Research On Optimal Control Strategy Of The Modular Multilevel Rectifier

With the increasing demand of electric energy in modern civilization,power conversion device has become ever more important.The converter has experienced the development stage of the two level to multilevel.In comparison with the two level rectifier,the multilevel rectifier has better fundamental characteristics,lower switching frequency and lower device voltage requirements.The modular multilevel converter is a relatively innovative multilevel topology.Due to its modular design method,easy power and voltage level expansion,it has attracted the interest of domestic and foreign scholars.Based on the modular multilevel rectifier,this paper aims to research on the operation principles in the ideal and non-ideal gird and puts forward some suggestions to promote the system control performance.It's the premise for the stable operation of MMC to charge submodule capacitor voltage to the desired value.In this paper,submodule capacitor voltage expression is derived.Based on this,the reason that submodule capacitor voltage cannot be charged to rated voltage is stated.In addition,a new selective charging method is put forward.In this paper,an optimal model predictive control strategy is put forward.In this algorithm,the optimal insert number of submodules can be obtained without the support of the cost function.So the problems of parameter adjustment of weight factors and complete trial-and-error do not need to be taken into account.It reduces in large part the complexity and computation of the controller.Based on the MMR mathematical model,the relationship between the arm voltage and the phase current,circulating current is derived.The number and duty cycle of submodule that would be into the system in the next time was obtained by using nearest level control or submodule unified pulse width modulated.In addition,it can achieve a stable submodule capacitor voltage by using submodule capacitor voltage sorting method.The mathematical model of MMR with ideal grid is established,and the design method of controller parameters for traditional double closed loop control strategy is studied.Instead of the traditional current loop feed forward decoupling controller,optimized model predictive control reduces the complexity of the current inner loop controller and improves the dynamic characteristics of the current inner loop.The mathematical model of MMR with non-ideal grid is established.This paper introduces the separation method of positive and negative sequence.A method based on optimal model predictive control is put forward to reduce DC side active power fluctuation under non-ideal grid or to reduce AC current imbalance under non-ideal grid.The effectiveness of these two methods is verified by results of simulation and experiment.