基于神经网络的故障电弧检测装置的研究

Research of Arc Fault Detctor Based on Neural Networks

提出一种以反向传播(BP)神经网络为基础的新型故障电弧辨识方法。电弧发生装置和数据采集装置分别提取电流的小波变换和时域变化的特征值,将提取的混合特征值输入到神经网络并采用自适应学习法进行学习,再将收敛的神经网络移植到故障电弧检测装置中,构成故障电弧辨识模块。在试验中装置可以准确地检测出已学习过的几种不同负载的故障电弧状态,并能及时切除故障线路。

： A new method for arc fault identification was proposed,which is based on back propagation （BP） neural network. Firstly, the current data of wavelet transform and time-domain are collected by arc generation and data acquisition device respectively. Then, it is input into BP neural network to learn by adaptive learning method. After that, the converged neural network is transplanted into the arc fault detection device to constitute the arc fault identification module. The detection device can accurately detect the are fault of several different loads and timely remove the fault line.