The use of machine learning techniques to assess experimental and fine-element analysis findings on the ortogonal cutting process

As in many industries, it is important to predict data that will affect the efficiency of production in the chip manufacturing industry in terms of reducing both cutting tool costs and production costs. In this context, besides the finite element method that has been used for years, machine learning and artificial intelligence methods used in many fields in recent years have emerged. When these learning methods are used as an aid to experimental and numerical studies, both experimental times and numerical calculation times are shortened significantly. Thus, it can be ensured that both the experimental costs and the numerical calculation times are shortened. In addition, it is possible to estimate unknown or untested parameters. In this study, primarily numerical modelling of the orthogonal machining problem was emphasized and the modelling results were compared with the experimental data. Then, experimental data and finite element data were taught to machine learning algorithms. After the teaching process, the cutting force and maximum chip temperature, which are unknown data, were estimated by entering only the cutting parameters. Support Vector Machine, Linear Regression and Gaussian Process Regression were determined as the machine learning algorithms that gave the most correct answer to the data obtained from the finite element analysis results with the experiments. In addition to these, the temperature, chip shape and stresses obtained in orthogonal cutting process are discussed.