PREDICTION METHOD OF SOIL CONDITIONS BY USE OF ARTIFICIAL INTELLIGENCE TO ANALYZE BORING-OPERATION DATA

Abstract

 It is required to have information on the soil conditions of fine fraction content and SPT-N value for estimating the liquefaction potential of ground, or for conducing soil improvement such as the chemical grouting method.

 In this study, machine learning techniques as the random forest method, the support vector machine and the neural network are applied analyzing boring-operation data obtained in the past soil improvement works at Tokyo International Airport. It is found out that the soil conditions of fine fraction content and SPT-N value can be accurately predicted by the multi-task learning model based on the neural network.