. Current design practice in California makes use of bi-linear load-deformation curve and does not account for the structure backfill properties. An experimental and an analytical research program were conducted at UCSD to further investigate such structure backfill interaction characteristics. In order to meet the objectives of this research project, a field investigation was conducted to develop a proper characterization of the soil types used for abutment structure backfills. The experimental program included five large-scale tests to examine the effect of structure backfill soil type, backfill height, vertical movement of the wall, and pre-existing cut slope in backfilling on stiffness and capacity of the abutments in the longitudinal direction. The study indicated that the response of bridge abutments in the longitudinal direction is nonlinear and a function of several influential factors which need to be considered. The passive resistance of the structure backfill is controlled by the soil shear strength and interface friction. It was shown that the vertical movement of the wall has a significant effect on the amount of maximum passive pressure.

This paper focuses on the preliminary results of an extensive 2D parametric study of time domain site response analysis for high frequency input motion. Such study in which direct boundary element used as numerical tool evaluates the effects of such motion on seismic behavior characteristics of various homogenous shapes surface topographies (semi-elliptical, semi-sine and trapezoidal shapes). It is assumed that the medium behaves as a linear elastic material. Also, it is considered that Ricker-typed wavelet input motion is Incident SV in-plane wave. Obtained results are presented in the form of dimensionless parameters that is Shape Ratio (the ratio of height to half width of the topography) and dimensionless frequency (the ratio of half width of the topography to wave length). Hence presented results could be generalized and used for various geometric and input motion values. Parametric study results show that in all geometric cases. Various distributed plane wave interference cause to fully disturbed wave field within the topography as well as a scattering of in-plane waves, where in such condition wave separation is very troublesome. Finally, the responses of 2D analyses at the top of topography for different-typed topographies versus related shape ratios (SR) are calculated and then, compared and correlated with the results of response spectra of a free field Ricker-typed wavelet.

The effects of earthquakes may be the most significant natural hazard that can impact the Delta levees, whose potential failure would be catastrophic to California's economy. A probabilistic seismic hazard analysis (PSHA) was performed to characterize the ground shaking hazard that could contribute to levee failure in the Delta. Based on a seismic source model that builds upon the characterizations of the major faults in the San Francisco Bay region from the Working Group on California Earthquake Probabilities and several newly recognized faults, a PSHA was performed for key sites throughout the Delta and probabilistic hazard maps were developed. Of particular significance are several seismic sources within the Delta, which were characterized as part of this study including the northern and southern Midland fault and two areal zones of deformation, which we call the Montezuma Hills and Thornton Arch zones. These potential seismic sources have been the subject of very little research and thus there is greater uncertainty in their characterization compared to the better studied major faults in the Bay region. Ground motion attenuation relationships from the Next Generation Attenuation (NGA) Project were used in the PSHA. Quasi-time-dependent hazard in 2005, 2050, 2100, and 2200 was calculated. The time-dependent behavior of the major faults was adopted from the WGCEP, which weighted both time-dependent and independent models in their forecasts. The hazard was calculated for a firm soil site condition (average shear-wave velocity in the top 30 m [V_s30] = 1100 ft/sec). For a return period of 500 years, the time-dependent probabilistic peak horizontal accelerations ranged from 0.2 to 0.4 g. Time-independent (Poissonian) hazard for the Delta differs little from the time-dependent hazard because the hazard is generally being controlled by the local Delta faults. The hazard as expects is higher at sites closer to the major faults.

Two large earthquakes that rattled the Hawaiian chain in October of 2006 were strong enough to cause minor to moderate damage to four nearby dams, including longitudinal and transverse cracks in the embankment fills, incipient slope instability, and possibly delayed internal erosion and piping that triggered emergency water releases from two of the reservoirs. Recent seismic studies by the first author indicate that many dams in Hawaii are potentially susceptible to severe damage from earthquake shaking due to an active seismic environment in some parts of the State. Adding to a general level of uncertainty are outdated construction methods used for dams as old as 120 years, relatively unknown dynamic characteristics of local volcanic soils, and a lack of sufficient understanding of strong motion propagation throughout the Islands.

Success Dam located in Southern California is seismically deficient and remediation design is currently in progress. The proposed remediation includes partial excavation of the existing embankment dam and construction of a buttressing embankment immediately downstream. Liquefiable alluvium will remain under the existing dam and will continue to affect the seismic response of the combined structure. In accordance with Corps of Engineers guidance, the design must be validated using a seismic deformation analysis. The deformation analysis was performed in steps of increasing complexity, starting with simplified procedures and continuing with more complex, state-of-the-art methods: (1) liquefaction assessment using the simplified procedure and QUAD4M coupled with limit equilibrium analyses assuming post-liquefaction residual strength in potentially liquefiable zones; (2) FLAC computer analysis using the decoupled effective stress constitutive model used by URS, and (3) FLAC analysis with the coupled effective stress model UBCSAND developed at the University of British Columbia, Canada.

The spillway on the small earth dam is made by reinforced concrete. Therefore, the weight could be the weak point in earthquakes. In this paper, shaking table tests were conducted to evaluate quake resistance of spillway focusing its attention to weight and stiffness of the spillway models, cross sectional shape of the models, and the effect of reinforced by geogrid. As the results, the proposed geogrid-reinforced spillway with lightweight material showed high earth protection.

