图学习与零序分量相结合的风电场集电线单相接地故障定位

doi:10.19595/j.cnki.1000-6753.tces.221031

摘要
图/表
参考文献
相关文章 (15)

全文: PDF (598 KB)
输出: BibTeX | EndNote (RIS)

摘要

风电场集电线路短、连接风机多且风机间距小,传统行波测距等故障定位方法难以适用。为此提出一种图学习与零序分量相结合的新型层次化单相接地故障定位方案。首先,计及集电线拓扑信息构建图数据,应用图卷积神经网络进行图分类建模判断集电线故障区域;为克服样本不平衡的影响,结合迁移学习进行模型串联,并引入代价敏感机制后处理分类结果。然后,利用风电场接地方式与箱变接线特点,基于双端零序分量推导风机无关的故障区域测距公式。所提方案可跨越风机实现定位,并能克服数据不同步影响,所需测点少,适用于含分支集电线。仿真表明,图卷积神经网络相对传统深度网络判断正确率更高,模型串联结合代价敏感机制可有效克服样本不平衡影响;故障区域判断与测距效果不受故障位置、过渡电阻、风速影响。

	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
	丁嘉
	朱永利

关键词 ：风电场, 集电线, 单相接地故障, 故障定位, 图卷积神经网络, 零序分量, 样本不平衡, 数据不同步

Abstract：

In wind farms, collecting lines are short with a large number of wind turbines connected to it, and fault characteristics of wind turbines are different from those of traditional power sources, so dense connection of special power sources is the typical characteristic of collecting lines. The elimination of influence of wind turbines, applicability of collecting lines with line branches and reduction of measurement points are key challenges faced by the fault localization for collecting lines. Given these above challenges, it is difficult for the fault localization method in transmission and distribution networks to apply in wind farms, and the existing fault localization method of collecting lines cannot effectively address all these challenges. Therefore, a new hierarchical fault localization scheme for single-phase-to-ground faults on collecting lines combining graph learning and zero-sequence components is proposed, which can achieve precise fault localization across wind turbines with the applicability of collecting lines containing line branches considered and a few measurement points used.
Firstly, the graph data is constructed according to electrical information at measurement points and collecting line topology, and GCN (graph convolution neural network) is used as base-models for graph classification modeling to obtain the identification model of faulty areas which consists of a primary classifier and a secondary classifier; in the process of identification model construction, in order to overcome the impact of sample imbalance, multiple classifier models are connected in series with transfer learning combined, and cost sensitive learning based on Bayes minimum risk is also introduced to post-process classification results. Then, considering the neutral-point grounding type of wind farms and the winding connection characteristic of box-type transformers, a wind-turbine-independent fault distance calculation formula is derived based on zero-sequence components, and thus the fault point in faulty areas can be obtained by zero-sequence components. Besides, a misidentification processing mechanism for faulty areas is also proposed to correct the misidentification which occurs rarely. This above proposed fault localization scheme does not require strict data synchronization, can take into account both of structural and attribute information of collecting lines to identify the faulty area with limited fault data provided by a few measurement points, and can overcome influence of wind turbines to achieve the fault distance calculation.
Verification results show that with the graph data constructed, faulty areas can be effectively identified by GCN, and the identification accuracy of all faulty areas reaches over 99 %, which is higher than traditional DNN (deep neural network); besides, in the process of dealing with sample imbalance, model series connection increases the total average accuracy of branch faults by 12 %, and further this above accuracy achieves another 6.1 % improvement with cost sensitive learning introduced, which verifies the feasibility of the sample imbalance treatment strategy in this paper. The fault distance calculation method proposed for faulty areas can overcome the influence of fault locations, fault resistance, wind speed and data asynchronization, and errors of fault distance is no more than 150 m. In addition, the proposed misidentification processing mechanism is also proved to be effective and efficient in correcting misidentification of faulty areas.

