Uncertainties in the estimation of unsaturated shear strength from soil-water characteristic curve

The shear strength of a soil is an important engineering property that is required in various geotechnical analyses involving saturated and unsaturated soils. The behaviour of unsaturated soils is governed by the Soil-Water Characteristic Curve that relates water content to soil suction. The contribution of shear strength due to soil suction has been proven to be substantial for certain soils and needs therefore to be accounted for to fully characterise the shear strength behaviour of a soil. Experimental procedures for the direct determination of the unsaturated shear strength in the laboratory are available. However, lengthy testing procedures, complexity of the equipment and the high level of expertise required are limiting the application for practical purposes. As a result, numerous prediction equations that utilise the Soil-Water Characteristic Curve to estimate the unsaturated shear strength have evolved over the past decades. Despite their advantage to swiftly produce the unsaturated shear strength, the estimated shear strength is underpinned by several assumptions and uncertainties that are unavoidable. Prediction equations to estimate the shear strength of unsaturated soils are widely used in the fields of research. However, limited studies have been carried out to address and quantify the various uncertainties in prediction equations that result in variability in the unsaturated shear strength estimation. This research focused on the estimation of the shear strength due to soil suction and the uncertainties associated with the prediction equations for the unsaturated shear strength. The shear strength behaviour for unsaturated soils was reinterpreted based on the typical desaturation behaviour of a soil along the Soil-Water Characteristic Curve and an equation to estimate the unsaturated shear strength was proposed. Uncertainties in the estimation of the unsaturated shear strength were assessed and subsequently quantified. Experimentally-obtained results from a soil tested in this research were used to illustrate the variability in the estimated shear strength of an unsaturated soil. The shear strength behaviour proposed in this research agreed well with the experimentally-obtained shear strength of the soil tested in this research. Additionally, based on an independent Database of unsaturated shear strength measurements, the proposed equation from this research indicated an improved predictive capability compared to existing equations for the estimation of the unsaturated shear strength. Uncertainty analyses, performed on the soil tested in this study, indicated that the variability in the estimated shear strength at the cohesion intercept was of the same magnitude as the reported variability of the effective cohesion from direct measurements.Doctor of Philosophy (CEE