MODELING OF CUTTING FORCE AND POWER CONSUMPTION USING ANN AND RSM METHODS IN TURNING OF AISI D3: COMPARATIVE STUDY AND PRECISION BENEFIT

The present work deals with the optimization of cutting pa-rameters when turning D3 steel using a CVD multi-layer coated carbide tool (Al2O3+TiC+TiCN). For that, response surface methodology (RSM) and ar-tificial neural network (ANN) were adopted for the modeling of cutting force (Fz) and cutting power evolutions (Pc). The applied precited approaches were also compared and their results were discussed. Moreover, a design of exper-imental (DoF) based on Taguchi L16 (4 perpendicular to 3 2 perpendicular to 1) method was adopted. This has helped to illustrate the relationship between cutting parameters (tool radius, cutting speed, feed rate and cutting depth) and selected responses which are cutting force and cutting power. The results revealed that the ANN and RSM exhibited very good accuracy with experimental data. However, the ANN pre-diction model provides the maximum benefit in terms of precision compared to the RSM model. For (Fz and Pc) the benefit is (7.5 and 16.3)%, respectively.