Design Of Fuzzy Controller Based On Genetic Algorithms And Application

The proper selection of fuzzy logic rules and fuzzy control membership functions, which determines the dynamic and static performance and control effect, is the key to design a fuzzy controller. To the deficiency of the traditional method, an algorithm is proposed to optimize fuzzy controller, which based on Genetic Algorithm (GA). The simulation results in MATLAB show that the algorithm is effective. Finally, an intelligent charging system that adopts such improved fuzzy controller is designed in this thesis.Firstly, fuzzy control and GA is introduced and discussed, such as the mach base of fuzzy control, the design of conventional fuzzy controller, the basic operation of GA and schemata theory and so on. Then, to avoid limitation of fuzzy controller, Simple Genetic Algorithm improved in following several aspects: a new method based on decimal-code is provide, which unites the membership functions and fuzzy logic rules into integrated one for whole-scope searching;the importance of the crossover and mutation operators is fully considered;in different phase of the evolution, each chromosome's possibilities of crossover and mutation can be changed. The result of simulation for second-order system indicates that fuzzy controller based on GA has better dynamic and static characteristic. Lastly, the design of the intelligent charging system is discussed in full detail. How to realize the power factor correction technology, the two-transistor forward converter of the system and the control circuit is introduced. The software design of the system is presented also, including the signal sampling and processing module, the fuzzy controller based on GA and charge termination module. A series of experiments have demonstrated that this system enjoys charging of efficiency and reliability. The temperature of battery is lower and the life of battery is longer.