The Relationship Between Permeability and the Morphology of Permeability and the Morphology of Diagenetic Illite in Reservoir Rocks Summary The permeability characteristics of two reservoir sandstones from the Magnus and West Sole fields in the U. K. North Sea are essentially similar and are strongly influenced by the occurrence of filamentous illitic clay. However, the amount of illite in the West Sole sandstone is 20 times that in the Magnus sandstone, suggesting that the distribution and form of the clay is more important than its actual amount. Scanning electron microscope (SEM) observations show that the clay mineral occurs as an open, tangled web of long, thin ribbons virtually filling the pore spaces, a morphology that is properly revealed only after critical point drying. Introduction Early work at the BP Research Center on the permeability characteristics of reservoir sandstones from the Magnus permeability characteristics of reservoir sandstones from the Magnus field in the U.K. North Sea revealed puzzling differences between results from well testing and those measured by routine core analysis. Core data always indicated a higher permeability, which, in the case of the water zone, permeability, which, in the case of the water zone, amounted to 20 to 30 times that found by well testing. Similarly, there was a very wide discrepancy between gas and brine permeabilities of sandstone plugs measured in the oven-dried state and in the wet, preserved state. A possible explanation for this anomaly was that there could possible explanation for this anomaly was that there could be dehydration and shrinkage of clay minerals within the sandstone pores. Indeed, conventional SEM observations did reveal the presence of diagenetic filamentous illite coating the pore walls, although at that stage it was thought to be unlikely that this mineral could be responsible for the large differences in permeability. Work at the Macaulay Inst., however, showed that the morphology of the illite in the Magnus sandstones, as inferred through SEM, depended crucially on the manner of drying the specimen. The in-situ morphology of the clay mineral was revealed only after critical-point drying, when it was observed as an open, tangled web of very thin, long ribbons virtually filling the pore spaces. In contrast, after air- or freeze-drying, the illitic mineral occurred in dense mats packed against the pore walls. Following these observations, it was demonstrated conclusively 1 that the "interface-sensitive" nature of the filamentous illite provided a complete explanation for the discrepancies between the well-test data and routine core permeabilities. it was further suggested that the permeabilities. it was further suggested that the indiscriminate adoption of routine core analysis procedures could lead to serious errors in the assessment of reservoir quality. Our investigation compares the permeability and clay mineralogy of Magnus sandstones with those of West Sole sandstones, also in the U.K. North Sea, to show that fluid flow characteristics cannot be predicted from the quantitative clay content of reservoir sandstones. The West Sole sandstones contain abundant filamentous illite, which amounts to about 10% of the rock, compared with about 0.5% in the Magnus sandstones. The SEM and X-ray diffraction characteristics of the separated illite from both sandstones are dissimilar, but despite these differences the fluid flow properties of the two sandstones are comparable. Permeability Measurements Permeability Measurements Core plugs were cut from preserved Magnus and West Sole cores using degassed, simulated formation water as the cutting fluid. The core plugs, 1 in. [2.54 cm] diameter and 1 in. [2.54 cm] long, were cut from the center of the drilled pieces to reduce possible mud contamination. They were stored under formation brine. Complete brine saturation was ensured before permeability measurements were made. The cores were evacuated under brine to remove any gas and a pressure of 1,000 psi [6895 kPa] was then applied for >72 hours to dissolve any remaining gas. After saturation, the initial brine permeability was measured, during which time the sample was confined in a Hassler core holder with a sealing pressure of 400 psi [2758 kPa] on the rubber sleeve. Then the samples were dried in a humidity oven. After drying, the gas permeability was measured. All the samples were then resaturated permeability was measured. All the samples were then resaturated with brine and the brine permeability redetermined. Table 1 shows that the permeability characteristics of the two sandstones are similar and that, in particular, the gas permeability measured after oven drying was 20 times higher than the initial brine permeability measured on preserved samples. The resaturated brine permeabilities are higher than the initial brine values, permeabilities are higher than the initial brine values, indicating a permanent effect caused by drying. JPT P. 2225

Petrophysical measurements on effectively clean, desaturated core plugs from different fields have shown that the assumption of a constant saturation exponent for a given sample frequently is violated, although the conventional bilogarithmic distribution of resistivity-index/water-saturation data is often approximately linear. These variations in saturation exponent are attributed to the effects of pore geometry, in particular the nonuniform distribution of electrolyte within a heterogeneous pore system as desaturation progresses. The data indicate that, if unrecognized, a variable saturation exponent can induce errors of more than 10 saturation units (s.u.) in the petrophysical evaluation of water saturation. Procedures are outlined for the identification of those physicochemical reservoir conditions that can give rise to a variable saturation exponent and for accommodating these variations in reservoir evaluation.

