Optimized Capacitor Voltage Balancing Algorithm Research For Modular Multilevel Converter

Modular multilevel converter (MMC) has a broad application prospect in the voltage source converter (VSC) based HVDC domain, because of its incomparable technical advantages. However, it requires a large amount of sorting computation, when applying the traditional capacitor voltage balancing strategy based on the sorting algorithm to maintain the capacitor voltage balance with excessive submodules(SMs). An improved strategy based on the grouping method is proposed, which can reduce the computational burden of sorting with the capacitor voltage balance. On the basis of improved strategy, a voltage balancing strategy based on prime factorization with the further improvement of grouping pattern is proposed, which can reduce the computational burden of sorting to the maximum extent. Moreover, the large-scale MMC-HVDC modeling method in the RT-Lab simulation enviroment is investigated, and the feasibility and effectiveness of two improved strategies are validated.First, the basic principles, modulation strategies and control methods of MMC are introduced. The mechanism of the traditional capacitor voltage balancing strategy applying the sorting algorithm is analyzed, and a reduced switching-frequency voltage balancing strategy is proposed.Second, an improved capacitor voltage balancing strategy is put forward. It reduces the number of SMs to sort by dividing the SMs into several groups, and a voltage balancing algorithm among groups is proposed to maintain the voltage balance among groups. As a result, the compuation burden decreases with the capacitor voltage balance. The sorting times is defined to contrast the computation burden of sorting between the traditional strategy and improved strategy quantitatively.Next, the optimization strategy based on the prime factorization is proposed inspired by the prime factorization principle to improve the grouping pattern further. The strategy divides the SMs within each arm based on the prime factorization with the prime factors arranged from the largest to the smallest and the optimized grouping pattern is proved by the mathematical process that the sorting times is the least.Finally, the improved strategy and the optimization strategy based on the prime factorization are both verified based on the 201-level MMC-HVDC in RT-Lab simulation enviroment. The simulation results indicate the two proposed stategies can both reduce the computational burden of sorting dramaticaly with the capacitor voltage balance, which validates the feasibility and effectiveness of two improved strategies.