Use of Neural Networks to Predict Ultimate Strength of Circular Concrete Filled Steel Tube Beam-Columns

Abstract

Artificial neural networks (ANNs) are useful computing system which can be trained to learn complex relationship between two or more variables. It learns from examples and storage the knowledge for future use. In this study, a model for predicting the ultimate strength of circular concrete filled steel tube (CCFST) beam-columns under eccentric axial loads has been developed in ANN. The available experimental results for 181 specimens obtained from previous studies were used to build the proposed model. The predicted strengths obtained from the proposed ANN model were compared with the experimental values and current design provision for CCFST beam-columns (AISC and Eurocode4). Results showed that the predicted values by the proposed ANN model were very close to the experimental values and were more accurate than the AISC and Eurocode4 values. As a result, ANN provided an efficient alternative method in predicting the ultimate strength of CCFST beam-columns.