Geosynthetic-reinforced soil walls under multidirectional seismic shaking.

摘要

This research study was conducted to determine the validity of the current seismic analysis of a free-standing simple geosynthetic-reinforced soil (GRS) wall. The case histories of GRS walls during the past several large earthquakes have indicated satisfactory seismic performances and suggested that the current design methodology is adequate. Despite successful cases, several GRS wall failures have been reported. Numerical simulation of GRS wall subjected to seismic loading can thus offer the opportunity to identify the discrepancy between the current design methodology and the more rigorous finite element method (FEM) solutions.;The predictive capability of the FEM computer program LS-DYNA was first validated against full-scale shaking table test walls. Material characterizations of the backfill and geosynthetic reinforcements were performed in the validation process. Material model parameters were determined from the available laboratory data. In particular, the backfill was simulated with a cap model with parameters dependent of stress level. Model calibration was also performed to fine tune the input parameters such as the viscous-damping constant, mass damping coefficient, and the soil-geosynthetic interface friction coefficient. The calibrated values along with the material characterization approaches were adopted for the subsequent parametric study. Prior to the parametric study, the extent of finite element model boundary was verified in order to minimize the boundary effect.;Results of parametric study were compared against the values determined using the Federal Highway Administration (FHWA) methodology. The parametric study results were presented as functions of peak horizontal acceleration (PHA), and the correlations of seismic performances with PHA were determined through single predictor variable regression analyses in order to indicate the trend of the calculated results. In general, the external stability analysis results predicted by FEM are higher than those determined using the FHWA methodology. However, for the internal stability analysis, FHWA methodology overestimates the reinforcement tensile load as compared to the results predicted by FEM.;Using FEM results, multivariate regression equations were developed for the various seismic performances based on multiple design parameters that are essential in the design of GRS walls. The prediction equations can provide first-order estimates of the performances for use in the preliminary analysis of free-standing simple GRS walls. The prediction equations are applicable for PHA greater than maximum limitation stipulated by the FHWA methodology.