Integrated Approach for Coastal Building Resilience to Compound Flooding Events

Context :

This PhD thesis will be carried out in the framework of the international project: SMart and Robust ClimAte Change Adaptation of Coastal Buildings (SMACHA) funded by the French National Research Agency and the University Grants Council from Hong Kong. The proposed PhD topic revolves around developing an integrated framework to enhance the resilience of coastal buildings facing compound flooding events. Compound flooding, resulting from the combination of storm surges, heavy rainfall, and sea-level rise, presents a significant challenge to coastal infrastructure.

Scientific proposal :

The first stage of the research lies in accurately predicting the behavior of coastal buildings under various flooding scenarios. Advanced numerical models will be implemented to simulate fluid-structure interactions (FSI) during compound flooding events. Experimental data provided by the project partners will be used to validate the FSI models. Recognizing the computational challenges associated with detailed numerical simulations, the second stage of the research integrates machine learning (ML) techniques with physics-based modeling. Physics-guided deep learning models will be trained using input-output datasets derived from simulations and experimental measurements. By embedding physics knowledge into ML models, the PhD student will develop surrogate models capable of predicting structural responses efficiently and accurately. These ML models will not only complement the numerical simulations but also enable rapid assessments of structural vulnerabilities and risks under various compound flooding scenarios. The insights gained from simulations and experimental data will inform the training of these ML models, ensuring a synergistic relationship between the two approaches.

Building upon the predictive capabilities established FSI and ML models, the research will also focus on developing flexible and robust adaptation strategies to enhance coastal building resilience. By assessing the effectiveness of multi-level adaptation measures, including component-level, building-level, and community-level interventions, the final stage will leverage insights gained from numerical simulations and ML predictions. The flexible management approach, integrating Bayesian statistical models and multistage stochastic dynamic programming, will enable adaptive decision-making in the face of evolving climate uncertainties and structural performance insights. Through case studies and benefit-cost analyses, the research will quantify the efficacy and financial viability of adaptation strategies, guiding the formulation of robust strategies capable of addressing deep uncertainties associated with climate change.

Profile required:

The candidate must possess strong skills in numerical simulation, particularly in fluid dynamics, machine learning, and probabilistic methods. He/she should hold a Master's degree in Civil or Mechanical Engineering and be motivated to learn new software and scientific programming languages. Knowledge of C++, Python, and OpenFOAM is considered a plus. The candidate should have excellent writing and speaking skills in English, with proficiency in French being optional but beneficial. Additionally, the ability to work well in a team is essential.

Laboratoire des Sciences de l’Ingénieur pour l’Environnement : LaSIE UMR - 7356 CNRS - Université de La Rochelle.
Les activités du laboratoire ont pour domaines applicatifs :

Durabilité et protection des matériaux sous contraintes environnementales,

Qualité des ambiances habitables,

Eco-procédés pour la qualité des produits et la valorisation énergétique des bio-ressources.

L’unité réunit un large spectre de compétences avec des approches intégrées depuis l’échelle atomique jusqu’au matériau, au bâti et son environnement à différentes échelles de temps et d’espace.
Elle établit un continuum du développement d’outils mathématiques aux applications et dépôts de brevets, en passant par des modèles et simulations numériques et expérimentales.

The candidate must possess strong skills in numerical simulation, particularly in fluid dynamics, machine learning, and probabilistic methods. He/she should hold a Master's degree in Civil or Mechanical Engineering and be motivated to learn new software and scientific programming languages. Knowledge of C++, Python, and OpenFOAM is considered a plus. The candidate should have excellent writing and speaking skills in English, with proficiency in French being optional but beneficial. Additionally, the ability to work well in a team is essential.