Closed-loop analytic filtering scheme of capacitor voltage ripple in multilevel cascaded H-bridge converters

• Main Authors: Rodriguez, Ezequiel Ramos, Farivar, Glen Ghias, Beniwal, Neha, Townsend, Christopher David, Dehghani Tafti, Hossein, Vazquez, Sergio, Pou, Josep
• Other Authors: School of Electrical and Electronic Engineering
• Format: Article
• Language: English
• Published: 2020
• Subjects: Engineering::Electrical and electronic engineering; Capacitance Estimation; Capacitor Voltage Filtering
• Online Access: https://hdl.handle.net/10356/139667
• Institution: Nanyang Technological University

This article proposes a closed-loop analytic filtering scheme for the low-frequency capacitor voltage ripple in cascaded H-bridge converters. Although this technique can be applied to a number of cascaded converter applications, this article focuses particularly on the low-capacitance cascaded H-bridge static compensator. Filtering of capacitor voltage signals is particularly important in this application, due to the presence of large low-frequency harmonic voltage ripple. The main novelty of the proposed algorithm in this article is the estimation of the capacitance value of each H-bridge. The closed-loop capacitance estimation method reduces steady-state error in estimated voltage ripple magnitude and phase, which is otherwise present in the available open-loop filters in the literature. Furthermore, fast dynamic response is achieved compared to filtering schemes based on low-pass and bandstop filters. Therefore, the proposed solution optimizes the tradeoff between filtering accuracy and transient response. It also mitigates parametric uncertainties, time delays, and harmonic contamination in the outer voltage control loop. Furthermore, the estimated capacitance is also useful to track the state of health of the dc-link capacitors. Experimental results on a seven-level 1-kVA cascaded H-bridge prototype are presented to demonstrate the comparative performance of the proposed filtering scheme and traditional approaches.