Estimation of ultimate bearing capacity of bored piles using machine learning models

Abstract

The ultimate bearing capacity of bored piles is an essential parameter in foundation design of structure. In the present study, three Machine Learning (ML) methods namely Adaptive Neuro-Fuzzy Inference System (ANFIS), Support Vector Machine (SVM) and Artificial Neural Network (ANN) were utilized to estimate bearing capacity of bored piles based on limited engineering parameters of pile and soil obtained from 75 test sites in Vietnam. These parameters include pile diameter, pile length, tensile strength of main longitudinal steel bar, compressive strength of concrete, average SPT index at the tip of the pile, average SPT index at the pile body. Validation of the methods was verified using standard statistical metrics namely Root Mean Square Error (RMSE) and Correlation coefficient (R). The results show that all the proposed models have good potential in predicting correctly bearing capacity of bored piles on training data (R>0.93) and on testing data (R>0.88) but performance of the SVM model is the best (R:0.985 (training) and R:0.958 (testing). Thus SVM model can be used for the accurate prediction of ultimate bearing capacity of bored piles for proper designing of the civil engineering structure foundation.