New Model Reference Adaptive Multi-parameter Identification Sensorless Control System For Permanent Magnet Synchronous Motor

For the closed-loop system of permanent magnet synchronous motors,sensorless control and a double closed-loop structure are the key to the control performance of permanent magnet synchronous motors.Among the mainstream sensorless control methods,the use of model reference adaptive systems for identifying speed and rotor position has the advantages of high estimation accuracy and structural simplicity,and is therefore widely used.When there is disturbance,due to the existence of PI structure in the speed adaptive law designed by the model reference adaptive system,the permanent magnet synchronous motor will not only lead to the decline of speed identification accuracy,but also have some inherent problems,such as poor robustness,weak anti-interference ability and poor dynamic tracking,because the current loop and speed loop of the traditional double closed-loop structure adopt PI control method.In order to improve the above problems caused by PI structure control,the specific work content of this thesis is as follows:(1)The PI control for the traditional speed loop has the problems of poor robustness and weak anti-interference ability.Firstly,the basis of sliding mode variable structure control is introduced.Although sliding mode variable structure has the characteristics of strong robustness and good anti-interference ability,the inherent problem of sliding mode variable structure is chattering.In order to improve this problem,the traditional sliding mode speed controller is improved by using variable exponential reaching law.Finally,a sliding mode speed controller based on a variable exponential reaching law is used to replace the conventional PI speed loop,laying the foundation for subsequent sections to expand the simulation verification and improve the speed adaptive law of the traditional model reference adaptive system with a high-order sliding mode algorithm;(2)The PI structure of the speed adaptive law in the traditional model reference adaptive system will lead to the problems of reduced speed identification accuracy and poor robustness.Firstly,the basic theory of the model reference adaptive system is introduced,and based on this,the speed adaptive law is derived.Then,the sliding mode variable structure is used to improve the PI structure of the speed adaptive law.In order to weaken the chattering of sliding mode variable structure,furthermore,the super-twisting sliding mode algorithm in the high-order sliding mode algorithm is proposed to improve the PI structure of the speed adaptive law in the traditional model reference adaptive system.Finally,a MATLAB/Simulink simulation platform is built to verify that the combination of the super-twisting sliding mode model reference adaptive system sensorless control and a sliding mode speed controller based on a variable exponential reaching law improves the speed identification accuracy,robustness and anti-interference ability;(3)The PI control for traditional current loop has the problem of poor dynamic tracking performance.Firstly,the basic principle and implementation process of the deadbeat predictive current control are introduced.At the same time,the influence of the stator resistance,stator inductance and rotor flux mismatch on the deadbeat predictive current controller is deduced theoretically,and the stable range of the deadbeat predictive current controller is deduced according to the closed-loop pole and discrete system performance analysis theory.Then,the parameter identification method based on model reference adaptive system is used to improve the deadbeat predictive current controller.Because the current state equation is two-dimensional,the identification of three parameters at the same time will have the situation of under-rank,thus reducing the accuracy of parameter identification.Therefore,the parameter stepwise identification method based on the new model reference adaptive system is proposed to solve this problem.Then,the parameter stepwise identification method based on the new model reference adaptive system is proposed to improve the deadbeat predictive current controller,and the improved deadbeat predictive current controller is used to replace the traditional PI current loop.Finally,an experimental platform is built to verify that the deadbeat predictive current controller based on the new model reference adaptive system parameter stepwise identification method can effectively improve the dynamic tracking performance.