The amount of water in soil plays very important role in defining its shear strength. Thus, when considering partially saturated slopes subjected to earthquake-induced ground shaking, it is important to quantify how such slopes respond to the induced dynamic stresses. For this purpose, undrained cyclic triaxial tests were performed on decomposed granite soil in order to quantitatively investigate the effects of degree of saturation and initial static shear stress on its earthquake response characteristics. The experimental results obtained were then used in conjunction with pseudo-static slope stability analysis and modified Newmark sliding block model to estimate the deformation of embankments which were under different degrees of saturation. The example calculations clearly showed the effect of saturation on the magnitude of seismic-induced displacements of embankments.

The effect of heterogeneous soil strength on a performance design procedure, based on permanent displacement according to safety factors, is examined. Soils are heterogeneous in nature. However, at present, the seismic design of embankments has been implemented assuming homogeneous soil strength. In this study, a seismic reliability analysis using a Monte Carlo Simulation and heterogeneous soil strength was applied to the Newmark Method. Input data of the reliability analysis in this study were based on in-situ soil test results of an actual airport embankment. The numerical results in this study suggest that the estimated displacement will increase if the heterogeneity of material is considered. His study also shows that the setup of the numerical analysis model to include the heterogeneity of the soil strength is important.

Success Dam is located on the Tule River, near the city of Porterville, California. The dam is a rolled earth-fill embankment approximately 44 meters (145 feet) high and 1036 meters (3,400 feet) long. The embankment is comprised of a central impervious core protected by upstream and downstream outer pervious zones. A majority of the dam is founded on a complex deposit of both Holocene and Pleistocene alluvium. Investigations and studies under the Corps of Engineers Dam Safety Assurance Program (DSAP) demonstrated that remediation is required to prevent dam failure from large earthquake shaking. The Corps of Engineers is designing a seismic remediation of Success Dam that consists of a modified earth-fill embankment centered and aligned generally along the downstream toe of the existing dam. The modified embankment will shift the alignment of Success Dam about 95 meters (310 feet) downstream onto a prepared foundation and will increase the footprint of the dam by 40%. Similar to the DSAP studies that identified the need for remediation, the new foundation area has been investigated in numerous phases to characterize the material properties and distribution for remedial design analyses. High-quality blowcount data ((N₁)₆₀) has been difficult to obtain at Success Dam. Conversely, a large volume of shear wave velocity data (Vs₁) has been collected. Liquefaction assessment has been conducted using both data sets and additional conservatism has been observed in the Vs₁-based procedure compared to the (N₁)₆₀-based procedures as reported in recent years by various researchers. A site-specific relationship between (N₁)₆₀ and Vs₁ is presented that reduces some of the additional conservatism. This paper provides an overview of the investigations, the geotechnical site characterization, and the relationship between field exploratory data sets. Discussions include the in-situ testing methods, the challenges of characterizing potentially liquefiable materials containing significant quantities of large particles, and the development of a site-specific data correlation for liquefaction potential analyses and remediation design.

This paper summarizes the characteristics of small seismic motions of four major compacted earth dams in Taiwan. Seismic data of four major dams recorded in the past thirty years were collected and analyzed. In general, ground motion characteristics can be expressed by peak ground acceleration (PGA), velocity (PGV), displacement (PGD) and predominant frequency of each motion history. The different statistical correlations between the PGA, PGV and PGD have been derived. The correlations are very consistent for different dams. The PGA amplifications from dam base to dam crest were also derived and show that the larger the base PGA, the smaller the amplification ratio, a phenomenon called non-linear effect of geo-material. The fundamental frequency of each earth dam was identified by the ratio of Fourier and response spectra of the seismic records at the dam crest and base. It was found that the larger the base PGA, the lower the fundamental frequency of the earth dam. The fundamental frequencies of the earth dams in direction transverse to the dam axis range from 1.5Hz to 2.0 Hz. It is interesting to observe that the fundamental frequency does not correlate well with the dam height as expected in the 2D dynamic analysis of the earth dam. This may suggest a possibility that a 3D dynamic analysis is required to more reasonably capture the real seismic response of earth dam.

Seismic performance of the Finished Water Reservoir (FWR) at the Robert B. Diemer water treatment plant was evaluated by performing dynamic nonlinear soil-structure interaction analyses with FLAC. The reservoir sits on a terrace cut into siltstone near the top of a ridge. For the north wall, which retains high ground above the reservoir roof, the critical failure mode is governed by bending and shear. The partially buried south wall is supported by caissons where it straddles filled-in ravines. The critical failure mode is loss of vertical support from caissons damaged by down-slope sliding fill within the ravines. It was analyzed considering the deep, out-of-plane wall beam bridging over caissons which might be damaged during shaking. The study concluded that the reservoir would maintain its structural integrity during MCE-level shaking, such that uncontrolled release of water would be prevented.

This paper presents the results of the seismic characterization of San Pablo Dam in support of the design of seismic upgrades. The limited implication of characterization is also presented. Previous three seismic evaluations of the hydraulically constructed dam concluded that much of the shell portions of the embankment and the foundation soils were potentially liquefiable; these observations were considered appropriate given the perceived nature of the materials and the process of construction. However, a recent stratigraphic and property characterization conducted on the dam indicated a consistently very clayey, very "lumpy", and over-consolidated nature for almost all the hydraulic shell materials. The foundation soils were also found to be very clayey, but less so, and less over-consolidated. The undrained shear strengths of these materials were derived from the available results of laboratory tests. The recent characterization indicates that the shell materials are unlikely to liquefy while liquefaction of the some portions of foundation soils cannot be discounted. The limited implication of this new characterization is also presented.

