基于人工鱼群算法和LS_SVM的变压器故障诊断

doi:10.16180/j.cnki.issn1007-7820.2020.11.007

摘要/Abstract

摘要：

针对变压器故障数据的特征信息不确定性以及传统诊断方法准确率较低的问题,文中采用人工鱼群算法和最小二乘支持向量机相结合的方法来进行变压器故障诊断。将IECTC10数据库中的DGA特征气体比值作为输入,建立基于最小二乘支持向量机的变压器故障诊断模型,并运用人工鱼群算法对最小二乘支持向量机的参数进行优化选取。然后根据诊断结果,选出分类效果最佳的多比值特征参量组合。实验验证结果显示,文中所提出的诊断方法准确率可达96.67%,拥有更高的故障诊断正确率。

关键词: 最小二乘支持向量机, 故障诊断, 人工鱼群算法, 变压器, DGA, IECTC10数据库

Abstract:

In view of the information uncertainty of transformer fault data and the low accuracy of traditional diagnostic methods, the combination of artificial fish swarm algorithm and least squares support vector machine is used to diagnose transformer fault. The DGA characteristic gas ratios of IECTC10 database is used as the input vectors, and the fault diagnosis model of transformers is designed based on LS_SVM. Meanwhile, the artificial fish swarm algorithm is utilized to optimize the parameters of the least squares support vector machine. Then, based on the diagnosis result, the multi-ratio characteristic parameter combination with the best classification effect is selected. The experimental verification results show that the accuracy of the proposed diagnostic method was up to 96.67%, and it has a higher accuracy rate of fault diagnosis.

Key words: least square support vector machine, fault diagnosis, artificial fish swarm algorithm, power transformer, DGA, IECTC10 database

中图分类号:

TP181

杨宇,曾国辉,黄勃. 基于人工鱼群算法和LS_SVM的变压器故障诊断[J]. 电子科技, 2020, 33(11): 36-40.

YANG Yu,ZENG Guohui,HUANG Bo. A Transformer Fault Diagnosis Method Integrating Artificial Fish Swarm Algorithm with Least Square Support Vector Machine[J]. Electronic Science and Technology, 2020, 33(11): 36-40.

图/表 11

图1

表1

图2

表2

图3

表3

表4

图4

图5

表5

图6

参考文献 16

[1]	Sherif S M G, Ibrahim B M T. A new approach of DGA interpretation technique for transformer fault diagnosis[J]. International Journal of Electrical Power and Energy System, 2016,81(7):265-274.
[2]	汪可, 李金忠, 张书琦, 等. 变压器故障诊断用油中溶解气体新特征参量[J]. 中国电机工程学报, 2016,36(23):6570-6578.
	Wang Ke, Li Jinzhong, Zhang Shuqi, et al. New features derived from dissolved gas analysis for fault diagnosis of power transformers[J]. Proceedings of the CSEE, 2016,23(36):6570-6578.
[3]	刘景艳, 王福忠, 杨占山. 基于RBF神经网络和自适应遗传算法的变压器故障诊断[J]. 武汉大学学报(工学版), 2016,49(1):88-93.
	Liu Jingyan, Wang Fuzhong, Yang Zhanshan. Transformer fault diagnosis based on RBF neural network and adaptive genetic algorithm[J]. Engineering Journal of Wuhan University, 2016,49(1):88-93.
[4]	谷凯凯, 郭江. 紧致融合模糊集和故障树的变压器故障诊断[J]. 高电压技术, 2014,40(5):1507-1512.
	Gu Kaikai, Guo Jiang. Transformer fault diagnosis method based on compact fusion of fuzzy set and fault tree[J]. High Voltage Engineering, 2014,40(5):1507-1512.
[5]	岑健, 李玉娜. 无量纲免疫支持向量机的复合故障诊断方法[J]. 计算机工程与应用, 2013,49(15):259-262.
	Cen Jian, Li Yuna. Complex fault diagnosis using dimensionless immune support vector machine[J]. Computer Engineeringand Applications, 2013,49(15):259-262.
[6]	Kim S W, Kim S J, Seo H D, et al. New methods of DGA diagnosis using IEC TC10 and related databases Part 1:application of gas-ratio combinations[J]. IEEE Transactions on Dielectrics and Electrical Insulation, 2013,20(2):685-690.
[7]	Duval M, dePabla A. Interpretation of gas-in-oil analysis using new IEC publication 60599 and IECTC10 databases[J]. IEEE Electrical Insulation Magazine, 2001,17(2):31-41.
[8]	Zheng H B, Liao R J, Grzybowski S, et al. Fault diagnosis of power transformers using multi-class least square support vector machines classifiers with particle swarm optimization[J]. IET Electric Power Application, 2011,5(9):691-696.
[9]	李赢, 舒乃秋. 基于模糊聚类和完全二叉树支持向量机的变压器故障诊断[J]. 电工技术学报, 2016,31(4):64-70.
	Li Ying, Shu Naiqiu. Transformer fault diagnosis based on fuzzy clustering and complete binary tree support vector machine[J]. Transactions of China Electrotechnical Society, 2016,31(4):64-70.
[10]	Benmahamed Y, Teguar M, Boubakeur A. Application of SVM and KNN to duval pentagon 1 for transformer oil diagnosis[J]. IEEE Transactions on Dielectrics and Electrical Insulation, 2017,24(6):3443-3451.
[11]	宗剑韬. 考虑供电区域差异的配电网可靠性成本效益分析与应用研究[D]. 北京:华北电力大学, 2017.
	Zong Jiantao. Cost-effectiveness analysis and application of distribution network reliability considering regional differences in power supply[D]. Beijing:North China Electric Power University, 2017.
[12]	韩平, 王天堃, 孟永毅. 基于LS-SVM的一次风机振动在线监测及故障预警研究[J]. 机电工程, 2016,33(5):629-632.
	Han Ping, Wang Tiankun, Meng Yongyi. Research of LS-SVM based method for online monitoring and fault prediction of primary air fan vibration[J]. Journal of Mechanical and Electrical Engineering, 2016,33(5):629-632.
[13]	高国磊, 李英娜, 段效琛, 等. 基于ACO优化LS_SVM的变压器故障诊断[J]. 电子科技, 2018,31(6):59-61.
	Gao Guolei, Li Yingna, Duan Xiaochen, et al. A fault diagnosis method for transformer based on least squares support vector machine optimized by ant colony optimization[J]. Electronic Science and Technology, 2018,31(6):59-61.
[14]	邱云飞, 李智义. 改进人工鱼群算法在SVM参数优化中的应用[J]. 计算机工程与科学, 2018,40(11):2074-2079.
	Qiu Yunfei, Li Zhiyi. An improved artificial fish swarm algorithm and its application for SVM parameter optimization[J]. Computer Engineering & Science, 2018,40(11):2074-2079.
[15]	郑含博, 王伟, 李晓纲, 等. 基于多分类最小二乘支持向量机和改进粒子群优化算法的电力变压器故障诊断方法[J]. 高电压技术, 2014,40(11):3424-3429.
	Zheng Hanbo, Wang Wei, Li Xiaogang, et al. Fault diagnosis method of power transformers using multi-class LS-SVM and improved PSO[J]. High Voltage Engineering, 2014,40(11):3424-3429.
[16]	洪洁, 王璐, 汪超, 等. 基于人工鱼群算法优化SVM的部动作sEMG识别[J]. 传感器与微系统, 2016,35(2):23-25.
	Hong Jie, Wang Lu, Wang Chao, et al. Recognition of sEMG hand actions based on artificial fish swarm algorithm optimized SVM[J]. Transducer and Microsystem Technologies, 2016,35(2):23-25.

