ABSTRACT
Joints provide paths for fluid flow through rock. Prediction and control of contaminant transport in ground water requires understanding the process of fluid flow in fractures.
In this study, a natural fracture was isolated from a low-permeability rock mass within a prototype in situ test block (volume 8 m3). Three-dimensional displacement monitors surrounded the test fracture and a borehole probe monitored the fracture1 s gas conductivity. The fracture was loaded by hydraulic flatjacks in the block’s boundary slots.
The results showed that rock displacement and fracture conductivity often changed in contradictory ways under load. The lower-than-expected correlation between mechanical and flow responses of the test fracture apparently was due to the presence of a large number of semi-independent rock subblocks; the test fracture behaved as two flexible tiled walls rather than two rigid plates. The subblocks were defined by seemingly minor, discontinuous joints. The instrumentation grid was too coarse and arbitrarily placed to allow satisfactory characterization of the behavior of the joint network.
Minor fractures thus appear to be a major cause of behavioral unpredictability. However, these fractures are extremely difficult, if not impossible, to map or characterize in detail. This leads to the fundamental question of whether confident modeling of jointed rock masses is feasible with current deterministic approaches to both modeling and field instrumentation.
