The Adaptive Fuzzy Control Of Brushless DC Motor Based On Genetic Algorithm

The brushless DC motor has a wide range of applications in the industrial fieldbecause of its small size, high power density, simple structure and good speedperformance advantages. Using intelligent control method which can make the systemretain high precision and stability is a hot spot in recent years.Brushless DC motor is a multi-variable, strong coupling nonlinear system, sotraditional PID control method is difficult to precisely control the system. Adaptivecontrol algorithm can identify parameter and estimate motor status, which are based onlinear model. It is difficult to satisfy the accuracy and real-time motor control. Fuzzycontrol is not required to grasp the exact model of the controlled object, but theselection of fuzzy rule and determination of membership function are still insufficient.Genetic algorithm is a global search algorithm. It can efficiently find the globaloptimal solution or near optimal solution, avoid falling into local optimal solution. Inthis paper, genetic algorithm, fuzzy control, PID control and adaptive control arecombined; a fuzzy adaptive PID controller based on genetic algorithm is designed.This paper briefly describes the basic structure and work principle of BLDCM,analyzes its mathematical model, then establishes double closed loop control system ofBLDCM. The fuzzy adaptive PID controller based on genetic algorithm is applied toouter ring speed ring, and PI controller is applied to inner ring. The simulation modelof double closed loop control for BLDCM is set up in Matlab2010b/Simulinkenvironment. It includes motor body’s main circuit module, current sampling module,logic commutation module, speed loop module as well as current PI controller module.The simulation results show that, the system which uses the fuzzy controller based ongenetic algorithm optimization has a short rise time and no overshoot, small steady-state error and other advantages. And the system has strong robustness and adaptability.In addition, this paper proposes two optimization methods. One method is that genetic algorithm optimizes the control rules and membership functions at the same time. Theother method is that genetic algorithm optimizes the control rules and membershipfunctions step by step. Both of these two optimization methods can improve theperformance of BLDCM. But the first method can obtain the membership functionsand control rules on the overall optimal, the optimization time is long. The secondmethod reduces the search time.