Mesoscopic scale model for predicting thermal conductivity of aeolian sand-rock-layer mixed medium in cold regions

The filling and accumulation of aeolian sand altered hydrothermal response of crushed-rock embankment (CRE) in the permafrost regions of the Qinghai-Tibet Plateau, significantly reducing its cooling capacity. Accurately determining the equivalent thermal conductivity of the mixed medium layer is the premise for evaluating the long-term thermal stability of CRE under aeolian sand conditions. Unlike traditional empirical thermophysical formulas, this paper established a mesoscopic scale model (ESE) to predict the equivalent thermal conductivity of the aeolian sand-rock-layer mixed medium, employing equivalent rules and mesoscopic three-phase theory. The ESE model was verified by laboratory experiments, demonstrating comparable accuracy to six other theoretical models. The errors of different prediction models and the main influencing factors were compared and analyzed. The results demonstrate that the predictions of the ESE model align well with experimental data. For moisture contents between 0.00 % and 15.00 %, the equivalent thermal conductivity ranges from 0.64 to 2.01 W·m⁻¹· °C⁻¹ in the frozen state (-5.00 °C) and from 0.64 to 1.61 W·m⁻¹· °C⁻¹ in the unfrozen state (5.00 °C). Among the six comparative theoretical models, the ESE model achieved the highest accuracy, with an average relative error of 5.70 %. The influence of each factor in the model ranked as follows: sand content (0.764) > porosity (0.613) > water content (0.598) > temperature (0.447). The research results provide a theoretical basis for selecting thermophysical parameters to predict cooling performance of CRE in sandy environments.