Abstrakt

Power system low frequency oscillation detection using local mean decomposition algorithm

Lei Wan-Zhong1, Li Tao2, Li Jian-Feng

In allusion to low frequency oscillation signal in Power System, the local mean decomposition (LMD) algorithm is applied to the low frequency oscillation signal detection in power system for the first time. Complex original signals can be decomposed into a number of product function component, Each layer of product function component is composed of the envelope signal and frequency modulation function, which contain all the instantaneous amplitude and instantaneous frequency information, Further combinations can get the original signal time frequency distribution. The typical power quality low frequency oscillation signal is selected and analyzed by LMD algorithm. This algorithm overcomes the incapability for the Fourier algorithm to deal with non-stationary signals, as well as the difficulty in choosing Wavelet. LMD algorithm can be accurate to abstract the dynamic oscillating performance and abundant transient fault information from the non-stationary signal, the amplitude and frequency curve, not only can accurately locate the disturbance moments, but also can detect the voltage fluctuation amplitude. The simulation waveform was influenced by "end effect" smaller. Simulation results show that LMD Algorithm is effective, and has better locate accuracy and computing speed than the HHT algorithm