Application of neural network in the prediction of pile bearing capacity

Abstract

ABSTRACT

This paper presents the results of applying the artificial neural network model in determining the pile load capacity. An artificial neural network model was developed and trained on EXCEL software and used to determine the pile load capacity. The results of the study are compared with the experimental formula according to TCVN 10304:2014, showing that the artificial neural network allows better prediction of the pile load capacity than the standard formula. The survey results also show that the model predicts the bearing capacity of bored piles better than that of pressed piles. In addition, the model determines the bearing capacity of concrete piles better than that of steel and composite piles. The results of the study are a premise for the application of artificial neural networks in solving other problems in the field of construction.

Keywords: pile bearing capacity, CPT results, artificial intelligence, neural network