A B S T R A C T : Petrographical studies can be undertaken to investigate the effects of production operations on sandstone reservoirs and to identify the nature of any rock-fluid interaction. In combination with diagenetic models the results may be employed to predict the effects of field development programmes on reservoir mineralogy, and to avoid costly damage to the reservoir's permeability. Acidization to remove drilling mud from Reservoir A samples was undertaken in two stages to avoid adverse rock-fluid interaction. HC1 dissolved siderite and chlorite but did not increase permeability. This is because the dissolved siderite released previously cemented clay particles into the pore space. A subsequent mixture of HCI and HF dissolved all the clays present in the treated area including the drilling mud, but also etched the quartz cement. Permeabifity increased significantly but the rock's strength decreased. This could cause an influx of loose sand during hydrocarbon production. Water injection into Reservoir B samples caused a variety of rock-fluid interactions. In finer-grained samples movement of siderite rhombs as well as clay particles blocked pores, and changes in porewater composition caused smectite to swell. Both effects caused permeability to decline. In the coarser-grained siderite-free samples, permeability improved after oil and clay particles had been displaced. In the reservoir these effects would combine to exacerbate the effects of the existing reservoir heterogeneity. Steam injection into Reservoir C samples caused mineralogical reactions. Amounts of dolomite and kaolinite decreased, and smectite and calcite were generated. This may affect the permeability of the reservoir and will determine whether oil can be produced through the affected sediment.

Diagenetic models and the results of laboratory experiments on cores from producing Brent Group fields (Cormorant and Tern) have been combined to aid prediction of reservoir properties in a deeply buried prospect (Pelican). In Cormorant and Tern illite particles are thin (20–40 Å), lath-shaped and liable to damage during drying before air permeability measurement, leading to erroneously high permeabilities. The magnitude of this drying effect ( K air conventional drying/ K air critical point drying) varies with illite morphology and the timing of hydrocarbon emplacement. In Tern pore-filling illites occur above and below the oil—water contact, and there is no significant difference in drying effect. In Cormorant, the drying effect is higher below the oil—water contact where a diversity of late diagenetic pore-filling illite morphologies occur. Injectivity tests on cores from Cormorant indicate that sea water injection will not lead to impairment of reservoir quality as a result of clay dispersal or other rock-related phenomena. Impairment occurred when flooding cores with produced water and was avoided by introducing divalent cations, saline brines or approximating in situ equilibrium conditions (pH 6.4). This equilibrium point is strongly influenced by the surface chemistry of feldspar and illite, rather than the more abundant kaolinite and quartz. Burial diagenetic replacement of kaolinite by illite will increase this equilibrium towards pH 7 but should not cause impairment during sea water injection into deeper reservoirs. Pelican contains abundant pore-filling illite. K-Ar age dates (44–25 Ma) suggest illite growth started after burial to over 8000 ft and palaeoburial temperatures of over 80–90°C. Compared with nearby reservoirs, variations in the nature and abundance of illite result from illitisation being arrested by hydrocarbon charge at different times. Optimal reservoir and aquifer properties occur in shallower structures charged from more deeply buried source rocks. Good reservoir properties occur with charge after short-lived illite growth, although later burial may reduce aquifer properties. The more prolonged illite growth experienced by deeply buried reservoirs such as Pelican, charged late in the Tertiary, has resulted in poor reservoir properties and poor aquifer permeabilities may be expected.