A Fast Steady-state Thermal Modeling of Power Modules Based on Fourier Analytic Series

  • Jiajun Zhou
  • , Hongchang Cui
  • , Laili Wnag
  • , Yongmei Gan
  • , Zijie Tang
  • , Kai Gao

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

With the increase of power flow, the thermal management of multi-chip parallel modules becomes a prominent problem. The steady-state thermal impedance network is used to analyze module thermal reliability quickly and provide layout iteration information. Based on the analytical model of cube structure, this paper proposes a fast method to extract the temperature variation in the process of layout simplification. The compensated analytical model has high precision and calculational efficiency. Compared with the finite element simulation, the proposed analytical model represents robust expression of thermal process under different working conditions. The maximum error is less than 2%, and Only 27% of the computational resources of finite element simulation are required.

Original languageEnglish
Title of host publication2024 IEEE 10th International Power Electronics and Motion Control Conference, IPEMC 2024 ECCE Asia
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages1825-1829
Number of pages5
ISBN (Electronic)9798350351330
DOIs
StatePublished - 2024
Event10th IEEE International Power Electronics and Motion Control Conference, IPEMC 2024 ECCE Asia - Chengdu, China
Duration: 17 May 202420 May 2024

Publication series

Name2024 IEEE 10th International Power Electronics and Motion Control Conference, IPEMC 2024 ECCE Asia

Conference

Conference10th IEEE International Power Electronics and Motion Control Conference, IPEMC 2024 ECCE Asia
Country/TerritoryChina
CityChengdu
Period17/05/2420/05/24

Keywords

  • analytic model
  • Fourier series
  • structure simplification
  • thermal coupling

Fingerprint

Dive into the research topics of 'A Fast Steady-state Thermal Modeling of Power Modules Based on Fourier Analytic Series'. Together they form a unique fingerprint.

Cite this