On Febuary 6, 2023, Nikolena Christofi defended her thesis at IRT Saint Exupéry in Toulouse. This thesis was prepared at LAAS-CNRS (Laboratoire d’Analyse et d’Architecture des Systèmes) in the embedded system speciality of the ED-AA (Ecole doctorale Aéronautique – Astronautique). After her defense, her doctoral degree was delivered by INSA (Institut National des Sciences Appliquées). While working on her thesis, Nikolena has been part of the S2C project at IRT Saint Exupéry.
ABOUT HER THESIS
“Improving operational diagnosis with an interdisciplinary modeling approach”
The proposed methodology aims to translate system functional and dysfunctional information, produced by the system design definition phase, into executable models, representative of the system operational activities, while offering a high fidelity fault detection and mitigation architecture. The methodology consists in two successive steps. The first step defines the creation of a pivot model, through the automatic transformation of Fault Trees –produced by system dysfunctional analyses, into Behaviour Trees. The pivot model expresses the system’s diagnostic objectives.
The second step consists in the elicitation of the pivot Behaviour Tree model into a second Behaviour Tree model, by integrating complementary data, produced during system design. The second model is an executable model, and is composed by the system operational activities themselves. This final model constitutes the operations-dedicated model, which can be integrated inside a user-interface equipped monitoring tool, to support operators’ diagnostics tasks. The reason the methodology consists in two steps, where dysfunctional and functional system information are gradually transformed and inserted inside the operations-dedicated model, is because information on system faults alone is not enough for the creation of a model dedicated to operations, to serve for fault detection and diagnosis. Both functional and dysfunctional system information is necessary to build a model for system monitoring.
The methodology to create the operations-dedicated model, along with the original proposal to adopt extended Behavior Trees for its representation, result from a thorough state-of-the-art analysis and substantial feedback provided by major actors of the national aerospace industry, such as Airbus Defence and Space (ADS) and the Centre National d’Etudes Spatiales (CNES). We also applied the proposed methodology on three different case studies, which allowed us to evaluate its feasibility in industrial contexts, along with its current benefits and future potential.
Future work consists in applying the proposed methodology on a real complex system, in order to evaluate its impact on operational diagnosis activities’ efficiency, as well as its scalability. Another perspective is the construction of digital twins based on the operations-dedicated model, for the simulation of operational scenarios ahead of the system’s deployment, and/or for the feeding back of operational data and return of experience information, during operations and maintenance, which could contribute to the amelioration of system architecture, within a digital data continuity system framework.
- N. Christofi, X. Pucel, C. Baron, M. Pantel, S. Guilmeau, C. Ducamp. Towards an Operations-Dedicated Model for Space Systems, Journal of Aerospace Information Systems – JAIS, AIAA Aerospace Research Central (ARC).
- Nikolena Christofi, Xavier Pucel, Claude Baron, Marc Pantel, David Canu, Jerome Golenzer, Christophe Ducamp. Introducing Operational Diagnosis Models for Ground Station Architectures using Behaviour Trees, 17th International Conference on Space Operations – SpaceOps 2023, Dubai, United Arab Emirates.
- N. Christofi and X. Pucel. A Novel Methodology to model Digital Twins for Spacecraft Operations using Fault Trees and Behaviour Trees, 2nd International Workshop on Model-Driven Engineering of Digital Twins – ModDiT’22, MODELS 2022 companion, Montreal, QC, Canada.
- N. Christofi and X. Pucel. From Safety Assessment Models to Operational Diagnosis Models, 33rd International Workshop on Principles of Diagnosis – DX 2022, Toulouse, France.
- N. Christofi, X. Pucel, C. Baron, M. Pantel, S. Guilmeau, C. Ducamp. Towards an agile, model-based multidisciplinary process to improve operational diagnosis in complex systems, 11th European Congress on Embedded Real Time Systems – ERTS 2022, Toulouse, France.
- N. Christofi, C. Baron, X. Pucel, M. Pantel, M. Machin, C. Ducamp. Adopting a model-based approach for satellite operations’ diagnosis, 13th International Conference on Modeling, Optimization and Simulation – MOSIM’20, Agadir, Morocco [held online].
ABOUT S2C PROJECT
Establishment of a numerical continuity between system definition and safety analysis.
|Mme Claude BARON||Thesis Director||LAAS-CNRS, INSA Toulouse, Toulouse Federal University|
|M. Marc PANTEL||Thesis Co-Director||IRIT, INP ENSEEIHT,|
Toulouse Federal University
|M. Xavier PUCEL||Thesis Co-supervisor||ONERA, ANITI,|
Toulouse Federal University
|Mme Besma ZEDDINI||Examiner||EISTI, CY Tech|
|M. Jean-Yves CHOLEY||Examiner||ISAE-SUPMECA|
|M. Jean-Michel BRUEL||Examiner||IRIT, INP ENSEEIHT,|
Toulouse Federal University
|M. Yvan BEAUREGARD||Reviewer||Superior school of technology (ÉTS) Montréal|
|M. Mahdi ZARGAYOUNA||Reviewer||Gustave-Eiffel university|
|M. Christophe DUCAMP||Guest||Airbus Defence & Space|
|M. Julien BACLET||Guest||IRT Saint Exupéry|