Reservoir Description Detail Required To Predict Solvent and Water Saturations at an Observation Well
- F.I. Stalkup (Arco E&P Technology) | S.D. Crane (Arco E&P Technology)
- Document ID
- Society of Petroleum Engineers
- SPE Reservoir Engineering
- Publication Date
- February 1994
- Document Type
- Journal Paper
- 35 - 43
- 1994. Society of Petroleum Engineers
- 4.1.5 Processing Equipment, 5.4.7 Chemical Flooding Methods (e.g., Polymer, Solvent, Nitrogen, Immiscible CO2, Surfactant, Vapex), 5.6.4 Drillstem/Well Testing, 5.3.2 Multiphase Flow, 5.6.1 Open hole/cased hole log analysis, 5.1 Reservoir Characterisation, 5.6.5 Tracers, 4.3.4 Scale, 5.5.8 History Matching, 5.5 Reservoir Simulation, 2 Well Completion, 5.4.9 Miscible Methods, 2.2.2 Perforating, 2.4.3 Sand/Solids Control, 5.1.5 Geologic Modeling, 1.6.9 Coring, Fishing, 4.1.2 Separation and Treating, 5.4.1 Waterflooding, 5.2.1 Phase Behavior and PVT Measurements, 6.5.2 Water use, produced water discharge and disposal, 1.8 Formation Damage, 4.6 Natural Gas
- 0 in the last 30 days
- 192 since 2007
- Show more detail
- View rights & permissions
|SPE Member Price:||USD 12.00|
|SPE Non-Member Price:||USD 35.00|
This paper describes the reservoir simulator analysis of an observation-well pilot test in the Prudhoe Bay miscible flood. It shows that 1-ft simulator layers and explicit representation of thin, discontinuous shales are required to predict solvent occurrences and log-derived water and solvent saturations at the observation well. The predictions show that solvent flows past the well in numerous tongues that are only a few feet thick.
The miscible process is an enriched-hydrocarbon-gas condensing/vaporizing drive. Zick1 was the first to describe this type of process, and others have elaborated on it.2-4
In 1983, Arco Alaska Inc., operator of the eastern half of the Prudhoe Bay field, began an enriched-gas miscible flood in the Flow Station 3 area (Fig. 1). In 1987, this initial flood was expanded (to the other crosshatched areas in Fig. 1) by Arco and BP Exploration, operator of the western half of the field.
The current combined projects are among the world's largest miscible floods. The Flow Station 3 project encompasses ˜400 million STB of initial oil, and the combined project areas encompass ˜3 billion STB of initial oil.5 Currently, injection of enriched-gas solvent into all project areas is ˜370 MMscf/D. Solvent used was manufactured by blending flash drum liquids with separator gas. Table 1 gives typical solvent compositions.
An observation-well pilot test was conducted in the Flow Station 3 area (Fig. 1) to evaluate the flood. Well 13-98 is ˜525 ft from the injector, Well 13-06, on a line between Wells 13-06 and 13-05, the side producer. The observation well is cased with fiberglass but is not perforated. Induction and epithermal-neutron logs were run monthly to monitor water and gas saturations vs. depth as first water and then alternate solvent and water slugs were injected into Well 13-06.
Pilot area oil gravity was 28°API and viscosity was ˜1 cp at the reservoir conditions of 200°F and 3,850 psi in the Flow Station 3 area when the miscible flood was started.
The pilot provides an unusual opportunity and challenge for reservoir simulation. An especially pertinent question is the degree of grid refinement and reservoir description detail required to predict the log-derived solvent and water saturation histories. Another question is whether laboratory-measured relative permeabilities are adequate for predicting reservoir-scale displacement. This paper describes the analysis that tries to answer these questions.
Pilot Area Geology
The Sadlerochit sandstone in the Flow Station 3 area was deposited by a complex of braided streams. The foot-by-foot horizontal permeabilities measured on Well 13-98 core (Fig. 2) show that the reservoir is highly stratified, with high-permeability sands interspersed with lower-permeability shaly sands and shales.
The total interval contains various shale types. Flood-plain shales are the most continuous and may be correlative over interwell distances. They are thought to be 0.1 to 2 ft thick, 600 ft wide, and 1,500 ft long. Bar drape shales are much thinner and less extensive areally. They are thought to be 0.05 to 0.2 ft thick, 100 ft wide and 800 ft long. Fig. 3 shows various shales identified in Well 13-98 core on sections of the gamma ray and compensated-density logs. These logs are reasonably good indicators of shale presence but not of shale type except for the thick flood-plain shales.
Fig. 4 shows a cross section through Wells 13-06, 13-98, and 13-05. The formation is ˜150 ft thick at the observation well and has a gentle dip of about 3°. An aquifer underlies the oil column, and an ˜20- to 30-ft-thick tar mat lies on top of the water/oil contact (WOC), severely restricting vertical water influx into the oil column. Although the injector is perforated over most of the 110-ft interval above the top of the tar mat, this perforated interval is only correlative with about three-fourths of the sand interval in the observation well.
General Log Response and Analysis
Base-induction and epithermal-neutron logs were taken before any fluid injection into Well 13-06. Because the observation well was cored with an oil-based fluid, the base-induction log should reflect log response to initial water saturation. The reservoir pressure in the vicinity of Well 13-98 had declined from an initial 4,300 to ˜3,850 psi before the epithermal-neutron base log was taken. Therefore, according to depletion calculations, a 2%- to 3%-PV free-gas saturation could have been present at that time.
Responses to fluid injection were determined by comparing subsequent logs with the base logs and locating wellbore intervals where departures occurred. The logs were analyzed for water and gas saturations for the 41 intervals indicated by the dotted horizontal lines in Fig. 2. These log-analysis intervals were selected to reflect high- and low-permeability intervals or portions of these intervals. Fig. 5 shows that water saturations increased in almost every log-analysis interval. Gas saturation increased in numerous intervals in approximately the upper three-fourths of the formation at Well 13-98, which approximately correlates with the perforated injection interval in Well 13-06. The closed symbols in Fig. 2 show intervals of definite gas saturation increases and illustrate the highly stratified nature of solvent sweepout at this location.
Quantitative Log Analysis.
The time-lapse log analysis used dual-induction, epithermal-neutron, and temperature logs. Quantitative analysis of these logs has evolved and improved over the years as a result of extensive work at the Prudhoe Bay Unit. The logs were measured through unperforated fiberglass casing. Water saturations were determined from dual-induction and temperature logs, and gas saturations were determined from epithermal-neutron and temperature logs.
Initial dual-induction-log water saturations were calibrated to the oil-based core saturations. Experimental results showed that the apparent Archie saturation exponent, n, increased with increasing water saturation. The water saturation calculations used a modified Archie's law that accounted for the variable apparent n.6 Conductive rock effects from pyrite streaks were removed with a parallel-conductivity model.6 Changes in reservoir temperature from cool-water injection were measured, and the corresponding changes in water conductivity were calculated. The deep-induction logs were deconvolved into 4-ft intervals with a forward-modeling technique similar to that Anderson et al.7 described. This allowed the dual-induction and neutron logs to have similar vertical resolutions.
|File Size||1 MB||Number of Pages||9|