We found a match
Your institution may have rights to this item. Sign in to continue.
- Title
基于迁移学习与残差网络的刀具磨损状态监测.
- Authors
周建民; 王云庆; 杨晓彤; 黄熙亮; 夏晓枫
- Abstract
To solve the problem that the existing tool wear monitoring methods based on deep learning require too many samples and have low identification accuracy, a tool wear monitoring model based on transfer learning (TL) and deep residual network (Res- Net) was established. The vibration monitoring signals during tool machining were converted into energy time-frequency diagram by continuous wavelet transform, which were used as the input of network model. The ResNet50 model trained on ImageNet dataset was used as the pre-training model, and was applied to the field of tool wear state monitoring by transfer learning method. The example verification shows that the recognition accuracy of TL-ResNet reaches 98. 52%, realizing intelligent recognition under different tool wear states. So the accuracy and stability of tool wear state monitoring are improved effectively
- Subjects
DEEP learning; WAVELET transforms; PROBLEM solving; MACHINE tools; CUTTING tools; RECOGNITION (Psychology)
- Publication
Machine Tool & Hydraulics, 2023, Vol 51, Issue 18, p215
- ISSN
1001-3881
- Publication type
Article
- DOI
10.3969/j.issn.1001-3881.2023.18.035