Study On Hybrid Vector Model Predictive Control For VIENNA Rectifier Optimization Considering Parameter Mismatch

VIENNA rectifier,as a typical three-level converter topology,has the advantages of unit power factor,no dead zone and low current stress.So is widely used in charging piles,electric vehicles and aviation power supplies.As a nonlinear multi objective Control algorithm,finite control set-model predictive control(FCS-MPC)has the advantages of no carrier and fast dynamic response,so it is more suitable for the strongly nonlinear VIENNA rectifier system.However,when the conventional FCS-MPC is oriented to the VIENNA rectifier with unidirectional energy flow,there are problems such as large ripple of grid connected current,difficulty in selecting the weight factor for midpoint potential balance control and poor parameter robustness.Based on this,a hybrid vector model predictive control strategy introducing inductor parameter identification is proposed in this paper.Firstly,the value function with active power and reactive power as constraints is established,and the set of optional voltage vectors in a single cycle is extended by linear synthesis of virtual voltage vectors with basic voltage vectors,which realizes the hybrid control of basic voltage vectors and virtual voltage vectors and improves the problem of large ripple of grid connected current in the traditional model prediction control.Secondly,by using the preferred redundant small vector to participate in the virtual voltage vector synthesis,the problem of difficult weight factor selection for DC side midpoint potential balancing in the traditional FCS-MPC is circumvented,and the midpoint potential balancing control without weight factor is realized.Finally,the model parameters as one of the key fuctors affecting the grid connected performance.In order to further improve the robustness of grid connection performance to model parameters,an inductance parameter identification method based on model reference adaption is proposed by constructing an analytical equation of model parameters and grid connection performance to reveal the mechanism of model parameters’ influence on grid connection performance.In order to verify the correctness of the theoretical analysis and control strategy,a complete VIENNA rectifier experimental platform is constructed,and a comparative experimental analysis between the proposed hybrid vector model prediction and the traditional model prediction control is conducted from transient,dynamic and mid-point equilibrium control dimensions.The results show that the proposed hybrid vector model prediction control has better dynamic and static performance compared with the conventional model control.Meanwhile,when the inductor parameters are mismatched,the hybrid vector model predictive control of VIENNA rectifier with inductor parameter identification still has good grid side current performance,which effectively improves the parameter robustness of the system..