No abstract

Summary Steamflood operations commonly encounter steam override problems. Close well spacing is one method used to reduce this effect. Blind spot areas still exist in the pattern, however, and leave substantial amounts of unrecovered oil after project termination. This study examines the effect of horizontal wells on improving steamflood conformance. A three-dimensional (3D) numerical steamflood simulator was used. The study involves the use of horizontal wells for mature steamfloods to reduce steady override that has already occurred and for prevention of steam override if used at the start of steamflood operations. The data used are from a typical California heavy-oil reservoir. Simulation indicates that horizontal wells can be effective for reducing or preventing steam override. For a mature 18.5-acre [7.5-ha] pattern, simulated horizontal wells increased ultimate recovery at the end of 11 years from 63.2 to 74.7 of original oil in place (OOIP). When used at the start of steamflood, a pattern of vertical and horizontal wells had a predicted recovery of 72.2% OOIP after only 7 years. Introduction Application of horizontal wells for mining and oil recovery was practiced by the Russians over 40 years ago. Horizontal wells were used for exploration in Russia and production in China with limited success in the 1960's. With the improvement in drilling technology, horizontal wells of 4,600 ft [1400 m] in Italy and at depths of up to 9,500 ft [2900 m] in France have been reported. The first modem application of a horizontal well for recovering heavy oil was a 2,045-ft [623-m] well drilled to a depth of 1,560 ft [475 m] at Cold Lake, Alta., Canada, in 1978. Subsequently, more horizontal wells for heavy-oil recovery were drilled in the Athabasca tar sands at Ft. McMurray. Alta. Recent interest, in horizontal wells has led to developments that make these wells more feasible. Drilling difficulties, such as deviation control, excessive dragging, cuttings removal, and wireline tripping, have been at least partially solved. Horizontal wells can be logged and completed. As technology in this area improves, horizontal-well drilling costs decrease, making horizontal wells more economically feasible. This study's purpose is to use reservoir simulation to provide some insight into the possible application of horizontal wells for steam-flood. Previous simulation work indicates that horizontal wellbores can improve oil recovery based on the comparison of a conventional vertical wellbore model and a horizontal wellbore model. Other analytical methods were developed for estimating productivities from horizontal wells. Advantages of Horizontal Wells One advantage of horizontal wells over vertical wells is the increase of direct contact between wellbores and pay zones. The perforated interval per vertical well is limited to the pay-zone thickness. For a horizontal well, the perforated interval could be more than 10 times that of a vertical wellbore-e.g.. a 400-ft [ 122-m] horizontal well in a 30-ft [9-m] -thick pay zone. The second advantage is the ability to complete several horizontal wells from a single location and to cover a large drainage area. This is essential for drilling in arctic or environmentally sensitive areas or offshore where preparation of drillsites is a major expense. A third advantage is that in the case of very thin pay-zone areas, vertical drilling would be uneconomical because of insufficient wellbore openings for production. Properly placed horizontal wells can solve this problem. For certain thin formations with bottom-water table, horizontal wells could defer and reduce water coning by providing a low-pressure area over a long distance rather than at a single point as with vertical wells. This can he achieved as long as there are no extreme differences in vertical permeability near the wellbore so that fairly uniform withdrawal of produced fluid over the length of the horizontal wellbore can be maintained. The fourth advantage is the ability to inject and/or to produce fluids orthogonal to those from a vertical well. This provides the potential of improving sweep efficiency of a flood and therefore recovery efficiency. Description In the steamflooding process, the low density of steam compared with other reservoir fluids causes steam to rise to the top of the formation and create the steam-override phenomenon. In the upper portion of the reservoir, which was swept by steam, the oil saturation can be reduced to residual oil saturation (ROS) to steam. In the lower portion of the reservoir, which was either swept by condensing hot water or bypassed by injection fluid, however, the oil saturation can still be rather high at the end of the steamflood. These oil pockets, sometimes called blind spots, can be reduced by either drilling infill wells to recover the oil or using limited entry at the producing wells to improve the vertical conformance of a steamflood. This paper investigates the use of horizontal wells for a mature steamflood project and for an undeveloped reservoir ready for steamflood operation. For a mature steamflood project, the steam-override situation already exists. Therefore, the investigation centers on the location and effectiveness of the horizontal wells in producing oil and the improvement in oil recovery efficiency. For an undeveloped reservoir, the proper design of a steamflood project to minimize steam override and to obtain maximum oil recovery is the study's objective. This study therefore considers some of the ways horizontal wells may be used to provide a better overall sweep efficiency in a steam-flood. It is not intended as a recommendation for horizontal wells but to present some potential alternatives for possible steamflood applications. The simulation work is based on the assumption that there are no production or injection problems with horizontal wells. Problems like sand control and the inability to do selective completions still have not been completely solved with horizontal wells and could adversely affect economic considerations. The THERM model developed by Scientific Software-Intercomp was used for the computer simulation. The model is a 3D numerical simulator developed for thermal recovery operations. The model accounts for three-phase flow described by the Darcy flow equation and includes gravity, viscous, and capillary forces. Heat transfer is modeled by conduction and convection. Relative permeability curves are temperature dependent. The model is capable of simulating well completions in any direction (vertical, horizontal, inclined, or branched). Reservoir properties used in the study are typical of a California heavy-oil reservoir with unconsolidated sand. A dead oil with a gravity of 13deg.API [0.98 g/cm3] was used in the simulation. Table 1 shows pertinent reservoir and fluid properties and Fig. 1 shows relative permeability curves. All steam injection was at a pressure of 425 psia [2930 kpa] and steam quality was 65 %. It was assumed that all wells were kept in pumped-off condition with a maximum pumping rate of 1,000 B/D [159 m3/d] for vertical wells and 3,000 B/D [477 m3/d] for horizontal wells. The 125-ft [38-m] formation was divided into five equal layers. Mature Steamflood Reservoirs Typical steamfloods operate from 2.5- to 20- acre [1- to 8-ha] patterns. SPERE P. 69^

For multiple-oil-zone reservoirs, steamflooding two zones at the same time is advantageous. Three methods for dual injection through a common wellbore were tested at Texaco's Kern River field in Bakersfield, CA. Because dual injection results in an excessive amount of heat transfer between injection streams, it was imperative to predict steam quality or water temperature at the sandface for each injection stream. To evaluate the merits of each method adequately, a computer model was developed to predict downhole temperatures, steam qualities, and pressures for each injection stream. Field tests of the three injection configurations provided measurements necessary for validation of the model. This paper describes the computer model, the field test, and the results from the three dual-injection scenarios.