Research On Impact Of Wind Power On Power System Low-frequency Based On EEMD-Robust ICA And Prony Algorithm

Because of its advantages of non-emission of waste and environmental protection,wind power generation technology stands out from many renewable energy technologies and has become a relatively fast-growing and relatively mature technology.However,the specific operation experience shows that a large number of wind turbines are connected to the power system,resulting in the change of the power supply structure,which makes the low frequency oscillation modes change and has a great harm on the system stability.Therefore,studying the influence of large-capacity wind turbines connected to grid on the low-frequency oscillation characteristics of power system has become one of the most popular topics.In this thesis,the Doubly Fed Induction Generator(DFIG),which is the most representative,is taken as the research object to analyze the influence of DFIG wind turbines connected to grid on low-frequency oscillation of power system under different working conditions.Firstly,the basic principle and characteristics of the Prony algorithm are studied in detail.The algorithm can accurately identify the mode parameters without noisy signal,but for noisy signals,there is a big error for the identification results.Then,the basic principle of the Ensemble Empirical Mode Decomposition(EEMD)method and the denoising principle are analyzed,although the EEMD method can decompose the noisy signal into a series of Intrinsic Mode Functions(IMF),the traditional EEMD denoising method has the shortcomings of incomplete denoising and the useful signal being filtered out as noise.Based on this,an analysis method combining EEMD and Robust Independent Component Analysis(Robust ICA)algorithm with Prony algorithm is proposed to identify mode parameters of low-frequency oscillation signals with noises,this method not only overcomes the defaults that Prony algorithm is sensitive to noise,but also solves the problem that the traditional EEMD denoising method is poorly effective on denoising.Finally,on the basis of establishing wind turbines simulation model,mechanical transmission simulation model and generator simulation model,an IEEE power system model with 8 machines 24 nodes including grid-connected DFIG wind turbines is established by Matlab / Simulink simulation software.By using the combination of EEMD-RobustICA and Prony algorithm,the influence of different transmission distances and outputs of grid-connected DFIG wind farms on the low-frequency oscillation of power system is studied.The research results show that the combination of EEMD-Robust ICA denoising method and Prony algorithm proposed in this thesis can remove the noise better and identify the key oscillation modes accurately for the low frequency oscillation signal.Moreover,the influence of wind turbines connected to grid on low-frequency oscillation characteristics of power system vary with the transmission distance and output of wind turbines.It can be seen from the analysis that the increase of transmission distance will worsen the stability of the power system and when the output of wind turbines increases within a certain range,the stability of the power system can be improved.However,after a given output,the damping ratio will decreas,meanwhile the stability of the power system will reduce.