Time-to-Failure Analysis of Short-Duty Cycle, Inverter-Fed Electrical Machines Exposed to Prevailing Electrical Stress

Yatai Ji, Paolo Giangrande, Huanran Wang, Weiduo Zhao, Vincenzo Madonna, He Zhang, Michael Galea

Research output: Journal PublicationArticlepeer-review

1 Citation (Scopus)

Abstract

Compact and efficient electric drives are desirable in transport applications and an increased dc-link voltage value combined with fast-switching devices represents a promising solution for achieving light and energy-efficient systems. Nonetheless, the higher dc bus voltage and wide-bandgap device greatly raise the electrical stress (i.e., maximum interturn voltage) of inverter-fed electrical machines (EMs) insulation jeopardizing the overall reliability. Thus, the lifetime prediction of EMs gained more attention recently. Short-duty cycle EMs are commonly employed in more electric aircraft (MEA) applications, and different from EMs for continuous operations, as they could operate under higher insulation stress due to their short duration of operation. In this article, the time-to-failure analysis on short-duty cycle, inverter-fed EMs is presented relying on the physics of failure approach. Electrical stress is assumed to be the dominant aging factor and the lifetime model based on the inverse power law is developed and tuned according to the collected experimental data. Two typical case study scenarios for MEA applications are discussed and the time-to-failure is calculated via the built lifetime model with a special focus on interturn insulation.

Original languageEnglish
Pages (from-to)9368-9378
Number of pages11
JournalIEEE Transactions on Aerospace and Electronic Systems
Volume59
Issue number6
DOIs
Publication statusPublished - 1 Dec 2023

ASJC Scopus subject areas

  • Aerospace Engineering
  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'Time-to-Failure Analysis of Short-Duty Cycle, Inverter-Fed Electrical Machines Exposed to Prevailing Electrical Stress'. Together they form a unique fingerprint.

Cite this