Investigation, modeling, and optimization of cutting parameters in turning of gray cast iron using coated and uncoated silicon nitride ceramic tools. Based on ANN, RSM, and GA optimization

A comparative study is undertaken in terms of the surface roughness criterion (Ra), the tangential cutting force (Fz), the cutting power (Pc), and the material removal rate (MRR) in turning of EN-GJL-250 cast iron using both coated and uncoated silicon nitride ceramics (Si3N4). The experimental procedure is carried out according to L27 Taguchi design process, and the analysis of variance ANOVA approach used to identify the cutting parameters that most influence the responses gathered. The artificial neural network approach (ANN) and the response surface methodology (RSM) were adopted to developing the mathematical prediction models applied in the optimization procedure that used genetic algorithm (GA). The predictive capabilities of the ANN and RSM models were further compared in terms of their mean absolute deviation (MAD), mean absolute error in percent (MAPE), mean square error (RMSE), and coefficient of determination (R-2). It has been found that the ANN method provides more precise results compared to those of the RSM approach. Moreover, the coated ceramic tool has been found to lead to a better surface quality and a minimum cutting force compared to those obtained by uncoated ceramic. The wear tests undertaken show that, when the flank wear reaches the admissible value of [Vb]=0.3mm, the ratios (tool life (CC1690)/tool life (CC6090)), (Ra-CC1690/Ra-CC6090), and (Fz(CC1690)/Fz(CC6090)) are found to equal 0.88, 1.4, and 0.94, respectively.