Modelling thermal conductivity on salt-affected soils and its modification

• Published in: International journal of thermal sciences 2023-03, Vol.185, p.108071, Article 108071
• Main Authors: Ju, Zhaoqiang, Guo, Kai, Liu, Xiaojing
• Format: Article
• Language: English
• Subjects: Aqueous solutions; Predictive model; Saline soil; Soil thermal conductivity
• ISSN: 1290-0729; 1778-4166
• DOI: 10.1016/j.ijthermalsci.2022.108071

Accurate prediction of soil thermal conductivity (λ) is essential for the stability of architectural and civil engineers and for studying coupled heat and water transport in saline regions. However, the present λ predictive models don't consider salt effects on the λ estimation for salt-affected soils. In this study, we compare the measured and predicted λ as a function of water content in salt-affected soils. The thermal conductivities of dry soil (λdry) and saturated soil (λsat) were evaluated with previously developed models on various salt-affected soils. Due to the unsatisfactory performance of the existing models, a modified λ model considering the effects of soil salt was developed, and applied to the measured data. The results indicate that three existing models overestimate λ at intermediate water content for most of the salt-affected soils. The Côté and Konrad λdry model can be employed to accurately estimate λdry of salt-affected soils. The thermal conductivity of the aqueous solution (λw) in saline soil was 0.717 times that of salt-free water. The new modified λ model is based on the Lu 2007-model, estimates λdry with the Côté and Konrad 2005 model, and uses the actual λw of the soil solution. So, the predicted errors result lower. An independent validation indicates that the λ values predicted with the modified λ model well agrees with the known λ values. This study provides references for the application of the λ model in studying the thermal regime and the coupled heat and water transfer in salinized soil. •Three existing models overestimate λ at intermediate θv for most of the salt-affected soils.•The λ of the aqueous solution in saline soil was 71.7% of λ of pure water.•The modified λ model decrease the predicted errors in the salinized soils.