Research On Optimal Control Strategy Of The Modular Multilevel Converter

With the increasing application of power electronics in the field of reactive power compensation, active power filter and high voltage direct current transmission systems, high voltage power electronic equipment has become hot research topic. It is the most effective method for multilevel converter to solve the high voltage and high power conversion at present because of its advantages such as low voltages on switch devices, low harmonics in output voltage. Based on the modular multilevel converter, this paper aims to research on the modulation strategy, capacitor voltage balancing method and the optimal control strategy.Firstly, the topology and working principle of MMC are introduced, and the mathematical model of MMC is deduced. The carrier phase shift modulation strategy and the nearest level modulation strategy are analyzed. Then the simulation results are verified. Based on the theoretical calculation and simulation verification, the advantages and disadvantages of the corresponding modulation strategy and its application range are pointed out from the aspects of the output harmonic content, fundamental wave characteristics and dynamic characteristics.In order to reduce the low order harmonic content of the nearest level modulation when the number of submodule is little, the submodule unified pulse width modulation is used. However, this method still uses the traditional sorting algorithm, which does not consider reducing the switching frequency and the heavy burden of the controller. In order to further optimize the algorithm, this paper adopt the optimization method based on submodule unified pulse width modulation. This method neither requires the individual sorting of the SM capacitor voltage nor causes unnecessary switching actions.In order to keep the capacitor voltage balanced and suppress the circulation, the conventional control methods are generally combined with the PI controller. It causes the control system to have the difficulty of setting the parameter, the control delay and so on. This paper introduces the model predictive control. This method has the advantages of fast response, good tracking performance, multi-variable control and so forth. However, there is a large amount of computation to the controller. This paper used the optimization model predictive control. Through the sub-module capacitor voltage sorting and current direction to determine the possible combination of switch states, which can effectively reduce the possible combination of switch states and the computational complexity of the system. Both these two methods need to be carried out in the process of setting up a complex weighting factor. So this paper proposes satisfying-optimized predictive control based on MMC. Simulation results show that SOPC has good steady and dynamic characteristics.Finally, the optimal control methods used in this paper are verified in the MMC experimental prototype, and the experimental results show that the proposed methods are correct and effective.