Theoretical aspects of cap-rock and fault seals for single- and two-phase hydrocarbon columns

Cap-rock seals can be divided genetically into those that fail by capillary leakage (membrane seals) and those whose capillary entry pressures are so high that seal failure preferentially occurs by fracturing and/or wedging open of faults (hydraulic seals). A given membrane seal can trap a larger oil column than gas column at shallow depths, but below a critical depth (interval), gas is more easily sealed than oil. This critical depth increases with lower API gravity, lower oil GOR and overpressured conditions (for the gas phase). These observations arise from a series of modelling studies of membrane sealing and can be conveniently represented using pressure/ depth (P/D) profiles through sealed hydrocarbon columns. P/D diagrams have been applied to the more complex situation of the membrane sealing of a gas cap underlain by an oil rim; at seal capacity, such a two-phase column will be always greater than if only oil or gas occurs below the seal.These conclusions contrast with those for hydraulic seals where the seal capacity to oil always exceeds that for gas. Moreover, a trapped two-phase column, at hydraulic seal capacity will be less than the maximum-allowed oil-only column, but more than the maximum gas-only column. Unlike membrane seals, hydraulic seal capacity should be directly related to cap-rock thickness, in addition to the magnitude of the minimum effective stress in the sealing layer and the degree of overpressure development in the sequence as a whole.Fault-related seals are effectively analogous to membrane cap-rocks which have been tilted to the angle of the fault plane. Consequently, all of the above conclusions derived for membrane cap-rocks apply to both sealing faults sensu stricto (fault plane itself seals) and juxtaposition faults (hydrocarbon trapped laterally against a juxtaposed sealing unit). The maximum-allowed two-phase column trapped by a sealing fault is greater than for equivalent oil-only and gas-only columns, but less than that predicted for a horizontal membrane cap-rock under similar conditions. Where a two-phase column is present on both sides of a sealing fault (which is at two-phase seal capacity), a deeper oil/water contact (OWC) in one fault block is associated with a deeper gas/oil contact (GOC) compared with the adjacent fault block. If the fault seal is discontinuous in the gas leg, however, the deeper OWC is accompanied by a shallower GOC, whereas a break in the fault seal in the oil leg results in a common OWC in both fault blocks, even though separate GOC's exist. Schematic P/D profiles are provided for each of the above situations from which a series of fundamental equations governing single- and two-phase cap-rock and fault seal capacities can be derived. These relationships may have significant implications for exploration prospect appraisal exercises where more meaningful estimates of differential seal capacities can be made.The membrane sealing theory developed herein assumes that all reservoirs and seals are water-wet and no hydrodynamic flow exists. The conclusions on membrane seal capacity place constraints on the migration efficiency of gas along low-permeabiligy paths at depth where fracturing, wedging open of faults and/or diffusion process may be more important. Contrary to previous assertions, it is speculated that leakage of hydrocarbons through membrane seals occurs in distinct pulses such that the seal is at or near the theoretically calculated seal capacity, once this has been initially attained.Finally, the developed seal theory and P/D profile concepts are applied to a series of development geological problems including the effects of differential depletion, and degree of aquifer support, on sealing fault leakage, and the evaluation of barriers to vertical cross-flow using RFT profiles through depleted reservoirs. It is shown that imbibition processes and dynamic effects related to active cross-flow across such barriers often preclude quantitative analysis and solution of these problems for which simulation studies are usually required.