Abstract: The seismic responses for nuclear facilities are required to consider soil-structure interaction (SSI) effects. The upcoming revision of ASCE 4 establishes a goal of developing seismic responses with an 80% probability of non-exceedance. It is important to ensure the widely used deterministic analysis results to be in an 80th-percentile level response or higher as compared to the probabilistic analysis. The SSI analysis of the Lotung quarter-scale containment model is repeated using the probabilistic method using the computer program SASSI2000. The experimental Lotung results are used to evaluate the merits of the probabilistic approach and to make an assessment of the 80th-percentile responses with the responses obtained from recoded motion. The analysis is subsequently repeated at three sites with different soil depths and with modifications to the input parameters to observe their effects on the probabilistic analyses. Probabilistic analysis considers the variation of soil properties,layer thicknesses, and structural parameters. For all cases, a comparison of the seismic loads and in-structure response spectra (ISRS) at the 80th-percentile level is made with the envelope of three deterministic SSI analyses, demonstrating that the deterministic method envelopes the probabilistic ISRS, but probabilistic structural forces may exceed the deterministic. This paper further examines the effects of Monte Carlo and Latin Hypercube sampling methods and of the number of samples on the accuracy of results for the probabilistic analyses.
Probabilistic and Deterministic Seismic Soil Structure Interaction Analyses for the Lotung Quarter-Scale Containment Model
April 29, 2011
Publication: 2011 Structures Congress, Proceedings of the 2011 Structures Congress, April 14-16, 2011, Las Vegas, Nevada, ASCE
Services: Structural Investigation