Subtraction of Inconsistence Sensor Data to Improver the Chip Form Classification and Monitoring Efficiency

This research focuses on sensor based monitoring and control of chip formation during machining operations. In particular, the control for favorable chip formation has been carried out using real time cutting force sensor signal spectrums. The concerned machining process is longitudinal turning of carbon steel with coated carbide inserts, yielding different chip forms. In order to design automotive control for favorable chip formation, real time sensor signal spectrums are estimated through a parametric method that facilitates the corresponding chip form signal characteristics numerically. Moreover, presented approach advocates the refinement of sensor data before applying over the SOM (Self Organizing Map) methodology. The subtraction of inconsistent sensor data improves the preciseness of SOM NN learning. Hereby, it helps to enhance the performance of SOM to classify and identify the chip forms. Performance of SOM using refined data set has been compared to previous researches.