Predictive modeling of Fi and Ri in hard turning of 100Cr6 steel using a MLR and ANN approach

What is it about?

This research presents a comparative investigation and predictive modeling of cutting forces (Fx ,Fy ,Fz ) and surface roughness (Ra ,Rt ,Rz) in the hard turning of 100Cr6 steel (60 HRC) using CC650 ceramic insert. Utilizing a 2^3 full factorial design, Multiple Linear Regression (MLR) and Artificial Neural Network (ANN) models were developed to analyze the impact of cutting speed (Vc ), feed rate (f), and depth of cut (ap ).

Featured Image

Why is it important?

Both modeling approaches provided satisfactory predictive performance. Linear regression, despite its sim-plicity, achieved remarkable accuracy, with a negligible difference on the order of 10 − 10 compared to ANN. This indi- cates that, in well-structured systems with limited variability, linear models can perform as well as more complex ones, while offering greater interpretability, faster computation, and easier implementation. Nevertheless, ANN remain valu- able for capturing nonlinear relationships, but their complex-ity, the need for extensive parameter tuning, and their lack of transparency may restrict their industrial applicability.