RELIABILITY ASSESSMENT OF MULTIVARIATE SOIL PROPERTIES UNDER OFFSHORE AIRPORT USING LARGE-SCALE GENERIC DATABASE AND HIERARCHICAL BAYESIAN MODEL

Abstract

In this study, we investigated the feasibility of using a Hierarchical Bayesian Model (HBM) to estimate geotechnical parameters and provide engineering interpretations of soil properties. The study focused on estimating geotechnical properties of soft foundation soils at an offshore airport in Japan. We compared a basic HBM utilizing a global geotechnical investigation database (Generic database, GDB) with an HBM that employed a refined GDB based on similarity. The HBM with similarity refinement, incorporating geotechnical engineering knowledge at the soil layer level, demonstrated superior estimation accuracy and enhanced geotechnical insights. However, we observed that the accuracy of the estimates is contingent upon the quality of the GDB and may significantly decrease under certain conditions. We emphasized the importance of incorporating geotechnical engineering knowledge into the labeling and management of the database and highlighted the need for expanding the database in future endeavors.