编号	DGA比值	编号	DGA比值
1	CH₄/C₂H₂	6	H₂/C₂H₆
2	C₂H₂/CH₄	7	H₂/C₂H₂
3	C₂H₄/CH₄	8	H₂/C₂H₄
4	CH₄/C₂H₆	9	C₂H₆/C₂H₄
5	C₂H₆/C₂H₂	10	H₂/CH₄

特征选择结果	特征组合
特征选择结果	1	2	3
DGA比值	H₂/C₂H₄	H₂/C₂H₄	H₂/C₂H₂
	H₂/C₂H₆	H₂/C₂H₂	C₂H₂/C₂H₄
	CH₄/C₂H₂	CH₄/C₂H₂	CH₄/C₂H₄
	C₂H₆/C₂H₄	CH₄/C₂H₆	CH₄/C₂H₆
	CH₄/C₂H₆	——	——

特征	DGA比值	特征	DGA比值
1	H₂/C₂H₄	4	C₂H₆/C₂H₄
2	H₂/C₂H₆	5	CH₄/C₂H₆
3	CH₄/C₂H₂

故障类型	训练集	测试集	类型编号
PD	14	4	1
D₁	25	2	2
D₂	25	11	3
T₁	16	6	4
T₂	22	7	5

实例编号	油中溶解气体体积分数/10^-6					诊断结果
实例编号	H₂	CH₄	C₂H₄	C₂H₆	C₂H₂	LS_SVM	本文方法	IEC三比值	实际故障
1	155.00	581.00	215.00	293.00	142.00	2	4	1	1
2	49.00	41.60	124.00	25.10	15.70	5	3	2	3
3	21.00	7.60	96.30	4.70	26.73	3	3	3	3
4	9.00	16.70	85.30	9.90	3.40	2	2	3	2
5	250.00	63.00	66.00	2.80	120.00	3	3	3	3
6	4.10	3.50	1.20	0.68	5.20	2	4	4	4
7	12.10	4.40	3.50	40.00	4.10	4	5	5	5
8	44.20	101.00	11.80	13.00	0.45	2	2	2	2
9	6.84	11.30	12.60	20.00	4.01	5	5	5	5
10	58.00	103.00	251.00	51.00	6.00	1	1	2	1
11	72.00	116.00	70.00	36.00	19.00	3	3	3	3
12	9.00	16.70	85.30	9.90	3.40	2	2	1	2