Two experiments, Tests SSMJ01 and SSM02, involving steam foam drive processes have been performed in 3-D physical models. The physical models which simulated 118th of a 2-ha five-spot pattern for a Cold Lake oil sands deposit with a pay zone thickness of approximately 20 m. were designed using a thermal high-pressure scaling approach.
Both experiments were carried out successfully with a steam only injection period followed by steam-foam injection (i.e. steam-surfactant-nitrogen coinjection) periods. The formation of foam in situ was identified by the increase in pressure drop between the injection and the production wells after steam foam injection. Evidence of the injected steam being diverted from the steam swept zone into the oil rich zone due to the blocking effect of foam could be observed in the measured temperature profiles. Enhancement in oil production due to the diversion of steam was observed.
In addition, the effect of suffactant injection strategy on the propagation of foam was studied in Test SSM02. [I was found that after the foam has been fanned, it was more able to propagate away from the injection well by reducing the injected surfactant concentration.
The stearn drive process is a proven technique for the recovery of heavy ail and bitumen from underground reservoirs. In an effort to increase the efficiency of this process. considerable emphasis has been placed on developing methods to optimize reservoir conformance, particularly with regard to the problems of stearn channelling and gravity override. Foam is well known as a selective blocking agent, and has shown promise for the diversion of steam under conditions of poor reservoir conformance.
Steam-foam drive processes are very complex. 1-D cores, 3-D physical models and numerical models have been used as tools to understand the mechanisms of these processes. 3-D physical models can play a valuable role in extending the level of understanding from 1-D core size experiments to 3-D geometries and in developing data for validation of numerical models. From an assessment of a 3-D steam-foam drive experiment performed on a large-scale 90 cm physical model by Ridley et al. 1 and the subsequent numerical simulation conducted by Law and Ridley2 it became apparent that many improvements could be made in the design of this unscaled experiment The most significant problems encountered were: (l) controlling heat losses from the physical model:
estimating initial saturations of the test bed which was packed with mined Athabasca oil sands;
defining heat loss boundary conditions for Dumerical simulation; and
ensuring significant foam effects.
The 3-D physical models used in this study were designed by Law and Kimber3 in a numerical study based on a thermal high-pressure scaling approach. Compared to the earlier unscaled experiment1. The scaled experiments have:
better control over the heal losses from the physical model (i.e. heat losses which are more representative of those in the field); and
increased control over the initial saturation conditions in the physical model.