随着我国经济的高速发展，社会用电需求持续攀升，配电网规模不断扩大。现代配电系统呈现出网络拓扑复杂化、线路类型多样化等特征，导致系统的故障发生率显著提高。因此，快速精准地判定故障性质并确定故障位置，对于缩短供电中断时长、降低线路巡检强度具有重要意义。由于电力系统故障分量具有显著的非线性特征，同时故障数据量庞大，传统诊断方法在特征提取和数据分析环节容易产生误差传递问题，诊断效果往往不尽人意，通过将机器学习算法与先进信号处理技术相结合，构建配电网故障智能诊断系统，可以实现故障线路选择、故障分类和故障测距的协同优化，从而显著提升诊断系统的精确度和稳定性。

当配电线路出现微弱故障时，各线路的电压电流信号会表现出显著的非线性和非平稳特征，并受到噪声干扰严重，这导致故障线路的准确辨识面临重大挑战。针对这一问题，本文在行波传输机理的基础上，提出基于自适应噪声完备集合经验模态分解（CEEMDAN）结合小波阈值去噪（WTD）的电压电流行波极性选线方法。首先通过CEEMDAN对含噪信号进行分解，利用皮尔逊相关系数（PCC）计算得到有效IMF分量，结合小波阈值去噪对其进行分解重构，从而消除信号中的噪声影响，突出有效特征。随后利用降噪后故障线路电压电流行波的极性差异进行故障选线，计算各线路行波模极大值得到其极性，从而对故障线路进行识别。最后针对高阻接地故障、过零点故障和不同故障类型进行适应性分析，证明了所提方法泛化性、适应性优秀，为下文故障类型识别及精确定位奠定坚实基础。

在确定故障线路后，需进一步判别故障具体类型，而现有的配电网故障类型识别方法依赖人工构建特征向量进行故障识别，存在主观性强、特征可解释性差等问题，导致对高相似度故障样本的辨识精度显著降低。针对这一问题，提出了一种基于卷积神经网络（CNN）与最小二乘向量机（LSSVM）融合的故障类型识别模型。基于上文所得故障线路，首先通过短时傅里叶变换（STFT）采集故障电压电流基波和2-4次谐波幅值作为故障特征，利用相关性分析验证特征参数的有效性，并基于验证后的特征数据集，采用LSSVM替代传统CNN的Softmax分类器，构建CNN-LSSVM混合故障类型识别模型，实现故障类型的智能识别。仿真实验证明所提方法能对各类型故障进行精确识别，识别准确率达到98%，并在并在不同拓扑结构、过渡电阻和噪声干扰下保持高辨识精度，有良好的泛化性与鲁棒性。

在已知故障线路和类型的基础上，计算故障点距测量点的精确距离，而目前单端行波测距技术仍受到行波波速确定困难，行波波头识别困难等难题的影响。针对这一问题，提出了一种基于雾凇优化算法（RIME）优化最小二乘向量机的行波故障测距方法。通过小波包变换计算故障电压电流反向行波的前三个模极大值的突变时间及其Lipschitz指数，将其作为RIME-LSSVM测距模型的输入样本，利用最小二乘向量机拟合样本特征与故障距离之间的关系，搭建基于最小二乘向量机的测距模型，得到故障距离。并采用雾凇优化算法对最小二乘向量机的惩罚系数和核函数参数进行优化，以提高其测距精度及泛化能力。仿真实验证明所提方法不受故障馈线的不同、过渡电阻、故障类型及过零点故障的影响，在高阻接地、过零点故障等微弱故障情况下仍保持较高测距精度。

With the rapid development of China's economy, the social demand for electricity continues to rise, leading to the continuous expansion of distribution network scale. Modern distribution systems exhibit characteristics such as complex network topology and diversified line types, resulting in a significant increase in system failure rates. Therefore, rapid and accurate determination of fault characteristics and location is of great significance for reducing power outage duration and minimizing line inspection intensity. Given that power system fault components exhibit pronounced nonlinear characteristics and the vast volume of fault data, traditional diagnostic methods often suffer from error propagation issues during feature extraction and data analysis, leading to unsatisfactory diagnostic outcomes. By integrating machine learning algorithms with advanced signal processing techniques, an intelligent fault diagnosis system for distribution networks can be constructed, enabling collaborative optimization of fault line selection, thereby significantly improving the accuracy and stability of the diagnostic system.

