Applying Machine Learning Algorithm for a Milling Process Simulator for Process Modelling and Optimization

Abstract

Manufacturing industry is currently the heart of data driven revolution which is the cornerstone in transforming the traditional manufacturing systems to highly automated smart manufacturing by embedding new technologies such as Internet of Things (IoT), cyber-physical systems and cloud computing in physical advanced manufacturing processes to measure and monitor real time data. It is necessary to manage data and apply big data analytics to extract meaningful pattern. Also, it is helpful to adopt artificial intelligence techniques in manufacturing context to increase process efficiency within the framework of industry 4.0. The aim of this paper is implementing neural network algorithm for a milling process simulator in order to model and optimize the process. MATLAB was used to build the model based on input and target milling machine data which are material removal rate and spindle load respectively, the objective of the model was to predict spindle load. Then, the network has been tuned to improve the results, the first model achieved a good performance with overall mean squared error value of 0.948 while the second one had bed better performance of 0.879. The tuned model had lower average error value of 0.580 compared to 0.606 in the original model.