An ADALINE based capacitor voltage estimation algorithm for modular multilevel converters

摘要

This paper presents a novel capacitor voltage estimation using the adaptive linear neuron (ADALINE) algorithm for monitoring of MMC sub-module voltages. The proposed estimation unit requires only three voltage sensors per phase for the arm reactors and the output phase voltages. Measurements of sub-module capacitor voltages and associated communication links with the central controller are not needed. The proposed strategy can be applied to MMC systems that contain a large number of sub-modules. The method uses PSCAD/EMTDC, with particular focus on dynamic performance during boost operation.