Reinforced Concrete Beam with Fiber Reinforced Polymer Deflection Prediction Using Machine Learning

Abstract

RC beam is composite materials that is commonly used in construction. The conventional method to predict the RC beam deflection is always consuming a greater amount of time and the deflection value obtained may be affected by human error. The aim of this study is to develop a predictive machine learning model based on the beam deflection historical data. The data set is obtained from published articles and analyze using PYTHON. The data then feed to the multiple linear regression algorithm to train and evaluate the model. For the machine learning development process, in involves processes such as data preparation, data pre-processing, features selection, features scaling, data partitioning, and evaluation of the model performance. R-squared value and correlation between the predicted displacement value and actual displacement value is used to evaluate the performance of the model. Predictive machine learning model is highly recommended to be used in civil engineering field as the computational is much more efficient compared to conventional testing method.