Abstract

Records show that ongoing global warming has changed the thermal condition of the ground in seasonally frozen areas of Canada, causing widespread ground surface settlement and harm to infrastructure, particularly embankments. In the current study, finite element numerical analysis is conducted to evaluate how climate change may influence the thermal-mechanical (TM) regimes in road embankments that are under climate conditions in two Indigenous communities in Saskatchewan, Canada, namely Yellow Quill and James Smith. This evaluation includes the analysis of embankment on the climate data from 1975 till 2100, where the data is divided into different time periods of Historical (1975-2000), Future-1 (2023-2048), Future-2 (2049-2074) and Future-3 (2075-2100. From each period, 5 representing years including extreme cold, extreme hot, expected hot, expected cold, expected mean years are considered to simulate the TM regimes with and without traffic loads. The relation for temperature-dependent thermal expansion coefficients of soils is derived and included in the modeling based on the ice content and a mixture theory. Temperature dependent mechanical properties are also involved to account for freeze-thaw induced stress redistribution and the related potential plastic deformation. According to the coupled thermal-mechanical study, which takes into account the Linear Drucker-Prager yield criterion for the stress analysis at critical location of embankments, cases with an extreme cold climate indicates the worst effect on the embankment foundation. It is reflected by the more significant temperature variations causing larger plastic zones in the toe of the embankment when compared with other climate scenarios. The extreme hot cases tend to generate more displacement on the road surface as the climate is getting warmer. The present study only sheds light on the thermal-mechanical aspect, and it does not include pore water flow behavior due to frost actions. Therefore, the result on the heave or settlement of embankment surface is not significant. Nevertheless, the inclusion of temperature-dependent thermal expansion coefficients considering the ice contents provides a better estimation of thermal-mechanical responses.