关键词: 铝电解槽, 阳极电流, 频谱, 小波包, 神经网络

Abstract:

As the object of anodic current signal of 160 kA prebaked anode cells, these signals of different conditions were analyzed by the method of “spectrum  wavelet packet  neural network”.The results show that the anode current signals of the different status have different peak frequency value, so it is possible to extract energy characteristics vectors of anode current signals in different cell states using wavelet packets decomposition and wavelet packets reconstruction.According to the wavelet packet energy characteristics vectors extracted from anode current signals, diagnosis model based on BP neural network was established and verified.The simulation results show that the model of network identification is simple in construction, high accuracy in recognition, and convenient to realize online monitoring and realtime identification for the anode current signals of the different status.

Key words: 铝电解槽, 阳极电流, 频谱, 小波包, 神经网络