Portal de Periódicos da CAPES

Design of a Tertiary Hydrocarbon Miscible Flood for the Mitsue Reservoir

1988; Society of Petroleum Engineers; Volume: 40; Issue: 02 Linguagem: Inglês

10.2118/13270-pa

1944-978X

J.P. Frimodig, V. Sankur, Claire Chun,

Hydraulic Fracturing and Reservoir Analysis

Summary. A large-scale hydrocarbon miscible flood has been designed and is being conducted in the Mitsue Gilwood Sand Unit No. 1 in Alberta, Canada. The horizontal miscible flood occurs mostly in previously waterflooded zones and is a true test of the tertiary displacement mechanism. An area of 17,900 acres 17250 ha] is being flooded with a first-contact-miscible (FCM) hydrocarbon solvent followed by a lean chase gas, both alternated with equal volumes of water. Inverted five-spot injection patterns are used. This paper reviews the evaluation and design process by use of recently developed methodologies, involving innovative laboratory techniques and simulation processes used in planning the miscible flood. Laboratory work was used to develop planning the miscible flood. Laboratory work was used to develop correlations for determining the hydrocarbon solvent enrichment required for both FCM and multiple-contact miscibility (MCM) as a function of solvent C2+ fraction and molecular weight at the reservoir pressure of 2,500 psia [17.2 MPa]. Simulation studies indicate that pressure of 2,500 psia [17.2 MPa]. Simulation studies indicate that optimum solvent and chase-gas slug sizes of 15 and 25%, respectively, will yield an incremental 12.2% of original oil in place (OOIP), or 21.3 MMSTB 13.4 × 10(6) stock-tank m3]. Solvent and chase-gas recoveries were estimated to be 73 and 44%, respectively. Early performance of the miscible flood indicates that the solvent is successfully mobilizing tertiary oil. Wells previously shut in because of high water production have experienced an increase in oil rates and a decrease in WOR's. Introduction A hydrocarbon miscible flood is being conducted in portions of the Mitsue Unit located in Alberta, Canada (Fig. 1). This flood is designed to be expanded in stages to encompass ultimately the majority of the unit. This paper is concerned with the evaluation of the first stage, which includes 27 patterns encompassing 174 MMSTB [27.7 × 10(6) stock-tank m3] OOIP, or approximately 23% of the unit's OOIP. Most of the miscible flood project area was water-flooded to near the residual oil saturation. The miscible flood recovery in these areas is through a tertiary displacement mechanism. The remaining areas will be miscibly flooded as a secondary process. The evaluation program described here was designed to determine optimum operating parameters and performance for the miscible flood while differentiating between areas of tertiary and secondary displacement. The evaluation program was composed primarily of laboratory work and simulation studies. Laboratory work included pressure/ composition, slim-tube, and coreflood experiments, with various hydrocarbon solvents and Mitsue reservoir oil. Results were used to construct FCM and MCM correlations. Compositional simulation studies included areal and cross-sectional runs used to construct a miscible-flood forecast model. This model was used to estimate optimum operating parameters and to predict performance. Project Description Project Description The first stage of the miscible flood was sized to use all the available unit solution-gas production for injection into the project area as miscible solvent. Solution gas is enriched with liquefied petroleum gas (LPG) imported to the unit through existing product pipelines. Using all the unit solution gas allows the Mitsue gas plant to be shut down during the solvent-slug injection phase. Lean-chase-gas injection will follow the solvent. During this period, the restarted-gas-plant residue gas production will satisfy much of the chase-gas requirements. Both solvent and chase-gas injections are designed to be alternated on a 1:1 reservoir volume ratio with water in 2-month cycles. Geology. The Mitsue Unit produces from the Gilwood sandstones of the Middle Devonian Watt Mountain formation. This shallow dipping stratigraphic trap has an extensive aquifer system on the western (downdip) edge and regions of free gas on the eastern (up-dip) edge. The approximate area of the reservoir is 140,000 acres [56 650 ha], and its maximum length is 42 miles [67.6 km]. The estimated OOIP is 770 MMSTB [122 × 10(6) stock-tank m3]. The sands were deposited fluvially from the northwest to the southeast in thin channels. Depositional studies indicate that the sands can be grouped into three layers: Layers 1 (Upper), 2 (Middle), and 3 (Lower). Fluid flow between layers is prevented by shale. The three layers can be further subdivided into six correlatable sands (or channels): two in Layer 1, three in Layer 2, and one in Layer 3. Channels are often separated by shale barriers. Only Layers 1 and 2 are to be miscibly flooded. Layer 3, which is of limited area extent and contains less than 5% of the OOIP, will not be flooded. Fig. 2 shows a typical cross section. Average reservoir rock properties for the channels in Layers 1 and 2 are shown in Table 1. Reservoir Development. Irregular development drilling in the Mitsue reservoir since its discovery in 1964 has resulted in an average well spacing of 425 acres [172 ha]. Primary depletion occurred until 1968. During this period, reservoir pressure declined from the discovery pressure of 2,620 to 2,225 psia [18.1 to 15.3 MPa]. with a cumulative oil production of 13.7 MMSTB 12.2 × 10(6) stocktank m3]. The Mitsue Gilwood Sand Unit No. 1 was formed in 1968, and a peripheral water-injection pressure-maintenance scheme was instituted along the downdip edge of the pool. This pressure-maintenance scheme was designed to augment the natural pressure pressure-maintenance scheme was designed to augment the natural pressure support from the existing aquifer. Ultimate waterflood recovery is anticipated to be at least 348 MMSTB [55.3 × 10(6) stock-tank m3], or 45% of the recognized OOIP. Reservoir response to waterflooding has been excellent; the flood front has moved through the contacted reservoir in an almost ideal (piston-like) displacement mechanism. Water breakthrough usually kills the well unless the wet channel can be shut off. Flood front advancement has varied between layers because of their differences in reservoir properties and injectivities. This has created the potential for crossflow in some producing wells that could trap oil between flood fronts in a layer. producing wells that could trap oil between flood fronts in a layer. Therefore, flooded-out zones are shut off as quickly as possible. The highest waterflood oil recovery, as estimated by volumetrics, is in the downdip portion of the reservoir. Waterflood oil recovery is expected to decrease updip, however, where the reservoir sands become less continuous and have lower permeabilities. JPT P. 215