基于多特征融合的交流系统串联电弧故障诊断

发布时间：2018-06-21 02:23

本文选题：电弧故障 + 小波变换　；参考：《电网技术》2014年03期

【摘要】：为实现对用电系统低压端串联电弧故障的准确诊断,根据交流系统中低压串联电弧故障的奇异性、能量特性及不确定性,通过自主搭建的电弧故障模拟实验平台及不同负载下的串联电弧故障模拟实验,提出一种基于多特征融合的串联电弧故障诊断方法。该方法根据信号不同特性,结合小波变换理论对经降噪预处理后的采样信号进行主成分分析,提取各频段特性对信号的贡献率,并以信号3种特性中最大贡献率所在频段的空间位置关系作为特征向量构成1?3阶信号特性分布矩阵;将此矩阵作为网络的输入向量,利用改进多层前馈神经网络构建特征向量与电弧故障之间的映射关系。测试结果表明,该方法可减小电弧燃烧对诊断结果的影响,实现对串联电弧故障的诊断分类。
[Abstract]:In order to accurately diagnose the series arc fault of low voltage end of electric power system, according to the singularity, energy characteristic and uncertainty of low voltage series arc fault in AC system, A fault diagnosis method for series arc based on multi-feature fusion is proposed by using the self-built arc fault simulation platform and the series arc fault simulation experiment under different loads. According to the different characteristics of the signal and the wavelet transform theory, the method carries out the principal component analysis of the sampled signal after de-noising preprocessing, and extracts the contribution rate of the characteristics of each frequency band to the signal. The spatial position relation of the largest contribution rate in the frequency band of the three characteristics of the signal is taken as the eigenvector to form the signal characteristic distribution matrix of order 1 and 3, and the matrix is used as the input vector of the network. An improved multilayer feedforward neural network is used to construct the mapping relationship between eigenvector and arc fault. The test results show that this method can reduce the influence of arc combustion on the diagnosis results and realize the diagnosis classification of series arc faults.
【作者单位】：沈阳工业大学电气工程学院;
【基金】：国家自然科学基金项目(51377106,51337001)~~
【分类号】：TM501.2

【参考文献】