When weak faults occur in distribution lines, the voltage and current signals of each line exhibit significant nonlinear and non-stationary characteristics, accompanied by severe noise interference, posing major challenges to accurate fault line identification. To address this issue, this paper proposes a voltage and current traveling wave polarity-based line selection method combining Adaptive Noise Complete Ensemble Empirical Mode Decomposition (CEEMDAN) with Wavelet Threshold Denoising (WTD) based on the traveling wave transmission mechanism. First, the noisy signal is decomposed using CEEMDAN, and effective IMF components are obtained by calculating the Pearson Correlation Coefficient (PCC). Wavelet threshold denoising is then applied for decomposition and reconstruction to eliminate noise interference and highlight effective features. Subsequently, the polarity differences of voltage and current traveling waves in the fault line are utilized for fault line selection. The polarity of each line is determined by calculating the modulus maxima of the traveling waves, enabling fault line identification. Finally, adaptability analyses for high-impedance grounding faults, zero-crossing faults, and different fault types demonstrate the excellent generalization and adaptability of the proposed method, laying a solid foundation for subsequent fault type identification and precise localization.

After identifying the fault line, the specific fault type needs to be determined. Existing fault type identification methods for distribution networks rely on manually constructed feature vectors, which suffer from strong subjectivity and poor feature interpretability, leading to significantly reduced accuracy in identifying highly similar fault samples. To address this issue, this paper proposes a fault type identification model based on the fusion of Convolutional Neural Network (CNN) and Least Squares Support Vector Machine (LSSVM). Based on the identified fault line, the fundamental and 2nd-4th harmonic amplitudes of fault voltage and current are collected as fault features using Short-Time Fourier Transform (STFT). The effectiveness of the feature parameters is verified through correlation analysis. Using the validated feature dataset, LSSVM replaces the traditional Softmax classifier in CNN to construct a CNN-LSSVM hybrid fault type identification model for intelligent fault type recognition. Simulation experiments demonstrate that the proposed method achieves an identification accuracy of 98% for various fault types and maintains high recognition accuracy under different topologies.

With the fault line and type identified, the precise distance from the fault point to the measurement point needs to be calculated. However, existing single-terminal traveling wave fault location techniques still face challenges such as difficulty in determining traveling wave velocity and identifying waveheads. To address this issue, this paper proposes a traveling wave fault location method based on the Rime Optimization Algorithm (RIME)-optimized Least Squares Support Vector Machine (LSSVM). The arrival times of the first three modulus maxima of the reverse traveling waves of fault voltage and current and their Lipschitz exponents are calculated using wavelet packet transform and used as input samples for the RIME-LSSVM location model. The relationship between sample features and fault distance is fitted using LSSVM to construct a fault location model. The RIME algorithm is employed to optimize the penalty coefficient and kernel function parameters of LSSVM to improve location accuracy and generalization capability. Simulation results demonstrate that the proposed method is unaffected by fault feeder type, transition resistance, fault type, or zero-crossing faults, maintaining high location accuracy even under weak fault conditions such as high-impedance grounding and zero-crossing faults.

[1] 文劲宇, 周博, 魏利屾. 中国未来电力系统储电网初探[J]. 电力系统保护与控制, 2022, 50(07): 1-10.

[2] Hou S, Guo X, Research on fault location of distribution lines based on the standing wave principle[J]. Processes, 2021, 9(8): 1-20.

[3] 刘科研, 叶学顺, 李昭, 等. 基于多分辨率小波变换的配电网高阻接地故障检测方法[J]. 高电压技术, 2022: 1-12．

[4] Kim C, Bialek T, Awiylika J. An initial investigation for locating self-clearing faults in distribution systems[J]. IEEE Transactions on Smart Grid, 2013, 4(2): 1105-1112.

[5] Alamuti M, Nouri H, Ciric R M, et al. Intermittent fault location in distribution feeders[J]. IEEE Transactions on Power Delivery, 2012, 27(1): 96-103．

[6] 张玉玺, 王增平, 李振钊, 等. 基于特征频带暂态无功功率的配电网故障选线新方法[J]. 电力系统保护与控制, 2023, 51(1): 1-11．

[7] 吴佳享, 孙云莲, 陈楚昭. 微弱行波下多分支配电网故障定位[J]. 电机与控制学报, 2023, 27(05): 20-27.

[8] 刘亮, 曾祥君, 邓名高, 等. 基于线路归一和网络解耦的配网故障测距方法[J]. 仪器仪表学报, 2018, 39(01): 243-249.

[9] Zhang W, Wang D, Hou M Q. Travelling wave fault location approach for hybrid LCC-MMC-MTDC transmission line based on frequency modification algorithm[J]. International Journal of Electrical Power & Energy Systems, 2023, 147: 108-862.

[10] 姜博, 董新洲, 施慎行, 等. 自适应时频窗行波选线方法研究[J]. 中国电机工程学报, 2015, 35(24): 6387-6397.

