Automated Cam Programming for Machining of Turned Components Utilizing Autodesk Fusion 360 CNC Module and ANN

Abstract

This article is a new fusion of Computer-Aided Manufacturing (CAM) and Artificial Neural networks (ANN) to estimate the parameters of the surface roughness (Ra and Rz) and the forces cutting (Fx, Fy, Fz) during the cut 3-axis milling on EN-31 steel. A detailed dataset about different toolpath strategies and different cutting parameters with Fusion 360 CAM software was made, and this dataset was used to train ANN models. A lot of exploratory data analysis (EDA) and training were done using standardized machine learning methods, and models were measured using certain metrics including Mean Squared Error (MSE), Mean Absolute Error (MAE) and R-squared (R²⁾. It was found that the ANN models showed good predictive precision with R 2 values of 0.817 and 0.860 and all the components of the cutting forces with values greater than 0.99. These outcomes mean that ANN can be successfully used to model nonlinear intricate relations between inputs of machining and outputs of performances. The described CAM-ANN hybrid solution provides a huge step towards automation of surface integrity and correct force determination, which minimizes the necessity of physical tests and makes the production more data-wise and intelligent. The described approach is in line with the goals of Industry 4.0 to introduce predictive intelligence and favor process optimization in new-age machining facilities.