document

Summary

An improved compositional chemical-flood simulator has been used for the study of the low-tension pilot project at the Big Muddy field near Casper, WY. Both the tracer injection conducted before injection of the chemical slug and the chemical-flooding stages of the pilot project were analyzed. Using a compositional simulator, we successfully matched not only the oil recovery but also the tracers, polymer, alcohol, and chloride histories. Simulation results indicate that for this freshwater reservoir, the salinity gradient during preflush and the resulting calcium pickup by the surfactant slug played a major role in project success. In addition, the analysis of the effects of crossflow on the performance of the pilot project indicate that for the well spacing of the pilot, crossflow does not play as important a role as it might for a large-scale project.

Introduction

Computer simulation is an integral part of the design and scale-up of any process as complex as micellar/polymer flooding. One- and two-dimensional (1D and 2D) simulation studies of micellar/polymer flooding and comparison with corefloods have been under way for some time. However, 2D and 3D simulations have been limited by the speed and core storage of available computers. Until recently, we were not able to use our 3D chemical-flooding simulator (UTCHEM) for performance prediction on a field scale. The advent of supercomputers with large storage capacity and high speed, however, together with the application of vector processing to our chemical-flooding simulator, has made large-scale multiwell simulation applications a possibility. One objective of this study, therefore, is to validate UTCHEM by comparing its predictions with field results. Other objectives are to analyze the results of the pilot by identifying the major factors affecting the oil recovery efficiency and to determine how these results relate to the scale-up problem. Some new features in the simulator were required for this type of simulation and have already been added; still other features and modifications may be necessary as a result of this and other large-scale simulation studies and will be incorporated as they are identified. The results of this field-scale simulation and other optimization studies should aid in the design and operation of field projects. The Big Muddy pilot was selected for this simulation study. Reservoir description, completeness and quality of the available data, and the size of the project were the main criteria for selecting this pilot project. Extensive physical property data on a TRS 10-410 isobutyl alcohol (IBA)/decane/brine formulation have been measured at the U. of Texas during the past several years, and this system is sufficiently close to the one used at Big Muddy that these data could be used to advantage. The only significant variance was the oil. Overall, the Big Muddy pilot served our purpose of simulating a surfactant field pilot in more detail than previously reported for other field pilots, while using more laboratory and reservoir property data than previously possible. This enabled us to identify and study in detail process mechanisms such as cation exchange that played a key role in performance. The resulting interpretations are subject to critical examination and possibly systematic improvement because our UTCHEM simulator is extremely well documented in the literature and anyone interested can determine how each property was modeled or any other feature of the simulator. Very few cases comparing field results with simulation of surfactant pilots have been reported. The simulations of two projects operated by Marathon Oil Co. reported by Kazemi and MacMillan and Roszelle come the closest to our effort in their approach, but our case is much more detailed in terms of reported field data, laboratory data, and description of the simulator and how it was used.

Description of Big Muddy Pilot Project

Conoco Inc. initiated a low-tension pilot test in 1973 at the Big Muddy field east of Casper in Converse County, WY. The test sand was the Second Wall Creek reservoir of the Frontier formation, with an average depth of 3,150 ft [960 m], average porosity of 0.19, and net thickness of 65 ft [19.8 m] (see Table 1). Waterflooding in the reservoir started in 1953, and at the time of the pilot project, the oil cut was less than 1 %. Second Wall Creek is a freshwater sand with an average permeability of 52 md. The test pattern was a 1-acre [4047-m2] five-spot made up of five new wells drilled inside a 5-acre [20,235-m2] area defined by four existing waterflood wells (Fig. 1). All five new pattern wells were fractured and propped to improve low injection rates caused by a low parting pressure of 75 to 150 psi [517 to 1034 kPa] below hydrostatic pressure and to decrease the total pilot-test time. The average waterflood residual oil saturation (ROS) in the test area was about 32%. The pilot test started by injecting an 80% PV tracer preflush of 0.6 wt% solution of NaCl in the Sundance water supply. Different chemical tracers were injected into each of the four injection wells: tritium was injected in Well 14, ammonium thiocyanate in Well 31, methanol in Well 79, and ethanol in Well 27. Sulfate present in the injection water served as an overall tracer. Lack of production of ethanol was thought to be a result of bacterial degradation. A 25% PV surfactant slug consisting of 2.5% sodium petroleum sulfonate (equivalent weight 427), 3.0% IBA, 0.2 wt% sodium hydroxide to raise the pH to about 12 for bacterial control, and about 200 ppm xanthan gum polymer was injected. The surfactant slug was followed by a 50% PV polymer drive. Polymer concentration was tapered to minimize viscous instabilities. The polymer slug was followed by water injection. The tritium and thiocyanate tracers and the entire recovery process from the start in Aug. 1973 to the project's end in Feb. 1978 were simulated and are discussed below.

Brief Description of the Simulator

UTCHEM is an isothermal, slightly compressible, chemical-flooding compositional simulator. In the simulator, the material-balance equations are solved for up to 19 components: water, oil, surfactant, polymer, anions, divalent cations, Cosurfactant 1, Cosurfactant 2, water tracer, partitioning tracer, oil tracer, sodium dichromate, thiourea, trivalent chromium, gel, hydrogen, carbon, and organic acid species. Monovalent cations are given by electroneutrality condition. These components may form up to three phases -- aqueous, oleic, and microemulsion -- depending on relative amounts and effective salinity of the phase environment. The major physical phenomena modeled in the simulator are phase density, phase viscosity, phase behavior, dispersion, dilution effects, adsorption, interfacial tension (IFT), relative permeabilities, capillary pressure, capillary phase trapping, cation exchange, alcohol partitioning (constant or variable), and polymer properties such as permeability reduction, inaccessible PV, and shear-thinning effects. The solution scheme used is analogous to implicit pressure, explicit saturation formulation. First, the pressure equation is solved implicitly for the phase pressures and velocities using explicit dating of saturation-dependent terms. Then the conservation equations are solved explicitly for total concentrations. Phase concentrations and saturations are obtained by flash calculations.

SPERE

P. 24^