For several years Phillips Petroleum Company has been waterflooding portions of the Ekofisk Reid reservoir for purposes of enhanced oil recovery. Boreholes drilled in waterflooded portions of the reservoir have encountered poor core recoveries and highly fractured rock (poor core recoveries and highly fractured zones were not uncommon in the Ekofisk reservoir before waterflooding, however). Results of laboratory compression tests designed to simulate production-related compaction and subsequent waterflooding indicate that injection pressures currently used to inject seawater into the reservoir are high enough to induce shear failure in high porosity reservoir chalks. A model of chalk deformation explains brittle failure of chalk that has been subjected to stresses well in excess of yield stress.
Phillips Petroleum Company began producing from the giant Ekofisk Field (North Sea) chalk reservoir in 1971. Petroleum production peaked in the late 1970s at nearly 350,000 bbls of oil equivalent per day. A successful pilot seawater flood of the Cretaceous age Tor Formation reservoir in northern portions of the field to evaluate waterflooding for enhanced oil recovery led to waterflooding the northern two thirds of the Tor Formation starting in 1987. In 1988 the waterflood was expanded to include the remaining Tor Formation and the overlying Lower Danian age Ekofisk Formation reservoir. Filtered North Sea water is injected into the Ekofisk and Tor formation chalk reservoirs at bottomhole pressures ranging from 6,100 psi (42 MPa) to 7,500 psi (52 MPa) and at rates of 2,400 to 4,450 m3/day.
A well drilled in 1989 through the Ekofisk Formation within a waterflooded region of the field encountered poor core recovery, caliper logs revealed caving and hole enlargement, and recovered core was highly fractured and discontinuous. Striations on some fracture surfaces indicated that shear displacement along some of the fractures had likely occurred. These conditions raised suspicions that the apparent wellbore instability might be somehow related to the waterflood. A hydraulic fracture test conducted in a well drilled into a waterflooded portion of the reservoir showed that the effective minimum in-situ horizontal stress in the reservoir chalk is remarkably low (Teufel and Farrell, 1990). A K value (where K = effective horizontal stress/ effective overburden stress) as low as 0.04 was calculated for one horizon during the time water was injected into a nearby injection well.
Compaction of reservoir rock accompanies reduction in pore fluid pressures resulting from production of fluids from the reservoir. Uniaxial strain compression tests, in which the diameter of the specimen is maintained constant by continually adjusting the confining pressure as the axial stress is increased, are generally thought to provide good approximations of compaction of unfractured rock in a reservoir with fixed (immobile) lateral boundaries. Hence, significant compaction occurs only in the vertical direction while horizontal compaction is negligible. The second objective was to simulate the waterflood. This was accomplished by holding the total overburden stress and total horizontal stress constant, then gradually increasing the pore fluid pressure in the specimen until the rock failed.
