Numerical predictions and experimental measurements of the near-field of a large turbulent helium plume

Measurements of a large, 1 m in diameter, turbulent buoyant helium plume axe taken and compared to results from numerical simulation. The experiments are carried out in Sandia's FLAME (Fire Laboratory for Accreditation of Model by Experimentation) facility and consist of a low velocity (0.351 m/s) helium plume emitting upwards into ambient air with low velocity co-flow air supplied radially to minimize the effects of the test chamber walls. The measurement approach consists of measurements of time-resolved instantaneous velocity using Particle Image Velocimetry (PIV). The numerical simulations are based on the use of the Large Eddy Simulation (LES). Results from this study show excellent agreement between the LES and experimental data for puffing frequency and mean and RMS streamwise velocity for downstream locations of X/Dp > 0.5. For locations very near the base of the plume (i.e. X/Dp < 0.5) the LES over-predicts the streamwise mean and RMS velocity and under-predicts the mean and RMS crossstream velocities. In addition, the prediction of velocities near the base of the plume exhibit a strong sensitivity to grid resolution. This sensitivity suggests the need for modification of the current dynamic Smagorinsky and gradient diffusion turbulence models to account for the effects of subgrid scale buoyancy.