基于深度NCA与原型网络的MMC-HVDC故障诊断方法

黄志光, 连攀杰, 骆玮, 刘嘉宸, 时艳强, 刘伟成

天津工业大学学报 ›› 2026, Vol. 45 ›› Issue (4) : 62 -72.

PDF (2105KB)
天津工业大学学报 ›› 2026, Vol. 45 ›› Issue (4) : 62 -72. DOI: 10.3969/j.issn.1671-024x.2026.04.008
电气工程

基于深度NCA与原型网络的MMC-HVDC故障诊断方法

    黄志光1, 连攀杰2, 骆玮1, 刘嘉宸1, 时艳强1, 刘伟成1
作者信息 +

Fault diagnosis method for MMC-HVDC based on Deep NCA and prototypical networks

    Huang Zhiguang1, Lian Panjie2, Luo Wei1, Liu Jiachen1, Shi Yanqiang1, Liu Weicheng1
Author information +
文章历史 +
PDF (2154K)

摘要

针对模块化多电平换流器柔性直流输电(MMC-HVDC)系统的故障暂态过程复杂、浅层特征易混叠及在线诊断实时性要求高等问题,提出一种基于深度邻域成分分析(Deep NCA)与原型网络的故障诊断方法。首先,利用一维卷积神经网络从桥臂电流、子模块电容电压和接地电流等多通道暂态序列中自适应提取深层特征;其次,引入Deep NCA对嵌入空间进行监督度量约束,增强类内紧凑性和类间分离性;最后,基于类别原型实现快速距离判别。仿真结果表明:所提方法在5类运行状态识别中准确率达到95.20%,高于1D-CNN+Softmax、1D-ResNet和1D-Transformer等典型深度模型;负荷偏移条件下的平均准确率为91.73%,单类训练样本数减少至50组时准确率仍为86.93%,表明该方法具有一定的工况适应性和有限训练样本适应能力。

Abstract

To address the complex fault transient process, feature overlapping in shallow feature spaces, and high real-time requirements for online diagnosis in modular multilevel converter-based flexible high voltage direct current (MMC-HVDC) systems, a fault diagnosis method based on deep neighborhood components analysis (Deep NCA) and prototypical networks is proposed. First, a one-dimensional convolutional neural network is used to adaptively extract deep features from multi-channel transient sequences, including arm current, submodule capacitor voltage, and grounding current. Second, Deep NCA is introduced to impose supervised metric constraints on the embedding space, enhancing intra-class compactness and inter-class separability. Finally, fast distance-based classification is achieved through class prototypes. The simulation results show that the proposed method achieves an accuracy of 95.20% for five operating states, outperforming typical deep learning models including 1D-CNN+Softmax, 1D-ResNet and 1D-Transformer. Under load-shift conditions, the average diagnostic accuracy is 91.73%, and an accuracy of 86.93% is retained when only 50 training samples per class are used. Results demonstrates that the proposed method possesses favorable adaptability to varying operating conditions and the capability to learn from limited training samples.

关键词

柔性直流(HVDC)输电 / 模块化多电平换流器(MMC) / 故障诊断 / 深度邻域成分分析(Deep NCA) / 原型网络 / 度量学习

Key words

flexible high voltage direct current (HVDC) transmission / modular multilevel converter (MMC) / fault diagnosis / deep neighborhood components analysis (Deep NCA) / prototypical network / metric learning

引用本文

引用格式 ▾
黄志光, 连攀杰, 骆玮, 刘嘉宸, 时艳强, 刘伟成. 基于深度NCA与原型网络的MMC-HVDC故障诊断方法[J]. 天津工业大学学报, 2026, 45(4): 62-72 DOI:10.3969/j.issn.1671-024x.2026.04.008

