Predictions of ANN Model for Axial Strength of FRP-Confined CFST Columns

Abstract

The present investigation aims to predict theoretically the axial compression capacity of carbon fibre-reinforced polymer (CFRP)-wrapped concrete-filled steel tube compressive members (FCFST) by suggesting an artificial neural network (ANN) model. For achieving the aim of this study, a database of 216 FCFST compression members was engaged from a preceding investigation, and the training, validation, and testing of the ANN model were carried out. The training and validation of the suggested ANN model were performed for a different number of layers and neurons. The suggested model considers the interaction mechanism and interrelations of the various variables of FCFST compression members. Lastly, the improved ANN model depicted a good correlation with the database results of FCFST compression members.