Continuing to use severely worn cutting tools during milling processing would cause damage to the cutting tools, which can lead to a decrease in the quality of the machined parts and even result in a large number of unqualified products. If the situation is serious, it can also cause damage to the machine tool and threaten the lives of personnel. The application process of digital twin in tool wear monitoring includes tool parameter collection and sensor installation, establishment of digital twin model, model parameter update and calibration, data preprocessing and feature extraction, establishment of wear monitoring model, real-time monitoring and early warning, experimental design and data collection, monitoring effect evaluation, and cost-benefit analysis. Among them, the digital twin model was established based on the collected data, including the Geometric modeling of the tool, material properties and cutting force model. This article used the CNN+LSTM (Convolutional Neural Network+Long short term memory) method to establish a wear monitoring model, and analyzed and evaluated experimental data. The accuracy and reliability of the digital twin model and monitoring algorithms were verified, while their cost-effectiveness in real production was analyzed. In this paper, the wear monitoring time of digital twin monitoring method B was 25 hours, while that of traditional manual inspection method B was 80 hours. The method proposed in this article has high accuracy and stability under new operating conditions, which can help improve industrial production efficiency and reduce costs.