登录浏览全文

4963

注册一个新账户 忘记密码

参考文献

[1] 王佳, 吴任博, 肖健, 等. 新型电力系统能源调度与碳排放计价优化方法研究[J]. 电测与仪表, 2024, 61(10): 17-25.
Wang Jia, Wu Renbo, Xiao Jian, et al.Research on the optimization method of energy dispatching and carbon emission pricing for novel power system[J]. Electrical Measurement & Instrumentation, 2024, 61(10): 17-25(in Chinese).
[2] 雷顺广, 束洪春, 李志民. 基于桥臂功率特征的全-半混合型柔性直流输电线路保护[J]. 电工技术学报, 2023, 38(13): 3563-3575.
Lei Shunguang, Shu Hongchun, Li Zhimin.Full-half bridge hybrid VSC-HVDC transmission line protection method based on power characteristics of bridge arms[J]. Transactions of China Electrotechnical Society, 2023, 38(13): 3563-3575(in Chinese).
[3] 慕宗君, 邱俊宏, 李振兴, 等. 多类型故障影响下柔直线路故障机理分析及单端保护方法[J]. 中国电力, 2025, 58(1): 107-114.
Mu Zongjun, Qiu Junhong, Li Zhenxing, et al.Fault mechanism analysis and one-terminal protection method of flexible DC transmission line under the influence of multi-type faults[J]. Electric Power, 2025, 58(1): 107-114(in Chinese).
[4] 张博, 王聪博, 詹荣荣, 等. 计及控保协同的自适应电流差动保护方法[J]. 中国电力, 2025, 58(2): 1-8.
Zhang Bo, Wang Congbo, Zhan Rongrong, et al.Adaptive current differential protection method considering control and protection coordination[J]. Electric Power, 2025, 58(2): 1-8(in Chinese).
[5] 李景丽, 袁豪, 徐铭铭, 等. 小电阻接地系统高阻接地故障检测技术综述[J]. 电测与仪表, 2023, 60(6): 10-18, 51.
Li Jingli, Yuan Hao, Xu Mingming, et al.Overview of high impedance fault detection technology in small resistance grounding system[J]. Electrical Measurement & Instrumentation, 2023, 60(6): 10-18, 51(in Chinese).
[6] 茆美琴, 程德健, 袁敏, 等. 基于暂态能量流的模块化多电平高压直流电网接地优化配置[J]. 电工技术学报, 2022, 37(3): 739-749.
Mao Meiqin, Cheng Dejian, Yuan Min, et al.Optimal allocation of grounding system in high voltage direct current grid with modular multi-level converters based on transient energy flow[J]. Transactions of China Electrotechnical Society, 2022, 37(3): 739-749(in Chinese).
[7] 张心怡, 盖午阳, 孙巧玉, 等. 基于电流变异特性的直流配电系统单极接地差动保护方法[J]. 智慧电力, 2025, 53(6): 93-100.
Zhang Xinyi, Gai Wuyang, Sun Qiaoyu, et al.Single-pole grounding differential protection method for DC distribution systems based on current variation characteristics[J]. Smart Power, 2025, 53(6): 93-100(in Chinese).
[8] 韦延方, 吴凯航, 王晓卫, 等. 基于故障行波VS-DDTW距离的柔性直流输电线路保护方案[J]. 电力系统及其自动化学报, 2025, 37(12): 24-34.
Wei Yanfang, Wu Kaihang, Wang Xiaowei, et al.Protection scheme for flexible DC transmission line based on VS-DDTW distance of fault traveling waves[J]. Proceedings of the CSU-EPSA, 2025, 37(12): 24-34(in Chinese).
[9] 李杭泽, 戴志辉, 石旭, 等. 计及控制响应的多端混合直流输电系统短路电流近似计算方法[J]. 电工技术学报, 2024, 39(9): 2810-2824.
Li Hangze, Dai Zhihui, Shi Xu, et al.Approximate calculation method of short-circuit current of multi-terminal hybrid DC transmission system considering control strategy[J]. Transactions of China Electrotechnical Society, 2024, 39(9): 2810-2824(in Chinese).
[10] 王雪芹, 张大海, 李猛, 等. 基于小波能量谱和SSA-GRU的混合直流输电系统故障测距方法[J]. 电力系统保护与控制, 2023, 51(12): 14-24.
