Critical Load Prediction of Steel Compression Members using Neural Networks

摘要

The subject of the current research study is the prediction of the critical load of steel cold-formed Z-profile compression members, using neural networks with the purpose to use the trained network for optimization. In the critical load prediction a backpropagation neural network is developed. The test patterns, not included in training patterns, are estimated considering the value of the error. Parallel networks are applied to eliminate the inaccuracy of the results, came from the badly structured input data. The trained network gives very fast results, which are accurate enough for the practical application. Therefore this procedure can be the basis of a genetic algorithm based optimization method for cold-formed compression members.