No abstract

Summary This paper presents examples of xanthan corefloods visualized by X-ray tomography. Water and aqueous xanthan solutions are distinguished from each other by addition of sodium iodide (NaI) at different concentrations. One reservoir sandstone and one outcrop (Rosbrae) sandstone core samples were used. The reservoir sample was naturally divided into two longitudinal zones differing in permeability by about 20-fold. The Rosbrae sample was homogeneous, with a permeability of 450 md. Miscible xanthan/water displacement tests were performed on both plugs. Immiscible displacement of light refined oil by xanthan was performed on the homogeneous sample. Introduction Unrecoverable oil reserves for discovered fields in the North Sea are estimated to be about 3 × 10(9) m3. Improved oil recovery (IOR) offers some advantages over exploration activities. IOR studies are cheaper than drilling, implementation of IOR methods can use existing infrastructure, and the studies involve lower risk. Many North Sea reservoirs are stratified and have permeability contrasts. Injection of polymers can improve the areal sweep efficiency of the waterflood and thus increase the amount of oil recovered. This study was motivated by a field case evaluating the use of xanthan biopolymer. The aim of this work was to study biopolymer flood behavior in porous media by means of computer tomography (CT). CT has proved to be an excellent technique to visualize flow phenomena in porous media, especially in heterogeneous (layered) systems. Methods and Materials The CT scanner used was a Siemens Somatom DRH. The main scanning parameters were tube voltage, 125 kV; measuring time, 5 seconds; dose, 280 mAs; projections, 960; and slice thickness, 8 mm. The principles of CT and its applications in the oil industry have been described by several authors. The scanner is connected on line to a separate image-processing system, Context Vision GOP-300, to enable postprocessing on a higher level; for instance, filtering, color manipulation, contrast enhancement, and calculating fluid saturations and porosities. The coreholder was a Hassler type made of carbon fiber tube with polyamide end pieces pressure tested to 6 MPa. The two rock samples used were a North Sea stratified sandstone core and a homogeneous Scottish outcrop (Rosbrae) sandstone core. The reservoir core consisted of two longitudinal layers with a permeability contrast of 1:20. Table 1 summarizes the rock properties. The fluids were synthetic seawater, solutions of the biopolymer xanthan, glycerol, and light refined oil. The density difference between the xanthan solution and the seawater was less than 0.005 g/mL. In some floods, the fluids were doped with 40 or 200 g/kg NaI to increase the X-ray absorption to gain a better contrast between the fluids in the images. Note that whenever the polymer phase was doped, the dopant was added to the polymer solvent. Thus, the polymer itself was not traced. Table 2 gives the composition and fluid properties of the seawater. The xanthan was formulated in experimental batches at concentrations of 400 and 1,000 ppm with 500 glutaraldehyde added as a bactericide. Neither cell debris removal nor microgel filtration was performed. The viscosity was measured by a Contraves viscometer. Fig. 1 shows the rheological properties of the 400-and 1,000-ppm solutions. The samples were mounted in the coreholder with a net confining pressure of 2 MPa. All CT scans were made in a longitudinal plane of view from inlet to outlet (Fig. 2). All experiments were performed at room conditions. The injection flow rates were maintained at 4 mL/h unless otherwise stated. This flow rate corresponds to an overall apparent velocity of q/A =0.084 m/d.

A series of stable and unstable floods has been carried out in a well-characterized, heterogeneous concentric core assembly in order to study predicted recovery mechanisms in such systems. A combination of effluent analysis from the separate layers in the core and computer simulation was used to establish the detailed flows which were subsequently confirmed using computer x-ray tomographic (CT} scanning. Good agre~ment has been obtained between experiment and theory for unit mobility, and viscous floods in which the respective roles of transverse dispersion and viscous crossflow have been clarified. Some of the main qualitative features found in the unstable displacement (viscous fingering} experiments have also been reproduced using fine grid simulations. CT scanning has been shown to be extremely useful for studying flow mechanisms in systems of this type.