Experiences With Waterflooding Lloydminster Heavy-Oil Reservoirs

Abstract

Summary This paper compares observed and theoretical behavior of waterflood performance in heavy-oil reservoirs in the Lloydminster area of western Canada. Lack of reliable primary production history makes determination of primary recovery difficult and consequently makes additional oil recovery by waterflood difficult to quantify. Comparison of predicted and actual performance indicates that the floods ar-e behaving as well as, if not better than, expected. Extra oil recovery by waterflood is not expected to add to the primary recovery of 3 to 8% by more than 1 to 2% of original oil in place (OOIP). Introduction The heavy-oil sands of the Lloydminster area (Fig. 1) are the southern extension of the trend of Lower Cretaceous bitumen and heavy-oil deposits extending from Athabasca through Cold Lake to Lloydminster. Unlike the first two areas, many of the Lloydminster reservoirs produce under primary recovery mechanisms. Total oil-in-place (OIP) estimates vary from 8 × 109 to 11 × 10(9) m3 (50 to 70 billion bbl). Husky Oil Operations Ltd. leases cover nearly 8 × 109 M2 (2 million acres) in this area along the Alberta/Saskatchewan border, with current oil production of about 4000 m'/d (25,000 B/D) from some 2.000 wells. Total oil production from the area is about 9500 m3/d (60,000 B/D). The heavy oil is found in the Mannville Group (Fig. 2) of Lower Cretaceous age, and the reservoirs are typically very fine- to fine-grained, clean, highly porous, uncon- solidated quartz sand bodies that grade laterally into silty or shaly facies or are truncated abruptly by channeling. Some 80 % of the OIP is found in sands less than 5 m (16 ft) thick. High crude viscosity, low solution GOR, and low initial reservoir pressures result in primary recovery efficiencies from 3 to 8% of OOIP. Waterflooding has been carried out in most of the major reservoirs since the mid-1960's, with initial expectations of doubling primary recovery. This paper summarizes the performance of Husky's main reservoirs and compares actual with predicted waterflood performance. These predictions have been facilitated greatly by recent extensive core and log data obtained from drilling numerous wells on close spacing for tertiary pilot projects. Reservoir Characteristics Fig. 3 shows the major oil pools in the area. Two fields, Wildmere and Wainwright, that have slightly higher gravity and less viscous crude have not been included in this analysis. Reservoir characteristics of the remaining heavy-oil pools are summarized in Table 1. Reservoir dips are extremely low (less than 1/2 @), and many reservoirs are underlain by bottom water, with the oil/water contact varying considerably in the same reservoir because of varying sand quality and oil gravity. Disparities of 4 m (13 ft) are common, and differences of up to 10 m (32 ft) have been observed. Most reservoirs are thought to have been at the bubble point, and many initial gas caps were found. Individual well completion rates range up to 25 m 3/d (80 B/D) oil but average about 7 m3/d (23 B/D). Newly drilled wells typically have a sand cut of 15 %, which decreases over a 6-month period and then stabilizes at 0 to 1 % during the life of the well. The high viscosity, sand production, and low GOR result in numerous production problems. These have required development and application of specialized completion techniques to ensure efficient and profitable operation. During the first 6-month production period, a large volume of sand usually is produced. A peak productivity commonly is obtained that is an order of magnitude larger than would be predicted from the radial flow semisteady-state calculation. JPT P. 1643^