Establishment and Verification of a Thermal Conductivity Model for Fine-Grained Soils

Soil thermal conductivity (STC), which plays an important role in the analysis of soil heat transfer, is widely used in various fields. In this study, based on the McKee and Bumb model for the soil-water characteristic curve (SWCC), a new STC calculation model is suggested to address the problem of insufficient prediction accuracy in existing STC models, and 27 Canadian soils are used to establish the new model. The new STC model is evaluated and compared with three existing models using six soils. For fine-grained soils over the full degree-of-saturation range, the results show that the new STC model achieves the highest prediction accuracy (RMSE = 0.056 W/m·°C), outperforming the Johansen model (RMSE = 0.077 W/m·°C), the Coté and Konrad model (RMSE = 0.082 W/m·°C), and the Bi et al. model (RMSE = 0.062 W/m·°C). Furthermore, the prediction accuracy of the four models is discussed in stages, and the new STC model performs better in the degree-of-saturation (Sr) range of 0.1–0.8. Based on this, a segment calculation method combining the advantages of the four models is proposed. The segment calculation method can better simulate the variation of STC with Sr. It is established that the new model and the segment calculation method can improve the accuracy of predicting STC and can be used in numerical simulations to calculate STC (RMSE = 0.053 W/m·°C). The research results provide more reliable technical support for the analysis of STC of infrastructure in cold regions.