Friction force during machining process – Part 1: Development of optimized neural network architecture

Abstract

Abstract Resource management is one of the most important and critical criteria for any engineering manufacturing facility to function as per the predefined schedule for fulfilling the customer demands. Considering this as a basic requirement, maintaining the good working condition of the identified resources such as machine tools becomes the most prioritized objective of the organization. In this context, an effort has been made to propose Artificial Intelligence (AI) based predictive model to predict the condition of the machine tool by considering the frictional force between the work piece and the cutting tool. Construction of the prediction model is a data driven process in which the relationship between input and output parameters has been established. In order to generate the data to develop such a prediction model for friction force, a turning operation was performed on ferrous and non-ferrous material with different combinations of process parameters, insert nose radius, tool height setting, wet and dry cutting conditions. Friction force between the cutting tool edge and the work piece was computed with the help of forces measured through the lathe tool dynamometer by applying the principle of Merchant’s circle diagram. The friction force computed for 648 experimental trials was used to develop an optimized Artificial Neural Network (ANN) architecture prediction model to predict the friction force for a given set of input parameters.