Données Personnelles | Thématiques de recherche | Charges de cours | Publications |
Automatique et analyse des systèmes
Les activités de recherche du SAAS sont essentiellement orientées dans deux directions: - la supervision et les systèmes de diagnostic (basés sur des modèles mathématiques) pour des système technique; - la régulation sous contrainte de systèmes dynamiques non linéaires avec l'accent sur les méthodes de type "reference governor" . Les méthodes développées sont appliquées à un large spectre de domaines: - production , distribution et stockage d'énergie électrique, - systèmes mécatroniques (drones et robotique), - procédés industriels.
SPACE4RELAUNCH -HEALTH MONITORING OF ELECTROMECHANICAL ACTUATORS FOR REUSABLE LAUNCHERS
Due to the expensive nature of rockets, a major challenge consists in reducing their costs. Launchers in particular are quite expensive and the majority is expendable. Market competition brought aerospace enterprises to search for new ways to cut prices, notably by reusing their launch systems. During each stage of ascension, specific boosters constituting the rocket are decoupled, generally left to fall into the sea. One of the reason why launchers are not reused comes down to the lack of knowledge on the health status of their components, including their actuators. The latter typically serve to orientate the nozzle for propulsion. In case of reusable launchers, additional electromechanical actuators (EMAs) can be included to control stabilizing fins and to deploy legs during descent. The project aims at developing a health monitoring system for high performance electromechanical actuators to be used for reusable launchers. It is part of a WIN4EXCELLENCE programme involving 30 PhD students and dealing with innovative tools for earth observation and reusable launchers. An hybrid fault diagnosis approach is considered for detecting both electrical and mechanical faults, combining a physical model with data-driven techniques. This methodology will require data on both healthy and faulty EMAs. To produce synthetic data on faulty scenarios, the first step consists in establishing a physical-based simulation of the EMA’s health status. The second step is about realising the fault diagnosis. System states will be evaluated through an observer, potentially a Gaussian process.