电力变压器的协同训练故障诊断算法  

An Algorithm of Co-Training for Fault Diagnosis of Power Transformers

作　　者：刘君[1,2] 李黎[3] 刘道群[2,4] 

机构地区：[1]重庆市广播电视大学理工学院,重庆400052 [2]重庆大学计算机学院,重庆400044 [3]重庆市气象局,重庆400044 [4]重庆理工大学,重庆400050

出　　处：《重庆工学院学报（自然科学版）》2009年第12期98-101,共4页Journal of Chongqing Institute of Technology

摘　　要：基于半监督学习思想,将协同训练的机制引入到故障诊断的学习中,采用多个不同的分类算法协同训练,提高电力变压器故障诊断的准确率.采用BP+SVM的多个训练器协同工作,目的是克服单个算法的准确性不高的缺点.将实例分析的结果与传统的BP算法的结果进行比较.结果表明,该算法能有效地对电力变压器单故障样本进行分类,具有较高的诊断准确率.According to co-training paradigm, this paper trains BP classifier and SVM classifier and make these two learners to co-work in the same training set to promote the classification accuracy. A mass of fault samples are analyzed in the algorithm and the results are compared with those obtained by the traditional BPNN. The comparison result indicates that the co-training algorithm has better classifying capability for single-fauh samples as well as high diagnosis precision.

关 键 词：神经网络 支持向量机 协同训练 电力变压器 故障诊断 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TM407[自动化与计算机技术—控制科学与工程]