辽宁工程技术大学学报（自然科学版）2017，Vol.36Issue(9)：964-970,7.DOI:10.11956/j.issn.1008-0562.2017.09.013
一种基于改进RVM变压器故障诊断新方法
A new fault diagnosis method for transformer based on improved RVM
付华 1齐致 2任仁3
作者信息
- 1. 辽宁工程技术大学电气与控制工程学院,辽宁葫芦岛125105
- 2. 国网辽宁省电力有限公司葫芦岛供电公司,辽宁葫芦岛125105
- 3. 国网辽宁省电力有限公司朝阳供电公司,辽宁朝阳122000
- 折叠
摘要
Abstract
Compared with the Support Vector Machine (SVM),the Relevance Vector Machine (RVM) has obvious advantages in classification performance.Kernel Principal Component Analysis (KPCA) and Quantum Particle Swarm Optimization (QPSO) algorithm is introduced to optimize the fault diagnosis model of RVM power transformer.The main characteristic gas content of the transformer is set as the input quantity,and the fault diagnosis model of the transformer based on KPCA-QPSO-RVM is established by the classification method of the two fork tree.The example analysis and comparison with the SVM and RVM methods prove that the method can obtain a better fault diagnosis accuracy rate,the number of correlation vectors is significantly less than the number of support vectors,and the diagnosis rate significantly increased.关键词
相关向量机(RVM)/核主成分分析(KPCA)/量子粒子群算法(QPSO)/变压器故障诊断Key words
relevance vector Machine (RVM)/kernel principal component analysis (KPCA)/quantum particle swarm algorithm (QPSO)/transformer fault diagnosis分类
信息技术与安全科学引用本文复制引用
付华,齐致,任仁..一种基于改进RVM变压器故障诊断新方法[J].辽宁工程技术大学学报（自然科学版）,2017,36(9):964-970,7.