Coquitlam Dam is a 30 m high embankment dam constructed in the early 1910s. The dam site is situated in a region of high seismic hazard in British Columbia, Canada. The existing dam's core and shells, and part of the dam foundation are deemed to be liquefiable under the design earthquake loading. Seismic upgrade to the Coquitlam Dam is currently being implemented by constructing a new compacted earth core rockfill dam at the downstream toe of the existing embankment. The new dam will be mostly founded on dense glacial deposits. Due to site and constructability constraints, a portion of the upstream slope of the new dam will sit on part of the downstream rockfill toe of the existing dam that is underlain by liquefiable sand and gravel alluvium along the original river channel. This paper describes the analyses carried out to conservatively assess the seismic performance of the new embankment dam. Deformation analyses using the 2-D plane strain finite difference program FLAC showed that displacements of up to about 1.1 m vertically and 0.4 m horizontally could occur at the dam crest under the design earthquake loading. Treatment of the liquefiable alluvium is thus deemed not necessary. The anticipated movements can be safely accommodated by the new dam that has been designed with sufficient freeboard, and suitably wide upstream and downstream filter and transition zones.

During the 1976 Tangshan Earthquake, a liquefaction-induced slope failure occurred at the upstream slope of one main earth dams of the Miyun Reservoir in the suburb of Beijing, China. Later, the reservoir was dewatered and the dam repaired. Between 1998 and 2001, a major seismic rehabilitation project was carried out on several dams of the reservoir after potential seismic stability problems were found during a safety inspection. This paper reports the techniques and construction methods used in the two rehabilitations.

As part of the Delta Risk Management Strategy (DRMS) project, the vulnerability of the Sacramento-San Joaquin Delta (the Delta) levees was assessed under various stressing events such as normal conditions, floods, and seismic events, and the effects of future climate change and subsidence. The seismic hazard of the project site was evaluated using a time-dependant probabilistic seismic hazard analysis to estimate the seismic hazard over the next 200 years. This paper discusses the methodology and results of a probabilistic seismic levee vulnerability analysis. The system of levees in the Delta was divided into a series of vulnerability classes using factors that differentiate the performance of the levees subjected to same earthquake ground shaking. The probabilistic evaluation of levee vulnerability (fragility) for a given vulnerability class is a function of random (aleatory) variables as well as epistemic uncertainties.

The use of ShakeMap in conjunction with the U. S. Geological Survey (USGS) ShakeCast software provides an effective means to notify engineering and field inspection staff of the occurrence of a potentially damaging earthquake and its potential impact at all dam locations around the State of California. The ShakeCast software automatically downloads a ShakeMap immediately after it is generated, and evaluates whether selected sites have experienced ground motions exceeding predefined threshold levels. ShakeCast automatically sends a customizable notification to identified individuals via e-mail, PDA, or cell phone apprising them of the results. Instrumental Intensity, based on a combined regression of recorded peak acceleration and velocity amplitudes, is our chosen threshold parameter because it better indicates damage potential than PGA alone. In the absence of custom threshold values determined by engineering analysis, we have selected Instrumental Intensity 5 as the "damage possible" threshold and Instrumental Intensity 8 as the "damage likely" threshold. The latter was selected because a significant percentage of dams that experienced Instrumental Intensity 8 and greater in the Loma Prieta and Northridge earthquakes, were damaged.

Seismic performance of an earthfill dam called the Walnut Canyon dam was evaluated in terms of seismic deformation and transverse cracking potential under postulated shaking conditions. A statistical review of construction data along with field investigation and laboratory testing played a vital role in characterizing materials comprising the dam for use in the nonlinear deformation analysis. As a result of this study, the continued use of the dam was approved by the California Division of Safety of Dams (DSOD) with provision for additional instrumentation.

The amplitude and frequency content of non stationary earthquake records varies with time. For this reason, the classical signal processing techniques can't present the appropriate dynamic behavior of the dam under earthquake loadings and consequently, time-frequency distributions such as wavelet method is proposed. In this paper, the recorded earthquake signals will be processed by classical signal processing techniques, such as power spectral density, cross spectral density, coherency and phase difference of the signals. The wavelet method is used as one of the newest techniques of signal processing to extract the time dependent characteristics of earthquake records. Using the mentioned method, the benefits and accuracy of the new signal processing technique is described and its importance in evaluating dynamic characteristics of embankment dams is discussed.

A numerical study is presented for a rigid block supported through a Coulomb friction contact surface on a horizontal or an inclined plane, and subjected to horizontal or parallel excitation. The latter is described with idealized pulses and near-fault seismic records strongly influenced by forward-directivity or fling-step effects (from Northridge 1994, Kobe 1995, Kocaeli 1999, Chi-Chi 1999, Aegion 1995). In addition to the well known (ever since Newmark) dependence of the resulting block slippage on variables such as the peak base velocity, the peak base acceleration, and the critical acceleration ratio, our study has consistently and repeatedly revealed a profound sensitivity of both maximum and residual slippage on: 1 the (unpredictable) sequence and even the details of the pulses contained in the excitation, 2 the (also unpredictable) direction (+ or –) in which the shaking of an inclined plane is imposed. By contrast, the slippage is not affected to any measurable degree by even the strongest vertical components of the accelerograms! Moreover, the slippage is often poorly correlated with Arias Intensity of the base excitation. These findings contradict some of the prevailing beliefs that emanate from statistical correlation studies.

