ANN Based Mechanistic Force Model for Face

Abstract

Due to increased global competition and increased calls for environmentally benign machining processes, there has been more focus and interest in making processes more lean and agile to enhance efficiency, reduce emissions and increase profitability. One approach to achieving lean machining is to develop a virtual simulation environment that enables fast and reasonably accurate predictions of machining scenarios, process output and provide access to needed information. Investigation on the utilization of artificial intelligence is carried out. Artificial Neural networks (ANNs) are employed to develop a smart data base that can provide fast prediction of cutting forces resulting from various combinations of cutting parameters. The data base can also predict the cutting coefficient usually predicted to calibrate the force models. The data base would be highly beneficial to the growing manufacturing industry in the United Arab Emirates (UAE), as it can be used to decide upon optimum parameters prior to carrying out cutting tests. With time, the force model can expand to include different materials, tools, fixtures and machines and would be consulted prior to starting any job. Predictions are compared to measured experimental results and are shown to be in good agreement. To address some of the difficulties encountered when using ANNs to predict cutting forces, the use of Polynomial classifiers (PCs) was also investigated to predict the cutting forces. A comparison between the predictions obtained using the PCs were found to be in good agreement compared to experimental results.