红水河2025，Vol.44Issue(3)：42-47,6.DOI:10.3969/j.issn.1001-408X.2025.03.008
基于长短期记忆网络的低水头永磁同步发电机微电流腐蚀故障检测技术研究
Stray-Current Corrosion Fault Detection Technology for Low-Head Permanent Magnet Synchronous Generators Based on Long Short-Term Memory Network
林琪 1杜兆龙 1张兴华 1李兴文 1何志雄2
作者信息
- 1. 广西右江水利开发有限责任公司,广西 百色 533000
- 2. 中能拾贝科技有限公司,广东 广州 510525
- 折叠
摘要
Abstract
To address the issue of stray-current corrosion fault detection in low-head permanent magnet synchronous generators,this study proposes a detection method based on a Long Short-Term Memory(LSTM)network.First,multi-region dominant frequency enhancement technology is adopted to improve the clarity of vibration signals.Subsequently,minimum entropy deconvolution is utilized to extract deep-level features from the vibration signals.Finally,the extracted features are input into the LSTM network for classification and identification.The kurtosis value of the enhanced vibration signals increases significantly,verifying the effectiveness of signal enhancement.Experimental results demonstrate that the proposed method significantly outperforms traditional approaches in terms of fault classification accuracy,recall rate,and F1-score for stray-current corrosion fault detection.It effectively enhances the corrosion fault classification performance of generators,enables timely identification of potential faults,reduces equipment wear and damage,prolongs service life,and ensures the stability and reliability of power generation operations.关键词
低水头永磁同步发电机/微电流腐蚀/长短期记忆网络/故障检测/特征提取Key words
low-head permanent magnet synchronous generator/stray-current corrosion/Long Short-Term Memory(LSTM)network/fault detection/feature extraction分类
信息技术与安全科学引用本文复制引用
林琪,杜兆龙,张兴华,李兴文,何志雄..基于长短期记忆网络的低水头永磁同步发电机微电流腐蚀故障检测技术研究[J].红水河,2025,44(3):42-47,6.