To provide evidences for the aseismic evaluation of the soil mountain slopes of the Beijing Olympic Forest Park, the seismic response and failure process of mountain slopes during a strong earthquake were investigated using centrifuge shaking table model tests. The results showed that magnification factors and peak displacements of the slope increased with increasing altitude. The vertical displacement increased monotonically during shaking. The residue horizontal displacements pointed from the axis to both sides after the earthquake. The strain localization occurred mainly at the upper and middle of the slope and can lead to a final landslide. Water content has a significant effect on the strength and deformation modulus of the soil, and as a result significantly affects the aseismic stability level, magnification factors and deformations of the slopes.

The Metropolitan Water District of Southern California (Metropolitan) has undertaken the addition of ozone disinfection facilities at all five of its water treatment plants as part of its Oxidation Retrofit Program (ORP). The ORP Site Preparation Project for the Robert B. Diemer Water Treatment Plant (Diemer Plant) is an integral part of this work. This project will prepare the plant site for construction of the ozonation facilities, i.e., ozone contactor, ozone generation building, and liquefied oxygen tank (LOX) farm. The Diemer plant, with a rated treatment capacity of 520 million gallons per day (mgd), is one of Metropolitan's five treatment plants to be upgraded with ozonation facilities. The plant is located about 37 miles east of Los Angeles and provides treated water by gravity flow to Metropolitan's central pool area, which serves approximately 3.5 million people in the greater Orange and Los Angeles County areas. The plant was constructed during 1961 to 1963 in the Chino Hills 820 feet above sea level to allow for gravity flow to Metropolitan's service area without resorting to pumping. As such, the site was created by removing the top of a broad ridge and placing the excavated materials in adjacent ravines to provide a large level pad for construction of various treatment facilities. Since the completion of the initial construction work, the plant has experienced a major expansion, which includes the construction of many smaller facilities, and currently the addition of ozonation facilities. Design and construction of the new facilities are challenging due to stringent seismic criteria caused by the nearby Whittier fault located 0.5 kilometer north of the plant area and capable of generating an earthquake event with a moment magnitude of about 7. Based on site-specific deterministic seismic hazard analysis, the 84^th percentile peak-horizontal ground acceleration is approximately 1.15g. Specifically, this paper describes the challenging issues and major design considerations that were associated with the preparation of construction documents for the Diemer Site Preparation Project. Major project elements that will be discussed in this paper include: the limited area of the Diemer site; the geology of the Diemer site and its nearby seismogenic sources; selection of site specific acceleration design criteria; alternative slope remediation methods investigated; selected slope remediation method and the design considerations; shoring and construction concerns during the design process; large excavation and backfill with roller compacted concrete; and current status of the construction project.

In the present study, the most simplified planer failure surface is assumed for the seismic stability analysis of soil slopes and the behaviors of this simple failure surface over the existing complex surfaces are studied. A model reinforced soil slope is divided into a number of finite horizontal slices for the seismic analysis using a limit equilibrium method (LEM) with pseudo-static approach. The variations of factor of safety with respect to horizontal seismic acceleration coefficient under different soil and seismic conditions are shown. It is found that for the slope angle β = 25⁰; angle of friction φ = 30⁰ and for the increase in the horizontal seismic acceleration coefficient k_h from 0.2 to 0.3, the respective decrease in the percentage of the factor of safety is about 21%. It is concluded that the stability of slope decreases with increase in both the horizontal and vertical seismic acceleration coefficients. The present results are compared with the available results in the literature. Also, the displacement of slope for different soil and seismic conditions and time of cycle of earthquake loading are evaluated. As expected, the displacement of the slope is found to increase significantly with increase in the seismic accelerations.

Slope failures during the 2004 Niigata-ken Chuetsu earthquake and their mechanisms are discussed herein. Slopes are composed of weak sedimented rock of alternating layers of strongly weathered sand stones and mud stones. Bedding planes had a strong effect on the failure mechanisms. An energy approach for run-out distance of failed soil mass in view of earthquake energy, together with potential energy, is applied to representative slope failures during the earthquake. It is found that the equivalent friction coefficient back-calculated by the energy approach is strongly dependent on the initial slope inclination. However, the absolute value of the former is smaller than the latter, indicating the failed soil mass initially accelerated.

Seismic stability of above-ground landfills is an issue of extreme engineering interest, and the proper implementation of the corresponding analyses requires dealing with several uncertainties related mainly to the special features of such large-scale geostructures. This study aims at providing an insight into the role of these parameters on the development of the potential failure modes that may take place during a seismic event. Additionally, the different failure modes are compared with respect to their seismic instability potential. Furthermore, an insight is provided on phenomena related to the simultaneous generation of more than one failure surfaces within the landfill. This is examined by developing a simple model based on the Newmark's sliding-block approach. Results indicate that the circular surfaces may provide a more critical mode of failure than the base sliding.

In recent years, notable progress has been made with respect to the understanding of the mechanics of Municipal Solid Waste (MSW) response, and hence, generic MSW properties for static and seismic design of MSW landfills have been recommended. However, it is not clear if seismic design of MSW landfills based upon these generic material properties is conservative, and if it is, to what degree. An attempt is made to evaluate how adequate (i.e., conservative) seismic design of solid waste landfills based upon published generic material properties is. The basis for evaluation is one of very few well documented landfill case histories (the Oil Landfill Case History) and selected generic material parameter sets. The results indicate that the use of generic material parameter sets at this site results in either reasonable prediction or slight over-prediction of recorded ground motions. The results of the parametric studies documented in this paper provide further insight on how to approach seismic design of MSW landfills.

