Investigation on thermal characteristics and prediction models of soils

Zhang, Tao，Cai, Guojun*，Liu, Songyu，Puppala, Anand J.

INTERNATIONAL JOURNAL OF HEAT AND MASS TRANSFER

Volume 106, March 2017, Pages 1074-1086

DOI:10.1016/j.ijheatmasstransfer.2016.10.084

http://dx.doi.org/10.1016/j.ijheatmasstransfer.2016.10.084

 Abstract：

This paper presents details of a study that deals with evaluation of thermal characteristics of soils across a range of soil types and saturation levels. Investigations were carried out with respect to the effect of moisture content, dry density, degree of saturation, particle size, and mineralogy composition on thermal resistivity of the soils. In addition, the prediction models developed by previous researchers for estimating the thermal conductivity of various soils at unfrozen state were reviewed and evaluated. The study reveals that the moisture content, dry density, and mineralogy composition have a considerable influence on the thermal resistivity of a soil. The critical moisture content is an important index to characterize the rates of change in thermal resistivity with increasing moisture content. An exponential decrease in thermal resistivity with degree of saturation is observed for fine-grained soils, whereas a linear correlation between thermal resistivity and dry density is shown for soils regardless of the moisture contents. The differences in thermal resistivity of samples with the same dry density are mainly attributed to variations of mineralogy composition. Based on the concept of normalized thermal conductivity, a simple step by step method to calculate the thermal conductivity is proposed and the empirical soil parameters for fine-grained soils in different areas are also presented. It has been demonstrated that the normalized thermal conductivity is able to correlate the actual thermal conductivity with thermophysical parameters of the soils (viz., moisture content, degree of saturation, porosity, and thermal conductivity of solid particles). The accuracy of the prediction method can be improved by employing the empirical parameters to fit the experimental data.      全文链接