Status
Scientific disciplines
Research direction
Applied Physico-chemistry and Mechanics
Affiliate site
Rueil-Malmaison
Wind farm flows are complex and difficult to predict, although their understanding is crucial for proper wind farm design and layout optimization. During the design phase, so-called “engineering” flow models, whose purpose is to provide an analytical representation of wind turbine wakes, are used to estimate wind farm power production. However, the accuracy of these models is still limited. Several phenomena are still poorly understood, such as overlapping wakes, turbulence generation in wakes, wake deflection, thermal effects, etc.
To improve these models, researchers rely on high fidelity numerical simulations, using the Large Eddy Simulation framework. Indeed, several Navier-Stokes based solvers are capable of dealing with both the presence of the wind turbines, using Actuator-Line (AL) techniques, and the Atmospheric Boundary Layer (ABL). But the computational cost remains too high: It is a challenge to apply these solvers for parametric studies and complete wind farm flow configurations. For such a study one must deal with both the atmospheric eddies (kilometric scale) and the wind turbine wake eddies (metric scale).
To overcome these limitations, we propose to pursue, in this Ph.D. thesis, the development of the ABL functionality in the open-source Lattice-Boltzmann-based flow solver waLBerla (FAU), able to perform accurate simulations in almost real-time. We target similar functionalities and physical representation compared with SOWFA, which is currently the most widely used Navier-Stokes-based (NS) flow solver in the wind energy community, but with much lower computational times. While the AL method is already implemented and validated, the ABL functionality is missing. After one year in Erlangen (Germany), where the Ph.D. student will get familiar with the solver, the student will be based in Rueil-Malmaison (France) for the last two years.
Keywords : Wind energy, wind farm, Atmospheric-Boundary-Layer, Lattice Boltzmann, HPC