This paper addresses large-scale testing of pipeline response to earthquake-induced ground rupture and pipeline system performance after earthquakes. Described in detail are results of large-scale and select centrifuge tests on high density polyethylene (HDPE) and steel pipelines using the facilities of the George E. Brown, Jr. Network for Earthquake Engineering Simulation (NEES). Pipelines composed of HDPE are able to accommodate substantial amounts of abrupt ground displacement at strains and cross-sectional deformation well below those associated with pipe wall rupture and/or loss of serviceability. Measurements of lateral soil reaction with tactile force sensors compare favorably with independent load cell measurements, and demonstrate the effectiveness of tactile force sensor technology. Lateral forces generated by soil-pipeline interaction during ground rupture are similar for dry and partially saturated sand, and a log spiral model is described that predicts maximum horizontal forces on pipelines during ground rupture that agree very closely with large-scale test results. Pipeline system performance is evaluated through modeling the water supply system operated by the Los Angeles Department of Water and Power. The models are used to support decisions regarding the use of recently retired reservoirs on an emergency basis after a future earthquake. System models, calibrated by observed earthquake performance, provide the means for extending physical and numerical modeling results for components and individual pipelines to system behavior and the attendant planning and operational decisions to increase network reliability.

A uniform confidence hazard evaluation provides a methodology to assess earthquake hazards so all pipes in a water system are designed consistent with their intended function and with a uniform confidence that design forces are greater than or equal to actual forces a pipe may experience during an earthquake. The method also provides design uniformity between different water systems. This paper summarizes: (1) how to define a pipe function in relation to its importance in aiding a community in post-earthquake recovery, and (2) how to identify transient, fault, landslide, and liquefaction induced lateral spreading ground movements with uniform confidence for pipeline design. The uniform confidence methodology is also useful for building and other infrastructure and lifeline systems designs.

Uplift behavior of a buried pipe is a typical damage to structures in liquefied ground. Since uplift of pipes could causes the functional loss of emergency supply of fire extinguish water, and the ground surface deformation due to these uplift of pipe could be a barrier for rescue transportation, an evaluation scheme of uplift deformation level in various ground conditions should be proposed. A series of shaking table tests was conducted to investigate the uplifting velocity of a pipe during shaking. Test results revealed the uplift velocities in various soils were dependent on the soil characteristics; especially the effect of fines contents. In comparison with the case with viscous waters, these effects might be not based on the effect of permeability. However, hollow-cylindrical cyclic shear tests indicated that the difference should be based on the difference of stress-strain relationships. Thus, the importance of laboratory test to evaluate the behavior of soil during liquefaction is clarified.

This paper presents Shake-Table testing and FLAC numerical modeling of liquefaction-induced large ground lateral movement and settlement, and pipeline behavior induced by soil movement. A series of laboratory Shake-Table experiments were performed using a model slope ground and model pipe buried under the crest of the slope in a box 1,800 mm long by 600 mm wide by 800 mm high. A typical 2H:1V slope was prepared with its crest at the center of the box. The model pipe had an outside diameter of 25 mm and was buried crossing the full box width and 100 mm below the crest. The above model ground was repeated for different amplitudes of input sine waves varying from O.1g to 0.25g. Observed and recorded ground failures included lateral spreading and settlement, sand boiling, time histories of excess pore water pressure buildup, input acceleration and model pipe responded acceleration. The above experimental cases were modeled using a nonlinear effective-stress modeling approach with computer code FLAC coupled with a practical-oriented pore pressure model based on cyclic stress method by Seed and co-workers (Seed, 1979). Pore water pressure buildup and dissipation were modeled using alternative stepping of fluid flow-mechanical interaction. The computed ground failures reasonably agree with the observations in the Shake-Table test. The FLAC modeling further revealed the soil liquefaction mechanism and induced ground deformation and pipe deflection.

Permanent ground deformation (PGD) is one of the most damaging hazards for continuous buried lifelines. This hazard is especially severe when the PGD results in net compression in the pipe. In that case, buckling of pipe material can occur. In this paper, centrifuge tests of buried pipelines subject to abrupt ground failure in the form of surface faulting are presented. The fault movement results mainly in compression in the pipe. The test results are compared with a case history of pipe failure in the 1999 Izmit, Turkey earthquake and also with the results from the centrifuge tests which result in net tension in the pipe. The experimental setup, procedures, and instrumentation are described in detail. Suggestions for design practice are offered based on the analysis of results from both the 1999 Izmit case history and the centrifuge modeling.

This paper describes an emerging methodology for seismic evaluation and design of port structures. The methodology is based on minimum life-cycle cost principle through probabilistic evaluation of performance. While conventional seismic design of port structures are based on a particular return period specified for design, the methodology based on life-cycle cost allows for consideration of ground motions with all (or varying) return periods and uncertainty in geotechnical conditions. The methodology based on life-cycle cost also allows the probability of failure being evaluated rather than prescribed by an authority. Since ordinary port structures are for commercial use, the method based on the life-cycle cost has potential advantages over conventional design. An example is presented. In particular, the simplified design charts that are based on a series of parametric studies of effective stress analyses have great benefit for implementing the proposed methodology in practice. Apart from the ordinary port structures, higher priority should be assigned for safety as a performance objective. This issue is especially relevant if port structures are essential parts of post-earthquake emergency strategies for recovery and restoration of urban areas. The paper also discusses the importance of this issue and the future direction of study.

