Optimisation des simulations haute-fidélité du champ de vent dans un parc éolien par l’apprentissage automatique

Wind energy plays a crucial role in the global transition to clean and sustainable energy. Maximizing wind farm efficiency requires accurate prediction of wind behavior. This project aims to enhance wind field simulations by combining high-fidelity computational fluid dynamics (CFD) models with machine learning techniques. While CFD simulations provide detailed insights into wind-turbine-terrain interactions, they are computationally intensive, limiting their use for real-time decision-making.
The project uses machine learning to learn from CFD data and build surrogate models that predict wind conditions much faster without losing accuracy. This approach enables rapid assessment of wind patterns, supports optimal turbine placement, improves operational strategies, and enhances maintenance planning.

Conducted at École de technologie supérieure (ÉTS) in Montreal, the research involves a collaborative team led by Dr. Reda Snaiki, with access to high-performance computing resources and expert guidance from PhD students. The intern will develop machine learning models and validate them against CFD simulations and, when possible, real wind data.
The expected outcomes include more informed operational decisions, increased energy production efficiency, and reduced costs. This methodology could be applied to various renewable energy systems, advancing sustainable and resilient clean energy technologies worldwide.

Faculty Supervisor:

Reda Snaiki

Student:

Partner:

Mohammed VI Polytechnique University

Discipline:

Engineering

Sector:

Education

University:

École de technologie supérieure

Program: