基于改进神经网络算法的变压器油色谱故障检测方法

Oil Chromatography Fault Detection Method of Transformer Based on Improved Neural Network Algorithm

针对变压器油色谱故障分析方法预测能力不足,诊断评价准确率低的缺陷,提出一种基于神经网络算法和灰色关联度方法的变压器故障识别组合方法,通过对变压器绝缘油色谱中H_(2)、CH_(4)、C_(2)H_(6)、C_(2)H_(4)、C_(2)H_(2)等特征气体的检测,并将其作为神经网络算法的输入变量,同时采用灰色关联方法对变压器绝缘故障的放电、高温、接地等12类故障进行关联度分析,制定合理的故障类别排序表用于神经网络的故障诊断。仿真结果表明:采用本文方法与传统三比值方法相比,其诊断时间缩短、准确度提高,具有一定的理论意义与实用价值。

Aiming at the defects of insufficient prediction ability and low accuracy of diagnosis and evaluation of transformer oil chromatography fault analysis method, a combined method of transformer fault identification based on neural network algorithm and grey relational grade method is presented. The characteristic gases H_(2),CH_(4),C_(2)H_(6),C_(2)H_(4),C_(2)H_(2)in the chromatogram of transformer insulation oil are detected, it is regarded as the input variable of neural network algorithm, and 12 kinds of faults such as discharge, high temperature and earthing of transformer insulation fault are analyzed by grey correlation method, a reasonable sorting table of fault classes is developed for fault diagnosis of neural network. The simulation results show that compared with the traditional three-ration method, this method has shorter diagnosis time and higher accuracy, and has certain theoretical significance and practical value.