Effect of Prior Knowledge on Site-Specific Selection of Regression Model for Characterization of Geotechnical Properties

Abstract

During geotechnical site characterization, many geotechnical properties might be difficult to measure directly and have to be estimated using indirect measurement and regression models. For example, when there is no possibility of direct compression test, geotechnical engineers and practitioners may utilize regression models (i.e. equations) to estimate the uniaxial compressive strength, UCS, of rock from point load index, Is(50). However, there are many equations relating Is(50) to UCS in the literature. This leads to the problem of how to select the most appropriate model for a particular rock deposit out of the numerous models available. This study presents a method that rationally compares different regression models and selects the most appropriate model for a specific site or deposit considered herein. The most appropriate model is selected using only a limited number of site-specific Is(50) data. The selected model is then used in a Bayesian framework to integrate the prior knowledge about UCS with the limited number of site-specific Is(50) data available for probabilistic characterization of UCS. The approach is shown to perform properly, particularly when the prior knowledge reflects information from the site.