Reliability Assessment of Liquefaction Prediction Based on Ensemble Learning of a Database of Geotechnical Information

Abstract

In this paper, we use AI to predict the ground information at unknown points from the ground information at known points in Urayasu City, which has a large amount of reclaimed land and a very high liquefaction potential. The paper also introduces a liquefaction hazard map and a method for indicating liquefaction hazard using the PL method based on the predicted ground information. Damage to structures due to liquefaction was widely recognized after the 1964 Niigata and Alaska earthquakes. Since then, analysis of the mechanism of liquefaction has been conducted based on field surveys and experiments, and research has been conducted on the occurrence of liquefaction and countermeasures against it. Given the concern about the occurrence of large-scale earthquakes such as the Nankai Trough Earthquake in the near future, the evaluation of liquefaction hazards will become even more important. In order to establish ground investigation and countermeasure methods against liquefaction and subsidence, it is essential to obtain detailed information in the ground. It is essential to develop and establish a method to predict unknown points or unknown areas in the ground with high accuracy based on the limited results of ground investigation.