摘要
选择合适的电弧模型,对建筑配电系统发生的故障电弧进行了仿真。基于小波的时—频分析特点和人工神经网络(ANN)的学习能力,提出了一种分辨故障电弧和正常负荷电流的方法。该方法通过小波变换对信号进行多分辨率分析,提取信号的特征矢量,利用人工神经网络对输入特征矢量进行故障识别。仿真实验的结果表明,该方法具有良好的故障识别性能。
The model of arcing was selected to sumulink the fault arcing produced in building power distribu- tion system. Based on time-frequency analysing character of wavelet and learning ability of artifical neural network （ANN）, a method of dectecting fault arcing and normal load current was presented. The current signals were decomposed with wavelet transform multi-resolution analysis and the feature vectors were extracted. Then the ANN was used to achieve the automatic fault detection according to the input feature vectors. Test results showed that this method had good performance for fault detection.
出处
《低压电器》
北大核心
2007年第10期38-42,共5页
Low Voltage Apparatus
关键词
建筑配电系统
故障电弧
小波变换
人工神经网络
故障检测
building power distribution system
fault arcing
wavelet transform
artificial neural network （ANN）
fault detection
