Numerical simulation of an industrial radiant tube burner using OpenFOAM

Abstract

This study presents a numerical investigation into the effects of physical models on the prediction accuracy of the wall temperature distribution in an industrial radiant tube burner. Utilizing a reacting flow solver based on OpenFOAM, we explored the effects of various physical models, including those for chemistry, combustion, heat transfer, and radiation properties. The choice of combustion model significantly influences prediction accuracy, playing a more dominant role than the chemistry mechanism. Moreover, the simulations captured a distinctive triple flame structure inside the burner, representing the coexistence of rich premixed, non-premixed, and lean premixed flame structures. Conditional scatter plots displayed the development of both premixed and non-premixed flame structures, converging on the fuel-lean side. Notably, accurate prediction of wall temperature distribution depends on the incorporation of a precise heat transfer model, coupled with a detailed radiation property model. Regarding the distribution of tube surface temperature in the main radiation zone (a distance from the burner nozzle greater than 1 m), the most accurate prediction exhibits a maximum deviation of less than 56 K and an average deviation of 24 K compared to experimental results. The simulation closely matched experimental data for exhaust concentration of NO within an error margin of 20 ppm. However, discrepancy was observed in the CO concentration, which was attributed to the simplified representations of fuel chemistry and composition, as well as the difficulties in accurately capturing the unsteady flame dynamics near the wall.