矿井低压电网极易发生漏电故障，引发瓦斯煤尘爆炸以及人身触电事故，给工矿企业的供电安全带来了巨大挑战。通常情况下，漏电信号较为微弱，容易受到噪声污染、采集误差等因素影响，其特征难以提取，使得矿井电网选漏变得十分困难。因此，深入研究矿井电网选漏判据对于提高矿井电网供电安全有着重要的意义。

       本文以矿井低压电网为研究对象，建立了矿井电网单相漏电模型，首先从理论和仿真两方面对漏电机理进行了分析，得到零序全电流和零序全电压的变化规律。其次利用改进集合经验模态分解算法（EEMD）对漏电信号降噪，针对单一选漏判据的适用性问题，采用最小二乘法和信号距离法构造暂态选漏判据，采用零序有功法和五次谐波法构造稳态选漏判据，从而全面地提取漏电信号的暂态电阻电流、稳态电流电压等暂稳态特征。然后建立基于深度置信算法（DBN）的融合选漏判据模型以融合暂稳态特征，并利用改进遗传算法（IGA）优化DBN模型的各隐含层神经元个数，提升特征提取能力，得到基于IGA-DBN融合选漏判据的最优模型。最后在MATLAB/Simulink中建立660V矿井低压电网仿真模型，验证本文构造的暂态与稳态选漏判据的可行性；进一步使用矿井电网模拟实验平台，搭建系统验证该融合选漏模型的实用性。结果表明本文提出的IGA-DBN融合选漏判据模型的准确率高于DBN算法、GA-DBN和GA-BP等算法，且抗干扰性强。

       本文提出的融合选漏判据模型，可以快速准确地选出漏电线路，从而保证矿井电网的安全运行，是对矿井电网供电安全技术的完善与补充，具有一定的理论研究意义与工程应用价值。

       The low-voltage power grid in the mine is prone to leakage fault, which leads to gas, coal dust explosion and personal electric shock accidents, which brings great challenges to the power supply safety of industrial and mining enterprises. Usually, the leakage signal is relatively weak, which is easily affected by noise pollution, acquisition error and other factors. Its characteristics are difficult to extract, which makes it very difficult to select leakage in mine power grid. Therefore, the in-depth study of leakage selection criteria of mine power grid is of great significance to improve the power supply safety of mine power grid.

        In this paper, the mine low-voltage power grid is taken as the research object, and a single-phase leakage current model of the mine power grid is established. Firstly, the leakage mechanism is analyzed from the theory and simulation, and the variation laws of zero-sequence full current and zero-sequence full voltage are obtained. Secondly, the improved ensemble empirical mode decomposition (EEMD) algorithm is used to reduce the noise of leakage signal. Aiming at the applicability of single leakage selection criterion, the transient leakage selection criterion is constructed by using the least square method and signal distance method, and the steady-state leakage selection criterion is constructed by using the zero sequence power method and the fifth harmonic method, so as to comprehensively extract the transient resistance current, steady-state current and voltage of leakage signal. Then, a fusion leak selection criterion model based on depth confidence algorithm (DBN) is established to fuse transient and steady-state features, and the number of neurons in each hidden layer of DBN model is optimized by using improved genetic algorithm (IGA), so as to improve the feature extraction ability, and the optimal model based on IGA-DBN fusion leak selection criterion is obtained. Finally, the 660V mine low-voltage power network simulation model is established in MATLAB/Simulink to verify the feasibility of the transient and steady-state leakage selection criteria constructed in this paper; Further, the mine power grid simulation experiment platform is used to build a system to verify the practicability of the fusion leakage selection model. The results show that the accuracy of IGA-DBN fusion leak selection criterion model proposed in this paper is higher than DBN algorithm, GA-DBN, GA-BP algorithm, etc. And it has strong anti-interference.

        The fusion leakage selection criterion model proposed in this paper can quickly and accurately select the leakage line, so as to ensure the safe operation of mine power grid. It is the improvement and supplement of mine power supply safety technology, and has certain theoretical research significance and engineering application value.

[1] 曹建文. 矿井低压电网选择性漏电保护装置设计[J]. 煤矿机电, 2021, 42(3): 15-18.

[2] 赵建文, 侯媛彬. 工矿电网漏电保护[M]. 西安: 西安电子科技大学出版社, 2013.

[3] 国家安全生产监督管理总局. 煤炭安全规程[S]. 北京: 中国煤炭出版社, 2022.

