Prediction of Mechanical Properties of Hot Rolled Strips by BP Artificial Neural Network

Based on analysis of artificial neural network model theory and modeling methods, combining with parameters of a certain factory strip research unit and mechanical performance inspection data, Through choosing Traingdm to train network, then, the determination of the input and output parameters, the hidden layers of the network, cell numbers of hidden layers, learning rate lr, momentum factor α and training accuracy called goal, this paper established the three layers of BP artificial neural network performance forecast model. The analysis of experimental results showed that it had the high coincidence between prediction results of yield strength, tensile strength, elongation through the training and measured data. Therefore, the BP artificial neural network performance forecast model had higher forecast precision and practicability. Therefore, it can be used in forecast calculation in the production process of strip steel.