The hydromechanical behaviour of a fracture: an in situ experimental case study

Relationships between flow distributions and mechanical opening in a single natural fracture are investigated in situ through field experiments. at a scale of about I m, in a granitic quarry. Experiments have been conducted at various injection flow-rates while the normal stress applied to the fracture was controlled by hydraulic flat jacks. Variations of the collected flow-rate (monitored through contiguous flow collectors distributed along the fracture periphery) with the injection pressure are fully reproducible. They show that the fracture opens only above a threshold pressure which increases with the externally applied stress. This threshold is non-zero with no applied flat jacks pressure which raises questions on the reliability of hydraulic jacking techniques for the measurement of the normal stress on preexisting fractures. It is shown that equivalent hydraulic aperture and mechanical opening are comparable only above a critical mean fracture opening estimated to be around 15-20 mum for the tested granite. For mean fracture openings Smaller than this value. the standard time scale used for stress measurements distorts the results. It is also shown that channeling effects may control flow away from the injection hole so that the hydraulic jacking stress measurement technique may overestimate the mean normal stress acting oil the fracture plane by as much as 4 MPa. It is concluded that hydraulic testing techniques for normal stress measurements should not include results from the fracture opening phase. Moreover, criteria should be established for validating results from the closing phase in order to demonstrate the absence of channelling effects. Finally, it is shown that, because of the elastic response of the rock, water injection in a fracture system decreases the interstitial pressure ahead of the increasing pressure front associated with the water flow. (C) 2003 Elsevier Ltd. All rights reserved.