Optimizations of Cutting Coefficients in Milling with Population Based Meta-Heuristic Algorithms and Modelling of Cutting Forces

Abstract The milling process needs to be optimized to improve production outputs. Therefore, it is essential to investigate the cutting forces that have significant effects on the performance of the milling process. In the literature, researchers have developed many models based on predicting the forces generated during milling. However, the developed models require many experimental data. In this study, a hybrid method, including metaheuristic algorithms and semi-analytical methods, is developed to predict milling cutting forces. The presented method uses a semi-analytical mechanistic prediction model and optimizes the cutting constants (Ktc, Kte, Krc, and Kre) with population-based swarm intelligence algorithms. Genetic Algorithm (GA), Differential Development Algorithm (DE), and Particle Swarm Optimization (PSO) Algorithms are chosen to solve the problem. The analytical prediction model based on truncation coefficients is set as the objective function. Each algorithm was run thirty times for the solution, and their performances were compared. DE was the best-performing algorithm in terms of stability and speed. Experimental measurements were performed to verify the results. The actual cutting force components and the predicted force components Fx and Fy were evaluated by the coefficient of determination (R2). The results for the force components Fx and Fy were observed to be similar to the actual measurement results, with an accuracy of 88.8% and 92.7%, respectively. The method presented in the study proved that the cutting coefficients can be calculated much faster with only one experiment. The findings are promising for future studies on cutting conditions and autonomous execution of the process.