Experimental Investigation of Soil Thermal Conductivity Over a Wide Temperature Range

The results are reported of an experimental investigation of the soil thermal conductivity over a wide temperature range, for various water contents and two soil types. The results are particularly important in predictions of underground heat transfer, which require a quantitative understanding of the coupled dependence of the soil thermal conductivity on texture, temperature, and water content. In the research, comprehensive sets of thermal conductivity for Ottawa sand (coarse soil) and Richmond Hill fine sandy loam (medium soil) are experimentally obtained using the guarded hot-plate method, for temperatures ranging from 2∘C\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$2\,^{\circ }\mathrm{C}$$\end{document} to 92∘C\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$92\,^{\circ }\mathrm{C}$$\end{document} and water contents varying from complete dryness to full saturation. For both soils, the thermal conductivity is observed to vary in three stages with respect to increasing water content: a very minor increase as water content increases to the permanent wilting point, a steep increase as water content further increases to field capacity, and a minor increase (for temperatures less than 72∘C\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$72\,^{\circ }\mathrm{C}$$\end{document}) or decrease for (temperatures greater than 72∘C\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$72\,^{\circ }\mathrm{C}$$\end{document}) when the field capacity is exceeded. Then, on the basis of gathered datasets, a similar Ke(Sr,T)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$Ke(S_{\mathrm{r}},T)$$\end{document} form of the soil thermal conductivity model by Tarnawski et al. is used to empirically fit the data. The resulted correlations fit the data well with their overall root-relative-mean-square percentage errors of 4.7 % and 6.1 % for Ottawa sand and Richmond Hill fine sandy loam, respectively, and are suitable for most engineering applications.