Surface Finish Monitoring in Taper Turning CNC Using Artificial Neural Network and Multiple Regression Methods

Abstract

On-line monitoring systems eliminate the need for post-process evaluation, reduce production time and costs, and enhance automation of the process. The cutting forces, mechanical vibration and emission acoustic signals obtained using dynamometer, accelerometer, and acoustic emission sensors respectively have been extensively used to monitor several aspects of the cutting processes in automated machining operations. Notwithstanding, determining the optimum selection of on-line signals is crucial to enhancing system optimization requiring a low computational load yet effective prediction of cutting process parameters. This study assess the contribution of three types of signals for the on-line monitoring and diagnosis of the surface finish (Ra) in automated taper turning operations. Systems design were based on predictive models obtained from regression analysis and artificial neural networks, involving numerical parameters that characterize cutting force signals (Fx, Fy, Fz), mechanical vibration (ax, ay, az), and acoustic emission (EARMS).