基于小波变换和GA-BP神经网络的电力电缆故障定位  

Cable Fault Location Based on Wavelet Transform and GA-BP Neural Network

作　　者：徐先峰[1] 马志雄 姚景杰 李芷菡 王轲 XU Xianfeng;MA Zhixiong;YAO Jingjie;LI Zhihan;WANG Ke(College of Energy and Electrical Engineering,Chang’an University,Xi’an 710064)

机构地区：[1]长安大学能源与电气工程学院,西安710064

出　　处：《电气工程学报》2024年第2期146-155,共10页Journal of Electrical Engineering

基　　金：国家重点研发计划资助项目(2021YFB2601300)。

摘　　要：由于电力电缆敷设于地下,当发生故障时难以快速且准确定位,出现了故障定位问题。因此,提出一种基于小波变换和遗传算法反向传播(Genetic algorithm back propagation,GA-BP)神经网络的电力电缆故障定位方法,在分析对比各小波能量集中程度和波动次数的基础上,选择多贝西小波(Daubechies wavelet 6,Db6)作为小波基函数,对于各故障位置,采集正向故障行波的α模分量,并对其进行小波分解。选取在d1尺度下的模极大值点作为特征值,同时将故障距离作为标签值,从而构建了训练和测试样本数据集;利用遗传算法(Genetic algorithm,GA)的种群进化和全局最优搜寻能力来改善误差逆传播(Back propagation,BP)网络对初始权重敏感的缺点,并使用优化后的权值、阈值重新对BP神经网络进行训练和预测,最后通过与传统双端行波定位算法、BP算法、粒子群优化BP算法(Particle swarm optimization BP,PSO-BP)相比较,证明了所提方法在测距性能方面的优越性。In view of the difficulty of locating power cables quickly and accurately when faults occur due to the fact that power cables are laid underground,a locating method for power cable faults based on wavelet transform and genetic algorithm back propagation(GA-BP)neural network is proposed.On the basis of analyzing and comparing the energy concentration degree and frequency of fluctuations of each wavelet,the Daubechies wavelet 6(Db6)is selected as the wavelet basis function.The wavelet decomposition is carried out on theαmode component of the forward fault traveling wave collected at each fault location.The maximum value of the mode under d1 scale is taken as the characteristic value,and the fault distance is taken as the label value.Using the population evolution and global optimal searching ability of genetic algorithm(GA)to improve the error back propagation(BP)network’s sensitivity to initial weight.The optimized weights and thresholds of the BP neural network are given for re-training and prediction.Finally,the proposed method is compared with the two-ended traveling wave positioning algorithm,BP algorithm and particle swarm optimization-BP(PSO-BP)algorithm to prove its excellent ranging performance.

关 键 词：小波变换 模极大值 双端测距 BP神经网络 PSO-BP神经网络 GA-BP神经网络 

分 类 号：TM743[电气工程—电力系统及自动化]