An Application of BP Neural Network to the Prediction of Compressive Strength in Circular Concrete Columns Confined with CFRP: Fid-Bau Portal

An Application of BP Neural Network to the Prediction of Compressive Strength in Circular Concrete Columns Confined with CFRP

AL-Bukhaiti, Khalil / Liu, Yanhui / Zhao, Shichun / Abas, Hussein

Abstract

The neural network comprises many neurons with extensive interconnections operating parallel and performing specific functions. This paper establishes a BP neural network prediction model for the compressive strength of CFRP-confined concrete based on a large number of experimental data to study the predictive ability of the BP neural network on the compressive strength of CFRP-confined concrete and the output performance of the neural network model. The model is based on a BP neural network that has been trained using many experimental data. An investigation is being conducted on the effect of different data combinations on the accuracy of the predictions made by the neural network model. The high-precision BP network model is created into generic and simplified formulae for application convenience. These formulas are developed based on the theory of neural networks. The neural network models’ findings and the empirical formulae for making predictions are compared and discussed. The BP neural network accurately predicts the compressive strength of CFRP-confined concrete, with over 90% of its data points having less than 15% error. In comparison, the regression model shows less accuracy, with less than 70% of its data points having an error within 15%. Compared to traditional regression models, the simple linear equation derived using Purelin instead of Sigmoid as the transfer function only adds a constant term. The average value of prediction/test results is 1.011. The analysis results show that BP neural network can extract the input and output parameters’ data information well and obtain a high-accuracy prediction model. The coefficient of variation is 0.112, which indicates that the prediction accuracy and stability are greater than average.