[4] 安东亮, 陈涛, 李军, 等. 基于半波傅氏算法的小电流接地选线装置设计[J]. 电力系统保护与控制, 2020, 48(9): 157-163.

[5] 李科, 随晓娜, 张俊, 等. 矿井电网故障选线方法研究[J]. 工矿自动化, 2018, 44(5): 71-75.

[6] 王彦文, 卢丹. 基于零序电流差动量的矿井低压电网选择性漏电保护新方案[J]. 工矿自动化, 2014, 10(9): 53-56.

[7] Cheng W, Shi X, Wen J. Single Phase Grounding Fault Line Selection Method of Coal Mine Power Grid Based on Smooth Decomposition Modified Threshold Denoising and Waveform Diversity[C]//2021 5th International Conference on Power and Energy Engineering (ICPEE). IEEE, 2021: 14-21.

[8] 王清亮. 基于等效电导的矿井电网智能漏电保护方法[J]. 工矿自动化, 2020, 46(6): 59-64.

[9] 赵建文, 李科, 随晓娜, 等. 矿井电网漏电保护新方法[J]. 工矿自动化, 2016, 42(2): 47-51.

[10] 邹有明, 张根现, 刘士栋, 等. 工矿企业漏电保护技术[M]. 北京: 煤炭工业出版社, 2004.

[11] 王星星, 孙俞年, 佟磊, 等. 新型漏电保护装置在井下供电系统中的应用[J]. 现代矿业, 2020, 36(10): 152-154.

[12] 毛少军. ZBL-L型低压选择性漏电保护装置在王村煤矿的应用[J]. 陕西煤炭, 2014, 33(03): 126-127.

[13] Niu L, Wu G, Xu Z. Single-Phase Fault Line Selection in Distribution Network Based on Signal Injection Method[J]. IEEE Access, 2021, 9: 21567-21578.

[14] 熊婷婷, 曾祥君, 王媛媛, 等. 非有效接地电网单相接地故障选线技术综述[J]. 电气技术, 2013, 14(5): 1-6.

[15] 杨蓬. 煤矿集中式选择性漏电保护装置的研制[J]. 机电信息, 2019(23): 36-37.

[16] 高鹏, 莫兴洋, 徐晋勇, 等. 煤矿井下中性点不接地单相漏电故障选线方法[J]. 煤矿安全, 2015, 46(1): 85-88.

[17] 仲康. 海上多平台互联电力系统单相接地故障选线方法研究[D]. 杭州: 浙江大学, 2020.

[18] 吴乐鹏, 黄纯, 林达斌, 等. 基于暂态小波能量的小电流接地故障选线新方法[J]. 电力自动化设备, 2013, 33(5): 70-74.

[19] 陈星, 黄天啸, 吴翔宇, 等. 考虑系统故障响应轨迹的交直流混联电力系统暂态能量计算方法[J]. 电力自动化设备, 2020, 40(9): 102-111.

[20] 王旭强, 杨青, 张耀, 等. 基于多源信息融合的配电网故障选线新方法[J]. 智慧电力, 2019, 47(9): 97-103.

[21] 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.

[22] 贾清泉, 石磊磊, 王宁, 等. 基于证据理论和信息熵的消弧线圈接地电网融合选线方法[J]. 电工技术学报, 2012, 27(6): 191-197.

[23] 兰华, 朱锋. 基于模糊粗糙集和GA-BP神经网络的配电网故障选线方法[J]. 南方电网技术, 2013, 7(1): 90-94.

[24] Shu H C, An N, Yang B, et al. Single Pole-to-Ground Fault Analysis of MMC-HVDC Transmission Lines Based on Capacitive Fuzzy Identification Algorithm[J]. Energies, 2020, 13(2): 319.

[25] 赵峰, 尹德昌. 小波包与改进BP神经网络的配电网故障选线[J]. 自动化仪表, 2013, 34(9): 4-8.

[26] 王晓卫, 魏向向, 侯雅晓, 等. 基于小波去噪与改进RBF神经网络的小电流接地系统故障选线方法[J]. 工矿自动化, 2014, 40(04): 46-50.

[27] 彭湃, 周羽生, 高云龙, 等. GA优化LVQ网络的配电网接地故障选线方法[J]. 电力系统及其自动化学报, 2015, 27(12): 64-69.

[28] 王磊, 曹现峰, 骆玮. 基于GA优化T-S模糊神经网络的小电流接地故障选线新方法[J]. 电测与仪表, 2016, 53(17): 6-11.

[29] 翟二杰, 舒征宇, 汪俊, 等. 基于VMD-LSTM的小电流接地系统故障选线方法[J]. 电工电能新技术, 2021, 40(01): 70-80.

