Research On Single Neuron Fuzzy Adaptive Control System Of Asynchronous Motor

With the development of computer science and control theories, the AC drive system has gradually become the mainly used electric drive. As the most common used AC motor,induction motor has got widely application and attention. Until now, there are a lot of mature control methods about induction motor speed control system, the vector control method is one of the classical control methods among them. However, the traditional vector control method cannot meet the control requirements of induction motor speed system when it is under different operating conditions due to its complex nonlinear characteristics. Therefore, designing a high performance induction motor speed control scheme based on the vector control is significant.Aiming at the poor adaptive ability of PI controller in the vector control, a single neuron fuzzy adaptive controller is proposed in this paper. The designed controller has combined the advantages of single neuron control and fuzzy control. It not only has the fast learning ability of single neuron control, but also has the effective nonlinear adjusting ability of fuzzy control. To further improve the control performance of the system, a load torque observer and a rotor resistance estimator are also designed in this paper, which are both adopted in the single neuron fuzzy adaptive control system.Firstly, the research overview of induction motor drive system and several commonly adopted control strategies are briefly introduced in the first part. The characteristics, advantages and development status of the neural network and fuzzy control are systematically presented in order to provide a theoretical basis on single neuron fuzzy adaptive controller.Secondly, the mathematical model of induction motor is directly given and the principle of vector control is briefly descripted in this part. Then, an improved single neuron fuzzy adaptive controller is proposed to overcome the shortcomings of vector control. And both the speed PI controller and flux PI controller are replaced by the improved controller. Finally, the simulation results show the superiority of the developed control method.Thirdly, in order to improve the anti-disturbance ability of the system, a reduced order load torque observer is designed based on the reduced order principle. Then combining feed-forward control strategy, using the estimated value of the load torque observer to compensate the expected torque current. The simulation results show that the anti-disturbance ability of the control system is more excellent after compensation.Fourthly, the system performance can be influenced by the change of the motor parameters in addition to the load disturbance, especially the most sensitive rotor resistance. Rotor resistance is related to the field oriented and vector control is established based on the rotor field orientation. Thus rotor resistance estimator is designed to accurately identify the real-time value of rotor resistance.Fifthly, the simulation model of the above designed scheme is established in Matlab/Simulink. By contrast with conventional PI control, the superiority of the proposed scheme is highlighted.The control performance of induction motor can effectively be improved by the single neuron fuzzy adaptive controller through the comparison of simulation results.Combined with the load torque observer and the rotor resistance estimator, the anti-disturbance and the parameter perturbed ability of the system is further enhanced.The overall control scheme is scientific and reasonable, the performance of the system is greatly improved.