Description

The energy transition relies in part on the massive deployment of marine renewable energy. Among emerging solutions, floating photovoltaic (FPV) systems in marine environments (coastal and offshore) offer considerable potential where land availability is limited. These power plants consist of interconnected modular structures forming large floating islands.

However, the hydrodynamic behavior of these multibody structures under wave action remains challenging to predict, particularly regarding the complex interactions between modules and the accurate assessment of mechanical loads at the connections. Current numerical approaches (such as Boundary Element Methods) are extremely computationally expensive when simulating tens or hundreds of connected modules, rendering them impractical for design optimization phases.

The originality of this project lies in overcoming this bottleneck through the development of an innovative hybrid numerical framework. It will combine fast physics-based modeling with machine learning models trained to predict the hydrodynamic behavior.

Skills required

Education: Holder of (or expecting to graduate by the end of 2026 with) a Master’s degree or an Engineering degree in marine hydrodynamics, applied mathematics, scientific computing, or fluid mechanics. Technical Skills: Strong foundation in numerical modeling, scientific computing, and algorithmics. Excellent programming skills (Python, Matlab, or equivalent) are essential. Interest in or prior experience with Machine Learning / data science is a plus. Key Attributes: Scientific rigor, autonomy, strong collaborative skills, and a strong motivation for innovation in offshore renewable energy.