Evaluation of the performance of pile supported wharf structures under seismic conditions has evolved in recent years through the use of numerical analysis programs (for example, PLAXIS and FLAC). Before the use of numerical analysis programs, the performance of a pile supported wharf structure was evaluated using a decoupled approach, where the stability and movement of the embankment is first evaluated using the limit equilibrium and Newmark's method, and the deformation and pile forces and moments are then estimated using structural analysis. However, unlike numerical analysis programs, this decoupled analysis method does not allow precise evaluation of the interaction between the soil mass and the embedded structure. This paper presents the results of seismic analyses of a pile-supported wharf on a fill embankment in Panama. Both the limit equilibrium/Newmark and numerical analysis methods were used to evaluate the response of the new wharf under two levels of earthquakes: an operating level earthquake (OLE) and a contingency level earthquake (CLE). The stability of three different embankment sections (compacted fill soil with a slender zone of rock, intermediate zone of rock and thick zone of rock) was first analyzed for slope stability (limit equilibrium method) and embankment deformation (Newmark's method) for each of the two design earthquake levels. These same embankment sections were then evaluated for deformation using PLAXIS V8.5 with the inclusion of the pile-supported wharf structure using earthquake time histories representative of the design earthquake levels as input. The factor of safety of embankment stability from limit equilibrium analysis and deformations of the embankment from Newmark's method were compared to the wharf performance (deformation, pile shear and moment) calculated by PLAXIS. The results of these comparisons indicate that the conventional decoupled method of analysis can provide a general evaluation of wharf performance but may underestimate the performance of a wharf that has irregular structure geometry or eccentric loading, or if the structure is sensitive to small movements under larger earthquake events. Recommendations for selection of appropriate pseudostatic coefficients for limit equilibrium analyses that compare well with expected slope/wharf deformation and pile shear and bending moments are provided.

The recently approved seismic code for the Port of Los Angeles (POLA) requires performance evaluation for two levels of earthquakes: the Operating Level Earthquake (OLE) which has 50% probability of exceedance in 50 years and the Contingency Level Earthquake (CLE) having 10% probability of exceedance in 50 years. For the OLE event, structures should remain operational with insignificant structural damage, while for the CLE event structures should undergo controlled inelastic structural behavior and limited permanent deformations which can be repaired within an acceptable period of time. For the case of pile-supported wharves, the expected level of performance is based on the estimation of the displacement capacity of the structure. Therefore, soil-pile-interaction has to be evaluated with a higher degree of accuracy. One of the most important factors is to properly define key soil parameters, such as soil stiffness, commonly characterized by p-y curves. Unfortunately, little information is available in the literature on the behavior of laterally loaded piles embedded in the large-diameter quarry-run material typically used in port dikes. In light of this, a series of full-scale load tests on piles in quarry-run material were carried out at University of California, San Diego's Englekirk Center from April through August, 2007, in order to obtain better understandings of the behavior of the wharf-foundation-dike system during lateral loading. This paper presents test setup, instrumentation, and the results of a full-scale test for a single pile.

ASCE 7-05 sets the minimum life-safety standards for designing buildings and other structures. However, it does not directly address all of the issues designers typically face in the seismic design of piers and wharves, especially with respect to dealing with kinematic loading from permanent ground deformation. In 2005, ASCE's Codes and Standards Committee and Coastal, Oceans, Ports, and Rivers Institute formed a formal Standards Committee to develop a national consensus standard for seismic design of pile supported piers and wharves. Within this Standards Committee, several Task Committees were formed to address various aspects of seismic design of piers and wharves. Task Committee 7 was charged with developing standards for the geotechnical and soil-structure interaction aspects pertaining to piers and wharves. In addition to recommending different levels of ground motions for different performance levels and importance categories of the structure, the ground motions per ASCE 7-05 are being adopted to maintain consistency with the minimum standards of this document. This paper discusses the efforts to develop the standards and commentaries for the ground motion, geotechnical, and soil-structure interaction aspects of the proposed standard, especially those that are unique to piers and wharves.

This paper presents the general framework for the development and calibration of simplified nonlinear Winkler-type mechanical models via finite element simulations for the analysis of seismic response of pile foundations in soil conditions remediated against liquefaction. These models are currently being developed for implementation in Monte Carlo simulations for risk assessment analyses of waterfront structures, to allow computationally efficient seismic response predictions of 3D pile-supported wharves for alternative scenarios of ground motion intensity and remediation configurations. The macro-elements are formulated to account for the multitude of soil resistance mechanisms mobilized at the foundation of pile-supported structures by considering both material and soil-pile interface (geometric) nonlinearities during dynamic loading of single piles, thus retaining the physical soundness of the mechanical configuration. This paper describes work under progress, and focuses on the alternative formulations investigated to identify the optimal macroscopic model that will allow credible yet efficient simulations of the seismic performance of pile-supported structures required for the quantification of damage states in liquefiable and treated-against sites.

Guam has experienced several large earthquakes in the past several decades. The 1993 earthquake with a moment magnitude of 7.7 caused significant damage throughout the island. Liquefaction and lateral spreading caused damages to commercial and naval port facilities at Apra Harbor. The Alpha-Bravo wharf, built in 1947 without proper seismic considerations, suffered some damage during the 1993 earthquake, including backland liquefaction. The U.S Navy commissioned a seismic improvement program that included upgrade of the Alpha-bravo wharf. The improvements to the wharf included increasing the water depth from approximately 10.7 m to 12.8 m, and retrofitting the existing wharf to meet current seismic and future operational requirements. A two-level seismic performance criterion was used: under the Level 1 earthquake associated with a 72-year return period, wharf elements sustain only minor or no structural damages with no interruption to operations; under the Lever 2 earthquake events with 475-year return period, the wharf elements undergo controlled inelastic structural behavior with repairable damage. Sheet piles and tie-back system were designed to accomplish the seismic retrofit and improvement of the wharf. In order to withstand high tie-back forces, the anchor blocks in the backlands were supported on battered, steel H piles. During construction, the test piles driven at anchor blocks showed softer driving, and dynamic pile analyses results indicated lower axial capacities than the design capacities. Adjustments were made by adding lugs to piles and increasing pile lengths to incorporate the findings from test piles. Partnering meetings and proactive participation of all interested parties including the owner, designers, and contractor throughout the construction of the project helped minimize project delays and cost overruns.

