摘要
目前国际上对变压器的故障诊断主要是通过检测变压器油中气体的成分来进行的。针对目前大部分故障诊断算法都存在数据量大、计算复杂、准确率低等问题,提出了利用半监督分类方法来诊断变压器故障。通过建立简单的变压器模型,推导算法公式,最后进行数据验证。通过实际的模拟实验发现,该方法能有效地减少计算数据的处理和计算过程,计算结果的准确率也比较高,为今后同类型的变压器故障研究提供了新的思路。
At present,the fault diagnosis of the transformer is mainly carried out by detecting the components of the gas in the transformer oil. In view of the fact that most of the algorithms have large data volume,complex computation and low accuracy,a semi supervised classification method has been proposed to diagnose transformer faults. Through the actual simulation experiment,it is found that this method can effectively reduce the computation data and simplify the computation process,and with high accuracy of the calculation results. It provided a new way of thinking for the research of the same type of transformer fault in the future.
引文
[1]店勇波,桂卫华,欧阳伟.基于相对重构贡献的变压器故障诊断力法[J].高电压技术,2012,38(11):2964-2971.
[2]陈小青,刘觉民,黄英伟,等.采用改进人上角群优化粗糙集算法的变压器故障诊断[J].高电压技术,2012,38(6):1403-1409.
[3]刁郑元兵,孙才新,李剑,等.变压器故障特征量可信度的关规则分析[J].高电压技术,2012,38(1):82-88
[4]杜文霞,句希源,吕锋.基于模糊聚类算法的变压器故障诊断研究[J].变压器,2009,46(8):65-69.
[5]付强,陈特放,米佼佼.采用自组织RBF网络算法的变压器故障诊断[J].高电压技术,2012,38(6):1368-1375.
[6]李霜,王朗珠,张为,等.基于DUA的改进BP神经网络的变压器故障诊断方法[J].变压器,2010,47(12):61-65.