Compressive Strength of Confined Concrete in CCFST Columns

This paper presents a new model for predicting the compressive strength of steel-confined concrete on circular concrete filled steel tube (CCFST) stub columns under axial loading condition based on Artificial Neural Networks (ANNs) by using a large wide of experimental investigations. The input parameters were selected based on past studies such as outer diameter of column, compressive strength of unconfined concrete, length of column, wall thickness and tensile yield stress of steel tube. After the learning step, the neural network can be extracted the relationships between the input variables and output parameters. The criteria for stopping the training of the networks are Regression values and Mean Square Error. After constructing networks with constant input neurons but with different number of hidden-layer neurons, the best network was selected. The neural network results are compared with the existing models which showed the results are in good agreement with experiments.