Prediction of grain size of Al-7Si Alloy by neural networks

Neural networks, which are known for mapping non-linear and complex systems, have been used in the present study to model the grain-refinement behavior of Al-7Si alloy. The development of a feed forward neural network (FFNN) model with back-propagation (BP) learning algorithm has been presented for the prediction of the grain size, as a function of Ti and B addition level and holding time during grain refinement of Al-7Si alloy. Comparison of the predicted and experimental results shows that the FFNN model can predict the grain size of Al-7Si alloy with good learning precision and generalization. © 2004 Elsevier B.V. All rights reserved.