基于概率神经网络的油浸式变压器失效判别仿真研究  

Simulation Research on Failure Discrimination of Oil Immersed Transformer Based on Probabilistic Neural Network

作　　者：钟明灯[1] 陈冬冬 ZHONG Mingdeng;CHEN Dongdong(College of Optoelectronic and Electromechanical Engineering,Minnan University of Science and Technology,Quanzhou Fujian 362000,China;College of Electronic and Electrical Engineering,Minnan University of Science and Technology,Quanzhou Fujian 362000,China)

机构地区：[1]闽南理工学院光电与机电工程学院,福建泉州362000 [2]闽南理工学院电子与电气工程学院,福建泉州362000

出　　处：《佳木斯大学学报（自然科学版）》2022年第3期65-69,共5页Journal of Jiamusi University：Natural Science Edition

基　　金：辽宁省自然科学基金重点领域联合开放项目(20-19KF2307)。

摘　　要：采用传统方法判别油浸式变压器失效故障时存在有效率低、性能差的问题,提出了基于概率神经网络的油浸式变压器失效判别仿真方法。利用时频定义了油浸式变压器失效信号,根据失效信号的能量分布情况提取出其模态特征,结合希尔伯特变换完成变压器失效信号的识别。采用概率神经网络连接拓扑结构,将失效故障值转换为网络错误帧,根据变压器的连接特性计算变压器失效的响应特性,引入变量分析法建立变压器失效数据关联函数,通过排列运算失效数据的关联变量构建失效数据统计量,以此进行油浸式变压器的失效判别。仿真实验结果表明,该方法不仅可以提高失效判别的有效率,还可以提升判别效率,具备一定的实际应用价值。In order to solve the problems of low efficiency and poor performance when traditional methods are used to identify the failure of oil immersed transformer,a simulation method for identifying the failure of oil immersed transformer based on probabilistic neural network is proposed.The failure signal of oil immersed transformer is defined by time-frequency,its modal characteristics are extracted according to the energy distribution of the failure signal,and the identification of transformer failure signal is completed combined with Hilbert transform.The probabilistic neural network connection topology is adopted to convert the failure fault value into network error frame,calculate the response characteristics of transformer failure according to the connection characteristics of transformer,introduce the variable analysis method to establish the transformer failure data correlation function,and construct the failure data statistics by arranging and calculating the correlation variables of failure data,so as to judge the failure of oil immersed transformer.Simulation results show that this method can not only improve the efficiency of failure identification,but also improve the efficiency of failure identification,and has a certain practical application value.

关 键 词：概率神经网络 信号识别 油浸式变压器 响应特性 失效判别 构建统计量 

分 类 号：TM41[电气工程—电器]