An improved method to sub‐module voltage balancing in modular multilevel converters with two voltage sensors

Abstract

Summary The modular multilevel converter (MMC) is an innovative structure that has expanded the usage of multilevel converters (MLCs) in the electrical industry. Since the MMC initially required a large number of sensors to manage the capacitor voltages, various studies have been done to reduce their number to improve the converter's reliability and reduce construction costs. In this paper's proposed method, merely two voltage sensors are used to balance the converter's capacitor voltages, and voltages are estimated using the exponentially weighted recursive least square (RLS) method in each sampling time. Also, the estimated voltages are sorted according to the capacitor current sign determined by the voltage sensors, and the current sensors are entirely eliminated. Furthermore, three cases for compensating the capacitor estimated voltages are presented, and the last case proposes the sub‐module force algorithm (SFA) for system protection. The efficacy of the proposed method has been assessed via simulation and experimental implementations.

A modified modular multilevel converter to reduce the second order ripples in the submodule capacitor voltage: Design and analysis

“…4 Depending on the application, MMCs structure is constructed by utilizing SMs such as half-bridge and full-bridge SMs and also many other SMs are preferred. 5 However, MMC's structure faces ripples in the SMs capacitor voltage because of the grid unbalance 6,7 low-frequency operation of the variable frequency motor drive 8,9 and unequal switching sequence for the SM's switches. 10 Interestingly, amplitude of load current and the load frequency of MMC are directly proportional to the magnitude of the ripples.…”

Arm‐current sensorless model predictive control for grid‐interfacing modular multilevel converter with reduced switching frequency

“…2 Numerous control techniques for the MMCs have been reported in the literatures, such as classical linear control, sliding mode control, and model predictive control (MPC). [3][4][5][6] MPC-based control methods have attracted considerable attention from researchers as they provide superior dynamic performance. 7,8 Among the various MPC-based control methods, finite control set (FCS) MPC is a widely adopted one for the control of MMCs as it does not require modulators.…”