基于变分模态分解和改进灰狼算法优化深度置信网络的自动转换开关故障识别

doi:10.19595/j.cnki.1000-6753.tces.222143

摘要
图/表
参考文献
相关文章 (13)

全文: PDF (829 KB)
输出: BibTeX | EndNote (RIS)

摘要

自动转换开关(ATSE)是保证系统的连续供电的设备,对其进行健康监测和故障诊断对系统的稳定运行具有重要意义。为了实现对ATSE的非侵入式故障识别,本文提出一种基于电流信号变分模态分解(VMD)的特征提取和改进灰狼算法(IGWO)优化深度置信网络(DBN)相结合的故障诊断方法。该方法首先利用样本熵确定VMD分解次数并对故障电流进行分解;其次对分解后得到的本征模态函数进行小波包能量的提取,并利用IGWO对DBN网络结构参数进行优化;最后通过DBN将电流能量特征与ATSE的故障类型建立起映射关系从而完成最终的故障识别。所提IGWO采用了分段调节与非线性递减的衰减因子相结合的策略,以平衡算法全局搜索和局部搜索能力;并采用莱维飞行更新探狼的移动位置,来避免算法陷入早熟收敛。实验结果表明该算法不仅能显著提高前期对参数寻优的训练速度,后续泛化实验的故障分类准确率也有着98.78%的良好表现。

	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
	刘帼巾
	刘达明
	缪建华
	杨雨泽
	王乐康
	刘琦

关键词 ：优化灰狼算法, 深度置信网络, 自动转换开关, 故障识别

Abstract：

Automatic Transfer Switching Equipment (ATSE) is a device that ensures the continuous power supply of the system. Fault diagnosis of ATSE is of great significance for the continuous operation of the system. The current data of solenoid coils can be used to identify the electrical and mechanical faults of ATSE. However, the noise contained in current and the network structure parameters of intelligent fault diagnosis method are difficult to confirm. To overcome the above problems, an ATSE fault diagnosis method based on variational mode decomposition (VMD) feature extraction and optimized depth belief network (DBN) is proposed. In this method, Wavelet Packet Energy, (WPE) is extracted as feature vector from the mode components decomposed by VDM and the network structure parameters of DBN are set by improved grey wolf algorithm (IGWO) to complete ATSE fault identification.
Firstly, the fault current signal is decomposed by VMD and its sample entropies are obtained for each mode component after decomposition. The decomposition number corresponding to the lowest value of sample entropy is taken as the final decomposition number of VMD. Secondly, principal component analysis is used to select the WPE of each mode component after decomposition and form the final feature vector space input to DBN. Meanwhile, for avoiding premature convergence and local optimum problems during fault classification, IGWO is used to optimize the network structure parameters of DBN to accurately recognize various faults of ATSE.
ATSE fault simulation experiments are designed and carried out in this paper. After determining the decomposition frequency of VMD, perform modal decomposition on the coil current of the simulated fault experiment. The superiority of VMD decomposition is demonstrated by the absence of modal aliasing in the intrinsic mode functions at different frequencies after decomposition. The feature vectors are input into the optimized DBN algorithm for fault classification experiments, and the experimental results are as follows: (1) The difference of classification accuracy between the optimized DBN and the training set is only 0.4% which indicates that there is no over-fitting problem in the modified model. (2) The optimized DBN achieves a classification accuracy of 98.78% for four common faults. Next, the comparison between the proposed method and the non-optimized methods shows that, the proposed method has the highest accuracy and the best stability in fault diagnosis of ATSE.
The following conclusions can be drawn from the experimental analyses: (1) After determining the decomposition times based on the sample entropy, the VMD can effectively avoid the mode overlap, thus completing the extraction of fault features from ATSE current signals. (2) Compared with shallow neural network, DBN has powerful mapping ability and can accurately characterize complex mapping relationship between original current signal and ATSE fault type. (3) After using IGWO to optimize the network structure parameters of DBN, the optimized DBN has no over-fitting phenomena compared with other comparison algorithms and improves the classification accuracy of ATSE faults.

Key words： improved wolf optimizer deep belief network automatic transfer switching equipment fault diagnosis

PACS:

TM564.3

基金资助:

河北省自然科学基金项目（项目编号：E2020202221）和河北省省级科技计划资助项目（项目编号：215676146H和225676163GH)资助

通讯作者: 刘达明男,1999年生,硕士研究生,研究方向为电器可靠性及检测技术。E-mail：liudaming2021@126.com

作者简介: 刘帼巾女,1972年生,教授,研究生导师,研究方向为电器设备可靠性。E-mail：liuguojin72@163.com

引用本文:

刘帼巾, 刘达明, 缪建华, 杨雨泽, 王乐康, 刘琦. 基于变分模态分解和改进灰狼算法优化深度置信网络的自动转换开关故障识别[J]. 电工技术学报, 0, (): 20235421-20235421. Liu Guojin, Liu Daming, Miao Jianhua, Yang Yuze, Wang Lekang, Liu Qi. Fault Identification of Automatic Transfer Switching Equipment Based on VMD-WPE and IGWO Optimized DBN. Transactions of China Electrotechnical Society, 0, (): 20235421-20235421.

