The Nonlinear Parameter Identification Of Excitation System Based On Artificial Intelligence

The accuracy of the parameters of excitation system model has significant influence on the analysis and control of power system. The conventional method is the linearization approach to identify the linear parameters of the excitation system only, in fact, the excitation system is nonlinear, this paper adopts two methods of neural network and genetic algorithm in artificial intelligence field to identify the nonlinear excitation system. Using MATLAB as the simulation tool, first establish the excitation and the power system model, adopt a step signal to the excitation system, and then collect corresponding data to identify the model parameter. The results prove that, the neural network trained by rich sample data can simulate the output of the original excitation system. The genetic algorithm may be used either each block of the excitation system is identified separately or the overall block is identified once according to the measurable data. The former has a better effect than the latter, however, the latter identified model can be used to simulate all the same. Separate the block as much as possible according specific conditions to identify may has better identification effect.