What is it about?
Thermal issue has been one of the bottlenecks for the performance and reliability of GaN High-Electron-Mobility-Transistors(HEMTs), which underscores the importance of accurate thermal modeling. Here, we propose a GP (Gaussian process)-RC (Resistor-Capacitor) compact thermal model integrated with the Ensemble Kalman filter (EnKF) to handle the nonlinear problems attributed to the temperature-dependent properties of GaN HEMTs under large-signal working conditions. The machine learning tech, GP, brings our model the ability and extendibility to handle various temperature-dependent relationships without restricting to some prescribed function forms.
Featured Image
Photo by Vadim Bogulov on Unsplash
Why is it important?
We propose a GP (Gaussian process)-RC (Resistor-Capacitor) compact thermal model integrated with the Ensemble Kalman filter to handle the nonlinear problems attributed to the temperature-dependent properties of GaN HEMTs under large-signal conditions. The GP brings our model the ability and extendibility to handle varied temperature-dependent relationships. An efficient model extraction scheme is devised based on the EnKF. Furthermore, it is capable of realizing the in-situ and timely model update via real-time data sequence.
Perspectives
This work could be helpful for thermal management of GaN devices through improving the compact thermal modeling with nonlinear effect. Furthermore, thanks to the extendibility of GP predictors, it is also possible to consider other influencing factors (like bias-dependence and size effect ) in the thermal model of GaN devices.
Yuchao Hua
Universite de Nantes
Read the Original
This page is a summary of: Nonlinear compact thermal modeling of self-adaptability for GaN high-electron-mobility-transistors using Gaussian process predictor and ensemble Kalman filter, Journal of Applied Physics, January 2024, American Institute of Physics,
DOI: 10.1063/5.0180835.
You can read the full text:
Contributors
The following have contributed to this page
