Method For Identification Of Low Frequency Oscillations Mode Based On EEMD And Matrix Pencil Algorithm

For a long time,low frequency oscillation problem is one of the important factors to threaten the operation security and stability of the power system,and with the increasing of power system scale,low frequency oscillation also surfaces from time to time.How to effectively suppress low frequency oscillation has become a focus in the research of many scholars in recent years,and the accurate identification for model parameters of low frequency oscillation is a difficulty.So the further study for analysis method of low frequency oscillation is very necessary.This paper mainly studies the matrix pencil algorithm,and compares it with the Prony method and the improved Prony method,which prove the effectiveness of the matrix pencil algorithm in the model parameters identification of low frequency oscillation.Although the matrix pencil algorithm can effectively identify model parameters of the oscillation,by adding the gaussian white noise with different SNR size to the ideal signal,the results show that the traditional matrix pencil algorithm still has certain limitation,namely for the traditional matrix pencil algorithm,the pole extraction accuracy is not high and the numerical calculation is instable under low SNR.These are greatly limits for the application of matrix pencil algorithm,and it has been difficult to meet the need of practical engineering.To solve the problems existing in the traditional matrix pencil algorithm,this paper proposes a new method for identification of power system low frequency oscillation mode based on ensemble empirical mode decomposition(EEMD)and matrix pencil algorithm.The EEMD is used to deal with the unstable signal.This not only retains the advantages of EMD,but also effectively solves the problems of mode mixing in EMD,which has good noise reduction effect.The EEMD filter and cross-correlation coefficient are used to get real IMF components in the range of low frequency oscillation,and signal energy weight is used to select the dominant mode component.The dominant mode component is identified by using the matrix pencil algorithm,and the spectral norm form index function is used to determine the actual modal order of the matrix pencil algorithm.The method combines EEMD and matrix pencil algorithm,which can effectively suppress the noise on the influence of the parameter identification process,thus overcoming the problems of the traditional matrix pencil algorithm.This makes it still has high accuracy of parameter identification under low SNR and strong anti-noise ability,and provides a new idea for identification of power system low frequency oscillation mode and research and design of damping controller,and has high value in practical engineering application.Simulative results of case study also show the feasibility and superiority of the method.