The use of a geostatistically derived facies probability map as an aid to locating infill wells is described. Short radius sidetracks have been designed due to difficulty in correlating thin channel sands away from the mother wellbore, and geostatistical based simulation models have been used to predict the likely production performance, and allow assessment of geological risk. Refined geostatistical grids have been used to better model local heterogeneity around the horizontal well locations. The use of formation damage and geomechanical studies in designing horizontal wells in depleted reservoirs is outlined.
The B-8 reservoir lies close to the bottom of the "B" Eocene sand units in the Eastern part of Lake Maracaibo, Venezuela (figure 1). It comprises generally poor quality delta front sands and shales, and is isolated from adjacent blocks by sealing faults, and shales at the top and bottom of the unit. Figure 2 shows a top structure map for the study area in the southern portion of the reservoir, approximately 4 km × 3 km in extent. This area is considered to be mainly isolated from the northern part by a reservoir section with low net to gross. Cumulative production to date from the study area has been 5.4 mmstb from a STOIP estimated at 63 mmstb. Reservoir pressure is currently about 1400 psia, approximately 1000 psia below the bubble point; fluid properties are shown in table I. The reservoir produces 24.5 API oil by natural depletion, with no evidence of aquifer support. Localised gas caps in some of the sand lenses have been reported, but poor sand continuity has prevented the formation of an effective secondary gas cap, and precluded the economic use of injection for pressure support.
Although production commenced in 1954, only three wells have achieved more than six years production, and well rates have been below 100 stb/d in recent years. Engineering studies in 1993 identified infill drilling as the preferred depletion option, and a horizontal well infill program started in 1994. Two medium radius sidetracks and a new horizontal well were subsequently drilled, but two of these wells had very poor production rates (30–50 stb/d) and it became obvious that the geology of the area was poorly understood; the infill program was suspended. To date, a total of 16 wells have produced from the zone.
This paper reports the work performed by a multidisciplinary team, using a geostatistical approach to gain a better understanding of the reservoir heterogeneity and to define probabilistically the performance of new infill wells.
The B-8 reservoir has been identified as a transitional delta front zone between the pro-deltaic B-9 unit, and more continental B-7 and B-6 channel sediments; reservoir quality improves upwards through this progradational sequence, deposited in Middle Eocene time. The B-8 unit, containing approximately 200 feet of sediments, was deposited in a mainly shallow marine environment, and comprises marine and coastal bars and shales, coastal plain silts, and channel sediments. The simple structural setting consists of a fault bounded monocline, with average strike NNW-SSE, and a dip of 3 degrees to the Southeast.
Analysis of logs from the horizontal wells indicated that continuity of channel sands was much poorer than originally thought, and that previously correlated sands didn't appear to be stratigraphically equivalent. Detailed stratigraphic and sedimentological studies were prepared to better define the reservoir geology.