History Matching the Steam/Foam Injection Process In a Thick Athabasca Tar Sand Reservoir

Abstract

Abstract Throughout the life of the Gregoire Lake In-Situ Steam Pilot (GLISP), a number of steps (i.e., steam stimulation, steam flooding, steam/foam injection, surfactant concentration reduction, pressure cycling, injection rate reduction and air/steam injection) were taken to develop and optimize the steam/foam injection process. After completing the project, a history match study was initiated to understand the process mechanisms and sensitive parameters. The ultimate purpose was to establish a steam/foam forecast model for the Athabasca tar sands. This paper presents the history match results and highlights the important process considerations that led to the final match. The model was validated by obtaining satisfactory predictions for other pilot producers. Overall, the empirical foam model was found to be adequate for field scale steam/foam simulation. The history match results confirmed that the communication path in the upper low oil saturation zone had a strong impact on the GLISP production performance and supported some of the operational decisions made by the project team during the pilot operation. The results also suggested that the steam/foam injection strategy could be further optimized. Introduction The Gregoire Lake In-Situ Steam Pilot (GLISP), a joint venture between AOSTRA. Amoco Canada and Petro-Canada (Phase A only) with Amoco being the operator, was the first steam/foam project conducted in the thick Athabasca tar sands. It successfully demonstrated the diversion of steam to the lower tar rich zone and resulted in improved bitumen recovery. The detail field history has been published by Sander et al.(1)This paper presents the most recent history match results based on techniques established in prior simulation studies.(2,3&4) A pilot well layout map is shown in Figure 1. The pilot consists of one injector (H6) and three producers (H3, H4 and H5) drilled in an inverted 4-spot pattern with an injector to producer distance of 46m. Three temperature observation wells (HO7, HO8 and HO9) were also drilled to obtain temperature information. The project was operated from October 1986 to July 1991. It was originally designed to create horizontal fractures in the rich tar zone to optimize steam flood heating. However, it was later found that a multiple fracture system was induced(5) and had led to steam overriding in the upper low oil saturation zone. A steam/foam diversion was thus initiated in October 1988 to mitigate the problem. Subsequently, additional tests were conducted to evaluate and optimize the process. The following summarizes the key pilot activities:Initial steam cycling to preheat the reservoir (October 1986 to July 1987).Steamflooding and occasional steam stimulation at production wells (July 1987 to September 1988).Steam/foam injection to mitigate the steam override effects (October 1988 to July 1989).Foam concentration reduction to optimize operating costs (August to December 1989).Injection pressure cycling, injection rate changes and air/steam injection to optimize the steam-oil ratio (January 1990 to July 1991). The objective of this paper was to match, as much as possible, the injection well bottomhole pressure, producer fluid rates and the observation well temperatures, with special emphasis given to the steamflood and foam injection periods.