Key words： Wind farm collecting line single-phase-to-ground fault fault localization graph convolutional neural network zero-sequence component sample imbalance asynchronous data

收稿日期: 2022-06-06

PACS:

TM76

基金资助:

国家自然科学基金（51677072）和中国国电集团公司科技项目（GDDL-KJ-2017-02）资助

通讯作者: 丁嘉男,1997年生,博士研究生,研究方向为电力线故障定位。E-mail：jia_ding_0132@163.com

作者简介: 朱永利男,1963年生,教授,博士生导师,研究方向为电力设备故障诊断、电力设备大数据分析与智能电网等。E-mail：yonglipw@163.com

引用本文:

丁嘉, 朱永利. 图学习与零序分量相结合的风电场集电线单相接地故障定位[J]. 电工技术学报, 0, (): 230608-230608. Ding Jia, Zhu Yongli. A Fault Localization Scheme for Single-Phase-to-Ground Faults on Collecting Lines in Wind Farms Combining Graph Learning and Zero-Sequence Components. Transactions of China Electrotechnical Society, 0, (): 230608-230608.

链接本文:

https://dgjsxb.ces-transaction.com/CN/10.19595/j.cnki.1000-6753.tces.221031 https://dgjsxb.ces-transaction.com/CN/Y0/V/I/230608

