Research On Model Predictive Control Of Flux-Switching Permanent Magnet Motors

With the advantages of high efficiency,large torque and high power density,permanent magnet(PM)synchronous motor(PMSM)has attracted considerable attentions in the field of new energy vehicle.As a special category of stator-PM brushless motor,flux-switching PM(FSPM)motor is one of research hotspots in motor field over the recent two decades due to the merits of simple structure,suitable for highspeed operation and convenient heat dissipations.To improve the control performance and extend the application,model predictive control(MPC)of FSPM motor is deeply and systematically investigated in this thesis,including basic theory of MPC,normal control of three-and five-phase FSPM motors,faulttolerant control of five-phase motor under single-phase open-circuit fault,and torque ripple reduction,etc..The main research contents and achievements of this thesis are as follows:1.The work principle,structural characteristic and mathematical modelling of a classic three-phase FSPM motor is introduced,which is employed as the research object.Two conventional MPC methods,namely,model predictive current control(MPCC)and model predictive torque control(MPTC),are investigated based on two-phase synchronous rotating coordinate.The inherent link between the two MPC schemes is also deduced by analyzing predictive model and cost function.Simulations and experiments are conducted to verify the two MPC methods.2.In order to alleviate the torque ripple of MPC-controlled FSPM motor,different strategies are proposed from the perspectives of control theory.Firstly,to reduce the impact of cogging torque on electromagnetic torque,harmonic current injection and iterative learning control methods are proposed on the basis of torque generation.Secondly,a vector synthesis method is proposed in conjunction with MPTC to improve the stead-state behavior due to the traditional application of only single voltage vector in each control period.The output pulse sequence of pulse-width-modulation(PWM)is optimized by replacing only single vector with multiple vectors.Consequently,the harmonic current and torque ripple are decreased.Simulated and experimental tests are carried out to validate effectiveness of proposed methods.3.The MPC of multi-phase drive is investigated based on a five-phase FSPM motor.With the vector space decomposition(VSD),a fully-decoupled mathematical model is built.Accordingly,a MPCC scheme considering the regulations of both fundamental and harmonic subspaces are designed.The selection of basic voltage vectors is discussed in detail.To suppress harmonic voltage vector,the concept of virtual voltage vector is developed by synthesizing large and medium vectors.Further,the duty cycle control technique is employed to optimize the duration of virtual vector.As a result,the steady-state performance of MPCC is enhanced for five-phase FSPM motor under normal operation.Simulations and experiments are performed to verify the developed schemes.4.The fault-tolerance of a five-phase FSPM motor under open-circuit fault(OCF)condition is investigated based on MPC technique.A fault-tolerant MPTC algorithm with harmonic suppression and voltage selection is developed.To achieve the complete regulation of both subspaces,torque,stator flux amplitude and harmonic currents are adapted as control variables to construct predictive model and cost functions.The weighting factor between torque and stator flux amplitude is eliminated by investigating the relationship between two variables.For the computational cost reduction,the stator flux vector is used to pre-select voltage vector candidates.Finally,the fault-tolerant control of the faulty drive is implemented with an increased sampling frequency.5.On the basis of space vector modulation technique,a novel space vector pulse width modulation(SVPWM)method is proposed and applied together with fault-tolerant MPTC scheme for a five-phase FSPM motor under OCF.The voltage vectors in the quadrant where the referenced voltage vector locates are selected as candidates for cost function optimization.With this manner,not only the control of fundamental and harmonic subspaces is realized but also the tracking error of voltage vector is minimized.Consequently,the fault-tolerant performance of the five-phase FSPM motor drive under OCF operation is improved.6.In order to validate proposed control schemes,two digital drive and control experimental setups,which are based on the real-time d SPACE 1104 and DSP-TMS320F28335,are built for three-and fivephase FSPM motors,respectively.Experimental results have verified the effectiveness of proposed methods,providing theoretical basis and experience guidance for further study in FSPM motor and other PMSM drives.