Summary. This paper presents the use of X-ray tomography to inspect a series of laboratory core-displacement tests. Saturation profiles in thin longitudinal planes of view (slices) in the core are followed as a function of time. Immiscible and miscible displacements in single core plugs and butted composite plugs are presented. Effects of heterogeneity of the core, dispersion of flood fronts, and end effects are observed and discussed. The technique has a large potential to improve the understanding of fluid displacement processes in laboratory core experiments. Introduction Data from laboratory displacement tests on reservoir rock samples are widely used as input parameters in reservoir simulators. Questions on the validity of these data are frequently asked concerning sample heterogeneity, dimensions, end effects, analysis procedures, and conditions. One way to study these effects would be to follow the movement of the fluids inside the reservoir rock sample. Several methods are used to visualize or to quantify saturations during laboratory displacement experiments. Some of these are based on radioactive tracers or an external source, nuclear magnetic resonance, microwaves, and electrical properties. All these methods give only a one-dimensional presentation of saturation. Computerized tomography of X-rays for medical purposes was introduced in the 1970's. Because this tool gives a two-dimensional (2D) image of the interior of an object, it should also provide information on reservoir rock. In our evaluation of the technique and commercial scanners, we performed several displacement tests using different displacement methods. Some of these tests are presented in this paper. However, no systematic study was undertaken to analyze the different parameters' influence on displacement behavior. Principle of Computerized Tomography of X-Rays Radiation of high energy can penetrate material of moderate density, such as reservoir rock. The problem is to reconstruct a detailed image contained in the attenuated high-energy rays transmitted. In 1917, Radon showed mathematically how to unscramble the information in the attenuated energy rays. A geometric interpretation of Radon's theorem emerged in 1972 with a technique called computerized tomography (CT). The device is called a CT scanner. Any object of moderate density is placed between an X-ray source and an array of collimated detectors. The source and detectors are mounted on a "yoke" that rotates. Power is pulsed to the X-ray tube, creating a fan of beams that traverses a thin slice (I to 8 mm [0.04 to 0.3 in.]) of the object (Fig. 1). Between each power pulse the yoke moves, and the next X-ray fan beam traverses the same slice from a slightly different angle. As the yoke is rotated 6.3 rad [360 degrees], as many as 700,000 different X-ray projections are available for mathematical processing. A presentation of the mathematical treatment is found in Ref. 10. The mathematical processing, performed by a computer, gives a synthesized image. The basic synthetic unit is the volume element. The CT slice is composed of many volume elements, each with its own characteristic attenuation, which are displayed as a 2D image matrix of picture elements (pixels) (Fig. 2). Because X-ray attenuations are related to density, the CT image gives the density distribution within every point of the object scanned. Because it is impractical to deal with the X-ray attenuation coefficient, A, in CT scanning, a new relative scale was defined as (1) where mutc = calculated X-ray attenuation coefficientandmuH2O = X-ray attenuation coefficient for water. SPERE P. 148^

TX 75083-3836, U.S.A., fax 01-972-952-9435. AbstractA two-phase (oil-water) relative permeability correlation for mixed-wet reservoir rock is developed and validated in this paper, including bounding drainage and imbibition processes and scanning hysteresis loops, all integrated with the corresponding changes in capillary pressure.The Corey-Burdine type relative permeability correlation is widely used in the industry. It was originally developed for water-wet reservoirs from a Brooks-Corey power-law capillary pressure correlation in combination with a bundle-oftubes model of the pore network.We have adjusted the Brooks-Corey capillary pressure correlation to be valid for mixed-wet rock and now present the ensuing Corey-Burdine relative permeability correlation for mixed-wet reservoirs.The functional form of the relative permeability correlation is symmetric with respect to fluid-dependent properties since neither fluid is privileged in a mixed-wet environment. It reverts to the standard Corey-Burdine correlation for the completely water-or oil-wet case. A water-wet behavior is displayed at low water saturations and an oil-wet behavior at low oil saturations, in accordance with experiments. The correlation provides an inverted S-shape oil relative permeability curve with an inflection point, and closed hysteresis scanning loops, as observed.The correlation is validated by comparison with measured relative permeability curves and simultaneously measured capillary pressure and relative permeability curves from the literature.The correlations and hysteresis logic are easily programmed, and we suggest that the Killough hysteresis model, employed in many numerical reservoir simulators, should be updated with the new scheme.

