Effective thermal conductivity (λ) has essential applications in disciplines such as geotechnical and geoenvironmental engineering, agriculture science and vadose zone hydrology. Based on the similarity between water retention and heat transfer in soils, some models have been established to estimate λ from soil suction (s), providing a unique way to study the coupled water flow and heat transmission in soils. However, few data are available in literatures for evaluating such relationship, and existing models are generally restricted to specific soil types and s ranges, without considering void ratio (e) effect common in engineering practice. In this study, a series of water retention tests and λ measurements were conducted on 6 typical Chinese soils with different e over a wide s range. Results indicated that for different soils, e could exert considerable impacts on both the soil water retention curves in the low and middle s regimes, and especially λ-s relations over a wide s range. According to the normalized concept, the relationship between normalized λ (λn) and natural logarithm of s (lgs) was also not characteristic for different soils with identical e or a soil with different e. Besides, increasing e or decreasing plasticity could cause delayed response of soil λ to variations in s, especially in the high and middle s regimes. A Weibull-based formular was established to describe the λn-lgs relations, giving a new approach to estimate soil λ from s over the full moisture range based on the normalized concept. By comparing with other existing models in literatures, the proposed λ(s) model was proved feasible and effective, giving satisfactory estimations of soil λ over the full moisture range (MAE ranging 0.014–0.086 W m−1 K−1, AIC < − 165.0 to − 290.7, centered RMSE < 0.020–0.102 W m−1 K−1 and R > 0.97) that outperformed all the other models investigated. The new λ(s) model has 3 parameters with clear physical meanings, i.e., φ accounting for the effect of e is a material (constant) parameter, B is related to water retention ability, and C is correlated with pore size distribution. Future efforts on measurement of λ(s) should be conducted under various soil textures (e.g., sandy and silty soils) and conditions (e.g., hysteresis effect), nevertheless the present work provides a new insight and framework for comprehensively understanding the linkage between water retention and heat transfer in soils.

Effect of void ratio on the relationship between thermal conductivity and soil matrix suction over the full moisture range

期刊论文
发表日期：

2026年6月23日

查看全文

Fan Peng,
Chen Bo,
Yunzhi Tan,
De’an Sun,
You Gao,
Yang Zhao

摘要

Effective thermal conductivity (λ) has essential applications in disciplines such as geotechnical and geoenvironmental engineering, agriculture science and vadose zone hydrology. Based on the similarity between water retention and heat transfer in soils, some models have been established to estimate λ from soil suction (s), providing a unique way to study the coupled water flow and heat transmission in soils. However, few data are available in literatures for evaluating such relationship, and existing models are generally restricted to specific soil types and s ranges, without considering void ratio (e) effect common in engineering practice. In this study, a series of water retention tests and λ measurements were conducted on 6 typical Chinese soils with different e over a wide s range. Results indicated that for different soils, e could exert considerable impacts on both the soil water retention curves in the low and middle s regimes, and especially λ-s relations over a wide s range. According to the normalized concept, the relationship between normalized λ (λ_n) and natural logarithm of s (lgs) was also not characteristic for different soils with identical e or a soil with different e. Besides, increasing e or decreasing plasticity could cause delayed response of soil λ to variations in s, especially in the high and middle s regimes. A Weibull-based formular was established to describe the λ_n-lgs relations, giving a new approach to estimate soil λ from s over the full moisture range based on the normalized concept. By comparing with other existing models in literatures, the proposed λ(s) model was proved feasible and effective, giving satisfactory estimations of soil λ over the full moisture range (MAE ranging 0.014–0.086 W m⁻¹ K⁻¹, AIC < − 165.0 to − 290.7, centered RMSE < 0.020–0.102 W m⁻¹ K⁻¹ and R > 0.97) that outperformed all the other models investigated. The new λ(s) model has 3 parameters with clear physical meanings, i.e., φ accounting for the effect of e is a material (constant) parameter, B is related to water retention ability, and C is correlated with pore size distribution. Future efforts on measurement of λ(s) should be conducted under various soil textures (e.g., sandy and silty soils) and conditions (e.g., hysteresis effect), nevertheless the present work provides a new insight and framework for comprehensively understanding the linkage between water retention and heat transfer in soils.