Methodological Framework for Optimizing the Reliability Improvement of Power Electronic Converter Systems: Mathematical Problem Formulation and Topology-Agnostic Optimization Approaches

Felix K.S. Adabla, Masoud Davari, Frede Blaabjerg, Yongheng Yang

Research output: Contribution to journalArticlepeer-review

Abstract

Given the ubiquity of power electronic converter (PEC) systems, they often represent the weakest link in terms of reliability within modern power and energy systems in different applications. Extensive research has established a direct correlation and interplay between the reliability of electrical systems in consumer electronics, nano/microgrids, and modern grids and the PECs they deploy. As a result, deficiencies in the PEC reliability can lead to diminished system-level performance, while enhancements in the PEC reliability correspondingly improve overall system reliability. While various PEC reliability improvement strategies exist, determining “optimal” approaches and the “extent” of such improvements remains challenging for finding the optimal design for the PEC reliability. This paper proposes a methodological framework for optimizing the PEC reliability improvement to fill this gap by employing an optimization framework that targets component-level reliability; to this end, different topology-agnostic optimization algorithms are proposed, studied, and compared. Utilizing the reliability models established for critical PECs’ components (switching devices and capacitors), four single-objective and two multi-objective optimization problems are formulated. These formulations incorporate key electrical and mechanical parameters, including component sizing, thermal management, and switching frequency, while also considering practical constraints. The proposed problem formulations and optimization-driven analysis systematically explore the design space to identify optimal solutions under various reliability constraints, demonstrating an acceptable trade-off for different improvement strategies and providing insights into achieving the reliability enhancements targeted in PEC systems. The methodological framework proposed in this paper is experimentally tested on an integral part of PECs, i.e., a half-bridge module, to preserve the generality and practicality of the methodological approach proposed in this paper.

Original languageEnglish
JournalIEEE Transactions on Consumer Electronics
DOIs
StatePublished - 2025

Scopus Subject Areas

  • Media Technology
  • Electrical and Electronic Engineering

Keywords

  • Artificial bee colony (ABC)
  • Harris hawks optimization (HHO)
  • design for reliability
  • genetic algorithm (GA)
  • hybrid GA-differential evolution (GA-DE)
  • hybrid GA-particle swarm optimization (GA-PSO)
  • mean-time-to-failure
  • reliability improvement optimization of power electronic converter systems
  • single-/multi-objective optimization
  • topology-agnostic optimization approach

Fingerprint

Dive into the research topics of 'Methodological Framework for Optimizing the Reliability Improvement of Power Electronic Converter Systems: Mathematical Problem Formulation and Topology-Agnostic Optimization Approaches'. Together they form a unique fingerprint.

Cite this