[30] 但扬清, 赵伟, 朱艳伟. 基于ABC-DNN的小电流接地故障选线方法[J]. 智慧电力, 2019, 47(4): 46~52.

[31] 王玉梅, 张家康. 基于卷积神经网络多判据融合的井下电网故障选线方法[J/OL]. 电源学报, 2022, 1-12.

[32] 陈霄. 基于改进深度置信网络的小电流接地系统单相接地故障选线研究[D]. 南京师范大学, 2021.

[33] 徐建军, 刘建宇, 闫丽梅. 基于原子稀疏分解的矿井电网故障选线方法研究[J]. 国外电子测量技术, 2018, 37(02): 90-94.

[34] 韦莉珊, 贾文超, 焦彦军. 基于5次谐波与导纳不对称度的配电网单相接地选线方法[J]. 电力系统保护与控制, 2020, 48(15): 77-83.

[35] 赵建文, 侯媛彬, 尹项根. 矿井电网单相漏电瞬时序网络模型[J]. 电网技术, 2011, 35(3): 119-123.

[36] 王清亮. 单相漏电的故障暂态特性[J]. 煤炭学报, 2010, 35(1): 160-164.

[37] 郭金龙, 吕景江, 王以磊. 矿井电网谐波治理及无功补偿的研究与应用[J]. 工矿自动化, 2012, 38(11): 64-66.

[38] 李洪涛, 任卫德. 基于计算机仿真的矿井专用电网谐振、谐波的抑制方法研究[J]. 煤炭技术, 2014, 33(2): 50-51.

[39] 杨勰颖, 孙曼, 贾峰. 含噪电能质量信号小波降噪的方法与改进[J]. 实验室研究与探索, 2017, 36(06): 9-12.

[40] 周登登, 刘志刚, 胡非, 等. 基于小波去噪和暂态电流能量分组比较的小电流接地选线新方法[J]. 电力系统保护与控制, 2010, 38(7): 22-28.

[41] 董丽梅, 舒勤, 夏书银. 基于遗传算法的EMD电力信号去噪[J]. 计算机仿真, 2014, 31(10): 123-127.

[42] 肖洒, 陈波, 沈道贤,等. 改进VMD和阈值算法在局部放电去噪中的应用[J]. 电子测量与仪器学报, 2021, 35(11): 206-214.

[43] Mariyappa N, Sengottuvel S, Patel R, et al. Denoising of Multichannel MCG Data by The Combination of EEMD and ICA and Its Effect on The Pseudo Current Density Maps[J]. Biomedical Signal Processing and Control, 2015, 18: 204-213.

[44] 赵世华, 巢亚锋, 孙利朋, 等. 基于小波变换与EEMD的绝缘子泄漏电流去噪方法研究[J]. 电瓷避雷器, 2019(6): 216-220.

[45] 宋美. 基于集合经验模态分解和小波收缩算法的自适应心电信号去噪问题研究[J]. 生物数学学报, 2015 (4): 629-638.

[46] Gaci S. A New Ensemble Empirical Mode Decomposition (EEMD) Denoising Method for Seismic Signals[J]. Energy Procedia, 2016, 97: 84-91.

[47] 陈仁祥, 汤宝平, 吕中亮. 基于相关系数的EEMD转子振动信号降噪方法[J]. 振动. 测试与诊断, 2012, 32(4): 543-546.

[48] Cheng X, Mao J, Li J, et al. An EEMD-SVD-LWT Algorithm for Denoising A Lidar Signal[J]. Measurement, 2021, 168: 108405.

[49] 汤涛, 黄纯, 江亚群, 等. 基于高低频段暂态信号相关分析的谐振接地故障选线方法[J]. 电力系统自动化, 2016, 40(16): 105-111.

[50] 李文宏, 陈景龙.基于最小二乘法的矿井单相漏电系统保护的研究[J]. 煤炭技术, 2018, 37(3): 258-260.

[51] Borden A R, Lesieutre B C. Variable Projection Method for Power System Modal Identification[J]. IEEE Transactions on Power Systems, 2014, 29(6): 2613-2620.

[52] 赵建文, 侯媛彬. 信号距离度模型及其波形复杂信号辨识[J]. 电子测量与仪器学报, 2013, 27(6): 485-491.

[53] 赵建文, 侯媛彬. 基于信号距离的矿井低压电网漏电快速识别[J]. 电工电能新技术, 2011, 30(01): 24-27.

