Conventional steamflooding of heavy oils and tars in thin reservoirs generally suffers from excessive heat losses to the over-and-under burden. High rate steam injection is often constrained by limited reservoir injectivity. A recently developed FAST process (Fracture Assisted Steamflood Technology) allows high rates of steam injection in thin shallow sands through hydraulical1y created horizontal fractures. This process limits the hear losses by shortening the steamflood life for a 5 to 7.5 acre (2 to 3 ha) pattern to less than two years.
Conoco demonstrated the technical feasibility of the FAST process in two pilot tests in south Texas where -2 ° API (l 093 kg/m3) tar was successfully produced. In an attempt to understand the dominant flow mechanisms in the FAST process and to develop a technique for predicting performance of future FAST steamfloods, numerical simulation was used to achieve a comprehensive history match of the second pilot test at the Saner Ranch in south Texas.
Despite the complexity of this recovery process, a commercially available steamflood simulator was used to match the performance of one-twelfth symmetry element of the inverted 7-spot, 7.5 acre (J hal pattern. In addition to matching the tar and water production, the temperature and pressure histories, temperature profiles. water cuts and volumetric tar recoveries were also matched.
One of the key elements was the modeling of multiple phase flow through an open fracture. This was accomplished by a simple technique which did not require reprogramming of the conventional steamflood simulator. The numerical simulation of the FAST process revealed dominant flow mechanisms in the process. The calibrated model is now being used for on-going predictive and process optimization studies.
This paper briefly describes the FAST process, pilot history, details of the technique used to model flow in an open fracture, and results of the numerical modeling study.
Conoco conducted two pilot tests of the FAST steamflood process in the south Texas tar sands deposits (Figure I). These formations contain -2 ° to +3 ° API (l 093 to I 052 kg/m 3) tar with viscosities of 1 to 20 million centipoise at reservoir temperatures. The first pilot was conducted from November 1977 to June 1980 at the Street Ranch lease where 169,040 barrels (26 875 m3) of -2 ° API (I 093 kg/m3) tar were recovered from a. 5-acre (2 ha) pattern after injecting 1,847,500 barrels(293 730 m3) of steam (1). The second pilot test was conducted at the Saner Ranch from April 1981 to January 1983. A total of 133,260 barrels (21 190 m3) of tar were recovered from a 7.5-acre (3 ha) pattern after injecting only 1,069,000 barrels (169 950 m3) of steam (2). The main objective of the first pilot was to prove the feasibility of the FAST process [or recovering -2 ° API(1 093 kg/m 3) tar. The objectives of the second pilot test were to demonstrate the FAST process in a different part of the reservoir and to improve performance.
The Saner Ranch pilot was a 7.5-acre (3 ha) inverted 7-spot pattern. As shown In Figure 2, one central injection well was surrounded by six producing wells arranged in a hexagonal configuration. The reservoir properties of the test pattern are illustrated in Figure 3. The reservoir is 65 feet (19.8 m) net thick with the sand divided into two zones of approximately equal thickness. The average tar saturation in the lower zone is 49 percent compared to 40 percent in the upper zone. This is the main reason pilot testing at the Saner Ranch was conducted in the lower