Statistical Evaluation of CFD Predictions of Measured Mixing Properties of Hydrogen and Methane for Lean Premixed Combustion
Hydrogen is a fuel of interest to the combustion community research as a promising sustainable alternative fuel to replace fossil fuels. The combustion of hydrogen produces only emission of water vapor and NOx. To alleviate the NOx emission, lean combustion has been proposed and utilized in last three decades for natural gas. Therefore, evaluation of mixing properties of both methane and hydrogen in lean combustion technology such as premixers is crucial for design purposes. Increased capability of computational systems has allowed tools such as computational fluid dynamics to be regularly used for purpose of design screening. In the present work, systematic evaluation of different CFD approaches is accomplished for axial injection of fuel into non swirling air. The study has been undertaken for both methane and hydrogen. Different Reynolds Averaged Navier Stokes (RANS) turbulence models including k–ε and RSM, which are relatively attractive as being computationally efficient, are evaluated. Further, the sensitivity of RANS models to different turbulent Schmidt number (Sct), as an important parameter in mass transport analysis, has been investigated. To evaluate the numerical results, fuel concentration is measured in different locations downstream of the injection point. This is accomplished by means of flame ionization detector (FID). Finally, a comprehensive comparison has been made between numerical and experimental results to identify the best numerical approach. To provide quantitative assessment, the simulations follow a statistically design matrix which allows analysis of variance to be used to identify the preferred simulation strategies. The results suggest high sensitivity of numerical results to different Sct and relatively low sensitivity to turbulence models. However, this general trend is the opposite for radial fuel injection.