Prediction of Effective Thermal Conductivity of Refractory Materials at High Temperatures based on Synthetic Geometry Generation

摘要

In the present article, the numerical prediction of the effective thermal conductivity (k(eff)) of low-carbon refractory materials at high temperatures is investigated. The employed numerical methodology consists of computational geometry generation by a modified random sequential adsorption (RSA) algorithm and solution of a heat conduction problem in the generated material sample by the finite volume method (FVM). The probability distributions, employed for modelling the grain sizes, are reexamined. Several aspects are recognised as crucial for reasonable predictions of k(eff) in the considered range from the room to the coking temperature. First, an appropriate estimate of the equivalent thermal conductivity k(rest) of the unresolved fine-scaled material is required, which is obtained from the effective medium theory (EMT). Furthermore, modelling the thermal expansion of the coarse and medium grains, leading to the formation of air gaps between the grains and the continuous phase at lower temperature, is extremely important. The presence of these air gaps could be implemented in FVM. The numerical predictions of keff show reasonably good agreement with experimental data in the complete temperature range, only if this gap width is considered as a function of the operating temperature, along with k(rest) and the temperature-dependent thermal conductivities of the constituents.