Novel data-driven open-circuit fault diagnosis method for modular multilevel converter submodules based on optimized deep learning

IF 1.3 4区 工程技术 Q3 ENGINEERING, ELECTRICAL & ELECTRONIC
Yang An, Xiangdong Sun, Biying Ren, Xiaobin Zhang
{"title":"Novel data-driven open-circuit fault diagnosis method for modular multilevel converter submodules based on optimized deep learning","authors":"Yang An, Xiangdong Sun, Biying Ren, Xiaobin Zhang","doi":"10.1007/s43236-024-00877-3","DOIUrl":null,"url":null,"abstract":"<p>As the proportion of clean energy continues to increase, low carbon energy systems will be a significant way to achieve the goal of carbon neutrality. Therefore, the reliability of modular multilevel converters (MMCs) is particularly significant. However, conventional open-circuit fault diagnosis (OCFD) methods usually have a limited localization speed or are difficult to achieve in practical engineering. Therefore, a fast and simpled OCFD approach for MMC SMs based on an optimized deep learning is proposed in this article. In this approach, data on the of submodule capacitance voltages are input into a trained WOA-DKELM model without the manually settings. The problems of randomness in the regularization coefficient<i> C</i> and the kernel parameters <i>K</i> can be solved by DKELM with WOA optimization, which has a strong generalization capability and higher prognostic accuracy. The effectiveness of the proposed approach is verified by experiment results. This approach achieves an average identification probability of 0.96 within 20 ms of the fault.</p>","PeriodicalId":50081,"journal":{"name":"Journal of Power Electronics","volume":"15 1","pages":""},"PeriodicalIF":1.3000,"publicationDate":"2024-07-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Power Electronics","FirstCategoryId":"5","ListUrlMain":"https://doi.org/10.1007/s43236-024-00877-3","RegionNum":4,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"ENGINEERING, ELECTRICAL & ELECTRONIC","Score":null,"Total":0}
引用次数: 0

Abstract

As the proportion of clean energy continues to increase, low carbon energy systems will be a significant way to achieve the goal of carbon neutrality. Therefore, the reliability of modular multilevel converters (MMCs) is particularly significant. However, conventional open-circuit fault diagnosis (OCFD) methods usually have a limited localization speed or are difficult to achieve in practical engineering. Therefore, a fast and simpled OCFD approach for MMC SMs based on an optimized deep learning is proposed in this article. In this approach, data on the of submodule capacitance voltages are input into a trained WOA-DKELM model without the manually settings. The problems of randomness in the regularization coefficient C and the kernel parameters K can be solved by DKELM with WOA optimization, which has a strong generalization capability and higher prognostic accuracy. The effectiveness of the proposed approach is verified by experiment results. This approach achieves an average identification probability of 0.96 within 20 ms of the fault.

Abstract Image

基于优化深度学习的模块化多电平转换器子模块新型数据驱动开路故障诊断方法
随着清洁能源比例的不断增加,低碳能源系统将成为实现碳中和目标的重要途径。因此,模块化多电平转换器(MMC)的可靠性尤为重要。然而,传统的开路故障诊断(OCFD)方法通常定位速度有限,或在实际工程中难以实现。因此,本文提出了一种基于优化深度学习的 MMC SM 快速、简化 OCFD 方法。在该方法中,子模块电容电压数据无需手动设置,直接输入训练有素的 WOA-DKELM 模型。正则化系数 C 和核参数 K 的随机性问题可以通过 DKELM 与 WOA 优化来解决,它具有很强的泛化能力和更高的预报精度。实验结果验证了所提方法的有效性。该方法在故障发生后 20 毫秒内的平均识别概率达到 0.96。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
Journal of Power Electronics
Journal of Power Electronics 工程技术-工程:电子与电气
CiteScore
2.30
自引率
21.40%
发文量
195
审稿时长
3.6 months
期刊介绍: The scope of Journal of Power Electronics includes all issues in the field of Power Electronics. Included are techniques for power converters, adjustable speed drives, renewable energy, power quality and utility applications, analysis, modeling and control, power devices and components, power electronics education, and other application.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信