Analysis of round robin test on seismic site response analysis

Abstract

Site response analysis, widely performed in practice to estimate seismic site effects, requires multiple inputs regarding the dynamic soil properties. While comprehensive site-specific field and laboratory tests need to be performed for the proper characterization of a site, engineers are often given only limited information and are forced to work with uncertainties. A round-robin test on site response analysis was performed in Korea to evaluate the dispersion of the calculated responses originating from the uncertainties in soil characterization and from the method of analysis. The participants were intentionally given limited information, which represented a typical set of inputs given in practice. The participants were asked to predict the dynamic properties based on the given data, and to perform a series of site response analyses. This paper documents the variations in the predicted dynamic properties and the computed responses. The results showed that the prediction of the shear wave velocity was most important for the evaluation of the ground-motion propagation, while the type of analysis had only a secondary influence. Important guidelines for the prediction of dynamic properties are also given.