Variable Period Multi-Vector Robust Predictive Current Control For Permanent Magnet Synchronous Motor

As global warming leads to increased climate instability,the goal of "3060" was presented.As the most widely used secondary energy source in the world,electricity is an important bridge to achieve "carbon peaking" and "carbon neutral".On the power generation side,wind power,photovoltaic and other new energy sources account for an increasing proportion;on the electricity consumption side,electric vehicles and other products use electricity to replace traditional primary energy to power the whole system.The use of electricity is inseparable from the motor,and an efficient and energy-saving motor control system can improve the efficiency of energy utilization.In this thesis,the steady-state performance and robustness improvement of the predictive current control system for permanent magnet synchronous motor are studied,and the basic principles of different control methods are analyzed and their control properties are compared.Firstly,to address the problems of large torque ripple and poor steady-state performance of traditional model predictive current control,based on the analysis of its advantages and disadvantages,the concept of variable period is introduced,and the methods of variable period single vector model predictive current control and variable period two vector model predictive current control are proposed.By extending the freedom of control period and obtaining the optimal control period based on the principle of current error minimization,the steady-state performance of the system is effectively improved without increasing the switching frequency.Secondly,to address the problem of relatively large current harmonics at high modulation ratios such as high speed and heavy load in conventional deadbeat predictive current control with a single 7-segment vector sequence,a variable-sequence SVM deadbeat predictive current control strategy is proposed based on the analytical derivation of current harmonics of four voltage vector sequences.This method online selects the optimal voltage vector sequence with minimal current harmonics,which is simple and effectively reduces the current harmonics.Compared with the conventional deadbeat predictive predictive current control based on the fixed vector sequence,the current harmonics and total harmonic distortion of the proposed method are minimized at full modulation indices.Finally,to address the problem of poor robustness in predictive current control,an ultra-local model is introduced in this thesis.Consider control period variation and calculate the input voltage gain and the unknown part of the system by using the sampled currents and stored voltages of the past two control periods,which achieves the estimation of the input voltage gain and the unknown part of the system without relying on any motor parameters.Then,base on the estimated input voltage gain and the unknown part of the system to predictive current and it is combined with the model predictive current control and deadbeat predictive current control mentioned earlier in this thesis to achieve robust control of the whole motor system.In addition to the theoretical verification,the proposed method is verified in this paper by Matlab/Simulink and experimental platform based on DSP TM320F28335.Its results prove the effectiveness of the proposed method.