OPTIMIZATION OF CUTTING SPEED IN TURNING USING ARTIFICIAL INTELLIGENCE

Abstract

In the production of oil and gas equipment, the correct choice of cutting speed during high-precision turning crucially determines the quality and efficiency of the process. The cutting speed affects not only the surface roughness, but also the tool life, the stability of cutting forces and the reduction of production costs. The use of traditional empirical formulas has limitations in describing complex nonlinear dependencies.

This article suggests an approach based on artificial neural networks to optimize cutting speed. During the study, an experimental plan was developed that corresponds to real production conditions, and data was collected on the parameters of surface roughness (Ra), cutting forces (Fc, Ff) and tool wear (VB). The data obtained was processed by an artificial neural network model, a multi-criteria optimization was performed, and stable operating windows were determined based on the Pareto compromise. The results showed that artificial neural networks provide more accurate predictions compared to classical regression models and increase the reliability of the process.

The shortcomings identified in the existing literature — focusing on only one criterion, limited model portability, and a lack of reliability tests — were partially eliminated in this study. Thus, the proposed approach presents both theoretically and practically significant results for choosing the optimal cutting speed in production.