Abstract
Join a PhD project at the forefront of Aviation 4.0 and contribute to the development of advanced Artificial Intelligence solutions aimed at making aeronautical systems more reliable, smarter, and energy-efficient through the intelligent use of large-scale industrial data.
Keywords : Aviation 4.0, Predictive Maintenance, Industrial Big Data, Multi-Objective Optimization
This PhD project focuses on developing Artificial Intelligence and Big Data methods to improve the reliability, performance, and energy efficiency of aeronautical systems. By leveraging large volumes of heterogeneous and uncertain industrial data (from sensors, maintenance logs, and real operations), the goal is to anticipate failures, optimize maintenance, and extend equipment lifespan.
The approach combines machine learning and multi-objective optimization to support smarter decision-making, balancing operational performance, costs, and environmental impact. Ultimately, the project contributes to the transition toward a more sustainable and data-driven aviation industry within the Aviation 4.0 framework.
Rejoignez une thèse au coeur des défis de l’Aviation 4.0 et contribuez au développement de solutions d’intelligence artificielle capables de rendre les systèmes aéronautiques plus fiables, plus intelligents et plus sobres énergétiquement grâce à l’exploitation des données massives industrielles.
Mots clés : Aviation 4.0, Maintenance prédictive, Big-data industriel, Optimisation multi-objectifs
Ce projet de thèse vise à développer des méthodes en intelligence artificielle et en Big Data pour améliorer la fiabilité, la performance et l’efficacité énergétique des systèmes aéronautiques. En s’appuyant sur de grandes quantités de données industrielles hétérogènes et incertaines (issues de capteurs, de journaux de maintenance et des opérations réelles), l’objectif est d’anticiper les défaillances, d’optimiser la maintenance et de prolonger la durée de vie des équipements.
L’approche repose sur la combinaison de l’apprentissage automatique et de l’optimisation multi-objectifs afin de favoriser une prise de décision plus intelligente, en conciliant performance opérationnelle, coûts et impact environnemental. À terme, ce projet contribue à la transition vers une industrie aéronautique plus durable et pilotée par les données, dans le cadre de l’Aviation 4.0.
Skills
The candidate must hold a Master’s degree (M2) or an engineering degree with specialization in computer science, artificial intelligence, data science, or big data. The candidate should possess skills in one or more of the following areas:
- Strong foundations in applied mathematics, statistics, and optimization, with an interest in complex systems modeling.
- Proficiency in major artificial intelligence techniques (machine learning, deep learning); experience with frameworks such as PyTorch, TensorFlow (and possibly federated learning tools) would be appreciated.
- Skills in big data processing and heterogeneous data analysis, particularly time series from sensors.
- Good level in scientific programming (Python recommended) and knowledge of data manipulation and analysis tools (pandas, scikit-learn, etc.).
- Interest in distributed architectures, parallel computing, or real-time data processing.
- Awareness of data quality issues (noise, missing data) and their exploitation in complex industrial environments.
A good level of scientific English, both written and spoken, is required. The candidate must demonstrate autonomy, rigor, good organizational skills, initiative, and strong scientific curiosity supporting learning abilities. The ability to work in a collaborative academic and industrial environment is also expected.
Compétences
Le candidat doit être titulaire d’un diplôme de niveau Master (M2) ou d’un diplôme d’ingénieur avec une spécialisation en informatique, intelligence artificielle, data science ou big data. Le candidat devra disposer de compétences dans un ou plusieurs des domaines suivants :
•Bases solides en mathématiques appliquées, statistiques et optimisation, avec un intérêt pour la modélisation de systèmes complexes.
•Maîtrise des principales techniques d’intelligence artificielle (machine learning, deep learning); une expérience avec des frameworks tels que PyTorch, TensorFlow (et éventuellement outils d’apprentissage fédéré) sera appréciée.
•Compétences en traitement de données massives (big data) et analyse de données hétérogènes, en particulier séries temporelles issues de capteurs.
•Bon niveau en programmation scientifique (Python recommandé) et connaissance des outils de manipulation et d’analyse de données (pandas, scikit-learn, etc.)
•Intérêt pour les architectures distribuées, le calcul parallèle ou le traitement des données en temps réel.
•Une sensibIlité aux problématiques liées à la qualité des données (bruit, données manquantes) et à leur exploitation dans des contextes industriels complexes. Un bon niveau d’anglais scientifique est requis, à l’écrit comme à l’oral. Le candidat devra faire preuve d’autonomie, de rigueur et d’un bon sens de l’organisation, ainsi que d’un esprit d’initiative et d’une forte curiosité scientifique favorisant sa capacité d’apprentissage. Une aptitude à travailler en équipe dans un environnement collaboratif, à la fois académique et industriel, est également attendue
Research Work
This PhD opportunity forms part of the research activities of the C2A project, coordinated by a consortium bringing together key academic, industrial and regional stakeholders from the Adour-Pyrénées region. The project aims to develop new educational, scientific, and technological approaches to support the transition towards sustainable aviation. In this context, the doctoral research will focus on the use of big data applied to fault prediction and energy optimization in aeronautical and industrial environments [5].
