Prediction of climate specific vertical movement of pavements with expansive soils based on long-term 2D numerical simulation of rainwater infiltration

Abstract

Long-term 2D simulation of rainwater infiltration into road embankments built on expansive soils is lacking in most highway design strategies. This paper explores the application of Hydrus-2D to establish a finite element model capable of performing long-term simulations of water flow in road embankments subjected to seasonal rainfall. Toward that end, climate specific suction profiles are generated from the simulations using real time climate data, which are used to predict vertical movement at the pavement edge. Sensitivity analyses show that the initial moisture condition has the most significant effect on moisture distribution and suction profile patterns. This induces a maximum decrease in vertical movement of 3.0 cm when a dry initial moisture state is compared with a wet initial moisture state. Higher values of evaporation rate and hydraulic conductivity result in less long-term vertical movement, although the impact is less obvious compared to initial moisture condition. The results also indicate that an extreme rain event can double the vertical displacement of pavements when compared to normal rain events, which proves the importance of using real climate data when expansive soils are involved. A case study presented the advantages of using Hydrus-2D suction profiles to predict vertical movement of pavements.