APPLICATION OF ARTIFICIAL NEURAL NETWORKS TO THE PREDICTION OF ULTIMATE SHEAR STRENGTH OF HIGH STRENGTH CONCRETE COLUMNS REINFORCED WITH CONCRETE FILLED STEEL TUBE

This paper investigates the feasibility of using neural networks to evaluate the ultimate shear strength of high strength concrete columns reinforced with concrete filled steel tube in the core of section. A three-layer back-propagation network is trained using the experimental data of the columns in shearing test, among which 24 groups are used for training sample while the remaining 6 groups are used for testing sample, to predict the ultimate shear strength of the columns. The results agree well with the test results. In the further analysis about the influence of various parameters on the ultimate shear strength, it is found that it is significant to put concrete filled steel tube in the core of section for the high strength concrete columns in improving its ultimate shear strength. And it is shown that with the increase of volume stirrup ratio, the ultimate shear strength is also improved accrodingly.