PREDICTION OF MAXIMUM CRACK WIDTH BY MACHINE LEARNING USING CONCRETE CONSTRUCTION DATA IN YAMAGUCHI SYSTEM

Abstract

This study was intended to improve the prediction accuracy for the Artificial Neural Network (ANN) model presented in the recent past research to predict the maximum thermal crack width in RC abutment walls. FEM thermal stress analysis was used to filter the potentially high and low-risk zero-cracking lifts based on thermal cracking index. Several new input variables for ANN were added to the past model, and sophisticated input data set was used. The revised ANN model was utilized to simulate thermal cracking response to three different material countermeasures, such as use of expansive additive, glass fiber sheet reinforcement, and type of cement.