PLoS ONE, Journal Year: 2024, Volume and Issue: 19(7), P. e0305529 - e0305529

Published: July 12, 2024

This study investigates the thermal conductivity ( λ ) and volumetric heat capacity C of sandy soil samples under a variety conditions, including freeze-thaw cycles at temperatures both above below zero differing moisture levels. To estimate these properties, novel predictive model, EFAttNet , was developed, which utilizes custom-designed embedding attention-based fusion networks. When compared to traditional de Vries empirical models other baseline algorithms, demonstrated superior accuracy. Preliminary measurements showed that values increased linearly with content but decreased temperature, whereas exhibited rising trend freezing temperature. Following cycles, were positively influenced by The -based model proved highly accurate in predicting particularly effective capturing nonlinear relationships among influencing factors. Among factors, degree saturation had most significant impact, followed number subzero temperatures, porosity, content. Notably, dry density exerted minimal influence on likely due overriding effects factors or specific characteristics, such as particle size distribution mineralogical composition. These findings have implications for construction engineering projects, especially terms sustainability energy efficiency. accuracy estimating properties various conditions holds promise practical applications. Although focused types insights gained can guide further research development managing across diverse environments, thereby enhancing our understanding application this field.

Language: Английский