STATISTICAL EVALUATION OF RANS SIMULATIONS COMPARED TO EXPERIMENTS FOR A MODEL PREMIXER

This paper investigates the ability of computationally time efficient CFD models to predict mixing of hydrogen and air in a model premixer with a complex 3D geometry. Axial and radial injection of hydrogen into swirling and non-swirling air streams at standard and elevated pressures and temperatures conditions are considered. Statistical analysis is applied to the results of comparison between experimental measurements and numerical simulations. The analysis of variance suggests that, for non-swirling flows, reasonable numerical predictions can be made if appropriate modeling parameters, such as turbulence model and turbulent Schmidt number, are considered. Comparing the results at standard conditions to those at elevated pressures and temperatures suggests that the conclusions at standard conditions may hold if appropriate scaling parameters are utilized.