Revision as of 10:43, 10 June 2024

Abstract

In this study, artificial neural network, a popular machine learning technique, was used to estimate the consolidation properties of points where no soil investigations have been conducted. The Holocene clay layer at the construction site of Kobe Airport, a large-scale man-made island in Osaka Bay, was targeted to estimate the consolidation properties. The performance of an estimation model built with artificial neural network depends on the datasets used during the training phase. Therefore, the average of multiple estimation results can be used. Numerical simulations using the estimated consolidation properties can accurately reproduce the settlement behavior owing to reclamation during the construction of Kobe Airport. Artificial neural network can easily and objectively estimate the consolidation characteristics of any point based on the existing soil investigation results.

Full Paper