Study of Spindle Current Signals for Tool Breakage Detection in Milling

Abstract

Failure of cutting tools significantly decreases machining productivity and product quality, thus, tool condition monitoring is significant in modern manufacturing processes. In this paper, a novel method based on singular value decomposition (SVD) and Linear Discriminant analysis (LDA) is proposed for detection of different broken tooth via spindle-motor current signals generated in end milling process. First, SVD algorithm is adopted to extract the useful singular values as salient features for indicating the tool state. Then classify the tool breakage event based on the selected features through the LDA technique. The experiments on a CNC Vertical Machining Centre show that this method is effective and can accurately classify the different broken tooth in end mill process.