Surface roughness is important for evaluating the machined surface quality. In this work, an Artificial Neural Network (ANN) surface roughness prediction model was developed by coupling it with Response Surface Methodology (RSM). For this interpretation, advantages of statistical experimental design techniques, experimental measurements, and artificial neural network were exploited in an integrated manner. Cutting experiments were designed based on small centre composite design technique to develop a RSM model. The input cutting parameters were: cutting speed, feed, and axial depth of cut, and the output parameter was surface roughness. The predictive model was created using a feed-forward back-propagation neural network exploiting the experimental data. The network was trained with pairs of inputs/outputs datasets generated by end milling medium carbon steel with TiN coated carbide inserts. The model can be used for the analysis and prediction of the complex relationships between cutting conditions and surface roughness, in metal-cutting operations, with the ultimate goal of efficient production. The ANN model was verified with the optimized parameters predicted by a coupled genetic algorithm (GA) and RSM technique also developed by the authors.
Optimization of CNC operating parameters to minimize surface roughness of Pinus sylvestris using integrated artificial neural network and genetic algorithm