A modified indirect model predictive control for modular multilevel converters

Modular multilevel converter (MMC) is among the promising topologies used in high-voltage/power purposes, especially for high-voltage direct-current (HVDC) transmission systems. In this paper, a modified model predictive control (MPC) strategy for an MMC-HVDC system is presented. In the proposed MPC strategy, output current, circulating current and capacitors voltage are controlled separately. In order to reduce the computational burden of the controller, the output current is controlled directly without any optimization. For suppression of the circulating current, capacitors voltage variations are employed to settle the converter arms voltages to their reference values. In addition, the submodules (SMs) capacitor voltages are balanced by using a novel voltage sorting algorithm. In this method, the conventional voltage sorting algorithm is combined with the MPC-based strategy to determine which SMs should be turned on. SM capacitors are sorted based on an expression including their predicted capacitor voltage and the transition between the present and next states. Employing the SM transition states makes it possible to control the converter switching frequency. Performance of the proposed control strategy is verified using MATLAB/Simulink environment. The obtained results confirm the effectiveness of the proposed strategy in confining the converter switching frequency and reducing the circulating current and the controller computational burden.