Updated Prediction Model for Thermal Conductivity of Soil Considering Liquid Phase Division

The safety of geo-based infrastructures depends on the performance evolution of geomaterials in a cyclic temperature environment, where the thermal conductivity is a key factor. The thermal conductivity of soil, which is typically predicted using the Johansen model, varies widely because of the different thermal conductivities of solid minerals when neglecting the complex hydration of clay minerals. In this study, the water within the soil was divided into clay mineral-related bound water and gravity-related free water. Hence, considering the microstructure of saturated deposited and stabilized soils, two phases were identified. One phase consisted of solid mineral and bound water and the other was free water. The bound water content and thermal conductivity were determined using the thermogravimetric and thermal probe methods, respectively, and an updated liquid-phase division (i.e., LD) model was proposed based on the Johansen model. A comparison of the predicted and measured values showed that the precisions of the traditional Johansen and updated LD model were ± 15 % and ± 5 %, respectively. This study not only clarified the importance of soil thermal conductivity but also provided an understanding of the performance evolution of geomaterials under long-term service conditions.