Application of Artificial Neural Network for Modeling the Flash Land Dimensions in the Forging Dies

This paper proposes an approach to modeling the flash land by applying artificial neural network (ANN). A three-layer feed-forward ANN, with backpropagation algorithm for supervised learning is created. A sigmoid type of non-linearity is applied to neurons. In the reference literature there are many examples showing that the prediction model developed by means of ANN is more accurate than the one developed by the regression analysis. The trained ANN has shown a high level of prediction so that it can be used for designing and optimizing the conventional forging process.