摘要
为了保证电力变压器安全运行、准确判断故障检修模式,研究一种基于混合神经网络算法的故障检修模式识别方法。将混合神经网络与传感模块相连,联合径向基函数,分析样本数据集合间线性映射关系,求解混合神经网络中隐含层中心与输出层的权重值。初始化分类故障数据,使用最近距离聚类方法,获得同密度下节点路径。用等比例特征量分析电力变压器产生的不同特征气体,分类成不同级别,明确电力变压器故障数据的传播方向,独立训练、归一化处理原始故障数据。根据特征矩阵输出非线性矢量,输出层输出故障模式的后验概率,完成电力变压器故障检修模式识别。实验结果表明,所提方法识别准确率高,迭代效率快,识别效果好,为电力变压器故障检修策略决策提供了新思路。
In order to ensure the safe operation of power transformer and accurately judge the fault maintenance mode,a fault maintenance mode recognition method based on hybrid neural network algorithm is studied.The hybrid neural network is connected with the sensor module,and the radial basis function is combined to analyze the linear mapping relationship between the sample data sets to solve the weight value of the hidden layer center and the output layer in the hybrid neural network.The classified fault data are initialized,and the nearest distance clustering method is used to obtain the node paths under the same density.The different characteristic gases generated by power transformers are analyzed by the characteristic quantities of equal proportion,and are classified into different levels,the propagation direction of power transformer fault data is clarified,and the original fault data are trained independently and normalized.According to the characteristic matrix,the nonlinear vector is output,and the output layer outputs a posteriori probability of the fault mode to complete the power transformer fault maintenance mode recognition.The experimental results show that the proposed method has high recognition accuracy,fast iteration efficiency and good recognition effect.It provides a new idea for the decision-making of power transformer fault maintenance strategy.
作者
林森
王海涛
顾新桥
LIN Sen;WANG Haitao;GU Xinqiao(Shibei Power Supply Branch of State Grid Chongqing Electric Power Company,Chongqing 400000,China)
出处
《微型电脑应用》
2024年第11期178-181,共4页
Microcomputer Applications
关键词
混合神经网络
变压器故障
故障检修
故障识别
聚类分析
hybrid neural network
transformer fault
fault maintenance
fault recognition
cluster analysis
作者简介
林森(1982-),男,硕士,高级工程师,研究方向电力设备检测技术;王海涛(1977-),男,专科,助理工程师,研究方向电力设备检测技术;顾新桥(1982-),男,本科,工程师,研究方向信息化建设及管理。
