Model predictive control (MPC) methods are popularly employed in modular multilevel converters (MMCs) due to their fast dynamic response and multi-objective control capability. However, they present some inherent problems, such as computational complexity, variable switching frequency, poor steady-state performance and tedious weighting factor selection. This paper develops a deadbeat predictive current control method for MMCs. This method can realize the reference tracking of ac current and circulating current in one sampling period without error, and thus provide a fast dynamic response as conventional MPC methods. Besides, switching state or voltage level evaluation, cost function calculation as well as weighting factor selection are not required. Therefore, it has a very low calculation burden, which is independent of the number of submodules (SMs). Since a modulation stage is utilized, a fixed switching frequency and consequently a satisfactory steady-state performance are obtained. The effects of time delay, parameter mismatch and SM capacitor voltage ripple on the control algorithm are discussed. Also, the corresponding improvement measures are provided to further enhance the steady-state performance and system stability of MMCs. The effectiveness and performance of the developed control algorithm are verified by experimental results.

The increasing penetration of renewable energy resources facilitates the carbon footprint reduction process yet reduces the power system inertia. As a result, the grid frequency and the rate of change of frequency (RoCoF) might probably go beyond the normal range, resulting in unexpected load shedding, generator tripping, and even frequency instability. To address this problem, grid-connected inverters are designed to participate in frequency regulation and provide the equivalent inertial support. Nevertheless, the inertia emulation effect is affected by the inverter synchronization dynamic and high RoCoF events may occur as the result of poor synchronization dynamics. In view of this limitation, a synthetic inertia control is developed in this article considering the synchronization dynamics. The synthetic inertia principles and control design guideline are explicitly provided. Finally, hardware experimental results of a scaled-down power system prototype are provided to validate the effectiveness of the proposed approach.

Grid synchronization stability is an important research topic for grid-forming converters. To prevent instability and undesired oscillations, the small-signal modeling and stability analysis must be performed, which requires the line impedance information. Nevertheless, the line impedance may be unknown in most cases. As a consequence, the designed control parameters may fail to satisfy the stability requirement. To fill in this research gap, this article presents a geometrical approach to evaluate and improve the small-signal synchronization stability of grid-forming converters. The proposed analytical approach maps the line impedance variation on a two-dimensional plane and identifies the stability boundary in a geometrical way. As a benefit, the controller of grid-forming converters can be explicitly designed to achieve the robust synchronization stability. Finally, the feasibility of the proposed technique is verified by simulation and hardware experimental results.

Symmetrical ac-side current control is widely used for grid-connected modular multilevel converters (MMCs) under grid fault conditions. However, it is revealed that this control method can lead to unbalanced arm currents and consequently asymmetrical electrical stresses and non-uniform temperature distributions. This paper aims to analyze the mechanism of the unbalanced arm currents and propose an effective method to solve this problem. Firstly, the electrical quantities inside MMCs are analyzed under unbalanced grid fault conditions. The analytical expressions of the unbalanced arm currents are derived considering different fault severities and operating conditions to reveal the “side-effects” of the symmetrical ac-side current control. After that, a zero-sequence ac voltage injection method is proposed. The injected voltage is directly derived from negative sequence ac current controller and ac-side current with pure mathematical manipulations, which can be easily obtained without extra controllers. The effect of the injected voltage on the original modulation index is very limited according to an accurate quantitative analyzation. Moreover, the proposed method does not affect the already achieved symmetrical ac-side current control and the circulating current suppression. The effectiveness and performance of the proposed analysis and control method are verified by both simulation and experimental results.