Analysis of high-pressure fluid flow in fractures with application to Yucca Mountain, Nevada, slug test data

Abstract

Increases in water pressure, if large enough, can be sufficient to open existing fractures in a fractured rock-water system. Such large water pressure increases are expected during earthquake activity and are also produced artificially in high-pressure hydraulic fracturing and slug injection tests. In an in-situ slug injection test, the slug of water is initially released into a packed-off interval of the borehole where the contained fluid is at ambient pressures. We recognize three separate stages in the behaviour of the fracture as the water pressure decreases due to drainage, namely opening, closing and, finally, completely closed. In the first stage, the water will initially drain out along conduits between the faces of the opening fracture. As the fracture opens and grows, it can become long enough to intersect any pre-existing network of open joints and fractures. If this connection occurs, the opened fracture will stop growing and there will be further rapid flow of fluid out along these open conduits. Even if connection to an open conduit does not occur, as the pressure drops the fracture will stop growing and eventually halt. At the pressure where the fracture tip begins to close, drainage into any open conduits will cease and drainage during fracture closing will thus be by flow into the porous walls of the fracture and borehole. This effect of the conduit being shut off may manifest as a kink in the observed height decay of the slug at this pressure due to the effect of the different drainage terms. As the water pressure decreases, the fracture will continue to close until the water pressure drops below the critical pressure and the fracture becomes completely closed. This total closure occurs at the same pressure as that necessary to re-open the fracture, and has been taken equal to the minimum principal stress. This total closure of the fracture also may manifest as a kink in the observed height decay of the slug, as below this closure pressure is the third, and final, stage where drainage is at the usual low pressures controlled by the permeability of the rock just around the packed-off interval of the borehole. We develop the most general mathematical model for this phenomenon using conservation of mass of the water in the borehole-fracture system for each separate pressure stage. This model of fracture behaviour, with separate flow stages and pressure regimes, explains the physical basis for the increasingly used technique in high-pressure injection tests of measuring the so-called re-opening pressure at the sharp change in slope of the rate of change of water height vs. height. Analytic and numerical solutions yield the rate at which the water height drops during the slug test and comparison with high-pressure test data confirms the accuracy of the model. Analysis of data from Yucca Mountain, Nevada, gives values for the fracture opening pressures that indicate the presence of extensional stresses of significant magnitude and of zones of pre-existing open fracture networks in the vicinity of the proposed repository site.