An automated method for efficient multi-parametric analysis of porous radiant burner performance

Abstract

Combustion within porous inert media is an effective way in order to obtain high radiant outputs for a large range of power densities, while simultaneously reducing pollutant emissions. Due to the complex interactions between the different heat transfer modes involved, the parameters that influence the heat transfer process hold much potential in terms of optimization of the porous radiant burner performance. In this respect, advanced 3D-printing techniques allow for the manufacturing of non-conventional porous media configurations, with possibility for tailoring geometric and radiative properties to a specific design demand. A progressive multi-dimensional parametric investigation is conducted with the aim of characterizing the range of different geometric, radiative and operating parameters that yields optimal porous radiant burner operation, with focus on the radiant efficiency and peak solid temperature. A one-dimensional porous media combustion model is employed as a black-box, including coupled gas and solid energy equations, multi-step chemical kinetics and radiative heat transfer. The model input parameters are made independent as much as possible, leaving aside existing property correlations. An application code is developed in order to explore the parametric domain defined by the investigated parameters. It performs the autonomous management of the model simulations, assigning adequate starting estimates in order to facilitate the successful and efficient convergence of parametric simulations. It is found that, among the 8 input parameters investigated, the emissivity, excess air ratio, extinction coefficient and scattering albedo are most determinant parameters for achieving the best porous radiant burner performance, improving the radiant efficiency in more than 50% when compared to the typical values from literature. Moreover, the results from the multi-dimensional parametric study show that an individual sensitivity analysis is not sufficient for getting the best improvement, since the non-linear cross-influence of several parameter is relevant.