An Improved Model for Predicting the Thermal Conductivity of Sand Based on a Grain Size Distribution Parameter

Soil thermal conductivity is a critical parameter in various fields related to soil heat transfer, mainly af-fected by porosity and saturation. In this research, the thermal conductivity of three types of sands with different properties was measured. Eight soil thermal conductivity models (including two semi-empirical models and six empirical models) were compared and evaluated. By analyzing the inadequacies of the ex-isting models and microscopic mechanism of soil, a grain size distribution parameter was defined, and an improved model was proposed. The results showed that four empirical models from published literature provided relatively satisfactory performance but still required improvement. The improved model can ac-curately predict the thermal conductivity of different types of soils as the root mean square error for the three soils were 0.072, 0.073, and 0.056 W center dot m -1 center dot k -1, respectively. Moreover, compared with the existing models, the improved model provided more scientific explanations for the limiting cases of porosity and saturation. In addition, this model revealed a close linear relationship between the thermal conductiv-ity of soil and solid particles. The prominent nonlinear relationships between thermal conductivity and porosity or saturation were also depicted. Furthermore, the content of fine particles has a remarkable impact on the thermal conductivity of dry soil. It is suggested that the improved model can be used for predicting the complete thermal conductivity of soil in case there is no available field test data. (c) 2023 Elsevier Ltd. All rights reserved.