基于贝叶斯网络的变压器故障诊断研究

doi:10.16180/j.cnki.issn1007-7820.2021.03.008

摘要/Abstract

摘要：

针对传统变压器故障诊断比值法中存在比值编码过于绝对和不能及时发现早期潜伏性故障的问题,文中将改良的无编码比值法和变压器故障类型作为贝叶斯网络节点,利用改进的搜索策略和评分函数建立基于贝叶斯网络的变压器故障诊断模型。以LabVIEW为上位机建立了变压器在线故障诊断系统,实现变压器的数据采集、故障推理和历史信息查询等多种功能。最后通过实验测试证明,该系统不仅能够有效提高变压器的故障诊断效率,还可以提前发现故障从而延长变压器使用寿命,为变压器的安全稳定运行提供策略支持。

关键词: 变压器, 贝叶斯网络, 搜索策略, 评分函数, 故障诊断, LabVIEW

Abstract:

In the traditional transformer fault diagnosis ratio method, there are problems that the ratio code is too absolute and early latency failures cannot be found in time. In view of these problems, the improved non-coding ratio method and the type of transformer fault are used as the Bayesian Network nodes, and the improved search strategy and scoring function are employed to establish a transformer fault diagnosis model based on Bayesian network. The on-line fault diagnosis system of transformer based on LabVIEW is established, which realizes the functions of data acquisition, fault reasoning and historical information inquiry. Finally, the experiment proves that the system can not only effectively improve the efficiency of transformer fault diagnosis, but also detect the fault in advance to extend the service life of the transformer, and provide strategic support for the safe and stable operation of the transformer.

Key words: transformer, Bayesian network, search strategy, scoring function, fault diagnosis, LabVIEW

中图分类号:

TP183

仝兆景,秦紫霓,赵运星,芦彤,郑权. 基于贝叶斯网络的变压器故障诊断研究[J]. 电子科技, 2021, 34(3): 43-47.

TONG Zhaojing,QIN Zini,ZHAO Yunxing,LU Tong,ZHENG Quan. Research on Transformer Fault Diagnosis Based on Bayesian Network[J]. Electronic Science and Technology, 2021, 34(3): 43-47.

