Research On Three-phase Voltage Type PWM Rectifier Based On Model Predictive Control

With the development of economy,the proportion of electric energy in social energy consumption gradually increases,and the requirements for voltage,frequency,current distortion rate and stability of electric energy were gradually raised.As an important converter device,the rectifier exchanges energy with the power grid frequently.The harmonic and reactive power loss generated by rectifier will reduce the power consumption efficiency and power quality.Therefore,reducing the pollution of the rectifier device to the power grid has become a hot topic.PWM rectifier has been widely used in various electric energy conversion occasions due to its superior current conversion ability,and it is a research hotspot in the field of power electronics.This thesis takes the three-phase two-level voltage source PWM rectifier as the research object,and studies its control strategy in detail.Firstly,according to the working principle of three-phase two-level voltage-type PWM rectifier,the mathematical model is established and the model predictive control algorithm is introduced.According to the working principle and characteristics of the model prediction algorithm,the influence of the number of prediction steps on the control effect is analyzed and compared.Aiming at the delay problem in the model predictive control algorithm,the impact of the delay error on the system control is analyzed,and a two-step predictive compensation method is used to compensate the delay error,which can effectively reduce the current harmonics on the grid side.Secondly,the optimization process of the multi-step model predictive control is a mixed integer optimization problem.When the number of prediction steps increases,the solution time will increase exponentially with the increase of the number of integer optimization variables.Therefore,it is necessary to find an optimization method to improve the solution efficiency.As an important algorithm for solving mixed integer optimization,branch-bound method can reduce the number of alternative items in the process of solving.In this thesis,the spherical decoding algorithm in the branch definition method is used to solve the optimization process of multi-step model prediction,which reduces the number of optimization times in the rolling optimization process,greatly reduces the amount of computation and improves the efficiency of the processor.Through the derivation of the mathematical model,the mathematical form of the value function is transformed to adapt to the solution form of the spherical decoding algorithm.In addition,by combining the spherical decoding control algorithm with the two-step predictive compensation method,the current quality of the threephase voltage-type PWM rectifier can be improved significantly,and more steps can be achieved while reducing the switching frequency.Finally,the simulation model is constructed in Matlab/Simulink,and the experimental platform is built.The hardware circuit and software design are introduced in detail,and the effectiveness and feasibility of the control strategy proposed in this thesis are verified through simulation and experimental analysis.This thesis includes 52 figures,2 tables and 100 references.