Model Predictive Direct Current Control of Modular Multi-Level Converters

Modular Multi-level Converters (M2LCs) are mostly controlled by using a hierarchical control scheme, where at least two control loops are required for controlling the load currents and balancing the capacitor voltages. This paper proposes a single controller, which is based on Model Predictive Direct Current Control (MPDCC) with long prediction horizons, to directly control the load currents within tight bounds around their sinusoidal references and balance the capacitor voltages. MPDCC uses a model of the converter for an online optimization process to deliver the best possible performance during both steady-state and transient operating conditions. A conceptual description and control algorithm of the proposed controller are presented in this paper. To validate the proposed concept, simulated performance of a three-phase, three-level 2 MVA grid connected M2LC is presented with a discussion. A comparison with a vector control (VC) pulse width modulation (PWM) scheme is also carried out to demonstrate the improvements in performance associated with the MPDCC scheme.