The French collaborative Trio4CLF project aims to understand and control the cryogenic cooling of amplifiers for high power (~1 PetaWatt) and high repetition rate (~1 Hertz) lasers. In such amplifiers, diodes or flash lamps heat up the amplification plates which partly absorb the radiative power. To preserve the laser quality and coherency, it is crucial to maintain a homogeneous temperature distribution in the amplifier with differences below a few Kelvin. As the crystal thermal conductivity increases with decreasing temperature, a possible solution is to maintain the amplifier in the range 50 – 150 Kelvin. Moreover, some crystals have a better amplifying efficiency at low temperature. The project proposes to cool down a laser amplifier by a cryogenic gaseous helium flow. Due to the low viscosity of cold helium, turbulence will develop in the flow between the amplifier plates, which can affect the coherency of the laser beam through inhomogeneous refractive index distribution. Thus, a precise knowledge of the heat exchange and of the turbulent fluid flow in the amplifier is requested.

To address this problem, Trio4CLF combines experimental and numerical studies. On one hand, the heat exchange between the solid and the turbulent flow is investigated by Large Eddy Simulations. On the other hand, experimental measurements are carried out to validate the numerical results. The experimental setup is a closed-loop wind tunnel called TRANSAT. The airflow passes through a solid grid at the entrance, so that the flow characteristics are close to an isotropous homogeneous turbulent flow with a controlled turbulence intensity. Two horizontal plates, separated by a few centimeters, are put in this flow to represent the amplifier plates.

Large Eddy Simulations have been carried out in the TRANSAT configuration using a CFD code developed by the CEA: TrioCFD. The entrance flow is a homogeneous planar flow with a constant velocity at 10 m/s. Turbulent boundary layers develop from the plates edge. To correctly capture them, the full height of the wind tunnel is simulated. The mesh is coarsen above and under the plates and refined between the plates, resulting in a 27 million cells mesh. The subgrid-scale viscosity is computed using the WALE (Wall Adapting Local Eddy) model of Nicoud and Ducros, well adapted for wall-bounded flows. The time scheme is a third order Runge-Kutta. For the dynamic convection, a second order centered scheme is used.

In the present paper, the dynamic field is presented. The development of the turbulent boundary layers created by the plates is observed. By comparison with the experimental results, it will be possible to assess the quality of the simulations with respect to the numerical choices, calculation scheme, mesh refinement and boundary conditions. The results will also give indications on the geometrical key features of the problem.