Statistical Simulation of Conductor Lay and AC Losses in Multi-Strand Stator Windings

Research output: Chapter in Book/Report/Conference proceedingConference Contribution (Conference Proceeding)

Abstract

The multi-strand winding configurations adopted in high-speed electrical machines are susceptible to increased AC losses as a result of proximity effects. Due to the characteristics of the insert winding process, variability can arise in the conductor lay and consequently, the AC losses of as-manufactured windings. Current finite element analysis practices cannot account for this variability resulting in under- or over-conservative AC loss estimates. This paper presents the development of a statistical simulation methodology that can represent the variability in conductor lay and AC losses of a distributed winding. Initial calibration of the methodology is achieved following statistical characterization of experimental AC loss measurements from a batch of industrial machines. The initial calibration of the methodology highlighted that only a limited amount of inter-bundle strand mixing is to be expected in as-manufactured windings, resulting in reduced variability in the expected AC losses compared to other statistical estimation approaches.
Original languageEnglish
Title of host publication2021 IEEE International Electric Machines & Drives Conference (IEMDC)
PublisherInstitute of Electrical and Electronics Engineers (IEEE)
ISBN (Electronic)978-1-6654-0510-2
ISBN (Print)978-1-6654-4628-0
DOIs
Publication statusPublished - 15 Jun 2021
Event2021 IEEE International Electric Machines & Drives Conference (IEMDC) - Virtual
Duration: 17 May 202120 May 2021

Conference

Conference2021 IEEE International Electric Machines & Drives Conference (IEMDC)
Period17/05/2120/05/21

Keywords

  • AC Loss
  • Conductor Lay
  • Insert Winding
  • Variability
  • Probabilistic Methodology

Fingerprint

Dive into the research topics of 'Statistical Simulation of Conductor Lay and AC Losses in Multi-Strand Stator Windings'. Together they form a unique fingerprint.

Cite this