A one-step model predictive control for modular multilevel converters

Pulse Width Modulation (PWM) and Model Predictive Control (MPC) are two switching methods for Modular Mutlilevel Converter (MMC) presented in the literature [1], [2]. PWM switching cannot mitigate circulating currents in the converter bridges. An MPC method proposed in [2] compares all switching sequences and finds the optimal one to minimize the objective function (a weighted sum of the absolute values of the ac current deviation, capacitor voltage deviation, and the circulating current of the next time step). The computation effort is significant when the numbers of submodules increase. In this paper, a one-step model predictive control with minimum computing effort has been proposed to control the ac currents, keep the capacitor voltages nominal, and mitigate the circulating currents. The algorithm determines the ideal values of the lower-level and upper-level voltages on each phase based on the current state. The algorithm also determines which submodules to be switched on/off for the next time step according to the corresponding upper/lower current. The performance of the proposed method is evaluated via simulation in Matlab SimPowerSystems. The PWM and the proposed MPC are compared for their control effort and performance.