Medium voltage 4-level double-star multilevel converter using model predictive control

Ana M. Llor, Samir Kouro, Carlos Reusser, Venkata Yaramasu, Bin Wu

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

4 Scopus citations

Abstract

The double-star converter (DSC) is a recently introduced medium-voltage 4-level multilevel inverter topology, with a reduced number of active switches. This topology has several advantages over existing commercial topologies (single dc-source, no clamping elements, no bidirectional switches or flying capacitors), which can satisfy the needs of several medium voltage applictions such as pumps, fans and compressors. The main challenge with this topology is the voltage balance of the three dc-link capacitors across the modulation index range. Classic controllers require sophisticated PWM techniques with closed loop compensation to maintain the capacitor voltages balanced. In this paper, a control algorithm based on model predictive control (MPC) is applied to the DSC, allowing a direct regulation of the converter output currents while simultaneously ensuring the dc-link capacitors voltage balance. Simulation results show the effectiveness of MPC when applied to this converter.

Original languageEnglish
Title of host publication2017 IEEE International Conference on Industrial Technology, ICIT 2017
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages136-140
Number of pages5
ISBN (Electronic)9781509053209
DOIs
StatePublished - 26 Apr 2017
Externally publishedYes
Event2017 IEEE International Conference on Industrial Technology, ICIT 2017 - Toronto, Canada
Duration: 23 Mar 201725 Mar 2017

Publication series

NameProceedings of the IEEE International Conference on Industrial Technology

Conference

Conference2017 IEEE International Conference on Industrial Technology, ICIT 2017
Country/TerritoryCanada
CityToronto
Period23/03/1725/03/17

Fingerprint

Dive into the research topics of 'Medium voltage 4-level double-star multilevel converter using model predictive control'. Together they form a unique fingerprint.

Cite this