Fuzzy Optimization Control Of Oil Feeding Pump System Based Genetic Algorithms

Oil Feeding Pump System is a complex system, which exits serious non-linearity > time-delay,strong couple and some unascertained factors, so it can't achieve better results by traditional control methods. The paper analyzes the main reason of causing system oscillation when it is adjusted by original PID control method, and proposes control strategy of combining basic value of rotation rate with adjustment value to control pump rotation speed according to non-linear characteristic of the system and control request. In the control strategy, the basic value of rotation rate is calculated by corrected similar formulation and the adjustment value is ascertained by output of fuzzy controllers, the strategy can assure steady precision and to eliminate vibration. In order to realize simply and conveniently, two two-dimensional fuzzy controllers are designed, and the outputs of the controllers is used as speed adjustment, but the'steady precision and anti-interference ability of this fuzzy control method are not desirable.There exits many subjective factors during designing the fuzzy control, in order to make fuzzy control rules and membership functions correction ., perfection and adjustment according to system properties, it combines genetic algorithms with fuzzy control, detailed analyzes the problem of designing fuzzy controller and proposes two advanced schemes:First scheme: The change-of-variables are emerged into input variables of the simple fuzzy controllers of Oil Feeding Pump System as one variable, and one PI block is connected after output of fuzzy controllers, consequently the structure of the improved fuzzy controller is analyzed, finally genetic algorithms with adaptive probabilities of crossover and mutation is applied to optimize membership functions and fusing factors of the fuzzy controllers , and the simulation results of before and after optimization are compared.Second scheme: Fuzzy neural networks controller is constructed by employing the great structure knowledge description ability of fuzzy logic and the great studying ability of neural networks to tune structure and parameters of membership functions, and extracts effective fuzzy control rules from data. The paper designs fuzzy neural networks controller based on Sugeno fuzzy model and applies the improved genetic algorithms to tune the networks parameters and membership functions of the designed fuzzy neural networks controller.According to simulation research of two control strategies, the simulation results show that the improved fuzzy controllers have improved control and achieve desired control results of Oil Feeding Pump System. The paper adopts the first advance scheme to realize the fuzzy optimization control of Oil Feeding Pump System and eliminates the turbulence generated by changing control ways. The experiment results demonstrate that the fuzzy control method realizes coordination optimization control of technique operation parameters of the system, what's more , the system has the characteristics of fast adjusting speed, high steady precise, good steady ability and has the better adaptive ability and robustness.