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