Research On Model Predictive Control Strategy Of Static Var Generator

As a new type of reactive power compensation device,Static Var Generator(SVG)is widely used in power transmission and distribution network because of its good dynamic response performance and continuous compensation.Taking SVG as the research object,this paper discusses the application of model predictive control(MPC)algorithm in SVG.MPC has attracted more and more attention in the field of power electronics and motor drive because of its simple basic principle,fast dynamic response and easy implementation by digital processor.Firstly,the working principle of SVG is briefly introduced,and its mathematical model is established.On this basis,the traditional voltage and current double closed-loop PI control and traditional direct power control(DPC)are studied,and verified by simulation.The results show that the traditional voltage current double closed loop PI control has better steady-state performance,but the dynamic performance is poor,while the traditional DPC is opposite,both of which have obvious defects in the dynamic and static performance.Secondly,the model predictive control(MPC)strategy is studied to overcome the shortcomings of traditional control strategy.The single vector,double vector and three vector MPC are studied respectively,and the comprehensive and detailed analysis and comparison are carried out by combining model predictive current(MPCC)and model predictive direct power(MPDPC)control methods.In addition,this paper proposes a TV-MPC strategy without cost function(CFF-TV-MPC),which reduces the computational complexity and the computational burden of the controller.The simulation model is established in Matlab/simulink,and the results show that CFF-TV-MPDPC has the better steady-state and dynamic performance.Again,in view of the situation that MPC is prone to steady-state error and even affect stability due to parameter mismatch,the reasons for the influence of parameter mismatch on stability are analyzed in detail.Taking TV-MPC(Including MPCC and MPDPC)as an example,a discrete Luenberger disturbance observer is designed to estimate the disturbance caused by system parameter mismatch,and real-time compensation is used to reduce steady-state error.The design idea of the disturbance observer and the setting method of related parameters is given.Simulation results show that the disturbance observer can significantly enhance the robustness of the proposed MPC strategy.Finally,the basic hardware experimental platform of SVG is built and the software algorithm based on TMS320F28335 are designed,and the dynamic and static performance of the system is studied.The experimental results show that the proposed control strategy is correct and feasible.