Analysis And Suppression Of Low Frequency Oscillation In Power System

In recent years,with the interconnection of regional power grid in China,coupled with high-gain fast excitation device more and more apply in power system,resulting in power system weak damping or negative damping state,prone to low frequency oscillation in external disturbance.Low frequency oscillation can damage the electrical equipment in power system,reducing the transmission capacity of the system.If low frequency oscillation can not be effectively suppressed,not only impact the stable operation of power grid,but also become one of the bottlenecks restricting the transmission capacity of the interconnection of regional power grid,causing serious economic losses.Therefore,it is very important to realize the stable operation of power grid by studying the mechanism of the low frequency oscillation and realizing the fast and effective real-time identification of the low frequency oscillation modes and studying the corresponding low frequency oscillation suppression strategy.Based on this,the main contents of this thesis are as follows:Firstly,this thesis introduces the mathematical model based on the single machine infinite system and the multi-machine system,and analyzes the root cause of the low frequency oscillation on the basis of the negative damping.Then,the basic structure of the power system stabilizer(PSS)and the principle of suppressing low frequency oscillation are introduced.After that,this thesis introduces the influence of the high-gain fast excitation device in the system.The simulation system can effectively suppress the low frequency oscillation after installing PSS,which lay the foundation for the following research.Secondly,this thesis considers the influence of Gaussian white noise on the accuracy of low frequency oscillation identification of power system,and proposes a denoising method based on sparse representation of tunable Q factor wavelet transform(TQWT).The method uses TQWT to decompose the noisy signal,uses sparse representation to deal with decomposd signal,and then reconstructs the tunable Q inverse transform to achieve the purpose of denoising.The method has the advantages of less number of parameters and simple setting.The method can effectively separate the noise from the low frequency oscillation signal and improve the accuracy of the low frequency oscillation mode identification.Then,based on the denoising,this thesis proposes an improved matrix pencil algorithm to identify the denoised low frequency oscillation signal.By introducing the relative change rate of singular value order method to improve the matrix pencil algorithm,the improved matrix pencil algorithm has the advantages of adaptive setting order,small computation and high precision order determination in system.Through the numerical test simulation,the IEEE four-machine two-area system,New England 10 machine 39 node system and North American power grid simulation data analysis,verify the accuracy and feasibility of the method in the low frequency oscillation mode identification of the power system.Finally,the low frequency oscillation is suppressed by PSS.For PSS optimization configuration,uses retention method to select the PSS installation site;Aiming at the problem of PSS parameter setting,this thesis proposes a hybrid algorithm SMAFA algorithm that combining the simplex method(SM)with the adaptive firefly algorithm(AFA)to optimize the PSS parameter.The hybrid algorithm combines the advantages of the global search with the AFA algorithm and the local search advantages of the SM algorithm.The mathematical simulation results show that SMAFA has more advantages than the traditional algorithm in the convergence speed.The PSS after setting is applied to the four-machine two-area system,the mall interference simulation results show the rationality of PSS optimal configuration and the validity of PSS parameter setting;The big interference simulation results show that the proposed method can improve the robustness of the system effectively.