图/表 7

表1

表2

图1

表3

图2

图3

图4

参考文献 17

[1]	倪雪松. 基于动态阈值的牵引变压器故障诊断技术研究[D]. 成都:西南交通大学, 2016.
	Ni Xuesong. Study on traction transformer fault diagnosis technology based on dynamic threshold[D]. Chengdu: Southwest Jiaotong University, 2016.
[2]	许力. 电力变压器状态检修与在线监测的实际运用与探索[D]. 南京:东南大学, 2015.
	Xu Li. The practical application and explore of power transformer state maintenance and online monitoring[D]. Nanjing:Southeast University, 2015 .
[3]	马利东. 变压器油中溶解气体在线监测系统研究[D]. 沈阳:沈阳工业大学, 2018.
	Ma Lidong. Study on online monitoring system of dissolved gas in transformer oil[D]. Shenyang:Shenyang University of Technology, 2018.
[4]	梅红争. 基于油中气体分析的电力变压器故障诊断研究[D]. 长春:长春工业大学, 2019.
	Mei Hongzheng. Research on fault diagnosis of power transformer based on gas analysis in oil[D]. Changchun: Changchun University of Technology, 2019.
[5]	葛许良. 基于DGA的变压器在线监测与故障诊断方法研究[D]. 淮南:安徽理工大学, 2018.
	Ge Xuliang. Research on on-line monitoring and fault diagnosis of transformer based on DGA[D]. Huainan:Anhui University of Science and Technology, 2018.
[6]	李硕豪, 张军. 贝叶斯网络结构学习综述[J]. 计算机应用研究, 2015,32(3):641-646.
	Li Shuohao, Zhang Jun. Review of Bayesian networks structure learning[J]. Application Research of Computers, 2015,32(3):641-646.
[7]	沈佳杰. 基于改进混合进化算法的贝叶斯网络结构学习[D]. 杭州:浙江大学, 2013.
	Shen Jiajie. Structure learning of BN using improved memetic algorithm[D]. Hangzhou:Zhejiang University, 2013.
[8]	柯文远. 基于LabVIEW的电控发动机波形分析和故障诊断系统[D]. 广州:广东工业大学, 2017.
	Ke Wenyuan. Waveform analysis and fault diagnosis system for electronically controlled engine based on LabVIEW[D]. Guangzhou:Guangdong University of Technology, 2017.
[9]	蒲杰. 基于 LabVIEW 的变压器振动监测方法及应用研究[D]. 北京:华北电力大学, 2016.
	Pu Jie. The research of transformer vibration monitoring method based on LabVIEW and its application[D]. Beijing:North China Electric Power University, 2016.
[10]	仝兆景, 时俊岭, 高彩霞, 等. 基于贝叶斯网络的变压器在线监测与分析系统设计[J]. 测控技术, 2018,37(2):114-117.
	Tong Zhaojing, Sho Junling, Gao Caixia, et al. Design of transformer on-line monitoring and analysis system based on Bayesian networks[J]. Measurement and Control Technology, 2018,37(2):114-117.
[11]	Quan J, Ruan L, Xie Z, et al. The application of Bayesian network theory in transformer condition assessment[C]. Hongkong:IEEE PES Asia-Pacific Power and Energy Engineering Conference, 2013.
[12]	王艳霞. 基于改进贝叶斯网络的故障诊断优化算法研究[D]. 秦皇岛:燕山大学, 2018.
	Wang Yanxia. Study on optimization algorithm of fault diagnosis based on improved Bayesian network[D]. Qinhuangdao:Yanshan University, 2018.
[13]	胡春玲. 贝叶斯网络结构学习及其应用研究[D]. 合肥:合肥工业大学, 2011.
	Hu Chunling. Bayesian network structure learning and its applications[D]. Hefei:Hefei University of Technology, 2011.
[14]	金焱, 胡云安, 张瑾, 等. K2与模拟退火相结合的贝叶斯网络结构学习[J]. 东南大学学报(自然科学版), 2012,42(S1):82-86.
	Jin Yan, Hu Yunan, Zhang Jin, et al. Bayesian network structure learning combining K2 with simulated annealing[J]. Journal of Southeast University(Natural Science Edition), 2012,42(S1):82-86.
[15]	高晓利. 基于贪婪搜索的贝叶斯网络结构学习算法[D]. 西安:西安电子科技大学, 2011.
	Gao Xiaoli. Algorithms for Bayesian network structure learning based on the greedy search[D]. Xian:Xidian University, 2011.
[16]	刘浩然, 李轩, 马明, 等. 贝叶斯网络结构学习的简化贪婪算法[J]. 小型微型计算机系统, 2015,36(2):306-309.
	Liu Haoran, Li Xuan, Ma Ming, et al. Simplified greedy algorithm for Bayesian network structure learning[J]. Journal of Chinese Computer Systems, 2015,36(2):306-309.
[17]	葛春晓. 基于案例推理的变压器故障诊断专家系统的研究[D]. 长春:吉林大学, 2014.
	Ge Chunxiao. The research of transformer fault diagnosis expert system based on case-based reasoning[D]. Changchun: Jilin University, 2014.

编号	属性	编号	属性
A₁	H₂	A₇	C₂H₂/(C₁+C₂)
A₂	C₂H₂	A₈	H₂/(H₂+ C₁+C₂)
A₃	C₁+C₂	A₉	C₂H₄/(C₁+C₂)
A₄	CH₄/H₂	A₁₀	CH₄/(C₁+C₂)
A₅	C₂H₂/C₂H₄	A₁₁	C₂H₆/(C₁+C₂)
A₆	C₂H₄/C₂H₆	A₁₂	(CH₄+C₂H₄)/(C₁+C₂)

编号	故障类型
S₀	正常
S₁	中低温过热(<700 ℃)
S₂	高温过热(>700 ℃)
S₃	低能放电
S₄	高能放电
S₅	放电兼过热
S₆	局部放电

诊断方法及参数		故障类型
诊断方法及参数		S₀	S₁	S₂	S₃	S₄	S₅	S₆
	测试集	7	21	86	15	15	3	9
三比	正判数	6	13	64	14	13	2	8
值法	正判率	77.0%
K2	正判数	7	19	76	14	13	2	9
模型	正判率	89.7%
DSC	正判数	7	20	79	15	14	3	9
模型	正判率	94.2%