Diagnosis of Roundness Error in CNC Turning Operations through the Optimum Selection of On-Line Signals

The cutting forces, mechanical vibration and acoustic emission signals obtained using dynamometer, accelerometer, and acoustic emissions sensors have been extensively used to monitor several aspects of the cutting processes in automated machining operations. This study assesses the significance of these on-line signals for the real-time monitoring and diagnosis of the roundness error in automated cylindrical turning processes. The system developed is based on predictive models obtained by regression techniques employing the orthogonal components of the cutting forces, mechanical vibration and acoustic emissions, and combines all three types of sensors into one system. This monitoring system enables the on-line monitoring and diagnosis of roundness error by registering, visualizing, and characterizing the signals obtained during the machining process.