Research On Control Strategy For The Two-stage Matrix Converter

In Comparison with the conventional AC-DC-AC converters,the advantages of the satisfied input-output characteristics and bidirectional energy flow are reserved for the matrix converters(MCs),in addition,the power density of the MC has been increased due to omit the intermediate energy storage link.Compared with the direct matrix converter(DMC),the modulation strategy and commutation mode of the twostage matrix converter(TSMC)are simplified to a certain extent due to its independence of the rectifier stage and the inverter stage in physical.On the other hand,the existence of virtual DC bus makes the output of multi-inverters possible,thus,it is more suitable for multi-drive systems.TSMC is taken as the research object in this paper,the grid side power factor control and model predictive control are studied respectively.The chapter 2 of this thesis starts with the basic principle of TSMC,the switching function matrix between input and output is established based on its topologica l structure.On this basis,the double space vector modulation strategy is introduced,the rectifier stage,the inverter stage and the coordinated control among two stages are elaborated in detail.Finally,a low-power experimental platform of TSMC is built,and its hardware structure and control system architecture are briefly introduced.The chapter 3 is aimed at the problem that the power factor on the network side is susceptible to the influence of system parameters,the mathematical model of the system is established,and based on this,the power factor on the network side is analyzed in detail.In order to solve the problem of phase shift of the grid side current caused by input filter,a low-complexity input current control method is proposed,which reconstructs the voltage and current of grid side and makes closed-loop tracking between different physical quantities possible.The closed-loop control structure based on ?? static coordinate system is relatively simple and does not need design the complex voltage phase-locked loop.The algorithm complexity is reduced to a certain extent,while the satisfied control performance can be obtained.Aiming at the problem that the power factor of the network side is not equal to 1 under the three-phase input balance condition,a low-cost phase angle compensatio n method is proposed.In comparison with the existing algorithms,the phase difference between three-phase synthetic vectors is reflected by the phase difference between single-phase,only one current sensor is needed to collect single-phase information,thus the hardware cost of the system can be reduced.Finally,the related technologies of model prediction have been studied,the switching model and input-output prediction model of TSMC are established.In order to reduce the computational complexity of the existing modulated model predictive control(M2PC)algorithm,a hybrid modulated model prediction control(HM2PC)is proposed.The satisfied dynamic and steady-state performance can be reserved with the proposed HM2 PC,moreover,the computational complexity of the whole control system can be reduced to a certain extent due to the corresponding optimization of the rectifier stage and the inverter stage prediction algorithm.The effectiveness of the research contents above all are verified by the simulation and experiment.