Improved Monte Carlo methods for computational modelling of thermal radiation applied to porous cellular materials

Abstract

Abstract Highly porous cellular media such as plastic, ceramic and metal foams present specific characteristics that make them interesting for a number of applications related to thermal engineering. In many of these applications, thermal radiation heat transfer can have a significant, ranging from a 20–30% effect for plastic foam insulation to becoming the predominant mode of heat transfer at higher temperatures. To model radiation heat transfer in these media, accurate models of radiation and microscopic structure of the foam are required. Over time, the state of the art has evolved from basic approaches using radiative conductivity to multi-scale models based on equivalent homogeneous media and the Radiative Transfer Equation, with more recent techniques such as MPA and GRTE allowing to model multiple phases, non-Beerian behavior, and history effects. Still, comparisons of homogenized models with reference results are sparse in literature, and results mixed. Furthermore, what more advanced models afford in accuracy they lose in simplicity, leading to cumbersome calculations. In this paper, we combine in a novel way hybrid direct-inverse identification of radiative properties and accounting of history effects, to present modified and improved version of the standard Homogenous Phase Approach (HPA+) and Multi Phase Approach (MPA+), with the intent of increasing accuracy while keeping maximum simplicity. We then compare existing and new models with a reference model based on Monte Carlo Ray Casting. In the comparison, we consider a varied set of porous morphologies and boundary conditions. The improved models consistently outperform existing homogenized models.