Variably saturated flow through discrete open fractures: Experimental contributions using Geotechnical Centrifuge Modelling

Abstract

Implications of improved understanding of partially saturated flow in jointed rock masses are numerous, especially given the complexity, heterogeneity and anisotropy of the intermediate fractured vadose zone. One such implication is the quantification of water movement for engineering purposes. This thesis aims to contribute to the understanding of variably saturated flow through discrete open fractures as it applies to rock engineering and geotechnical engineering applications. The initial phase comprises the analyses of results from Lugeon tests conducted at De Hoop Dam that guide a subsequent experimental phase by identifying field parameters and relationships that influence variably saturated flow processes. The series of flow visualisation experiments are developed using transparent smooth parallel fracture replicas with differing flow rates and inclinations, modelled in a geotechnical centrifuge. The findings show that the geotechnical centrifuge is a viable experimental tool for the replication of variably saturated fracture flow mechanisms. Additionally, it was found that preferential flow occupies the minority of the cross-sectional area despite the flux, and that flow becomes a matter of the continuity principle requiring substantially higher flow rates given the very low degree of saturation. Furthermore, these preferential paths are characterised by non-Darcian flow, which can be successfully evaluated using the Forchheimer relationship. The relevance of how flow occurs through fractures at partial saturation in the engineering context is understated, especially given the quest to achieve better quantification of in-situ tests during site investigations. The results prove that using common Darcian-based empirical correlations to define hydraulic conductivities from insitu tests is cautioned. Furthermore, these volume-effective approaches do not contribute to fundamental research and require a deeper understanding of the small-scale processes in the intermediate fractured vadose zone. The design of infrastructure therefore cannot be optimised without a thorough understanding of the complex flow conditions in natural and engineered rock masses.