Although predictive maintenance in the aerospace sector increasingly benefits from the rapid growth of available data, current approaches remain largely characterized by a siloed treatment of objectives [10, 11]. Most existing methods primarily aim to maximize resource availability or improve fault detection accuracy [8], while overlooking the energy implications related to system health and maintenance actions [13]. As a result, optimizing a single objective may encourage more energy-intensive operational strategies, revealing a trade-off that is still insufficiently addressed in current decision-making frameworks. In this context, the development of multi-objective optimization methods is crucial to simultaneously account for diverse and potentially conflicting criteria [9, 12, 13].
From a methodological standpoint, two additional challenges emerge. First, integrating heterogeneous data from multiple sources remains a persistent difficulty [9], as issues related to semantic heterogeneity, scalability, and data quality are still unresolved—particularly in Big Data contexts where formats, units, and terminology vary widely. Second, properly accounting for uncertainty in forecasting models is a critical concern. Indeed, deep learning approaches such as standard LSTM autoencoders fail to adequately capture epistemic uncertainty in their predictions [8, 11]. These limitations highlight the importance of adopting a formal multi-objective framework capable of exploring the Pareto trade-offs between energy consumption, operational costs, and resource reliability. Such an approach is essential to foster genuine energy efficiency within the Aviation 4.0 paradigm, where operational performance can no longer be decoupled from environmental sustainability requirements [8, 10, 13].
To address these scientific and industrial challenges, the doctoral research will be structured around two main research axes:
•The first research axis will focus on the development of advanced predictive analytics models aimed at the early detection of faults in industrial and aeronautical equipment [1]. Modern production and maintenance systems continuously generate large volumes of data from diverse sources, including IoT sensors, maintenance logs, operational machine data, as well as thermal, vibrational, and electrical measurements. The objective is to design advanced artificial intelligence techniques capable of leveraging this heterogeneous data to detect early signs of degradation, anticipate failure risks, and generate robust predictive maintenance indicators [11]. Such insights will also support the optimization of equipment usage by adapting operating conditions to the actual health state of the system, thereby maximizing its operational performance while tracking degradation over time [4]. This approach ultimately seeks to simultaneously enhance system availability, extend equipment lifespan, and better control maintenance costs. Attention will be devoted to challenges related to data quality, missing data management, and scalability in complex industrial environments.
•The second research axis will focus on the analysis and optimization of maintenance scheduling. To achieve this, various artificial intelligence approaches can be mobilized, particularly machine learning techniques applied to time series data, enabling the study of equipment degradation dynamics and their future operational usage. The research may also explore hybrid strategies that combine machine learning models with optimization methods based on multi-objective heuristics or metaheuristics [6]. Such approaches make it possible to simultaneously consider multiple, sometimes conflicting criteria, including the reduction of energy consumption, the improvement of operational performance, the limitation of equipment wear, and the control of maintenance costs. In addition, they can integrate dynamic variables such as operating conditions, production constraints, environmental factors, and actual equipment usage, as well as their interactions with energy consumption.
In an industrial environment where data is often sensitive, geographically distributed, and governed by strict confidentiality constraints, enabling collaborative data usage represents a major challenge for the aerospace sector. In this context, the integration of the proposed models into a distributed digital architecture aligned with the requirements of Industry 4.0 will be investigated. This research may leverage approaches such as distributed computing, real-time data processing, and federated learning to ensure secure and collaborative exploitation of industrial data [2,7], with their integration into a secure distributed architecture. The developed models will be validated on representative datasets to assess their robustness and generalization capabilities. Through this evaluation, the thesis aims to contribute to the development of advanced decision-support systems that foster more efficient and intelligent predictive maintenance strategies.
Work program
The objective of this thesis is to develop innovative approaches for the analysis of large-scale industrial data, with the aim of enhancing the reliability and performance of aeronautical systems. This work will focus on:
•The design of artificial intelligence models dedicated to predictive maintenance, capable of leveraging heterogeneous data sources.
•The identification of strategies for optimizing maintenance planning within a multi-objective framework.
•The validation of the proposed approaches through a real industrial use case, along with their integration into a secure distributed architecture aligned with Industry 4.0 requirements.
•The dissemination of research outcomes through publications in peer-reviewed journals and presentations at international scientific conferences.
Context
Lab presentation
CESI LINEACT (UR 7527), Laboratory for Digital Innovation for Businesses and Learning to Support the Competitiveness of Territories, anticipates and accompanies the technological mutations of sectors and services related to industry and construction. The historical proximity of CESI with companies is a determining element for our research activities. It has led us to focus our efforts on applied research close to companies and in partnership with them. A human-centered approach coupled with the use of technologies, as well as territorial networking and links with training, have enabled the construction of cross-cutting research; it puts humans, their needs and their uses, at the center of its issues and addresses the technological angle through these contributions.
Its research is organized according to two interdisciplinary scientific teams and several application areas.
-Team 1 "Learning and Innovating" mainly concerns Cognitive Sciences, Social Sciences and Management Sciences, Training Techniques and those of Innovation. The main scientific objectives are the understanding of the effects of the environment, and more particularly of situations instrumented by technical objects (platforms, prototyping workshops, immersive systems...) on learning, creativity and innovation processes.
