Online Condition Monitoring of Sub-module Capacitors in MMC Enabled by Reduced Switching Frequency Sorting Scheme

Abstract

This paper presents a novel condition monitoring scheme for sub-module (SM) capacitors in the modular multilevel converter (MMC), where the fast-affine projection (FAP) algorithm is cooperatively embedded in a dedicated range-based sorting scheme with reduced switching frequency to estimate the capacitor parameters. The proposed sorting scheme not only reduces the switching events of SMs, but also manages to maintain the capacitor ripple voltage within an acceptable range. Whilst the use of FAP achieves an on-line parametric estimation with superior speed and accuracy. Compared to previous studies, this proposed approach has significant advantages in terms of estimation accuracy and speed without the need for extra hardware. The effectiveness of the proposed approach is validated by a 5-level MMC model in MATLAB/Simulink. From simulation results, the estimation speed is less than 1.5 ms and the error is less than 3%.