In order to improve the effect of power grid fault diagnosis,a power grid fault type judgment method based on SCA-DA alarm data is proposed.The improved wavelet neural network fault identification model with momentum term and adaptive learning rate is constructed.The lifting wavelet is used to decompose the positive sequence signals of lines at both ends of com-ponents through the processes of splitting,prediction and adjustment.It obtains the positive sequence signals of different scales,inputs them to the neural network,and the output result is the refined power grid fault type.The experimental results show that when the decomposition scale is 3,the mean square error of fault recording signal is the smallest,and the perform-ance of wavelet neural network is more stable.This method judges the fault type and fault phase according to the sudden change of phase-to-phase two-phase current,and can accurately judge the fault type and fault cause of power grid.
关键词
数据采集与监视控制系统/电网故障/类型判断/故障录波数据/小波神经网络/故障识别模型
Key words
supervisory control and data acquisition/grid failure/type judgment/fault recording data/wavelet neural network/fault identification model
维普
万方数据