Research On Model Predictive Control Optimization Of Two-stage Matrix Converter

As an AC-AC power converter,matrix converter has received widespread attention because of its good input and output characteristics,high power density,and two-way energy flow.Compared with the traditional direct matrix converter,the twostage matrix converter has a more compact size and a simpler commutation mode,and has become a more promising AC power converter,and its control strategy has been widely studied by scholars at home and abroad.Compared with the traditional control strategy,the model predictive control has the advantages of fast dynamic response and good robustness,and it does not need to adjust complex controller parameters in the control process,and can control multiple control targets only through a single closed loop.Therefore,this thesis takes the two-stage matrix transformer as the research object to further optimize the model predictive control strategy applied to this topology.This thesis first introduces the circuit topology of the two-stage matrix converter,establishes a mathematical model according to its topology,and then analyzes the inputoutput relationship between the rectifier stage and the inverter stage.On this basis,the dual-space vector modulation strategy of two-stage matrix converter is introduced from the three parts of rectifier-stage modulation,inverter-stage modulation and two-stage coordinated control,and lays a foundation for introducing the predictive control principle of two-stage matrix converter model.Aiming at the problem of large number of sensors in the predictive control of twostage matrix converter,this thesis designs a model predictive control strategy for gridside voltage sensor.Based on the virtual flux control technology,the grid-side voltage is reconstructed,and the modulation idea is introduced to optimize the predictive control of the traditional finite set model,and the simulation comparison analysis with the traditional modulation model predictive control verifies that the proposed strategy can still realize the control performance of the traditional modulation model predictive control under the premise of reducing three grid-side voltage sensors.In order to further optimize the steady-state performance of the traditional modulation model predictive control of the two-stage matrix converter,a model predictive control strategy for the optimal switching sequence is proposed,and the proposed strategy can achieve better input and output characteristics without affecting the dynamics of the system.By comparing with the simulation of predictive control of traditional modulation model,the effectiveness of the proposed strategy is verified.Finally,the corresponding experimental platform is designed for the model predictive control of the two-stage matrix transformer,and the experimental results verify the effectiveness and feasibility of the above research.