[1] 王恰. 中国风电产业40年发展成就与展望[J]. 中国能源, 2020, 42(9): 28-32, 9.
Wang Qia.Achievements and prospects of China's wind power industry development for 40 years[J]. Energy of China, 2020, 42(9): 28-32, 9.
[2] 王宾, 任萱. 中性点经小电阻接地风电场集电线路单相接地故障测距研究[J]. 中国电机工程学报, 2021, 41(6): 2136-2144.
Wang Bin, Ren Xuan.Single-line-to-ground fault location in wind farm collection line with neutral point grounding with resistor[J]. Proceedings of the CSEE, 2021, 41(6): 2136-2144.
[3] 刘鑫, 滕欢, 梁梦可, 等. 基于电流偏差2-范数的有源配电网故障距离定位[J]. 电工技术学报, 2019, 34(增刊2): 720-728.
Liu Xin, Teng Huan, Liang Mengke, et al.Fault location of active distribution network based on current deviation 2- norm[J]. Transactions of China Electrotechnical Society, 2019, 34(S2): 720-728.
[4] 王守鹏, 赵冬梅, 商立群, 等. 基于线路分段参数的非全程同塔双回线故障定位算法[J]. 电工技术学报, 2017, 32(20): 261-270.
Wang Shoupeng, Zhao Dongmei, Shang Liqun, et al.Fault location algorithm for non-complete double-circuit lines on the same tower based on line segmentation parameters[J]. Transactions of China Electrotechnical Society, 2017, 32(20): 261-270.
[5] 李振钊, 王增平, 张玉玺, 等. 基于升维线性规划的主动配电网故障区段定位方法[J]. 电力系统自动化, 2021, 45(24): 122-132.
Li Zhenzhao, Wang Zengping, Zhang Yuxi, et al.Fault location method of active distribution network based on ascending dimension linear programming[J]. Automation of Electric Power Systems, 2021, 45(24): 122-132.
[6] 王小君, 任欣玉, 和敬涵, 等. 基于μPMU相量信息的配电网络故障测距方法[J]. 电网技术, 2019, 43(3): 810-817.
Wang Xiaojun, Ren Xinyu, He Jinghan, et al.Distribution network fault location based on μPMU information[J]. Power System Technology, 2019, 43(3): 810-817.
[7] 张健磊, 高湛军, 陈明, 等. 考虑复故障的有源配电网故障定位方法[J]. 电工技术学报, 2021, 36(11): 2265-2276.
Zhang Jianlei, Gao Zhanjun, Chen Ming, et al.Fault location method for active distribution networks considering combination faults[J]. Transactions of China Electrotechnical Society, 2021, 36(11): 2265-2276.
[8] 邓丰, 徐帆, 曾哲, 等. 基于多源暂态信息融合的单端故障定位方法[J]. 电工技术学报, 2022, 37(13): 3201-3212.
Deng Feng, Xu Fan, Zeng Zhe, et al.Single-ended fault location method based on multi-source transient information fusion[J]. Transactions of China Electrotechnical Society, 2022, 37(13): 3201-3212.
[9] 于华楠, 马聪聪, 王鹤. 基于压缩感知估计行波自然频率的输电线路故障定位方法研究[J]. 电工技术学报, 2017, 32(23): 140-148.
Yu Huanan, Ma Congcong, Wang He.Transmission line fault location method based on compressed sensing estimation of traveling wave natural frequencies[J]. Transactions of China Electrotechnical Society, 2017, 32(23): 140-148.
[10] 邓丰, 梅龙军, 唐欣, 等. 基于时频域行波全景波形的配电网故障选线方法[J]. 电工技术学报, 2021, 36(13): 2861-2870.
Deng Feng, Mei Longjun, Tang Xin, et al.Faulty line selection method of distribution network based on time-frequency traveling wave panoramic waveform[J]. Transactions of China Electrotechnical Society, 2021, 36(13): 2861-2870.
[11] 胡明峰, 刘洋, 华斌. 量测装置优化配置下的分布式配电网故障定位方法[J]. 电网技术, 2021, 45(7): 2616-2622.
Hu Mingfeng, Liu Yang, Hua Bin.Distributed fault location for distribution networks under optimal configuration of measuring devices[J]. Power System Technology, 2021, 45(7): 2616-2622.
[12] 张大海, 张晓炜, 孙浩, 等. 基于卷积神经网络的交直流输电系统故障诊断[J]. 电力系统自动化, 2022, 46(5): 132-145.
Zhang Dahai, Zhang Xiaowei, Sun Hao, et al.Fault diagnosis for AC/DC transmission system based on convolutional neural network[J]. Automation of Electric Power Systems, 2022, 46(5): 132-145.
[13] 喻锟, 胥鹏博, 曾祥君, 等. 基于模糊测度融合诊断的配电网接地故障选线[J]. 电工技术学报, 2022, 37(3): 623-633.
Yu Kun, Xu Pengbo, Zeng Xiangjun, et al.Grounding fault line selection of distribution networks based on fuzzy measures integrated diagnosis[J]. Transactions of China Electrotechnical Society, 2022, 37(3): 623-633.
[14] 王浩, 杨东升, 周博文, 等. 基于并联卷积神经网络的多端直流输电线路故障诊断[J]. 电力系统自动化, 2020, 44(12): 84-92.
Wang Hao, Yang Dongsheng, Zhou Bowen, et al.Fault diagnosis of multi-terminal HVDC transmission line based on parallel convolutional neural network[J]. Automation of Electric Power Systems, 2020, 44(12): 84-92.
[15] 邢超, 高敬业, 毕贵红, 等. 基于集成神经网络的特高压直流输电线路初始电压行波小波变换模极大值比单端测距方法[J]. 电力自动化设备, 2022, 42(11): 128-134.
Xing Chao, Gao Jingye, Bi Guihong, et al.Single-ended location method based on integrated neural network for modulus maximum ratio of traveling wave wavelet transform of initial voltage of UHVDC transmission lines[J]. Electric Power Automation Equipment, 2022, 42(11): 128-134.
