FS-MPC Based Thermal Stress Balancing and Reliability Analysis for NPC Converters

Mateja Novak, Victor Ferreira, Markus Andresen, Tomislav Dragicevic, Frede Blaabjerg, Marco Liserre

Research output: Contribution to journalJournal articleResearchpeer-review

19 Citations (Scopus)
147 Downloads (Pure)

Abstract

Active thermal control has been introduced to regulate the steady state and reduce the transient thermal mechanical stress in power electronic modules. Specifically, it can equally distribute the temperature among the devices, thereby better distributing the stress among a set of devices and reducing the failure probability in the most thermally-stressed devices. This is of great importance for multilevel topologies and in particular for the Neutral Point Clamped (NPC) topologies, which have an inherent thermal unbalance among devices of the same phase. In hybrid structures, whereby Si and SiC devices are mixed to achieve a better trade-off between efficiency and cost, this problem is even worse due to technology differences. This paper investigates the advantages introduced by using Finite-Set Model Predictive Control (FS-MPC) algorithms designed for achieving a balanced device junction temperature in hybrid NPC and Active-NPC converters. Moreover, a novel setup composed by a hybrid ANPC based power modules is used to experimentally validate the presented FS-MPC algorithm. A lifetime estimation is performed for the two topologies to highlight the long-term benefits of FS-MPC algorithm in these hybrid topologies for UPS applications.
Original languageEnglish
JournalIEEE Open Journal of Power Electronics
Volume2
Pages (from-to)124-137
Number of pages14
DOIs
Publication statusPublished - 5 Feb 2021

Keywords

  • Active Neutral Point Clamped (ANPC) inverter
  • Neutral point clamped converter
  • Finite set model predictive control (FS-MPC)
  • Hybrid power stage
  • Multilevel Converter
  • SIC
  • lifetime
  • Reliability

Fingerprint

Dive into the research topics of 'FS-MPC Based Thermal Stress Balancing and Reliability Analysis for NPC Converters'. Together they form a unique fingerprint.

Cite this