Study On Ga-Based Fuzzy Neural Excitation Contoller Of Turbine Generator

Excitation controllers of turbine generators are very important and complex equipment of power system. With the larger and larger scale of power system, the problem that the traditional control theories of excitation controllers of turbine generator are not competent as before appears, however, the developing Intelligent Control (IC) theory and updated optimization theory supplies some solutions to this problem.At the beginning of this dissertation, the actuality and trend of the control scheme of excitation control system of turbine generator and soft-compute theory are stated. After briefly introduce the basic Genetic Algorithm(GA) theory, aimming at the "prematurity" of basic Genetic Algorithm, we put forward a new improved Genetic Algorithm, the basic Genetic Algorithm combine Simulate Anneal ing (GASA) , to meliorate the local search ability of basic Genetic Algorithm.because many design problems , such as the preliminary fuzzy rule and input and output membership fuction are hard to gain and the learning process of fuzzy neural network(FNN) is slow and local optimization, we design the fuzzy neural network excitation controllers of turbine generators with Genetic Algorithm combine Simulate Anneal ing (GASA). Firstly, we aquire the initial fuzzy rules filtrate the initial fuzzy rules through Genetic Algorithm combine Simulate Annealing(GASA);then confirm input and output membership fuction through Genetic Algorithm combine Simulate Anneal ing (GASA) ;finally, the fuzzy neural network(FNN) is trained by Genetic Algorithm combine Simulate Anneal ing (GASA). the result of sinmulink proved that:the GA-based fuzzy neural network excitation controllers of turbine generators has excellence dynamic capability.