Study on a new method to identify inrush current of transformer based on wavelet neural network

Focusing on the key problem of transformer protection malfunction caused by the inrush current, this paper analyses the transient mechanism, establishes mathematical model, studies inrush current quantitatively by derivation of equation firstly. On this basis, the simulation models of inrush current and short circuit current are established on the simulation platform PSCAD/EMTDC, the wavelet multiresolution analysis of the two currents is adopted by using the wavelet toolbox of matlab in this paper. According to the ´higher energy´ characteristic of the inrush current´s waveform after wavelet transform, this paper uses db5 wavelet to extract wavelet transform energy characteristic values of inrush current and short circuit current, takes these as feature spaces of improved BP neural network pattern recognition, uses the classificatory function of neural network to distinguish inrush current and short circuit current. At last, a new reliability criterion which is simple and more easily digital applied is proposed. A lot of simulations verify that the new method proposed in this paper has the advantages of high dependability, good sensitivity and quick acting speed. The action time of protection is generally around 14ms.