Application of Artificial Neural Networks in Predicting Surface Quality and Machining Time

This paper explores the use of artificial neural networks to optimize metal machining processes. It has two main components. The first component focuses on developing a neural network in EasyNN to predict quality errors of metal surfaces machined with toroidal milling. The input data are extracted from the experimental values of the machining operations, and the output data are the measured surface roughness values. The network created not only predicts the quality errors but also identifies the key parameters influencing their quality, such as cutting speed, feed rate, and tool geometry, which are crucial for process optimization. The second component underscores the accuracy of the neural network′s predictions by using MatLab to develop a neural network that estimates the machining time, a key factor in cost calculation and efficient order planning. The experimental data used to train the network come from a restricted set of machining jobs. The obtained results demonstrate an adequate estimation of the machining time, thus facilitating the optimization of the entire metal machining process, positively impacting the cost and efficiency of industrial operations.