A tool breakage monitoring method for end milling based on the indirect electric data of CNC system

Abstract

Automatic tool breakage monitoring (TBM) is a vital technology of the unmanned workshops and automatic production lines for CNC machining. The current solutions for TBM mostly rely on direct data retrieved from the external sensors mounted on a machine tool, e.g., force and vibration sensors, which complicates the entire TBM system and adds additional costs. In this paper, instead of relying on the external sensors, indirect data from the CNC system, e.g., the spindle power, is utilized for the TBM. As the spindle power is a comprehensive signal, which not only contains a component of the cutting force, the direct factor determining tool breakage, but also involves the Coulomb friction and viscous damping forces, which are thermally sensitive and nonlinear, as well as the other components, such as the inertial force. For tool breakage to be effectively and accurately recognized, the spindle power data are preprocessed based on a mapping relationship between the spindle power and block numbers, with the one affecting tool breakage isolated via empirical mode decomposition (EMD). From the extracted signal, a support vector machine (SVM)-based method is then proposed to identify tool breakage. With the proposed method, actual machining experiments were conducted on a CNC milling center to verify that the proposed method can successfully extract the signal associated with tool breakage from the spindle power and that tool breakage can be accurately detected whenever it happens.