Experimental verification of current seismic analysis methods of reinforced soil walls

Abstract

The paper investigates the accuracy of two limit equilibrium design approaches used for seismic stability analyses of reinforced soil walls. Responses from a series of reinforced soil wall models, tested on shaking table, are compared to similar responses predicted using both NCMA, and AASHTO/FHWA design guidelines. First, orientation of the soil slip surface measured during subsequent base shaking was compared to the values predicted analytically. Comparison results suggested that the current design methods tend to under predict the size of the soil failure wedge at different input base acceleration amplitudes. Therefore, the current equation proposed to predict the angle of soil slip surface is modified to match measured values. Second, the horizontal and vertical seismic coefficients used in the current seismic analysis methods are compared to values inferred from shaking table accelerometer responses. Results indicated that equations suggested to calculate horizontal acceleration coefficient for design of reinforced soil walls are non-conservative for input base acceleration larger than 0.3 g. Finally, dynamic earth force magnitudes and locations, and dynamic reinforcement force increments measured from shaking table model tests are used to identify sources of conservatism and non-conservatism of the current design methodologies. Different modifications have been suggested to reduce conservativeness or increase the safety of the design guidelines.