链接本文:

https://dgjsxb.ces-transaction.com/CN/10.19595/j.cnki.1000-6753.tces.222143 https://dgjsxb.ces-transaction.com/CN/Y0/V/I/20235421

[1] 马星光,王晓一,黄锐,等.基于复合特征的GIS振动信号特征提取方法[J].高压电器,2021,57(12):83-89.
Ma Xingguang, Wang Xiaoyi, Huang Rui et al. Extraction Method of GIS Vibration Signal Characteristic Based on Composite Feature[J]. High Voltage Apparatus, 2021,57(12):83-89.
[2] 赵莉华,吴迅,王泽龙,等. 基于线圈电流的断路器操作机构仿真研究[J].电测与仪表,2020,57(12):9-13.
Zhao Lihua, Wu Xun, Wang Zelong et al. Simulation research on circuit breaker operating mechanism based on coil current[J]. Electrical Measurement ＆ Instrumentation2020,57(12):9-13.
[3] 李宁,何复兴,马文涛,等.基于经验模态分解的门控循环单元神经网络的锂离子电池荷电状态估计[J].电工技术学报,2022,37(17):4528-4536.
Li Ning, He Fuxing Ma Wentao, et al. State-of-Charge Estimation of Lithium-Ion Battery Based on Gated Recurrent Unit Using Empirical Mode Decomposition[J]. Transactions of China Electrotechnical Society, 2022,37(17):4528-4536.
[4] 李兵,韩睿,何怡刚,等.改进随机森林算法在电机轴承故障诊断中的应用[J].中国电机工程学报,2020,40(04):1310-1319+1422.
Li Bing, Han Rui, He Yigang. Applications of the Improved Random Forest Algorithm in Fault Diagnosis of Motor Bearings[J]. Proceedings of the CSEE, 2020,40(04):1310-1319+1422.
[5] 崔芮华,李泽,佟德栓.基于三维熵距和熵空间的航空电弧故障检测与分类技术[J].电工技术学报,2021,36(04):869-880.
Cui Ruihua, Li Ze, Tong Deshuan.Arc Fault Detection and Classification Based on Three-Dimensional Entropy Distance and Entropy Space in Aviation Power System[J]. Transactions of China Electrotechnical Society, 2021,36(04):869-880.
[6] 蒋海峰,张曼,赵斌炎, 等.基于改进Hilbert-Huang变换的电网故障诊断[J].电工技术学报, 2019,34(S1):336-342+351.
Jiang Haifeng Zhang Man Zhao Binyanet al. Fault Diagnosis of Power Grid Based on Improved Hilbert-Huang Transform[J]. Transactions of China Electrotechnical Society, 2019,34(S1):336-342+351.
[7] Guo Wei, Tse P W.A novel signal compression method based on optimal ensemble empirical mode decomposition for bearing vibration signals[J]. Journal of Sound and Vibration, 2013, 332(2): 423-441.
[8] 陈勇, 梁洪, 王成栋, 等. 基于改进小波包变换和信号融合的永磁同步电机匝间短路故障检测[J].电工技术学报,2020,35(S1):228-234.
Chen Yong, Liang Hong, Wang Chengdong et al. Detection of Stator Inter-Turn Short-Circuit Fault in PMSM Based on Improved Wavelet Packet Transform and Signal Fusion[J]. Transactions of China Electrotechnical Society, 2020,35(S1):228-234.
[9] Wen Jin, Wen Hutang, Tong Qin, et al.Fault diagnosis of high-voltage circuit breakers using wavelet packet technique and support vector machine[J]. CIRE-Open Access Proceedings Journal, 2017, 2017(1): 170-174.
[10] 李睿彧, 刘飞, 梁霖, 等. 基于参数优化变分模态分解的交流变频电机转子断条故障识别方法[J].电工技术学报, 2021,36(18):3922-3933.
Li Ruiyu, Liu Fei, Liang Lin, et al.Fault Identification of Broken Rotor Bars for the Variable Frequency AC Motor Based on Parameter Optimized Variational Mode Decomposition[J]. Transactions of China Electrotechnical Society, 2021,36(18):3922-3933.
[11] 庄瑞锋,陈丽安,肖梁贤.基于变分模态分解的高压断路器合分闸线圈电流特征值提取[J].高压电器,2022,58(04):145-150+157.
Zhuang Ruifeng, Cheng Lian, Xiao Liangxian. Feature Value Extraction of High Voltage Circuit Breaker’s Opening Coil Current Based onVariational Modal Decomposition [J]. High Voltage Apparatus, 2022,58(04):145-150+157.
