Artificial Intelligence and Long-term Performance of Power Electronics Systems

Shuai Zhao, Saeed Peyghami, Daniel Gebbran, Tomislav Dragicevic, Huai Wang, Frede Blaabjerg

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

Abstract

This paper presents the application of artificial intelligence (AI) in the long-term performance assessment of power electronics. Long term performance of power converters depends on two major factors: intrinsic material strength and extrinsic stressors. In order to ensure acceptable performance, various life cycle management approaches can be employed to either reinforce the strength of components or control the stressors. Reinforcement can be performed within the design or by means of maintenance within the operation. Furthermore, stress management can be carried out by suitable control methods. Due to the complexity of power converters especially in large-scale systems, AI methods are employed to facilitate life cycle management of power converters. This paper presents the AI methods used for this purpose, implementation of AI methods, and basics of AI for maintenance in power converters. Furthermore, two case studies on the device level and system level are presented to illustrate the application of AI in the lifetime expansion of power electronics systems. Moreover, a perspective on challenges in the integration of AI in power electronics is provided.

Original languageEnglish
Title of host publicationProceedings of 12th International Conference on Integrated Power Electronics Systems
PublisherIEEE
Publication date2022
Pages5-14
ISBN (Print)9783800757572
Publication statusPublished - 2022
Event12th International Conference on Integrated Power Electronics Systems - Mercure Hotel MOA Berlin, Berlin, Germany
Duration: 15 Mar 202217 Mar 2022
Conference number: 12

Conference

Conference12th International Conference on Integrated Power Electronics Systems
Number12
LocationMercure Hotel MOA Berlin
Country/TerritoryGermany
CityBerlin
Period15/03/202217/03/2022

Fingerprint

Dive into the research topics of 'Artificial Intelligence and Long-term Performance of Power Electronics Systems'. Together they form a unique fingerprint.

Cite this