Seismic Performance of Geosynthetic-Soil Retaining Wall Structures

Thumbnail Image
Zarnani, Saman
Seismic , Retaining Wall , Geosynthetics , Shaking Table , FLAC Numerical Simulation , EPS Geofoam , Earthquake , Reinforced Soil
Vertical inclusions of expanded polystyrene (EPS) placed behind rigid retaining walls were investigated as geofoam seismic buffers to reduce earthquake-induced loads. A numerical model was developed using the program FLAC and the model validated against 1-g shaking table test results of EPS geofoam seismic buffer models. Two constitutive models for the component materials were examined: elastic-perfectly plastic with Mohr-Coulomb (M-C) failure criterion and non-linear hysteresis damping model with equivalent linear method (ELM) approach. It was judged that the M-C model was sufficiently accurate for practical purposes. The mechanical property of interest to attenuate dynamic loads using a seismic buffer was the buffer stiffness defined as K = E/t (E = buffer elastic modulus, t = buffer thickness). For the range of parameters investigated in this study, K ≤ 50 MN/m3 was observed to be the practical range for the optimal design of these systems. Parametric numerical analyses were performed to generate design charts that can be used for the preliminary design of these systems. A new high capacity shaking table facility was constructed at RMC that can be used to study the seismic performance of earth structures. Reduced-scale models of geosynthetic reinforced soil (GRS) walls were built on this shaking table and then subjected to simulated earthquake loading conditions. In some shaking table tests, combined use of EPS geofoam and horizontal geosynthetic reinforcement layers was investigated. Numerical models were developed using program FLAC together with ELM and M-C constitutive models. Physical and numerical results were compared against predicted values using analysis methods found in the journal literature and in current North American design guidelines. The comparison shows that current Mononobe-Okabe (M-O) based analysis methods could not consistently satisfactorily predict measured reinforcement connection load distributions at all elevations under both static and dynamic loading conditions. The results from GRS model wall tests with combined EPS geofoam and geosynthetic reinforcement layers show that the inclusion of a EPS geofoam layer behind the GRS wall face can reduce earth loads acting on the wall facing to values well below those recorded for conventional GRS wall model configurations.
External DOI