TRIF - Application of Artificial Intelligence in Sustainable Design of Geotechnical Infrastructure Subjected to Climate Change

Climate change affects the sustainability of infrastructure such as roads, railways and embankments. The soils forming such geo-infrastructure frequently undergo moisture variations due to climate changes. Changes in soil moisture affects its suction that governs soil strength. This can lead to premature failure and loss of serviceability of infrastructure. Nature-based stabilisation methods (e.g. surficial vegetation) have been proposed to prevent water infiltration and preserve suction within soil formations during precipitation. However, the efficiency of these methods requires understanding the interaction between climate and infrastructure and quantifying suction-dependency of soil strength. Experimental and theoretical methods have not been successful in accurate determination of such relationships due to the multivariable nature of soils. Hence, the role of suction in the design of geo-infrastructure has commonly been neglected. This can lead to overly conservative (uneconomic) design based on “null suction scenarios” and unnecessary carbon footprint due to use of non-environmentally friendly retaining structures.
The aim of this project is to apply Machine Learning models to predict the effect of moisture variations (induced by climate changes) on soil strength. The novelty of the proposed project is the application of data driven and ML approaches in Geoscience and Engineering to employ simple soil index properties (instead of cumbersome, costly and timely experimental measurements) for reliable predictions of suction-dependent soil strength. This will allow accounting for the impact of climate changes on stability of geo-infrastructure and promoting nature-based stabilisation approaches. The project output can be incorporated into design protocols of geo-infrastructure and their stabilisation systems to ensure sustainability and serviceability of civil infrastructure under changing climate are maintained.