岩土工程
压实
抽吸
结算(财务)
土壤压实
土壤水分
路基
保水性
含水量
压力(语言学)
地下水
地质学
变形(气象学)
环境科学
土壤科学
工程类
哲学
付款
计算机科学
万维网
海洋学
机械工程
语言学
作者
Yongsheng Yao,Junjun Ni,Jue Li
标识
DOI:10.1016/j.compgeo.2020.103835
摘要
Granite residual soil is widely used as a subgrade material in pavement engineering. Previous studies have not investigated the influence of stress on the soil water retention curve for this type of unsaturated soil or its significance for predicting ground settlement. Thus, the aims of the present study were: (i) to experimentally investigate the water retention curves for granite residual soil at various stress levels and under different degrees of compaction; and (ii) to numerically explore the influence of stress-dependent water retention on the suction distribution and ground settlement during both rainfall and evaporation periods based on coupled seepage and deformation analyses. The experimental results showed that the applied vertical stress (from 0 to 90 kPa) decreased the retained water content by up to 13% when the suction was less than 200 kPa. The reduction in the retained water content was much more obvious when the degree of compaction decreased from 100% to 85%. The effects of stress on the water retention curves were almost negligible when the suction was higher than 200 kPa, regardless of the degree of compaction. The numerical results showed that the current analysis considering the influence of stress on water retention could predict a higher ground heave and a larger ground settlement, compared with the conventional method without stress effects. This was because the current analysis captured the larger suction reduction during rainfall and higher suction increase during evaporation. The ground deformation became larger when the rainfall intensity increased while the corresponding rainfall duration decreased under the same return period of rainfall. This study implied that the method employed in the present study was more scientific and conservative for assessing ground settlement in pavement engineering.
科研通智能强力驱动
Strongly Powered by AbleSci AI