Expediting Transient Thermal Frequency Response Characterization and Sensitivity Analysis

Timothy A. Polom, Robert D. Lorenz

Research output: Conference proceeding/Chapter in Book/Report/Conference Paperpeer-review

Abstract

This paper develops a technique, requiring no dedicated temperature sensing calibration step, to rapidly characterize transient heat transfer in packaged, power semiconductor components. It is presented as an alternative to traditional step response characterization methods by exploiting the phase delay metric native to frequency response function (FRF) analysis in the field of system identification. The paper introduces electrothermal engineering steps needed to design the measurement system. It presents aspects of power device physics and dynamic analysis to identify design space in which FRF data extracted from experiments are insensitive to potential unknowns. Power electronic circuitry, providing the needed, periodic heat actuation, and probing, allowing for simple reconstruction of transient temperature data in post-processing, are introduced. The developed method is leveraged in-lab to make a key measurement confirming high-frequency-only thermal FRF sensitivity to component die-attach, highlighting the opportunity to achieve localized, power electronic converter degradation sensing in situ.
Original languageEnglish
Title of host publication2020 IEEE Energy Conversion Congress and Exposition (ECCE)
Pages3320-3327
Number of pages8
DOIs
Publication statusPublished - 15 Oct 2020
Event2020 IEEE Energy Conversion Congress and Exposition (ECCE) - Detroit, MI, USA
Duration: 11 Oct 202015 Oct 2020

Conference

Conference2020 IEEE Energy Conversion Congress and Exposition (ECCE)
Period11/10/2015/10/20

Keywords

  • characterization
  • electrothermal
  • frequency response
  • heat transfer
  • power semiconductor devices
  • packaging
  • sensitivity analysis
  • system identification

Fingerprint

Dive into the research topics of 'Expediting Transient Thermal Frequency Response Characterization and Sensitivity Analysis'. Together they form a unique fingerprint.

Cite this