Blocking capability for switching function and average models of modular multilevel converters

Abstract

Modular multilevel converters (MMC) for high voltage direct current (HVDC) application employ hundreds of sub-modules. Explicit modeling of sub-modules implies hundreds of nodes and semiconductor devices in the equivalent models for the system; this in turn poses a computational challenge for electromagnetic transient (EMT) simulation. Fast and accurate simulation models for MMC based on sub-module/arm's Thevenin's equivalent, switching function and average value models have been proposed in the literature. However existing switching function and average value models lack the capability to inherently capture all operating states of the converter. This paper proposes switching function and average value models that accurately model all operating states of the converter and presents a comprehensive comparison with detailed model of the converter in terms of accuracy and computational speed.