SummaryA two-phase relative permeability correlation for mixed-wet rock is presented and validated. It includes provisions for bounding drainage and imbibition processes and scanning hysteresis loops, and is inferred from a capillary pressure correlation.The well-known Corey-Burdine relative permeabilities were developed for water-wet rock from a Brooks-Corey power-law capillary pressure correlation and a bundle-of-tubes network model. We have extended this correlation to mixed-wet rock and now propose the ensuing relative permeability correlation for mixed-wet reservoirs. The functional form is symmetric with respect to fluid-dependent properties, because neither fluid has precedence in a mixed-wet environment. It reverts to the standard Corey-Burdine correlation for the completely water-or oil-wet cases, and exhibits the following characteristics in agreement with reported experiments: first, water-wet behavior at low water saturations and oil-wet behavior at low oil saturations; second, an inverted S-shape oil relative permeability curve with an inflection point; and, third, closed hysteresis scanning loops.

Seawater and a North Sea formation water has been mixed in glass and in porous media. Core tests showed severe permeability loss due to scaliIl:g The extent depended on the mixing ratio and the amount injected. Scanning electron microscopy and X-ray diffraction were used to identify and compare mineralogy of scales formed in cores.

This study aimed to understand the impact of natural open fractures and stylolites on flow in a porous sandstone reservoir by analysing fluid flow behaviour at small scale (core sample). This analysis was done both experimentally and with fluid flow simulations within full core samples. Two core samples (seal peels) from the Snøhvit field in the SW Barents Sea were selected. These contain natural open fractures mostly associated with stylolites. Since no test data exist from the intervals with highest fracture intensity, it remained uncertain whether these fractures will affect the flow in the reservoir. Two series of flooding experiments were conducted on these core samples, one by tracer injection and the other by injecting gas (Nitrogen) in the 100% brine saturated seal peels. Computed tomography (CT) scans were used to image the fracture/stylolites system, to monitor the flooding experiments and to build a consistent simulation model. Single and two phase flow simulations were performed to mimic the experiments. First, the fractures and stylolites effective properties were determined via tracer injection experiments by matching the tracer production profile. Then, to simulate the gas injection, relative permeability and capillary pressure (Kr-Pc) curves generated from special core analysis (SCAL) measurements were applied for the matrix medium. For the stylolites, the same Kr curve was used and modified Pc applied. For the fractures, a linear Kr curve and no Pc were used. As a result, a good quality match was obtained, giving an understanding of the influence of a fracture/stylolites system on the flow within porous sandstone. Introduction The presence of open fractures is known to have significant effects on permeability and flow anisotropy in hydrocarbon reservoirs (e.g. Nelson 1981, 2001). Natural open fractures are present in the Jurassic reservoir sandstones of the Tubåen, Nordmela and Stø Formations in the Snøhvit Gas Condensate Field (Walderhaug 1992, Wennberg et al. in press a). In order to improve the understanding of the impact of these fractures on flow in the Snøhvit Field, Computed Tomography (CT) scan imaging techniques have been used on representative core samples (e.g. Wennberg et al in press b). First, CT scanning was used to describe the 3D geometrical properties of the fracture network including orientation, fracture density and fracture connectivity. Two types of fractures were observed: F1 fractures are short and stylolite related, and F2 fractures are longer, crosscutting the core and without any obvious relationship to stylolites. Second, monitoring of single and two phase flow experiments on samples containing these two types of natural open fractures was performed under 10 and 80 bar net confining pressure while using CT-scanning. The flow experiments showed that the presence of open fractures has a significant local effect on fluid flow even in a case with relatively high matrix permeability (200–300 mD). The sample containing F1 fractures showed a complex flow pattern influenced both by open fractures and stylolites. The present study aimed to further improve the understanding of natural open fractures and stylolites on flow to define the fracture/stylolites hydraulic properties to build a consistent simulation model. First, this paper focuses on the modeling of the fracture/stylolite system within two representative core samples:–Core sample A contains one long F2 type fracture crossing the sample (Figure 1);–Core sample B contains several F1 type fractures (Figure 2), which are short stylolite related. After the conversion of the CT-scan image into a core sample simulation model, hydraulic properties were determined. Permeability measurements on the whole core samples as well as mini-permeability measurements were conducted. In addition, for the fracture/stylolite intrinsic permeability determination, single phase miscible flow experiments were used.