Wang Xueqin, Zhang Dahai, Li Meng, et al.Fault location method for a hybrid DC transmission system based on wavelet energy spectrum and SSA-GRU[J]. Power System Protection and Control, 2023, 51(12): 14-24(in Chinese).
[11] 张英敏, 张文馨, 李保宏, 等. 柔直电网拓扑对故障电流的影响机理分析[J]. 电力工程技术, 2022, 41(5): 94-102.
Zhang Yingmin, Zhang Wenxin, Li Baohong, et al.Influence mechanism of MMC-HVDC grid topology on fault current[J]. Electric Power Engineering Technology, 2022, 41(5): 94-102(in Chinese).
[12] Wang Q H, Yu Y X, Ahmed H O A, et al. Fault detection and classification in MMC-HVDC systems using learning methods[J]. Sensors, 2020, 20(16): 4438.
[13] 李伟杰, 郝亮亮, 叶雪辉, 等. 考虑放电耦合的MMC-HVDC系统直流线路单极接地故障电流的计算方法[J]. 南方电网技术, 2023, 17(9): 68-75.
Li Weijie, Hao Liangliang, Ye Xuehui, et al.Calculation method for pole-to-ground fault current of DC lines in MMC-HVDC system considering discharging coupling[J]. Southern Power System Technology, 2023, 17(9): 68-75(in Chinese).
[14] 杨贺雅, 邢纹硕, 向鑫, 等. 基于多元高斯分布异常检测模型的MMC子模块开路故障诊断方法[J]. 电工技术学报, 2023, 38(10): 2744-2756.
Yang Heya, Xing Wenshuo, Xiang Xin, et al.A sub-module open-circuit fault detection and location strategy for modular multilevel converters based on multivariate Gaussian distribution[J]. Transactions of China Electrotechnical Society, 2023, 38(10): 2744-2756(in Chinese).
[15] Ke L Z, Hu G Z, Yang Y Q, et al.Fault diagnosis for modular multilevel converter switching devices via multimodal attention fusion[J]. IEEE Access, 2023, 11: 135035-135048.
[16] Tang T, Wang J W, Yang T Y, et al.An improved prototypical network with L2 prototype correction for few-shot cross-domain fault diagnosis[J]. Measurement, 2023, 217: 113065.
[17] Wang C D, Yang J L, Zhang B Q.A fault diagnosis method using improved prototypical network and weighting similarity:Manhattan distance with insufficient noisy data[J]. Measurement, 2024, 226: 114171.
[18] Liang Y, Zhang J W, Shi Z, et al.A fault identification method of hybrid HVDC system based on wavelet packet energy spectrum and CNN[J]. Electronics, 2024, 13(14): 2788.
[19] Zhang Y Q, Yu Y, Yang G S.An ultra-fast MMC-HVDC fault location algorithm based on transient voltage features and regression neural network[J]. International Journal of Electrical Power & Energy Systems, 2024, 162: 110249.
[20] Zhai Z Y, Wang N, Lu S R, et al.A novel open circuit fault diagnosis for a modular multilevel converter with modal time-frequency diagram and FFT-CNN-BIGRU attention[J]. Machines, 2025, 13(6): 533.
[21] Kheireddine M, Merzouk I, Hafaifa A, et al.A novel fault diagnosis methodology in modular multilevel converters to enhance reliability and resilience[J]. Results in Engineering, 2025, 27: 105682.

基金资助

国家电网有限公司华东分部科技项目(52992425001L); 国家自然科学基金项目(52407079); 湖南省自然科学基金项目(2025JJ60372)

AI Summary AI Mindmap
PDF (2105KB)

0

访问

0

被引

详细

导航
相关文章

AI思维导图

/

〈 〉