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.
