Application of neural networks and genetic algorithm in knowledge acquisition of fuzzy control system

Often it is difficult to acquire the requisite knowledge in the design of rule-based control systems. In the study, an advanced knowledge acquisition technique was used to optimize fuzzy controller in the control of hydroelectric generating unit system of hydropower plant. The designed control system may select optimal scale factors, membership functions and control rules efficiently by using neural networks and genetic algorithm. Genetic algorithm was used to optimize the parameters and rules of fuzzy controller in operating. Dynamic identification model of control system based on RBF neural networks was designed to appraise the controlling performance of fuzzy controller. Simulation results show that the advanced knowledge acquisition technique makes scale factors, membership functions and control rules of fuzzy controller arrive to optimization and its control performance is superior to conventional controller.