Centrifuge model tests were performed to verify numerical methods being used to assess the stability of a cut-and-cover subway tunnel in a highly seismic area. Since the tunnel is less dense than the backfill soils and the backfill soils are liquefiable, there is concern that the tunnel could suffer large deformations due to buoyancy forces in the event that liquefaction does occur. Numerical analyses indicated that tunnel deformations would be limited due to the dilatancy of the backfill soils, the geometry of the problem, and dissipation of excess pore pressures. The design, testing, results of the centrifuge tests are presented in this paper. The experiments provided added confidence to the design and review team that the analysis is capturing the dominant deformation mechanism. The centrifuge model tests clarified the mechanisms of deformation and are playing an important role in evaluating if a very expensive retrofit project is necessary.

Tunnel and large cut-and-cover structure response to earthquakes is dominated by the surrounding ground, and, contrary to shallow-founded structures, does not depend on the inertial properties of the structures themselves. The focus of the underground design is on the free-field deformation of the ground and its interaction with the structure, i.e., a soil-structure interaction phenomenon. This paper summarizes the performance in recent past seismic events of deep, large underground structures, and then elaborates on their design approach: geotechnical, seismic criteria and structural factors. Analytical models are examined and conclusions drawn about their advantages and limitations. A general approach to the design of deep underground structures is illustrated through the on-going design of a particular project.

The Hatch Mott MacDonald / Bechtel Joint Venture (HMM/Bechtel) team is designer and construction manager for the 5.25-mile Central Area Guideway (Tunnel Segment) of the overall 16.1-mile Silicon Valley Rapid Transit (SVRT) Project. This proposed project extends the existing Bay Area Rapid Transit (BART) system southward from Fremont to San Jose. The Tunnel Segment includes twin bored tunnels, mined cross passages, deep excavations for three underground stations, two portals, and two ventilation structures, situated in an area with highly variable subsurface ground soils, high seismicity, shallow ground water table, and within a dense urban setting with high-rise buildings adjacent to the largest excavation. This paper discusses the seismic analysis and design of the cut-and-cover structures of the SVRT project.

Historically, the damaged cases of tunnels during earthquake are rare due to their good earthquake-resistant performances. However, the tunnel may still be damaged owing to the violent squeeze of surrounding ground when subjected to an extreme strong shaking. On 21 September 1999, the new Sanyi railway tunnel in central Taiwan was seriously damaged by the Chi-Chi earthquake. At that time, the tunnel has been in operation for about one year. The damaged conditions of the tunnel were carefully investigated and well-documented. This paper summarizes the background information related to the geology, strong ground motions, design and construction of the tunnel and presents the damage patterns of the tunnel. Based on the compiled data, the possible reasons why the damages only occurred at some special locations have been proposed. Numerical analyses have been conducted to analyze one of the typical damaged cross-sections in order to explore the damage mechanism of the tunnel. The analyzed damage pattern agrees very well with field condition. The numerical results show that the bad shape of tunnel, the imperfect backfill grouting, and the secondary lining without any steel reinforcement were main factors that caused the tunnel seriously damaged during the earthquake.

This paper presents a brief summary of seismic analyses for Caldecott Tunnel Improvement Project. It includes seismic hazard analysis, wave scattering analyses to evaluate soil structure interaction (SSI) for the proposed 4^th bore, and development of seismic coefficient for slope stability design. Probabilistic Seismic Hazard Analysis (PSHA) was used to establish the rock motion criteria; 1,500-year and 300-year return periods were adopted for the Safety Evaluation Earthquake (SEE) and Functional Evaluation Earthquake (FEE), respectively. The wave scattering analyses were performed using the spectrum-compatible time histories to account for localized geologic features including site specific topography and soil properties specific to the current project. The analyses enable evaluation of the degree of ground distortion that contributes to racking and ovaling of the tunnel liner. Seismic coefficients for several critical slopes near portals were also derived considering the concept of wave propagation within the slope mass.

The Bay Tunnel is a major water supply facility to be constructed between the San Andreas Fault and the Hay ward Fault in the San Francisco Bay Area. The design earthquake for the Bay Tunnel has a return period of 1,000 years and a horizontal peak ground acceleration of approximately 0.6g. A full dynamic analysis of the Bay Tunnel, considering the soil-structure interactions, was performed to evaluate the performance of the tunnel final lining system during the design earthquake. This paper discusses the methods employed to perform the dynamic analysis and the results obtained. A brief comparison of the results from the dynamic analysis with those from the closed-form solutions and numerical racking analysis is also presented.

This study presents results from one-dimensional seismic ground response analyses, performed for the Silicon Valley Rapid Transit (SVRT) project. Due to the large amplitude of the design input motion, large levels of shear strain are calculated in the soil column that are highly sensitive to modeling details. Critical issues affecting the results include: (1) the use of an innovative stress-strain (backbone) curve that captures expected behavior at small strains from empirical modulus reduction models along with expected large strain behavior as represented by material-specific rate-adjusted shear strengths; (2) assessment of velocity variability and randomness from variable stratigraphy (conditional on depth and soil type) on the computed response; (3) comparison of results computed from equivalent-linear and nonlinear codes; and (4) use of empirical amplification factors to help identify potential limitations of the results obtained from ground response analysis. Consideration of these issues indicates that our results are generally reasonable to use as input to subsequent soil-structure interaction analyses.

