A study of tool life in hot machining using artificial neural networks and regression analysis method

In this study, the high manganese steel specimens heated with liquid petroleum gas flame were machined on a lathe under different cutting conditions of feed rates, depth of cuts, cutting speeds and surface temperatures. A mathematical model for tool life was obtained from the experimental data using a regression analysis method. In addition, the tool life was estimated using artificial neural network (ANN) with backpropagation (BP) algorithm. Then, this program was trained and tested. Finally, the experimental data are compared with both the regression analysis results and the estimations of ANN.