基于模式识别技术的变压器故障诊断

当前方法变压器故障诊断方无法描述变压器故障的变化特点,导致变压器故障诊断时间长、诊断误差高等问题,为为了提高变压器故障效果,提出基于模式识别技术的变压器故障诊断方法.首先,分析当前变压器故障诊断的研究进展,找到各种方法的局限性;然后,采用粒子群算法确定故障类别门限值,利用模式识别技术根据峰值、诊断阈值、熵建立得到随机模...

Transformer fault diagnosis based on pattern recognition technology

The current method of transformer fault diagnosis can not describe the changing characteristics of transformer fault,which leads to the problems of long time and high diagnosis error.In order to improve the effect of transformer fault,a transformer fault diagnosis method based on pattern recognition technology is proposed.Firstly,the research progress of transformer fault diagnosis is analyzed,and the limitations of various methods are found.Then,the threshold value of fault category is determined by particle swarm optimization algorithm,and the random fuzzy characteristic factor is established according to the peak value,diagnosis threshold and entropy by pattern recognition technology.The fault category interval is obtained according to the characteristic factor,and the transformer fault diagnosis is completed according to the fault category matching results.The simulation results show that this method can effectively improve the speed of model training and the accuracy of transformer fault diagnosis,and obtain an ideal effect of transformer fault diagnosis.