Tool Wear State Monitoring Based on Long-Term and Short-Term Memory Neural Network

Abstract

The real-time detection of tool wear condition is the difficult point of machine tool processing status monitoring. The vibration signal of tool wear shows strong nonlinear and non-stationary phenomenon. The neural network model can be used to analyze the vibration signal of the tool to judge the wear state of the tool. To this end, an online monitoring method for tool wear status based on long-term and short-term memory neural network (LSTM) is proposed. Firstly, the tool vibration signal is collected by the acceleration sensor, and then the vibration signal is decomposed by wavelet packet transform to form the energy value corresponding to different frequency bands, which is used as the feature vector input of the long-term and short-term memory neural network. Finally, the long-short-term memory neural network model is used to process the tool vibration signal to judge the wear condition of the tool. In addition, the method is compared with BP neural network and multi-layer BP neural network fault diagnosis method. The results show that the LSTM network method is more effective for on-line monitoring of tool wear status.