Salt migration and accumulation are the main sources of salt expansion. To study the role of salt transport in soil, laboratory tests were conducted to simulate salt transfer under real conditions. Temperature, water content, salt content and soil displacement of a sodium sulfate soil were measured during the freezing process/freeze-thaw cycles. Meanwhile, Salt concentration was regressed bases on test data to investigate the movement of salt diffusion, in the process, the Pitzer ion model was employed to calculate the freezing point of saline soils to determine the frozen depth. Moreover, the amount of crystallization was estimated by the saturation curve of a sodium sulfate solution and the detected salt content. The results show that salt transfer in soil occurs as a result of numerous physicochemical processes and that the maximum salt transport occurred in the frozen fringe zone in the soil. Salt crystallization increases the effect of salt transfer in the soil unidirectional freezing process. In addition, salt expansion had an accumulative effect, and it increases as the number of freeze-thaw cycles increases. The quantity of salt that is transported increases as the salt content increases.

Saline soils is porous medium whose thermal conductivity is affected by components and particle shape. However, the particle shapes in geomaterials are various, and the component and its content are in dynamic change as the soil varies between unfrozen and frozen state. In this study, based on the theory of multi-phase medium mixed geothermal conductivity, a new model is developed to explore the dynamic variation characteristics of the thermal conductivity of sulfate soils. The developed model is validated by comparing the experimental results against the calculated ones. Compared with the other two models, the new model is more reasonable and effective, especially in low temperature.