Identification And Suppression Of Sub-synchronous Oscillation With Grid-integration Of Dfig-based Wind Farms Via Series Compensated Lines

Nowaday,as the new energy generation plays an more and more important role in the power grid,wind power is representing the new energy and being integrated into the original power system at a high proportion with the high permeability of power electronic equipment.Under this trend,the damping of the grid-connected wind power system is getting weakened or even being negative.In order to improve the wind power’s outgoing capacity and transient stability in a large-scale,the use of series compensating capacitor technology is seems like essential.But it also has the risk of causing Subsynchronous Oscilation(SSO).Once SSO occurs,it will certainly cause great damage to the operational stability and safety of the whole wind power grid-connected system.In this thesis,we are going to present following aspects of the SSO of doubly-fed wind farms,which connected to the grid through series complement:Based on the analysis of the mechanism of SSO,a mathematical model suitable for the SSO research of the grid-connected system of doubly-fed wind farm is established.Among them,we are focusing on analyzing the rotor-side converter and network-side converter based on double closed-loop control strategy and stator voltage-directed vector control in order to lay the foundation for the subsequent simulation model construction and the identification and suppression of SSO.In order to analyze the parameters of SSO modes in doubly-fed wind farms,a new noise-like oscillation mode parameter identification algorithm based on noise-like signal MCEEMD and NEx T-ARMA was proposed in combination with the wide-area measurement system to obtain the noise-like response characteristics of grid-connected wind farms under environmental excitation in real time.To solve the problem of noise-like signals containing a large amount of noise,Complementary Ensemble Empirical Mode decom-position(CEEMD)with entropy improvement was introduced to effectively screen out abnormal components in the identified signals and improve the signal-to-noise ratio(SNR).The free oscillation response component of the signal is solved by using the Natural Excitation Technique(NEx T).It is used as the input data for ARMA.The parameter estimation method is used to identify the modal parameters of SSO.Through the analysis of the test signals,the simulation signals of the grid-connected model of doubly-fed wind farm,and the calculation examples of the measured signals of doubly-fed wind farm,it is shown that the modal parameter identification algorithm combined with MCEEMD and NEx T-ARMA can accurately identify the modal parameters of the SSO,which verifies the feasibility of the proposed method in theory and practical engineering.In order to reveal the influence law of wind speed,series complement and rotor side inner loop control parameters on the SSO phenomenon of grid-connected system,the time-domain simulation analysis method combined with the proposed modal parameter identification method were used to analyze the IGE and SSCI phenomena.The result shows that the weakening of electrical damping is the root cause of SSO induced in grid-connected system of wind farm.In view of the problem that the traditional PID controller in the rotor side converter has poor adaptive adjustment ability of control parameters in the face of the complex and changeable working conditions of wind farm,the method of RBF neural network adaptive PID controller was applied.The RBF neural network was used to improve the PI control of the stator side active power outer loop of the rotor side converter,the PID controller can adjust the parameters adaptively according to the change of working conditions.By doing simulation analysis,the results show that the RBF neural network control can effectively suppress the SSO under different working conditions.There are 56 figures,16 tables and 95 references in this thesis.