[12] Long Wen, Xing Yuli, Liang Gao, et al.A new convolutional neural network-based data-driven fault diagnosis method[J]. IEEE Transactions on Industrial Electronics, 2018, 65(7): 5990-5998.
[13] 李世晓,杜锦华,龙云.基于一维卷积神经网络的机电作动器故障诊断[J].电工技术学报,2022,37(S1):62-73.
Li Shixiao Du Jinhua Yun Long. Fault Diagnosis of Electromechanical Actuators Based on One-Dimensional Convolutional Neural Network[J]. Transactions of China Electrotechnical Society, 2022,37(S1):62-73.
[14] 徐奇伟,黄宏,张雪锋等.基于改进区域全卷积网络的高压引线接头红外图像特征分析的在线故障诊断方法[J].电工技术学报,2021,36(07):1380-1388.
Xu Qiwei, Huang Hong, Zhang Xuefeng, et al.Online Fault Diagnosis Method for Infrared Image Feature Analysis of High-Voltage Lead Connectors Based on Improved R-FCN[J]. Transactions of China Electrotechnical Society, 2021,36(07):1380-1388.
[15] 张婷, 王海淇, 张认成, 等. 基于自归一化神经网络的电弧故障检测方法[J].仪器仪表学报,2021,42(03):141-149.
Zhang Ting, Wang Haiqi, Zhang Rencheng et al. An arc fault detection method based on the self-normalized convolutional neural network[J]. Chinese Journal of Scientific Instrument,2021,42(03):141-149.
[16] 刘甜甜, 朱熀秋. 基于改进SPSO算法优化LS-SVM的六极径向混合磁轴承转子位移自检测技术[J]. 中国电机工程学报,2020,40(13):4319-4329.
Liu Tiantian, Zhu Huangqiu.Self-sensing Modeling of Rotor Displacement for Six-pole Radial Hybrid Magnetic Bearing Using Improved Simplified Particle Swarm Optimization LS-SVM[J]. Proceedings of the CSEE, 2020,40(13):4319-4329.
[17] 汪光远,杨德先,林湘宁等.基于深度置信网络的柔性直流配电网高灵敏故障辨识策略[J].电力系统自动化,2021,45(17):180-188.
Wang Guangyuan, Yang Dexian, Lin Xiangning, et al.High-sensitivity Fault Identification Strategy for Flexible DC Distribution Network Based on Deep Belief Networks[J]. Automation of Electric Power
Systems, 2021,45(17):180-188.
[18] 李益兵, 王磊, 江丽. 基于PSO改进深度置信网络的滚动轴承故障诊断[J].振动与冲击,2020,39(05):89-96.
Li Yibing,Wang Lei,Jiang Li.Rolling bearing fault diagnosis based on DBN algorithm improved with PSO[J]. Journal of Vibration and Shock. 2020,39(05):89-96.
[19] Reddy, Vanga Karunakar, Ravi Kumar AV. Multi-channel neuro signal classification using Adam-based coyote optimization enabled deep belief network[J]. Biomedical Signal Processing and Control, 2022, 77: 103774.
[20] 崔江, 郭瑞东, 张卓然, 等. 基于改进DBN的发电机旋转整流器故障特征提取技术[J]. 中国电机工程学报, 2020,40(07):2369-2376+2415.
Cui Jiang, Guo Ruidong, Zhang Zhuoranet al. Generator Rotating Rectifier Fault Feature Extraction Technique Based on Improved DBN[J]. Proceedings of the CSEE, 2020,40(07):2369-2376+2415.
[21] Hu Chang Huaet al. A Prognostic Model Based on DBN and Diffusion Process for Degrading Bearing[J]. IEEE Transactions on Industrial Electronics, 2019,:1-1.
[22] 孙丽君,冯斌斌,陈天飞.基于鞅论的灰狼优化算法全局收敛性分析[J].控制与决策,2022,37(11):2839-2848.
Sun Lijun, Feng Binbin, Chen Tianfei.Global convergence analysis of grey wolf optimization algorithm based onmartingale theory[J]. Control and Decision, 2022,37(11):2839-2848.
[23] Xue Fenget al. A hybrid clustering algorithm based on improved GWO and KHM clustering[J]. Journal of Intelligent & Fuzzy Systems, 2022, 42(4) : 3227-3240.
[24] Liu Xinyang and Wang Yifan and Zhou Miaolei. Dimensional Learning Strategy-Based Grey Wolf Optimizer for Solving the Global Optimization Problem[J]. Computational Intelligence and Neuroscience, 2022, 2022 : 3603607-3603607.