摘要详情

ID / 提交时间

277 / 2025-06-15 22:36:58

标题

Identification of end mill wear state under variable operating conditions

关键字

current signals, pyramid networks, tool wear, condition recognition

主题及专题

4、System Monitoring, Fault Diagnosis and Prediction

状态

全文待审

作者

宏坤李 / 大连理工大学

摘要

As a key component of impeller machinery, the machining quality of the impeller directly affects the working performance and effective life of the impeller machinery. The manufacture of impeller involves complex geometries, which requires extremely high cutting performance of the tool, and accurate identification of the wear state can effectively ensure the machining accuracy and extend the service life of the tool. Wear states may have different effects on different time scales and frequency scales. In this paper, an improved feature pyramid network with excellent multi-feature extraction and fusion capability is proposed to achieve intelligent identification of tool wear state under variable working conditions. Firstly, ESPN is proposed as the basic skeleton of the feature pyramid architecture to capture potential features and multidimensional information in the current signal to generate multilevel feature samples; secondly, the ECA module is proposed to strengthen the expression of features at all levels and to enhance the fusion process of the FPN structure sequentially, so as to generate high-quality multilevel outputs; lastly, the method is validated through laboratory simulation experiments and spindle current data in the actual manufacturing process of the impeller. the effectiveness of the method in tool wear identification. The results show that the method can be applied to real-time tool wear monitoring with high recognition accuracy.