Tool Condition Monitoring and Maintenance Based on Deep Reinforcement Learning

Tool status monitoring requires collecting a large amount of data to complete analysis, and different types of tools may exhibit different wear and failure modes during processing, making tool status monitoring more difficult. Therefore, a tool condition monitoring method based on deep reinforcement learning is proposed. The feature of tool wear is extracted by wavelet packet analysis. Introduce regression network into deep reinforcement learning network, and complete tool condition monitoring by combining regression algorithm with deep reinforcement learning network algorithm. Finally, specific suggestions for tool status maintenance are provided. To verify the effectiveness of the proposed method, comparative experiments were designed. The results show that the accuracy of tool condition monitoring is high, the monitoring decision coefficient can be maintained above 0.95, and the mean absolute percentage error is smaller.