ABSTRACT
The absolute automation of manufacturing process experiences frequent breakage and corrosion of production tools during the process. To avoid this industrial automation requires optimization techniques in manufacturing process that includes the careful condition monitoring of the tool bit. It is preferred by the production standards to monitor the tool flank for deformities and wear without disrupting the production process so that the throughput does not suffer. The proposed method of such classification from the sound signals captured during the turning process exhibit a more efficient way of finding the degree of tool wear than other methods. The other methods can only do a rough estimation of tool condition and categorize them into slight, severe, and medium. The classification is done with the help of a neural network classifier that using logistic regression from the class average of different categories. The proposed procedure would be used to monitor the tool condition during the process of turning by classifying the emitted sounds by the logistic regression technique. The classification accuracy is confirmed with correlation result and regression analysis. An average of 94% accuracy is reached with our proposed technique.
