一种基于频谱感知生成的小样本电机故障诊断方法
By using a spectrum-sensing generative learning framework and a ResNet-SE model, the problems of insufficient and imbalanced data in motor fault diagnosis are solved, achieving efficient fault identification under small sample conditions and improving diagnostic accuracy and robustness.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- QIQIHAR UNIVERSITY
- Filing Date
- 2026-05-14
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies face problems such as insufficient labeled data and unbalanced distribution in motor fault diagnosis, resulting in insufficient diagnostic accuracy and robustness, especially under small sample conditions.
A spectrum-aware generative learning framework is adopted, which combines a deep convolutional neural network (ResNet-SE) with generative adversarial networks (GAN) and SE attention mechanism. Through data augmentation and frequency domain consistency constraints, the quality and diversity of generated samples are improved, and residual learning is combined to capture the local transient features and long-range dependencies of vibration signals.
It significantly improves the accuracy and robustness of motor fault diagnosis under small sample conditions, enhances the model's fault identification ability, especially the fault mode discrimination ability under conditions of scarce and imbalanced data.
Smart Images
Figure CN122413142A_ABST