An iterative procedure is proposed to estimate seismic-induced distortions in cut and cover rectangular structures. The procedure is based on an existing analytical solution for deep rectangular structures. The existing solution assumes elastic behavior of soil and structure, tied contact at soil-structure interface, and static loading. The new procedure builds on the analytical solution and approximates dynamic response with a pseudo-static analysis, and soil-stiffness degradation through an iterative procedure where the soil's shear modulus is changed in each iteration based on the shear strain of the soil obtained in the previous iteration. Predictions obtained from the analytical solution have been verified by a series of numerical tests which include the response of. the Daikai station during the 1995 Kobe earthquake. The relative errors in terms of deformation between analytical and numerical results are less than 15%.

The Niigata-ken Chuetsu-oki earthquake (M_j=6.8) occurred on July 16, 2007, affected the coastal areas of the southwestern Niigata prefecture, including Kashiwazaki city and Kariwa village, most of which were located on or near the sand dunes that run parallel to the coastline. The peak ground accelerations recorded were as large as 6.7 m/s² in Kashiwazaki city and 12.2 m/s² in the Kashiwazaki-Kariwa nuclear power plant. The damage to the city caused by ground shaking, however, appeared to be limited to very old buildings whereas that caused by ground failure extended to both old and new buildings as well as many infrastructures in alluvial deposits and sand dune slopes. The operation of the Kashiwazaki-Kariwa nuclear power plant was automatically shutdown and all the critical structures seemed to be in good condition; however, a fire broke out around a house electric transformer and many non-critical structures as well as incidental components such as pipes and ducts were damaged, retarding the re-operation of the plant. The major cause of the damage including the fire at the site seemed to be not only the near field strong ground motions but also the settlements of backfill around the critical buildings.

Great advances have been made over the 40-some years in which geotechnical engineers have actively been involved in the practice of earthquake engineering. Most recently, advances have come through the development of performance-based earthquake engineering, which seeks to predict the seismic performance of structures and facilities in ways that are useful to a wide variety of stakeholders. Performance-based earthquake engineering requires the integrated, collaborative efforts of several groups of earthquake professionals, including geotechnical engineers; as such, it will affect the practice of geotechnical engineering in seismically active areas. This paper reviews the evolution of performance-based earthquake engineering, discusses the notion of performance and its description, and describes a recently developed framework for performance evaluation. The nature and effects of the many uncertainties that apply to the prediction and description of ground motions, system response, physical damage, and loss are described. The paper gives examples of different manners in which performance-based earthquake engineering can be implemented into practice. Finally, a series of challenges and opportunities presented by performance-based earthquake engineering for geotechnical engineering practitioners are identified and discussed.

Geotechnical instrumentation using Micro-Electro-Mechanical Systems (MEMS) are relative newcomers to this field and, as such, require extensive validation testing. This paper presents the use of a Shape-Acceleration Array (SAA) to instrument full-scale laminar container tests at the University of Buffalo. The SAA is a sensor array based on MEMS accelerometer measurements of angles relative to gravity. The sensors are contained in 30 cm long rigid segments which are connected by composite joints that prevent torsion but allow flexibility in two degrees of freedom. These rigid segments and flexible joints are combined to form a sensor array which is capable of measuring three-dimensional (3D) ground deformations at 30 cm intervals and 3D accelerations at 2.4 m intervals to a depth of 100 m. Two of these MEMS-based sensor arrays were utilized in a series of tests as part of a liquefaction and lateral spreading study. Extensive instrumentation was possible in this laboratory setting, thus the acceleration and deformation measurements from the SAAs can be compared to the traditional instrumentation, such as accelerometers and potentiometers. The goal of validating and calibrating the measured accelerations and displacements of the SAA system was achieved through these full-scale tests.

The Method of Characteristics was introduced by Streeter, Wylie, and Richart (1974) over thirty years ago. The method demonstrated the ability to faithfully reproduce 1-dimensional, horizontally polarized, vertically propagating shear waves through soil. Since then, many adjustments and improvements have been introduced. This paper summarizes some of those improvements and presents further refinements to the approach. Such features as independent energy balance evaluation, interpolation along time (as opposed to space) lines, stress-reversal checking, and non-linear stress-strain table look-up are also included. These improvements have eliminated past problems with numerical damping and accuracy of stress-strain hysteresis. The newer version, WinMOC, employs a more user-friendly interface with a multitude of soil stress-strain models, full graphing capabilities, response spectra generation, and evaluation of many ground motion parameters. The Method of Characteristics is compared to an equivalent linear frequency domain method. Results generally show more accurate representation of strong motion with the same ease of use. This is especially demonstrated at sites with deep (> 150 ft) soil layers. A computational example is presented so that reader may judge the efficacy of the method.

This paper examines some characteristics of ground vibration as induced by high-speed trains using a wide variety of field-measured data. The measurements for analysis consist of different geological conditions, train speeds, foundation types, and superstructures. Based on these analyses, geological condition is found to be a very important factor which obviously influences initial mobilized vibration level and vibration attenuation. For the train speed factor, the vibration level increases with increasing train speed, but the incremental rate is more obvious on soft ground. The foundation type has an influence on the initial mobilized vibration level and its attenuation rate with distance. In addition, the trains on an embankment and a bridge with shallow foundation present relatively consistent vibration characteristics. Finally, some recommendations for further studies are proposed.