Modeling Surface Roughness and Cutting Forces in the Turning Process Using AI Tools: ANN, SVM, KNN, GPR, and GBOOST

The importance of modelling the cutting conditions in a material removal manufacturing process is reflected in the choice of optimized parameters. This paper focuses on developing an artificial intelligence model of the roughness of a machined surface and the resulting cutting force in an orthogonal cut, choosing the cutting speed Vc, feed rate per revolution <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$f$</tex>, depth of cut <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$a_p$</tex>, and tool wear Vb as input parameters. The K Nearest Neighbors k-NN, Support Vector Machine Regressor SVM-R, Gaussian Process Regressor GPR, Gradient Boosting GBoost, and Artificial Neural Network ANN are trained and optimized, the final model is a combination of three good models by choosing the average of the maximum and minimum output for the roughness model. For the force resultant modelling, it is found that the output closest to the measured data is that of ANN added to the average of the GPR, ANN, and GBoost output. The model is the basis for predicting the effect of cutting parameters on roughness and cutting force, especially the effect of tool wear and the depth of the pass, and also for optimizing these cutting conditions to achieve a good surface finish. Two optimization tools, namely the Genetic Algorithm GA and the Particle Swarm Optimization PSO are used to optimize the cutting parameters. The minimum value found in the study range is equivalent to Ra<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">min</inf>= 0.1352μm with the PSO tool.