Hydromechanical behaviour of two unsaturated silts: laboratory data and model predictions

Agostino Walter Bruno*, Domenico Gallipoli, Joao Mendes

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

19 Downloads (Pure)

Abstract

This paper presents the results from a campaign of unsaturated and saturated isotropic tests performed on two compacted silts of different coarseness, namely a clayey silt and a sandy silt, inside triaxial cells. Some tests involved an increase/decrease of mean net stress at constant suction or an increase/decrease of suction at constant mean net stress. Other tests involved an increase of mean net stress at constant water content with measurement of suction. During all tests, the void ratio and degree of saturation were measured to investigate the mechanical and retention behaviour of the soil. The experimental results were then simulated by the bounding surface hydromechanical model of Bruno and Gallipoli (2019), which was originally formulated to describe the behaviour of clays and clayey silts. Model parameters were calibrated against unsaturated tests including isotropic loading stages at constant water content with measurement of varying suction. Loading at constant water content is relatively fast and allows the simultaneous exploration of large ranges of mean net stress and suction, thus reducing the need of multiple experiments at distinct suction levels. Predicted data match well the observed behaviour of both soils, including the occurrence of progressive yielding and hysteresis, which extends the validation of this hydromechanical model to coarser soils. Specific features of the unsaturated soil behaviour, such as wetting-induced collapse, are also well reproduced.
Original languageEnglish
JournalCanadian Geotechnical Journal
Early online date28 Aug 2021
DOIs
Publication statusE-pub ahead of print - 28 Aug 2021

Fingerprint

Dive into the research topics of 'Hydromechanical behaviour of two unsaturated silts: laboratory data and model predictions'. Together they form a unique fingerprint.

Cite this