Hierarchical Bayesian Approach to Estimating Variability of Liquefaction Resistance of Sandy Soils Considering Individual Differences in Laboratory Tests

Kyohei Ueda

doi:10.1061/(asce)gt.1943-5606.0002749

Abstract

Hierarchical Bayesian Approach to Estimating Variability of Liquefaction Resistance of Sandy Soils Considering Individual Differences in Laboratory Tests

Highlights

In the organization of observation and experiment results, both the median trend and the variability due to uncertainty should be evaluated, for example, to derive an empirical equation
Cyclic undrained triaxial tests are commonly used in research and practical design to evaluate the liquefaction resistance of sandy soils
This paper proposes a methodology to evaluate liquefaction resistance by considering the variability or uncertainty associated with experimentation, using the concept of Bayesian statistics

Summary

Introduction

In the organization of observation and experiment results, both the median trend and the variability due to uncertainty should be evaluated, for example, to derive an empirical equation. Because boring and sounding are generally performed at certain intervals, the spatial distribution of soil properties is unknown at unexplored locations; this kind of uncertainty is called spatial variability (Phoon and Kulhawy 1999a; Vanmarcke 1977). Transformation uncertainty arises in the correlation equation (or transformation model) for designing soil parameters (e.g., friction angle of sand) based on test indexes [e.g., standard penetration test (SPT) blow count or cone penetration test (CPT) results] (Phoon and Kulhawy 1999b). Statistical uncertainty is due to the lack of quantity of observations and experiments. Measurement uncertainty comes from the precision and accuracy of measuring instruments, along with any other factor that might affect an experimenter’s ability to measure

Objectives

Discussion

Conclusion