Parameter Identification Of Generators For Large Power System Dynamic Equivalence

The complexity of large power system makes it difficult to analyze and calculate, so dynamics equivalence is an insistent demand. We should build up a equivalent system to replace the whole external system (the parts of system which we needn't pay attention to), and decrease the rank of the large power system. In the same time, the equivalent system must have the same dynamic influence for study system (the parts of system which requires detailed model) in transient period, just like before the equivalence.Coherency-based equivalence method is the primary method for power system transient stability analysis. It meets the requirements of non-linear system, and can suit the interface of the transient stability program. Coherency-based equivalence method includes two parts: one is coherency identification, the other is parameter identification of coherent generators.This paper mainly discusses parameter identification of coherent generators, especially about how to obtain the accurate parameter of the equivalent generator.The theory of parameter identification of coherent generators is the application of system identification. System identification has two methods: frequency-domain method and time-domain method. To deal with the parameter identification of coherent generators, we chose frequency-domain method. There are many large generators in large power system, and the model is complicated, and the generator equations are higher order differential equations, so it is too difficult to do generator parameter identification by time-domain method. Frequency-domain method only needs to know frequency response, which can be obtained easily. And it is simple to use transmission functions to describe the model of generator.Based on 5th order model of generator, this paper presents the methodof parameter identification of the equivalent generator by frequency-domain method, proposes the transmission functions of synchronous generator, excitation system, speed governor and deduce expressions of parameter estimate. The 5th order model of generator suits the practicality of large power system.This paper proposes some methods to resolve the problems of the algorithm for frequency-domain parameter identification. The author improves the algorithm based on Least Square Estimation (LSE) method, which can consider amplitude-frequency response and phase-frequency response in the same time, and use Gauss-Newton method to solve non-linear least square problem, which improves identification accuracy.At the last of this paper, the author takes the Center China Power System for example to do simulated experiments. The simulation results are given, and they prove that the method which be proposed in this paper is effective: the external system has the same dynamic influence to the study system after equivalence.