2025年01月10日 星期五 首页   |    期刊介绍   |    编 委 会   |    投稿指南   |    期刊订阅   |    统合信息   |    联系我们
计量学报  2023, Vol. 44 Issue (4): 582-590    DOI: 10.3969/j.issn.1000-1158.2023.04.14
  电磁学计量 本期目录 | 过刊浏览 | 高级检索 |
基于深度迁移学习和LSTM网络的微电网故障诊断
吴忠强,卢雪琴
燕山大学 工业计算机控制工程河北省重点实验室,河北 秦皇岛 066004
Microgrid Fault Diagnosis Based on Deep Transfer Learning and LSTM Network
WU Zhong-qiang,LU Xue-qin
Key Lab of Industrial Computer Control Engineering of Hebei Province, Yanshan University, Qinhuangdao, Hebei 066004, China
全文: PDF (178922 KB)   HTML (1 KB) 
输出: BibTeX | EndNote (RIS)      
摘要 提出一种基于深度迁移学习与长短期记忆网络相结合的微电网故障诊断方法,可对不同结构的微电网进行故障诊断。首先利用小波包变换提取故障特征组成特征向量作为网络输入;其次,利用源域数据样本对长短期记忆网络故障诊断模型进行预训练,并保存相关参数;然后采用迁移学习将预训练模型中的参数迁移至域自适应网络,得到深度迁移学习与长短期记忆网络相结合的模型;最后根据有标记数据(源域数据)和目标域数据对模型进行微调迁移训练,将单一微电网故障诊断模型迁移至其他不同结构微电网。测试结果表明,该方法能高性能地检测和识别不同结构微电网中任何类型的故障,识别结果均方误差为8.5905×10-5,相比于自适应调整前的长短期记忆网络模型小2个数量级,识别效果更好,诊断精度有明显提高。
服务
把本文推荐给朋友
加入我的书架
加入引用管理器
E-mail Alert
RSS
作者相关文章
吴忠强
卢雪琴
关键词 计量学微电网故障识别深度迁移学习长短期记忆网络小波包分解    
Abstract:A microgrid fault diagnosis method based on the combination of deep transfer learning and long-short-term memory network is proposed, which can be used to diagnose microgrids fault with different structures. First, wavelet packet transform is used to extract the feature vectors of fault as the input of the network; second, source domain data samples is used to pre-train the fault diagnosis model built by long short-term memory network, and relevant parameters is saved; Then transfer learning is used to migrate the parameters in the pre-training model to the domain adaptive network to obtain the combination model of deep transfer learning and long-short-term memory network; finally, the model is fine-tuned and migrated according to the labeled data (source domain data) and target domain data, and the single microgrid fault diagnosis model is transferred to other microgrids of different structures. The test results show that the proposed method can detect and identify any type of faults in microgrids of different structures. The mean square error of the identification result is 8.5905×10-5, smaller two orders of magnitude compared with the long-short-term memory network model before adaptive adjustment, the recognition effect is better and the diagnosis accuracy is obviously improved.
Key wordsmetrology    microgrid    fault recognition    deep transfer learning    long-short-term memory network    wavelet packet decomposition
收稿日期: 2021-04-26      发布日期: 2023-04-18
PACS:  TB971  
基金资助:河北省自然科学基金(F2020203014)
通讯作者: 卢雪琴(1995-),女,四川德阳人,燕山大学硕士研究生,主要研究方向为微电网故障识别。Email: 1031794826@qq.com     E-mail: mewzq@163.com
作者简介: 吴忠强(1966-),男,上海人,燕山大学教授,主要从事微电网故障识别方面的研究。Email: mewzq@163.com
引用本文:   
吴忠强,卢雪琴. 基于深度迁移学习和LSTM网络的微电网故障诊断[J]. 计量学报, 2023, 44(4): 582-590.
WU Zhong-qiang,LU Xue-qin. Microgrid Fault Diagnosis Based on Deep Transfer Learning and LSTM Network. Acta Metrologica Sinica, 2023, 44(4): 582-590.
链接本文:  
http://jlxb.china-csm.org:81/Jwk_jlxb/CN/10.3969/j.issn.1000-1158.2023.04.14     或     http://jlxb.china-csm.org:81/Jwk_jlxb/CN/Y2023/V44/I4/582
京ICP备:14006989号-1
版权所有 © 《计量学报》编辑部
地址:北三环东路18号(北京1413信箱)  邮编:100029 电话:(010)64271480
本系统由北京玛格泰克科技发展有限公司设计开发  技术支持:support@magtech.com.cn