[16] 彭华, 朱永利, 袁胜辉. 风电场集电线路单相接地故障组合测距[J]. 仪器仪表学报, 2020, 41(9): 88-97.
Peng Hua, Zhu Yongli, Yuan Shenghui.Combined fault location for single-phase grounding of wind farm collection line[J]. Chinese Journal of Scientific Instrument, 2020, 41(9): 88-97.
[17] 张科, 朱永利, 郑艳艳, 等. 风电场输电线路单相接地故障定位研究[J]. 太阳能学报, 2020, 41(5): 114-120.
Zhang Ke, Zhu Yongli, Zheng Yanyan, et al.Fault location of single-phase earth in transmission lines of wind farm[J]. Acta Energiae Solaris Sinica, 2020, 41(5): 114-120.
[18] 彭华, 朱永利. 基于apFFT频谱校正和XGBoost的风电场集电线路单相接地故障测距[J]. 电工技术学报, 2020, 35(23): 4931-4939.
Peng Hua, Zhu Yongli.Single phase grounding fault location for power lines of wind farm based on apFFT spectrum correction and XGBoost algorithm[J]. Transactions of China Electrotechnical Society, 2020, 35(23): 4931-4939.
[19] Hoerauf R.Considerations in wind farm grounding designs[J]. IEEE Transactions on Industry Applications, 2014, 50(2): 1348-1355.
[20] 马伟, 裘愉涛, 丁冬, 等. 基于阻尼最小二乘法的单相接地故障定位方案[J]. 电网技术, 2018, 42(9): 3049-3054.
Ma Wei, Qiu Yutao, Ding Dong, et al.Single-phase grounding fault location scheme based on levenberg-marquarat algorithm[J]. Power System Technology, 2018, 42(9): 3049-3054.
[21] 谢潇磊, 刘亚东, 孙鹏, 等. 新型配电网线路P MU装置的研制[J]. 电力系统自动化, 2016, 40(12): 15-20, 52.
Xie Xiaolei, Liu Yadong, Sun Peng, et al.Development of novel PMU device for distribution network lines[J]. Automation of Electric Power Systems, 2016, 40(12): 15-20, 52.
[22] 袁冰, 王宾, 陆元园, 等. 风电场并网线路单相接地故障单端测距误差特性分析[J]. 电力系统保护与控制, 2016, 44(19): 63-69.
Yuan Bing, Wang Bin, Lu Yuanyuan, et al.Error analysis of single-end fault location for single-line-to-ground fault in transmission line with wind farm connection[J]. Power System Protection and Control, 2016, 44(19): 63-69.
[23] 王晨清, 宋国兵, 迟永宁, 等. 风电系统故障特征分析[J]. 电力系统自动化, 2015, 39(21): 52-58.
Wang Chenqing, Song Guobing, Chi Yongning, et al.Fault characteristics analysis of wind power system[J]. Automation of Electric Power Systems, 2015, 39(21): 52-58.
[24] Zhang Dingcheng, Stewart E, Entezami M, et al.Intelligent acoustic-based fault diagnosis of roller bearings using a deep graph convolutional network[J]. Measurement, 2020, 156: 107585.
[25] 张翼, 朱永利. 图信号与图卷积网络相结合的局部放电模式识别方法[J]. 中国电机工程学报, 2021, 41(18): 6472-6481.
Zhang Yi, Zhu Yongli.A partial discharge pattern recognition method combining graph signal and graph convolutional network[J]. Proceedings of the CSEE, 2021, 41(18): 6472-6481.
[26] Lee J, Lee I, Kang J.Self-attention graph pooling[C]//Proceedings of the 36th International Conference on Machine Learning, Long Beach, America, 2019.
[27] 陈剑, 杜文娟, 王海风. 基于对抗式迁移学习的含柔性高压直流输电的风电系统次同步振荡源定位[J]. 电工技术学报, 2021, 36(22): 4703-4715.
Chen Jian, Du Wenjuan, Wang Haifeng.Location method of subsynchronous oscillation source in wind power system with VSC-HVDC based on adversarial transfer learning[J]. Transactions of China Electrotechnical Society, 2021, 36(22): 4703-4715.
[28] Correa Bahnsen A, Aouada D, Ottersten B.Example-dependent cost-sensitive decision trees[J]. Expert Systems With Applications, 2015, 42(19): 6609-6619.
[29] Bahnsen A C, Stojanovic A, Aouada D, et al.Cost sensitive credit card fraud detection using Bayes minimum risk[C]//2013 12th International Conference on Machine Learning and Applications, Miami, FL, USA, 2014: 333-338.
[30] 曾德辉, 王钢, 李海锋, 等. 小电阻接地配电网多回线故障分析与自适应零序电流保护[J]. 电力自动化设备, 2019, 39(5): 45-52.
Zeng Dehui, Wang Gang, Li Haifeng, et al.Fault analysis of multi-feeder grounding fault and self-adaptive zero-sequence current protection scheme for low-resistance grounding distribution network[J]. Electric Power Automation Equipment, 2019, 39(5): 45-52.
[31] 陈福锋, 赵谦, 兰金波, 等. 风电场汇流线路同步采样及差动保护[J]. 电力系统自动化, 2013, 37(14): 19-24.
Chen Fufeng, Zhao Qian, Lan Jinbo, et al.Synchronous sampling and differential protections of collector lines in wind farms[J]. Automation of Electric Power Systems, 2013, 37(14): 19-24.
[32] Wang Bin, Ni Jiang, Geng Jianzhao, et al.Arc flash fault detection in wind farm collection feeders based on current waveform analysis[J]. Journal of Modern Power Systems and Clean Energy, 2017, 5(2): 211-219.