• Sedighi, Majid
• Jivkov, Ap
• Yan, Huaxiang

Abstract

Desiccation of cohesive soils due to the loss of moisture is a geo-environmental challenge and it is increasingly<br/>becoming a wide spread problem due to climate changes. The understanding of desiccation in clay soils and how it<br/>leads to cracking has substantially been improved during the last 20 years through experimental investigations and<br/>theoretical developments. However, very limited predictive models for the initiation and propagation of cracks during<br/>clay desiccation have been presented in the literature. The slow progress in this field is essentially due to the<br/>limitations of the local (differential) formulations, solved by classical numerical methods. We will present a non-local<br/>(peridynamic) formulation of clay desiccation and fracture propagation, which couples weakly the moisture flow and<br/>the mechanical deformation and allows for the emergence and evolution of discontinuities. This model is<br/>incorporated into a multi-physics computational implementation of peridynamics (Pyramid). Two series of validation<br/>exercises by comparing the model predictions with experimental data with very different setups and consequently<br/>experimental outcomes will be presented. The validations presented will include the shrinkage and cracking of a soil<br/>sample in a ring test and desiccation-induced cracking of a long clay sample. It will be shown that the model<br/>captures accurately the experimentally observed behaviour for crack initiation to occur at a narrow water content<br/>range. The simulation results, including crack initiation and number of generated cracks, correlate well with the<br/>experimental observations, lending strong support for the predictive capabilities of the model and its computational<br/>implementation in the Pyramid code.

Topics

• impedance spectroscopy
• simulation
• crack