Capacitor Voltage Estimation Scheme With Reduced Number of Sensors for Modular Multilevel Converters

Abstract

This paper presents a new method to measure the voltage across the submodule (SM) capacitors in a modular multilevel converter (MMC). The proposed technique requires only one voltage sensor per arm. This reduces the number of sensors required compared with conventional sensor-based methods. Therefore, the cost and complexity of the system are reduced, which in turn improves the converter's overall reliability. The proposed method employs an exponentially weighted recursive least square algorithm to estimate the SM capacitor voltages through the measured total arm voltage and the switching patterns of each SM. There is thus no need for extra sensors to measure these control signals as they are directly provided from the controller. The robustness of the proposed method is confirmed via introducing deviations for the capacitance values, dynamic load changes, dc voltage change, and startup transient condition. Simulation and experimentally validated results based on a single-phase MMC show the effectiveness of the proposed method in both, steady-state and dynamic operations.