Full
Outline
The performance of a data-driven fault diagnosis method deteriorates due to insufficient number and unbalanced classes of inverter fault data in the case of weakly supervised learning in few-shot scenarios. To overcome the defect of traditional data augmentation methods which do not consider the data distribution characteristics, an inverter fault diagnosis method based on data augmentation with specific coefficients wavelet reconstruction was proposed. First, the raw small sample data was decomposed by wavelet packet, and a specific coefficient was obtained through calculation. After random fine-tuning, the sample was directly expanded by wavelet packet reconstruction. The process was repeated to obtain a sufficient and class-balanced training set. Then, convolutional neural networks were used to achieve “end-to-end” fault feature extraction and classification. Finally, experimental results show that the proposed method can effectively expand the fault data in the case of weakly supervised learning in few-shot scenarios and improve the fault diagnostic accuracy. Therefore, it can be used as data preprocessing in few-shot fault diagnosis of inverters.
56
5
| 科 Family | 属数 Number of genus | 种数 Number of species | 占总种数比例 Percentage of total species (%) | 属 Genus | 种数 Number of species | 占总种数比例 Percentage of total species (%) |
|---|---|---|---|---|---|---|
| 鹅膏菌科Amanitaceae | 2 | 11 | 5.26 | 鹅膏菌属 Amanita | 10 | 4.78 |
| 小菇科 Mycenaceae | 2 | 12 | 5.74 | 丝盖伞属 Inocybe | 5 | 2.39 |
| 多孔菌科 Polyporaceae | 8 | 14 | 6.70 | 蜡蘑属 Laccaria | 5 | 2.39 |
| 红菇科 Russulaceae | 3 | 23 | 11.00 | 小皮伞属 Marasmius | 6 | 2.87 |
| 小菇属 Mycena | 11 | 5.26 | ||||
| 光柄菇属 Pluteus | 5 | 2.39 | ||||
| 红菇属 Russula | 17 | 8.13 | ||||
| 栓菌属 Trametes | 5 | 2.39 |
关闭全屏
No. 86 Xueyuan South Road, Haidian District, Beijing
010-62662699(Operation and Technical Support)
qkjq@cast.org.cn
Copyright © 2025 China Association for Science and Technology. All rights reserved. For all open access content, the relevant licensing terms apply.
Sponsored by the Office of the Leading Group for Cybersecurity and Informatization of CAST, and supported by Science and Technology Review Publishing House
