Predicting a high reliability response for small sample size Soil Structure Interaction (SSI) analyses

June 29, 2013
Publication: Safety, Reliability, Risk and Life-Cycle Performance of Structures and Infrastructures - Proceedings of the 11th International Conference on Structural Safety and Reliability, ICOSSAR 2013, p 3555-3560

Abstract: Probabilistic analyses of seismic structural response are computationally intensive and therefore rely on small sample sizes. This paper considers an empirical method for determination of adjustment factors that allow calculation of reliable results for a desired probability of non-exceedance based on the sample size, desired confidence level, and allowable error. The use of the adjustment factor is applied to probabilistic soil-structure interaction analyses that consider multiple variables. The probabilistic SSI analyses were performed for various sample sizes and sampling methods to verify the methodology for predicting high confidence responses. Monte Carlo, Latin Hypercube, and partial orthogonal sampling are also examined for their effectiveness in improving the reliability of the probabilistic results. The results of these analyses form the basis of recommendations for computing high confidence responses from small sample size SSI analyses.