Modélisation et détermination des stratégies de conditionnement de puissance pour des réseaux de puissance répartis pour des applications embarquées aéronautiques et spatiales

La soutenance de thèse de Quentin HILPERT,  intitulée « Modélisation et détermination des stratégies de conditionnement de puissance pour des réseaux de puissance répartis pour des applications embarquées aéronautiques et spatiales » aura lieu le jeudi 2 mai 2024 à 9h30 en salle de thèses C002 à l’ENSEEIHT, 2 rue Charles Camichel.

Les travaux de thèse ont été réalisés sous la direction de Stéphane CAUX pour le LAPLACE et François BONNET du CNES.

Jury de thèse :

Mme Claire VALENTIN, professeure à l’Université Claude Bernard Lyon 1 (laboratoire LAGEPP), Rapporteuse

Mme Manuela SECHILARIU, professeure à l’Université de Technologie de Compiègne (laboratoire Avenues), Rapporteuse

Mme Antoneta-Iuliana BRATCU, maître de conférence à Grenoble INP (GIPSA-lab), Examinatrice

Mme Ada CZESNAKOWSKA, responsable R&T chez Airbus Defence & Space, Examinatrice invitée

M. Pascal MAUSSION, professeur à Toulouse INP / ENSEEIHT (Laplace), Examinateur

M. Stéphane CAUX, Directeur de thèse

M. François BONNET, Co-directeur de thèse

Résumé :

The electrical power system (EPS) is one of the most critical systems, if not the most critical, of a spacecraft considering that a spacecraft is nothing but an electrical system and the purpose of the EPS is to generate, condition, store and distribute electrical power to the entire system. In other words, it is the first link in the food chain and if it stops working, the whole system is lost. The resulting need for robustness and reliability, as well as the particularly isolated and hostile environment in which it must operate, have guided the design choices of the EPS and led to a system with very high efficiency, electrically speaking, but at the cost of a lack of flexibility. Nevertheless, the current context and recent developments give the opportunity to deeply rethink the design of the EPS and this is what CNES and Airbus have been doing for several years. The modular and distributed electrical architecture that has emerged from these developments offers many possibilities in terms of control, but also challenges in terms of stability and optimization. The objective of this thesis is to demonstrate that control laws integrated in the conditioning modules allow a regulation of the battery charge and a stability of the onboard network whatever the topology of the power bus and the consumption profiles.

To do this, a parametric modeling of the system was carried out by being inspired by the modeling methods used in the problems of terrestrial microgrids. The models of the various elements of the system were thus integrated under Matlab and Matlab Simulink in order to simulate a typical system inspired by satellites developed by the CNES. A thorough bibliographical study was carried out to identify control laws used in similar problems and which would be of interest to be adapted to space applications. These control laws have been tested in simulation, with particular attention to the stability constraints, which allowed to select the most suitable ones in order to validate them experimentally on a demonstrator set up on the same model. Based on these results, a global strategy is proposed.