Study Of Continuous-Time Approaches To Parameter Estimation Of Excitation System

The excitation system parameters of large synchronuous generators have great effects on modern power system stability studies. Accurate and sufficient detailed excitation system models are essential to presice calculations of power system control and stability limits. However, many parameters which are used in studies are either "manufacturer specified"or "typicaP'values which may be grossly inaccurate, we cannot but use Eq constant model or inaccurate model parameters for calculations of power system stabilities, which may obtain unreliable results in simulation and stability analysis of electric power system. So,it is necessary to identify excitation system parameters from the field data by means of modern identification technology. The excitation system model to be identified can be classified into Linear excitation system model under small disturbance dynamic tests and nonlinear excitation system model under large disturbance tests. This paper only discusses parameter estimation of the first linear excitation system model. In the literature available, various techniques have been proposed for parameter determination of linear excitation system models. The FFT/LST method in frequency-domain , the LSE method in discrete time-domain and the PLPF method in continuous time-domain all have been most commonly used in field tests . PLPF method belongs to continuous time identifaication approaches .Based on PLPF method , This paper will concentrate on the application of continuous time identification approaches for parameter identification of excitation system , and presents the Linear Integral Filter method and Digital Low-pass Filters method. In simulation studies, PLPF method and LIF method are both used for identification of linear simplified excitation system model which parameters would be tuned for three times , their results are compared with each other and with the nominal values . It is showed that not only LIF method has the same identification precision and capability of resistance to noise as PLPF method, but also has some advantages such as improving the accuracy of parameter estimates when / is chosen properly, converging rapidly relative to PLPF method when identificating high order models, eliminating completely the effect of the initial condition problems of the integral-equation. So, LIF method issuitable to be applied in excitation system field tests together with PLPF method .This paper is organized as follows:Chapter 1 discusses the significance of parameter estimation of excitation system via field test and various identification approaches which have been presented. In Chapter 2, the computer models of excitation system are described in detail and MATLAB/SIMULINK based excitation contol system for simulation studies is also built. Chapter 3 discusses Equation Error Model Structure and theLeast-Squares Method, The well-known Tusin transformation between the discrete-time model and continuous-time model is also given. Chapter 4 presents a birds eye view of continuous time identification approaches and focuses on PLPF method and LIF method .In Chapter 5. in simulation studies , PLPF method and LIF method are both used for identification of excitation system model which parameters would be tuned for three times , their results are compared with each other and with the nominal values.Chapter 6 concludes the paper.