No abstract

Summary Saturation exponents, determined from electrical measurements on core plugs desaturated with air as the displacing fluid, are influenced by pore size and pore-size distribution. This influence, which manifests pore size and pore-size distribution. This influence, which manifests itself through the resistivity index, is most pronounced in microporous rocks. Two controlling factors are electrical tortuosity and surface conduction effects, both of which increase as the electrolyte-filled pore volume (PV) is reduced. Theoretical and experimental data have indicated that the saturation exponent can increase or decrease during desaturation, depending on how these factors interrelate. With due caution, the risk of errors in estimates of hydrocarbons in place caused by the use of an inappropriate saturation exponent can be greatly reduced. Introduction Microporosity in reservoir rocks has been variously described as comprising pores with an average or minimum diameter that is less than 1/16 mm [0.0025 in.], as so small that it cannot be discerned at magnifications less than 50X, or as indicating a minimum pore diameter that is no more than 1 mu m. We have adopted and extended this last definition to include those parts of a pore space with typical dimensions of 1 mu m or less. Thus, subject to this criterion being met, we also incorporate such features as surface roughness and dead-end pores within the collective term microporosity. Most natural formations are microporous to some degree. The purpose of this paper is to provide preliminary insight into the influence of microporosity on the petrophysical determination of hydrocarbon saturation. A key factor in this evaluation is the saturation exponent, n, and this paper is concerned principally with evaluating the effect of microporosity on this parameter. The approach uses two-dimensional (2D) models of pore systems for which saturation exponents can be calculated from geometrical and electrical parameters. The computation process makes use of a derived algorithm that approximately relates the saturation exponent to electrical tortuosity, specific surface area, and surface conductance. The algorithm therefore is based on a geometrical model in which surface conduction is the electrical manifestation of electrochemical effects. Algorithm Derivation The Archie equation relating electrolyte (water) saturation, Sw, to electrical parameters can be written as (1) where Ct = conductivity of a partially saturated granular medium and Co = conductivity of this medium when fully saturated with identical (aqueous) electrolyte of conductivity Cw. Note that n is generally a function of Sw. To express Co and Ct in terms of calculable or specifiable parameters, we draw upon a relationship between electrical tortuosity, formation resistivity factor, FR (defined as Cw/Co), and porosity, phi, for a fully electrolytesaturated pore system: (2) where = ratio of mean electrical current path between two reference planes, Le, to the direct distance between these same planes, L--i.e., = (Le/L)2. This definition of, interprets Le to be the length of a single equivalent pore with a constant cross-sectional area in a direction parallel to the reference planes. Some authors have defined, = (Le/L)2. Eq. 2 can be extended to the more general case of partially-electrolyte-saturated media for which the tortuosity is designated, to distinguish from the particular case : (3) where GR = Cw/Ct. This analogy and the resulting equivalence of Eqs. 2 and 3 are apparent from Fig. 1, which depicts a fully-water-saturated granular medium of porosity phi 1 (Fig. la) and a hydro carbon-bearing granular medium of porosity phi 2 (Fig. lb), with identical electrolyte geometries so that phi 1 = phi 2 Sw. We note further that (4) from Eq. 1 and the definition of FR. By combining Eqs. 3 and 4, we have (5) Eqs. 2 through 5 have been used in conjunction with a surface conductance model for fully-electrolyte-saturated reservoir rocks, but modified so that different geometric factors might he used for the two components of current flow: (6) where F1 = formation resistivity factor associated with the bulk fluid conduction, F2 = formation resistivity factor associated with surface conduction, = surface conductance, and A = specific surface area expressed per unit PV of the system. Taking 1 =F1 phi and 2=F2 phi from Eq. 2, with the restrictive assumption that the porosity available for bulk-fluid conduction is the same as that porosity available for bulk-fluid conduction is the same as that bounded by the pore surface, and substituting into Eq. 6 yields (7) (8) This derivation is similar to, but less rigorous than, the surface-structure model of Pape and Worthington. The derivation of Eq. 8 can be paralleled for a partially-electro-lyte-saturated system, in which case the starting point is (9) Eq. 9 is a modification of Eq. 6, which has been extended to partially saturated conditions through reasoning outlined by Waxman partially saturated conditions through reasoning outlined by Waxman and Smits and explained in general terms by Worthington. In this case, we take and from Eq. 5, with the restrictive assumption that the electrolyte-filled porosity available for bulk-fluid conduction is the same as that bounded by the pore surface. By substituting into Eq. 9, we can show that (10) To a first approximation, (11) Substituting Eqs. 8 and 11 into Eq. 1 yields (12) Eq. 12 forms the basis of the method for computing saturation exponents for specified pore models, as these are subjected to different desaturation levels.

An independent method for estimating initial reservoir water saturation in the Gyda Field has been investigated. Further application of recent tracer technology was used in describing the initial fluid saturations in the reservoir. This was achieved by using a deuterium tracer technique to quantify aqueous mud-filtrate invasion to core and which enabled 'native' water saturation determinations to be made. The volume of extracted core-water was translated into native reservoir water saturations after correcting for mud-filtrate invasion, reservoir overburden, pressure and temperature effects.