Turbulence model choice for the calculation of drag forces when using the CFD method.

The aim of this work is to specify which model of turbulence is the most adapted in order to predict the drag forces that a swimmer encounters during his movement in the fluid environment. For this, a Computational Fluid Dynamics (CFD) analysis has been undertaken with a commercial CFD code (Fluent). The problem was modelled as 3D and in steady hydrodynamic state. The 3D geometry of the swimmer was created by means of a complete laser scanning of the swimmer's body contour. Two turbulence models were tested, namely the standard k-epsilon model with a specific treatment of the fluid flow area near the swimmer's body contour, and the standard k-omega model. The comparison of numerical results with experimental measurements of drag forces shows that the standard k-omega model accurately predicts the drag forces while the standard k-epsilon model underestimates their values. The standard k-omega model also enabled to capture the vortex structures developing at the swimmer's back and buttocks in underwater swimming; the same vortices had been visualized by flow visualization experiments carried out at the INSEP (National Institute for Sport and Physical Education in Paris) with the French national swimming team.