[54] 庄伟, 牟龙华. 基于零序电流有功分量的配电网接地故障定位[J]. 同济大学学报(自然科学版), 2014, 42(3): 468-473.

[55] Heydt G T. Fifth Harmonic Spectral Voltage Components in EHV Power Transmission Systems[J]. IEEE Transactions on Power Delivery, 2021, 37 (1): 458-463.

[56] 戚焕兴, 殷林飞, 万俊, 等. 基于深度置信网络状态最优反馈的智能发电控制策略[J]. 电力建设, 2021, 42(10): 78-88.

[57] 刘冬兰, 孔德秋, 常英贤, 等. 基于受限玻尔兹曼机的电力信息系统多源日志综合特征提取[J]. 计算机系统应用, 2020, 29(11): 210-217.

[58] 垠飞, 王力. 融合D-S证据理论的DBN电路故障诊断算法[J]. 辽宁工程技术大学学报(自然科学版), 2021, 40(05): 448-453.

[59] 孔维宇. 基于GA-DBN的组合导航故障诊断技术研究[D]. 哈尔滨: 哈尔滨工业大学, 2021.

[60] Jiacheng L, Lei L. A hybrid Genetic Algorithm Based on Information Entropy and Game Theory[J]. IEEE Access, 2020, 8: 36602-36611.

[61] Nose-Filho K, Lotufo A D P, Minussi C R. Short-term Multinodal Load Forecasting Using A Modified General Regression Neural Network[J]. IEEE Trans on Power Systems, 2011, 26(4): 2862-2869.

[62] Cheng P, Chen D, Wang J. Clustering of The Body Shape of The Adult Male by Using Principal Component Analysis and Genetic Algorithm-BP Neural Network[J]. 2020, 24(17): 13219-13237.

[63] 王森, 程春田, 武新宇, 等. 自适应混沌整体退火遗传算法在水电站群优化调度中的应用[J]. 水力发电学报, 2014, 33(5): 63-70.

[64] Bakhache B, Ghazal J M, El Assad S. Improvement of The Security of Zigbee by A New Chaotic Algorithm[J]. IEEE Systems Journal, 2013, 8(4): 1024-1033.

[65] 杨从锐, 钱谦, 王锋, 等. 改进的自适应遗传算法在函数优化中的应用[J]. 计算机应用研究, 2018(4): 1042-1045.

[66] 毕惟红, 任红民, 吴庆标. 一种新的遗传算法最优保存策略[J]. 浙江大学学报:理学版, 2006, 33(1): 32-35.

[67] Coates A., Lee H, Ng A Y. An Analysis of Single-Layer Networks in Unsupervised Feature Learning[C]//Proceedings of The 14th International Conference on Artificial Intelligence And Statistics. Fort Lauderdale, USA: Microtome Publishing, 2011: 215-223.

[68] 陈霄, 居荣. 基于PSO-DBN的配电网单相接地故障选线方法[J]. 电工技术, 2021(5): 29-31+36.

[69] Wang S, Zhang J, Liu M, et al. Large-Signal Behavior Modeling of GaN P-HEMT Based on GA-ELM Neural Network[J]. Circuits, Systems, and Signal Processing, 2022, 41(4): 1834-1847.

[70] Zamaruiev V V. The Use of Kotelnikov-Nyquist-Shannon Sampling Theorem for Designing of Digital Control System for A Power converter[C]//2017 IEEE First Ukraine Conference on Electrical and Computer Engineering (UKRCON). IEEE, 2017: 522-527.

[71] Wang B, Gu X, Ma L, et al. Temperature Error Correction Based on BP Neural Network in Meteorological Wireless Sensor Network[J]. International Journal of Sensor Networks, 2017, 23(4): 265-278.

[72] Cheng P, Chen D, Wang J. Clustering of The Body Shape of The Adult Male by Using Principal Component Analysis and Genetic Algorithm–BP Neural Network[J]. Soft Computing, 2020, 24(17): 13219-13237.

[73] 嵇文路, 赵晓龙, 张明, 等. 基于小波包全频带分析和OS-ELM的小电流单相接地故障选线[J]. 哈尔滨理工大学学报, 2021, 26(02): 110-117.

[74] 徐丙垠, 薛永端, 冯光, 等. 配电网接地故障保护若干问题的探讨[J]. 电力系统自动化, 2019, 43(20): 1-7.

[75]邱进, 崔鑫, 田野, 等. 小电流接地配电线路弧光高阻接地故障电压特征分析[J]. 电力系统保护与控制, 2019, 47(16): 115-121.