Rationalization of factors of safety in analysis of beams on geosynthetic reinforced random earth beds by Monte Carlo simulation

Abstract

The paper presents an approach for rationalizing the factors of safety in the analysis and design of foundations, wherein a single lumped value is assigned to an uncertain soil parameter on the basis of multiple field testing. A soil-foundation system, comprising a geosynthetic layer (idealized as a beam) placed over a random poor soil and overlain by a compacted sand layer and the foundation beam, has been modeled in a lumped parameter mode. The model parameters comprise among other, the relative stiffness of the random poor soil that has been treated as a lognormally distributed random variable. Monte Carlo simulation has been performed on the model at various levels of the coefficient of variation (<i>COV</i>) of the uncertain/random parameter to arrive at the probability distribution functions (PDF) of the state variables viz., the normalized mid-span deflection and bending moment in the foundation beam. These PDFs have been subsequently invoked to correlate the factors of safety to the <i>COV</i> and the risk of failure. It has been suggested that factors of safety should be introduced in the foundation design by considering <i>COV</i> of the uncertain soil parameters and the allowable risk.