Home
Scholarly Works
Time-Efficient Integrated Electrothermal Model for...
Conference

Time-Efficient Integrated Electrothermal Model for a 60-kW Three-Phase Bidirectional Synchronous DCDC Converter

Abstract

This article introduces an accurate and fast electrothermal model to calculate power loss and junction temperature of power electronic devices in a three-phase dcdc converter, and to simulate converter dynamics during a simulated drive cycle. The proposed electrothermal model calculates power loss and junction temperature by representing the three-phase dcdc converter dynamics within a drive cycle using a three-phase state-space average model. The three-phase state-space average model is modified to represent converter dynamics when one or two phases are off, in addition to the phase transition dynamics. The existing electrothermal models can either represent converter dynamics but are too slow for long simulations like drive cycles, or they reduce the simulation time at the expense of junction temperature accuracy. Furthermore, the existing electrothermal models in the literature are for inverters and single-phase dcdc converters. The proposed electrothermal model combines the advantages of fast simulation time with accurate estimation of junction temperature. The power loss calculated by the proposed method is verified by a switching Simulink/PLECS model, and the mosfet temperature estimation is validated experimentally using a 60-kW three-phase bidirectional boost converter under steady-state conditions and a custom drive cycle. The efficiency of the three-phase converter is measured experimentally and compared with the efficiency calculations from the proposed electrothermal model.

Authors

Azer P; Rodriguez R; Guo J; Gareau J; Bauman J; Ge H; Bilgin B; Emadi A

Volume

56

Pagination

pp. 654-668

Publisher

Institute of Electrical and Electronics Engineers (IEEE)

Publication Date

January 1, 2020

DOI

10.1109/tia.2019.2948804

Conference proceedings

IEEE Transactions on Industry Applications

Issue

1

ISSN

0093-9994

Contact the Experts team