-Team 2 "Engineering and Digital Tools" mainly concerns Digital Sciences and Engineering. The main scientific
objectives focus on modeling, simulation, optimization, and data analysis of cyber physical systems. Research work also focuses on decision support tools and on the study of human-system interactions in particular through digital twins coupled with virtual or augmented environments.
These two teams develop and cross their research in application areas such as
-Industry 5.0,
-Construction 4.0 and Sustainable City,
-Digital Services.
Areas supported by research platforms, mainly that in Rouen dedicated to Factory 5.0 and those in Nanterre dedicated to Factory 5.0 and Construction 4.0.
Links to the research axes of the research team involved
This This PhD thesis is integrated into Team 2 “Engineering and Digital Tools.”
Recuitement Process
Selection will be based on application review and interview.
The application must include:
•A detailed Curriculum Vitae;
•A motivation letter explaining the candidate’s motivation for pursuing a PhD;
•Academic transcripts and corresponding grade reports;
•Recommendation letters if available;
•Any additional supporting documents deemed relevant.
Applications will be processed on a rolling basis; therefore, this PhD offer will close once a suitable candidate has been selected.
FRAMEWORK & FUNDING
This PhD thesis is part of the C2A – Campus Aéro Adour project, a national initiative supporting the transition toward more sustainable and decarbonized aviation. Supported by the France 2030 program under the “Compétences et Métiers d’Avenir (CMA)” call for projects, this initiative aims to anticipate major technological, industrial, and environmental transformations in the aeronautics sector by developing a territorial ecosystem dedicated to low-carbon aircraft skills.
Duration: 3 years.
Location: CESI Campus of Pau, France.
Starting date: October 2026.
Supervisor(s):
Dr. Marwa DAAJI, Associate Professor and Researcher, CESI LINEACT, Pau
Dr. Ghita BENCHEIKH, Associate Professor and Researcher, CESI LINEACT, Pau
Dr. Amine BRAHMIA, Research Director, CESI LINEACT, Strasbourg
Bibliography:
[1]W. Zhang, D. Yang and H. Wang, "Data-Driven Methods for Predictive Maintenance of Industrial Equipment: A Survey," IEEE Systems Journal, vol. 13, no. 3, pp. 2213-2227, Sept. 2019.
[2]Llasag Rosero, R., Silva, C., Ribeiro, B. et al. Label synchronization for Hybrid Federated Learning in manufacturing and predictive maintenance. Journal of Intelligent Manufacturing, 35, 4015–4034 (2024).
[3]Suseelan, Sam. AI-based fuel optimization for sustainable aviation operations. International Journal of Engineering Technology Research & Management (IJETRM), 2026.
[4]Bencheikh, G., Letouzey, A., & Desforges, X. (2022). An approach for joint scheduling of production and predictive maintenance activities. Journal of Manufacturing Systems, 64, 546-560.
[5]Daaji, M., Benatia, M. A., Ouni, A., & Gammoudi, M. M. (2025). Predicting wind turbines faults using multi-objective genetic programming. Expert Systems with Applications, 281, 127487.
[6]Bencheikh, G. (2024). Metaheuristics and machine learning convergence: a comprehensive survey and future prospects. Metaheuristic and Machine Learning Optimization Strategies for Complex Systems, 276-322.
[7]Baahmed, A.-R.-e.-M., Dollinger, J.-F., Brahmia, M.-e.-A., & Zghal, M. (2026). HiFEL-OCKT: Hierarchical federated edge learning with objective congruence and multi-level knowledge transfer for IoT ecosystems. Internet of Things, 36, 101868.
[8]Duan, S., Sun, J., Yu, Z., & Liu, S. (2025). Uncertain data driven predictive maintenance: A cost-oriented implementation method on aircraft system. Reliability Engineering & System Safety, 264, 111278.
[9]Bensaci, M., Meftah, M. C. E., Meftah, E. H., Laouid, A., & Sheikh, S. M. (2025). Integrating Heterogeneous Data: A Systematic Review of Challenges and Evolution Solution. Proceedings of the 9th International Conference on Future Networks and Distributed Systems.
[10]Sanz, A. R., & Castán, J. P. (2018). Aviation 4.0: more safety through automation and digitization. Aircraft Technology, 25.
[11]Serradilla, O., Zugasti, E., Rodriguez, J., & Zurutuza, U. (2022). Deep learning models for predictive maintenance: a survey, comparison, challenges and prospects. Applied Intelligence, 52(10), 10934-10964.
[12]Cao, Q., Liu, S., Varghese, A. J., Darbon, J., Triantafyllou, M., & Karniadakis, G. E. (2025). Automatic selection of the best neural architecture for time series forecasting via multi-objective optimization and Pareto optimality conditions. arXiv preprint arXiv:2501.12215.
[13]Pinciroli, L., Baraldi, P., & Zio, E. (2023). Maintenance optimization in Industry 4.0. Reliability Engineering & System Safety, 234, 109204.
