Validation of a Reynolds-averaged numerical simulation environment to simulate high-pressure, auto-igniting hydrogen diffusion flames

The hydrogen (H2) fueled direct-injection (DI) compression-ignition (CI) argon power cycle (APC) is an attractive technology to counteract the mismatch between energy demand and supply from renewable energy sources. The development of the APC, as well as air-breathing DI CI H2 engines, can be advanced by computational fluid dynamics (CFD) models. Reynolds-averaged Navier-Stokes (RANS) combustion models are an effective approach to predict global in-cylinder variables such pressure and heat release rate. However, validity of this method to simulate the complex phenomena associated with high-pressure, auto-igniting hydrogen jets is not ensured due to the underlying model assumptions. In this study, a RANS CFD environment using two commonly used eddy-viscosity models and the Reynolds stress model (RSM) is extensively validated. Combustion is modeled with a detailed chemistry model. First, hydrogen distributions in non-reacting jets are compared against literature data to assess the accuracy of the models on turbulent mixing at high pressure. Subsequently, the capability of the model to simulate auto-igniting jets is assessed by comparison with reported measurement data of closed-vessel experiments at high pressure. Ignition delay times, pressure rise profiles, and heat release rate profiles are compared for different ambient temperature and oxygen concentration. Adequate agreement is found for the diffusive combustion phase for all models, despite the lack of turbulence-chemistry interaction in the combustion model. Trends with ambient temperature and oxygen concentration were well predicted and the best agreement is found for the RSM.