Artificial Intelligence-Based Neural Network Prediction Model for Predicting Multi-responses of Finishing Honing Process

Abstract

Finishing honing process is one of the most effective machining process which can improve the surface properties of the product as the main target of manufacturing industries is to achieve the predefined quality of the product. This paper presents artificial intelligence-based neural network approach with two hidden layers having eight and seven neurons, respectively, as an alternate to all the conventional approaches for developing a predictive model to find the best or optimal values of process parameters. In this study, the capability of the proposed artificial neural network prediction model of the surface and process parameters is checked by taking a real-time machining experiment of the finishing honing process. This study’s research shows that the proposed method is capable of predicting the highly precise values of the machining parameters for the finishing honing process with a correlation coefficient of 0.9904 and mean relative error between the predicted value and the experimental value of the average roughness, maximum height roughness, and machining time is found to be 3.90%, 1.13%, and 0.45%, respectively.KeywordsFeedforward backpropagation neural networkFinishing honingSurface roughness