Study On Parameter Identification For The Generator's Excitation System

To study power system stability issue, it's necessary to consider the accurate model and parameters of excitation system. However, many parameters which are used in studies are either manufacturer specified or typical data, or use E_q' constant model foranalysis the stability of power system. These seriously affect the accuracy and reliability of the calculation. So, it is necessary to identify parameters of excitation system.The thesis first studied the identification method called Fast Fourier Transform/Least Square Estimation, expatiated the principium of the method, the select criteria of test single parameter settings were summarized based on theoretical deduction, and analyzed its merits and defects in application.Because the Fast Fourier Transform/Least Square Estimation identification method can't identify the nonlinear part of excitation system, the paper introduced Particle Swarm Optimization (PSO) algorithm in the parameter identification that makes the nonlinear link parameters identification possible and only according to the sampling data of input and output, directly in the time domain to identify parameters of excitation system without going through FFT transform, and it's very simple.The paper used PSO algorithm to identify the parameters of first-order link, nonlinear link and so on which are common in excitation systems and analyzed the results of identification when inputs various forms of disturbance signal, such as Step signal, Random Signal and the signal added noise. When considering the impact between link part of excitation system, improved PSO algorithm, that maked PSO algorithm not only can identify single-link of excitation system but also can identify the overall excitation system. Simulation examples prove that PSO algorithm has a strong ability to identify the parameters of excitation system.