Skip to Main content Skip to Navigation
Theses

Model reduction for thermal management of high power electronic components for aerospace application

Abstract : In the transition to more electric aircraft, one of the technological locks is the overheating of electronic components, which affects deeply their reliability and their lifetimes. Therefore, it is necessary to control the temperature of the electronic components. This thesis aims to build two reduced thermal models to monitor in real time the temperature of electronic components. The difference between the direct reduced model (DROM) and the inverse reduced order model (IROM) lies in the input parameters and the mathematical formalism of their constructions. For the DROM, the input parameters are the boundary conditions. This model is developed in two stages. The first step is to build a reduced base representative of the solution by applying POD (Proper Orthogonal Decomposition) on a snapshots matrix. The snapshots matrix is obtained from the finite element model (FEM) solution. The second step is to calculate the coordinates of a new solution using the Galerkin projection of the FEM on the reduced basis. A DROM built with 10 modes decreases drastically the computational time and the obtained absolute error is less than 0.1 °C except during sudden power variations. For the IROM, the input parameters are the temperature of the sensors placed far from the electronic components. This model does not need to know the boundary conditions as the DROM. The first step is to build a reduced base that couples the temperature or electronic components and the temperature of sensors using the POD. The second step is to identify the coordinates of the electronic components temperature from the measurements using a minimization algorithm. The error of the IROM of order 3 does not exceed 0.6 °C except during sudden power variations.
Complete list of metadatas

Cited literature [66 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-02900474
Contributor : Abes Star :  Contact
Submitted on : Thursday, July 16, 2020 - 10:39:11 AM
Last modification on : Friday, July 17, 2020 - 5:11:42 AM

File

these.pdf
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-02900474, version 1

Citation

Hazem Ben Aissia. Model reduction for thermal management of high power electronic components for aerospace application. Thermics [physics.class-ph]. Université de Lyon, 2019. English. ⟨NNT : 2019LYSEI071⟩. ⟨tel-02900474⟩

Share

Metrics

Record views

36

Files downloads

31