This topic deals with turbulent mixing occurring at the interface between fluids of different densities, which is important in many industrial applications.
The Rayleigh–Taylor instability is a key mechanism underlying these phenomena.
It arises when pressure and density gradients are opposed.
Its evolution proceeds through several stages: linear, weakly nonlinear, and then fully turbulent.
A major objective is to develop RANS models capable of accurately reproducing this complex dynamics.
The project proposes to evaluate the GSG Reynolds stress model in this context.
High-resolution DNS datasets are available for different density contrasts.
Several calibration methods will be investigated, including analytical, statistical, and Bayesian approaches.
Finally, improvements to the model may be achieved through data assimilation and machine learning techniques.