[11] Chen J, Li H, Deng C, et al. Detection of single-phase to ground faults in low resistance grounded MV Systems[J]. IEEE Transactions on Power Delivery, 2021, 36(3): 1499–1508.

[12] 田书欣, 李昆鹏, 魏书荣, 等. 基于同步相量测量装置的配电网安全态势感知方法[J]. 中国电机工程学报, 2021, 41(02): 617-632.

[13] S. M. Chopdar, A. K. Koshti. Fault detection and classification in power system using artificial neural network. in: 2022 2nd International Conference on Intelligent Technologies (CONIT), Hubli, India, 2022: 1-6.

[14] 吴岩, 关石磊, 孟晓丽, 等. 综合多域特征及融合算法的配电网单相接地故障类型识别[J]. 高电压技术, 2023, 49(05): 2059-2067.

[15] T. Berghout, M. Benbouzid, S. M. Muyeen. Machine learning for cybersecurity in smart grids: A comprehensive review-based study on methods, solutions, and prospects. International Journal of Critical Infrastructure Protection, 2022, 38: 100547.

[16] Chen L, Tan M, Xu T. Research on the automation system of the intelligent analysis and evaluation platform for distribution network changes. in: 2022 IEEE Conference on Telecommunications, Optics and Computer Science (TOCS), Dalian, China, 2022: 1482-1486.

[17] 郭威. 配电线路故障类型辨识及故障选线定位方法的研究[D]. 北京: 华北电力大学(北京), 2022.

[18] Song G, Ma Z, Li G, et al. Phase current fault component based single-phase earth fault segment location in non-solidly earthed distribution networks[J]. International Transaction on Electrical Energy Systems, 2015, 25: 2713-2730.

[19] Chen R., Yin X et al. Computational fault time difference-based fault location method for branched power distribution networks[J]. IEEE Access, 2019, 7, :181972-181982.

[20] Lopes F V, Dantas K M, Silva K M, et al. Accurate two-terminal transmission line fault location using traveling waves[J]. IEEE Transaction on Power Delivery, 2018, 33(2): 873-880.

[21] 陈宇迪, 邓祥力. 基于行波模态分解的柔性直流配电网故障测距方法[J]. 上海电力大学学报, 2025, 41(01): 50-58.

[22] 李晔, 贾娜, 何佳伟, 等. 基于暂态极性比较的分布式光伏配电网快速方向判据新方法[J]. 中国电机工程学报, 1-14.

[23] Ye X, Hu Y, Shen J, et al. An adaptive optimized TVF-EMD based on a sparsity-impact measure index for bearing incipient fault diagnosis[J]. IEEE Transactions on Instrumentation and Measurement, 2021, 70: 1-11.

[24] Tong X, Wen H. A novel transmission line fault detection algorithm based on pilot impedance[J]. Electric Power Systems Research, 2020, 179: 1-9.

[25] 吴岩, 关石磊, 孟晓丽, 等. 综合多域特征及融合算法的配电网单相接地故障类型识别[J]. 高电压技术, 2023, 49(05): 2059-2067.

[26] 叶远波, 王吉文, 邵庆祝, 等. 配电网故障识别Transformer-联邦迁移学习算法设计[J/OL]. 电力系统及其自动化学报, 1-10[2025-04-20].

[27] Zhang S, Wang J, Liu H, et al. Prediction of energy photovoltaic power generation based on artificial intelligence algorithm[J]. Neural Computing Application, 2020, 33(5): 821-835.

[28] 熊晓东, 陈伟杰, 朱祥瑞, 等. 基于人工神经网络的配电网故障段识别和定位[J]. 电气开关, 2024, 62(04): 69-72+76.

[29] Chen K, Hu J, Zhang Y, et al. Fault location in power distribution systems via deep graph convolutional networks[J]. IEEE Journal on Selected Areas in Communications, 2020, 38(1): 119-131.

[30] 彭华, 王文超, 朱永利, 等. 基于LSTM神经网络的风电场集电线路单相接地智能测距[J]. 电力系统保护与控制, 2021, 49(16): 60-66．

[31] Guo M, Yang N, Chen W. Deep-learning-based fault classification using Hilbert-Huang transform and convolutional neural network in power distribution systems[J]. IEEE Sensors Journal, 2019, 19(16): 6905-6913.

[32] Belagoune S, Bali N, Bakdi A, et al. Deep learning through LSTM classification and regression for transmission line fault detection, diagnosis and location in large-scale multi-machine power systems[J]. Measurement, 2021, 177: 1-14.

