Research On Flux Adjusting Strategy Of Memory Machine Based On Active Disturbance Rejection Control

Permanent magnet synchronous machine(PMSM)is widely used in the field of new energy electric vehicles due to its simple structure,high efficiency and power density.However,the flux weakening control in the high-speed region leads to a serious decline in the overall efficiency.As a new kind of PMSM,variable flux memory machine(VFMM)can apply instantaneous current pulse to flexibly adjust the magnetization state(MS)of permanent magnet(PM),which can realize flux weakening with almost no loss,and greatly broaden the speed regulation range and high efficiency operation area.In this thesis,the flux adjusting technology of VFMM with d-axis current pulse is mainly studied.Firstly,based on the mathematical model and operation principle of VFMM,a method for obtaining the inductances and flux linkage models is proposed,and a hybrid-magnetic-circuit VFMM(HMC-VFMM)is used as an example to build the simulation model.Secondly,a linear extended state observers(LESO)is designed based on the mathematical model of VFMM,and a linear active-disturbance-rejection-based feed-forward current controller(LADR-FFCC)with superior performance and strong robustness is further proposed.The LESO can accurately estimate and in advance compensate the disturbance voltage caused by the variation of the PM flux linkage and machine parameters during the MS manipulation,which effectively improve the control accuracy and response speed of magnetization current.Then,a high-precision PM flux linkage estimation method is proposed based on the LESO.The d-axis inductance is estimated by using the recursive least square method,and the feedforward voltage is calculated to compensate the disturbance voltage caused by the change of the inductance parameter,reducing the estimation error of the steady-state PM flux linkage.Finally,a magnetization strategy for VFMM is designed,and an HMC-VFMM is taken as an example to perform experimental verification on the drive system designed based on STM32H743 and dual-in-line package intelligent power module(DIPIPM).The simulation and experimental results show that the proposed LADR-FFCC has the advantages of fast response,high control accuracy,simple parameter setting,strong robustness and easy implementation.The proposed flux linkage estimation method has fast convergence,and the steady-state error of PM flux linkage estimation can be maintained within 3% in the effective operation region of the machine.In addition,the proposed magnetization strategy for VFMM can successfully realize the high-precision on-line regulation of PM flux linkage,and effectively prevent the reversible demagnetization of low coercive force PM.