Core Acquisition and Analysis for Optimization of the Prudhoe Bay Miscible-Gas Project

Abstract

Summary Core data from a sidetrack well 250 ft from an old injection well showed intervals of excellent miscible-gas sweep, with miscible-flood residual oil saturations (ROS's), Sorm, of ˜5% PV. Less-well-swept intervals had intermediate Sorm values. Numerous intervals had not been contacted by miscible injectant (MI). Chromatographic analyses of the Dean-Stark core extracts quantified miscible-gas alteration of the oil and was especially valuable in the interpretation of intermediate oil saturations. Compositional reservoir simulations showed that high-permeability thief zones and a number of thin shales controlled sweep efficiency near the cored well. Background The Prudhoe Bay field on the north coast of Alaska is the largest oil field in North America, with total estimated reserves of ˜ 12 billion bbl and a current production rate of ˜ 1.1 million STB/D. The field is overlain by a large gas cap, and most of the field is being produced by gravity drainage. Waterflood operations at Prudhoe Bay, which are confined to the downstructure and peripheral areas of the field, are producing ˜ 400,000 STB/D. The Sadlerochit group, the major productive interval of the field, includes a thick section composed of high-permeability fluvial sands and interbedded shales.1 In the Flow Station 2 area, where the sidetrack core was taken, these shales create up to five completely isolated flow intervals. Prudhoe Bay EOR began in late 1982 with an 11-pattern pilot project. Expansion of EOR operations began in 1987, and now more than 100 patterns are operating. As Fig. 1 shows, the unit plans significant further expansion of the Prudhoe Bay Miscible Gas Project (PBMGP) over the next few years. The PBMGP currently has a solvent injection rate of ˜ 600 million scf/D at an average water-alternating-gas (WAG) ratio of ˜ 3 : 1. The solvent is manufactured by the central gas facility (CGF), which has a nominal processing capacity of 7.5 billion scf/D. Residue gas from the CGF is reinjected into the gas cap. The design and installation of the Prudhoe Bay gas-handling facilities have been described previously.2–4 EOR evaluation and optimization have been complicated by immature waterflood development, infill drilling, and extensive fracturing and workover programs. Numerous field measurements have been made to evaluate miscible-flood displacement and sweep efficiency, including water- and solvent-injection profiles, logging data from a fiberglass-cased observation well, single-well tracer-test data for waterflood and miscible-gas ROS's, and separator-gas-sample compositional analyses.5 The available field and laboratory data indicated favorable EOR response, but more conclusive data were needed to justify additional projects. Sidetrack Core Acquisition A very poor injection profile and operational problems in Drillsite 3 Well 18 (Well DS 3-18) led to a redrill of this injector in April, 1993, as Well DS 3-18A. This was the first sidetrack of an existing WAG injector at Prudhoe Bay, and it provided an ideal opportunity to obtain conclusive data on miscible-flood sweep efficiency and ROS's. Cumulative solvent injection into Well DS 3-18 was 8.4 Bscf before the redrill, which was 2.7% of the pattern total PV; cumulative water injection was 64 million res bbl. Table 1 shows the spinner profile data, and Fig. 2 shows a cross-section of Wells DS 3-18 and DS 3-18A. Most of the injection went into the high-permeability conglomerate of Zone 3; very little injection went into Zone 2. Most of the injection into Zone 4 was at the very top of the interval, where gravity segregation would prevent the MI from contacting new oil. Because of this inefficient profile, no solvent had been injected into Well DS 3-18 since May 1992. Compositional simulation results of a detailed near-well model with a similar reservoir description were used to determine the optimal placement of the sidetrack. These results showed that, after a similar volume of solvent injection, the maximum amount of "character" in oil saturation should occur ˜ 200 to 300 ft from the injection well. If the sidetrack were placed closer to the original well, solvent sweep would be more or less uniform and little would be learned about gravity segregation of the MI. If the sidetrack were placed farther from the original well, it would encounter only a few solvent fingers. We therefore planned to put the sidetrack 250 ft from the old injector to gather the maximum possible sweep efficiency and residual-saturation data. This modeling data proved critical in the proper location of the sidetrack well. Table 2 lists the MI and water throughput at the 250-ft radius. A total of 467 ft of conventional core was cut at the DS 3-18A sidetrack well over a 7-day period. Total core recovery was 459.5 ft, or 98.4%, from 12 cores. Typically, 60-ft aluminum core barrels were used; core diameter was 4 in. Directional surveys indicated that the sidetrack, which had a 35 ° hole angle, was within the target of 250± 50 ft from the old wellbore. Coring rates averaged ˜ 40 ft/hr in sandstones and ˜ 5 ft/hr in conglomerates and shales. A 30,000-ppm KCl mud with bentonite, starch, polymers, and calcium carbonate solids was used to obtain both low invasion and good shale stability. Mud weight was 9.8 lbm/gal. NaBr was used as a tracer to monitor filtrate invasion. Refs. 6 and 7 provide details of the coring procedures. Although overall core recovery was excellent, a significant fraction of the conglomeratic intervals were disaggregated in coring and/or surface processing. Whole core porosity and oil saturation could be obtained on almost every foot of core, but permeability data were unavailable in ˜ 20 ft of the conglomerate. Unfortunately, the most-permeable intervals probably were included in this 20 ft. Geologic and operational studies have shown that effective permeabilities of some conglomerates are much higher than have been measured in the cores. Open-framework conglomerates, which have very little sand matrix material and virtually no cementation, appear to be responsible for the extremely high permeabilities observed from injection and production profiles.8,9 These open-framework conglomerates appear to be a few inches to about a foot thick and could easily be represented in the numerous rubble intervals observed in the core.

Recovery rates, enhanced oil recovery and technological limits

“…When used after a water flood this has the effect of re-establishing a pathway for the remaining oil to flow through and results in a very low residual oil saturation (2% has been measured in reservoir cores recovered from gas swept zones [51]). The drawback of this process is that the gas is both less viscous and less dense than the oil.…”

Modeling of capacitance flow behavior in EOS compositional simulation

“…Flow-visualization experiments (Stalkup, 1970;Chatzis et al, 1983;Campbell and Orr, 1985;Bahralolom et al, 1988;Stern, 1991) showed that microscopic oil bypassing was related to pore structures with bimodal or wide pore-size distributions. Shielding of oil by water films in pores can also hinder the contact between the bypassed oil and injection gas in the region with high water saturations (Shelton and Schneider, few inches can cause marked oil bypassing by gas (McGuire et al, 1995). Front instability can also lead to oil bypassing, if the mobility ratio between the injection gas and reservoir oil is large (Gardner and Ypma, 1984;Brock and Orr, 1991).…”

Unconventional EOR Program Increases Recovery in Mature WAG Patterns at Prudhoe Bay

“…[3][4][5] The nonconventional EOR program began in 1996 initially as a MI stimulation (MIST) of an individual well. 4 The program evolved to utilizing a Miscible Injector sidetrack (MIST) to provide a horizontal completion with multiple injection points along the horizontal section of the wellbore (see Figs.…”

“…3,4 Prior to the start of the PBMGP, continuous and WAG floods were carried out on 16 plugs taken from a total of 4 preserved cores. These cores were injected with different miscible gas compositions and varying quantities of between 0.88 and 3.74 pore volumes.…”

“…6 The actual MI sweep efficiency in the Victor was determined by coring the sidetrack of well 03-18A. 3 A historymatched, fully compositional reservoir simulation of WAG in the Victor showed a very limited interval in which EOR oil was actually mobilized. Although the entire interval was open to injection, the bottom 100 ft of the interval was not contacted by solvent.…”