[33] 朱方博, 张俊林, 王瑞驰, 等. 基于SA-SAE的配电网故障分类方法[J]. 电气自动化, 2023, 45(02): 100-102.

[34] 许可, 范馨月, 张恒荣. 基于图卷积网络的配电网故障定位及故障类型识别[J]. 实验技术与管理, 2023, 40(01): 26-30.

[35] Gopakumar P, Reddy M, Mohanta D. Adaptive fault identification and classification methodology for smart power grids using synchronous phasor angle measurements[J]. IET Generation, Transmission Distribution, 2015, 9(2): 133-145.

[36] 傅文进, 赵云龙. 基于灰色关联度的配电网故障自动识别方法[J]. 自动化应用, 2021, (06): 103-105.

[37] Dehghani M, Khooban M, Niknam T. Fast fault detection and classification based on a combination of wavelet singular entropy theory and fuzzy logic in distribution lines in the presence of distributed generations[J]. International Journal of Electrical Power and Energy System, 2016, 78: 455-462.

[38] V. H. Bui, W. C. Su. Real-time operation of distribution network: A deep reinforcement learning-based reconfiguration approach[J]. Sustainable Energy Technologies and Assessments, 2022, 50: 101841.

[39] S. Belagoune, N. Bali, A. Bakdi, B. Baadji, K. Atif. Deep learning through LSTM classification and regression for transmission line fault detection, diagnosis and location in large-scale multi-machine power systems[J]. Measurement, 2021, 177(3): 1-14.

[40] Sun J, Zhang L, Chen M, et al. Feeder selection method of single-phase high impedance grounding fault in SRGS[C]. 2021 6th International Conference on Smart Grid and Electrical Automation(ICSGEA). 2021, 112–116.

[41] 肖瑞超, 徐涵. 基于零序电流-电压波形相似度的小电阻接地系统高阻接地故障检测方法[J]. 机电工程技术, 2021, 50(12): 64–67.

[42] Hao Q, Tianyou L. High impedance fault line selection method for resonant grounding system based on wavelet packet analysis[C]//2018 China International Conference on Electricity Distribution (CICED). IEEE, 2018: 1174-1179.

[43] 魏科文, 张靖, 何宇, 等. 基于VMD和相关性聚类的谐振接地系统单相接地故障选线[J]. 电力系统保护与控制, 2021, 49(22): 105-113.

[44] 洪翠, 付宇泽, 郭谋发, 陈永往. 基于卷积深度置信网络的配电网故障分类方法[J]. 电力自动化设备, 2019, 39(11): 64-70

[45] Stefanidou-Voziki P, Sapountzoglou N, Raison B, et al. A review of fault location and classification methods in distribution grids[J]. Electric Power Systems Research, 2022, 209: 108031.

[46] J. Doria-García, C. Orozco-Henao, L.U. Iurinic, et al. High impedance fault localization: A comprehensive review[J]. Electric Power Systems Research, 2023, 214: 108892.

[47] Doria-Garcia J, Orozco-Henao C, Iurinic L U, et al. High impedance fault location: Generalized extension for ground faults[J]. International Journal of Electrical Power & Energy Systems, 2020, 114: 105387.

[48] 黄智鹏, 秦飞翔, 朱革兰, 等. 基于改进阻抗法的复杂有源配电网快速故障测距方法[J]. 自动化技术与应用, 2023, 42(03): 80-84.

[49] J. U. N. Nunes, A. S. Bretas, N. G. Bretas, et al. Distribution systems high impedance fault location: A spectral domain model considering parametric error processing[J]. International Journal of Electrical Power & Energy Systems, 2019, 109: 227–241.

[50] 何晓, 雷勇, 周聪聪, 等. 消除零模波速影响的配电网单端行波故障测距算法[J]. 电力系统保护与控制, 2016, 44(23): 39-45.

[51] 曾哲, 邓丰, 张振, 等. 基于VMD-WVD的故障行波全波形时-频分析方法[J]. 电力系统保护与控制, 2022, 50(07): 49-57.

[52] Deng F, Zeng X, Pan L. Research on multi-terminal traveling wave fault location method in complicated networks based on cloud computing platform[J]. Protection and Control of Modern Power Systems, 2017, 2(2): 199-210.

[53] 盛万兴, 刘科研, 李昭, 等. 新型配电系统形态演化与安全高效运行方法综述[J]. 高电压技术, 2024, 50(01): 1-18.

[54] 丁佳立, 王昕, 郑益慧等. 利用线路中点电流检测的改进单端行波故障测距方法[J]. 高电压技术, 2020, 46(05): 1814-1822.

[55] 邓志祥, 张帆, 潘建兵等. 基于零模-线模波速差的配电网单端测距新方法[J]. 电网与清洁能源, 2023, 39(01): 78-84.

[56] 侯辉. 基于行波时频特性的配电网单相接地故障测距方法研究[D]. 西安: 西安科技大学, 2019.

[57] Deng F, Li X R, Zeng X J. Single-ended traveling wave protection algorithm based on full waveform in the time and frequency domains[J]. IET Generation, Transmission & Distribution, 2018, 12(15): 3680-3691.

[58] 李振兴, 程宜兴, 吴李群, 等. 基于初始波头广域传输路径的行波定位单元优化配置[J]. 电力系统自动化, 2017, 41(18): 60-66.

[59] Jin W, Lu Y, Tang C. A novel double ended method for fault location based on travelling wave time-series[C]//2016 IEEE International Conference on Power System Technology (POWERCON). IEEE, 2016: 1-6.

[60] 林洪, 王寒, 张康伟, 等. 基于改进双端行波法的配电网故障定位方法研究[J]. 电工技术, 2023, (05): 88-90.

[61] Zhong L , Kai Z , Jun Z , et al. Improved double-ended traveling wave ranging technology for DC distribution cable lines[J]. Journal of Physics: Conference Series, 2023, 2495(1).

[62] 胡冰颖. 基于零模线模时差的配电网双端行波故障测距[D]. 淮南: 安徽理工大学, 2021.

[63] 亓臻康, 王浩宗, 董新洲, 等. 不依赖GNSS的输电线路双端行波故障测距[J]. 中国电机工程学报, 2024, 44(10): 3766-3777.

[64] 李练兵, 孙腾达, 曾四鸣, 等. 基于多端行波时差的配电网故障定位方法[J]. 电力系统保护与控制, 2022, 50(03): 140-147.

[65] R. Dashti, M. Daisy, H. Mirshekali, H. R. Shaker, M. H. Aliabadi. A survey of fault prediction and location methods in electrical energy distribution networks[J]. Measurement, 2021, 184: 109947.

[66] A. S. Bretas, C. Orozco-Henao, J. Marín-Quintero, O. D. Montoya, W. Gil-González, N. G. Bretas. Microgrids physics model-based fault location formulation: Analytic based distributed energy resources effect compensation[J]. Electric Power Systems Research, 2021, 195: 107178.

[67] Qiao J, Yin X, Wang Y, Xu W, L M. Tan. A multi-terminal traveling wave fault location method for active distribution network based on residual clustering[J]. International Journal of Electrical Power & Energy Systems, 2021, 131(3): 107070.

[68] 邓丰, 史鸿飞, 冯思旭等. CNN-LSTM全景故障特征挖掘的配电网单端定位方法[J]. 中国电机工程报, 2023, 43(S1): 114-126.

[69] Mirzaei M, Vahidi B, Hosseinian S H, Accurate fault location and faulted section determination based on deep learning for a parallel-compensated three-terminal transmission line[J]. JET Generation, Transmission & Distribution, 2019, 13(13): 2770-2778.

[70] 钟建伟, 朱涧枫, 黄秀超, 等. 基于双态二进制粒子群优化算法的配电网故障定位[J]. 电力系统及其自动化学报, 2019, 31(03): 29–34.

[71] 林婷婷, 李玥, 高兴, 万玲. 基于改进短时傅里叶变换的磁共振随机噪声消减方法. 物理学报[J], 2021, 70(16): 134-146.

[72] Su H, Zhao D, Heidari A A, et al. RIME: A physics-based optimization[J]. Neurocomputing, 2023, 532: 183-214.

[73] Zhang C, Song G, Wang T, et al. Single-ended traveling wave fault location method in DC transmission line based on wave front information[J]. IEEE Transactions on Power Delivery, 2019, 34(5): 2028-2038.

[74] 李俊曹, 伍川辉, 李恒奎. 基于复Morlet小波的列车走行部轴承故障监测系统[J]. 仪表技术与传感器, 2022(08): 63-68.

[75] Chen Y , Li Y , Zhao Y . Sub-pixel detection algorithm based on cubic B-spline curve and multi-scale adaptive wavelet transform[J]. Optik-International Journal for Light and Election Optics, 2016, 127(1): 11-14.

[76] Ehsan A, Yang Q. State-of-the-art techniques for modelling of uncertainties in active distribution network planning: A review[J]. Applied energy, 2019, 239: 1509-1523.

[77] Wang X, Gao J, Wei X, et al. Single line to ground fault detection in a non-effectively grounded distribution network[J]. IEEE Transactions on Power Delivery, 2018, 33(6): 3173-3186.