Realistic History Matching of Cyclic Steam Stimulation Performance of Several Groups of Multilateral Wells in the Peace River Field, Canada

Abstract

With 8 billion barrels of bitumen in place and more than 30 years of thermal piloting and demonstration projects, Peace River offers an excellent growth opportunity for Shell's ultra-heavy oil portfolio. In support of this initiative, integrated geological and reservoir modeling of two project areas was conducted. The key objectives were toimprove predictive modeling capability of cyclic steam stimulation (CSS) projects by history matching two groups of CSS multilateral wells anddevelop a history matched physical representation that not only validates empirical models but can be deployed to optimize CSS designs for full field development. Detailed geological models were created over two pad areas providing a geological framework large enough to have realistic boundary conditions, including impact of surrounding wells. The geological models were imported into CMG's STARS thermal reservoir simulator, and a relatively fine grid was extended over each project area. All available historical production, injection, pressure and temperature data were used in history matching. Steam-induced reservoir dilation, explicit fracturing, and relative permeability hysteresis were important aspects of the overall physical representation. Common physical parameters for dilation/re-compaction, fractures, permeability/porosity transforms, vertical to horizontal permeability ratios, and relative permeability hysteresis were used for both pads. Each pad area maintained its own unique geological, petrophysical, and fluid properties, in line with observed field trends. Excellent history matches (aided by experimental design) of injection and production volumes, injection wellhead pressures, estimated production bottom-hole pressures and temperature profiles were achieved not only for the entire Pad A and B groups of wells, but also for the individual wells. In summary, a predictive CSS simulation model has been developed and validated by history matching two areas of the Peace River field. The model is suitable for sensitivity studies of geological, petrophysical, and fluid properties. It is also capable of assessing impact of well configuration, spacing, steam quality, and steaming strategy. Introduction Peace River is 100% Shell owned heavy oil property located in north-western Alberta, Canada, approximately 700 km northwest of Edmonton (Fig. 1). It holds approximately 8 billion barrels of 7°API oil in place trapped in approximately 30 m thick semi-consolidated sand layer (Fig. 2a) buried at a depth of about 600 m, and spread over approximately 370 km2. The Bluesky reservoir has been broadly classified into two intervals, poorer quality Estuarine and good quality Deltaic (Fig. 2b). CSS is employed to extract the oil, most recently using closely spaced multi-lateral horizontal wells drilled from a central pad. Modeling Objectives Growth plans being considered cover development across the entire field. Optimizing such a development plan requires investigating a host of sensitivities, including well configuration, placement, spacing, steam quality, steam slug sizes, production cycle length etc. Thermal reservoir simulation is a viable tool that can be deployed to explore for the most optimum case. The challenge with such models is for them to be reasonably well history matched, while at the same time retaining their predictive capability. Injection above fracture pressure adds to physical complexities. Addressing this challenge became the primary modeling objective.

Reservoir Simulation of Steam Fracturing in Early-Cycle Cyclic Steam Stimulation

“…Moreover, one interesting result from this study is that the case without a predefined fracture layer yielded results that were more similar to the behavior observed in the field. This implies that earlier approaches, as performed by Lebel and Moriyama (1997), Koci and Mohiddin (2007), and Walters et al (2000) to model CSS with the use of predefined fracture layers, should not be used.…”

“…In another study conducted by Koci and Mohiddin (2007) for Shell's Peace River property, the authors outline two different methods in a widely applied simulator that can be used to include geomechanical effects in the model. The first is a reservoir-dilation and -recompaction model, also known as the Beattie-Boberg model (Beattie et al 1991).…”

“…Because the location, depth, and orientation of these hydraulic fractures are diffi-cult to determine, it is difficult to specify the location of these fractures. In the study conducted by Koci and Mohiddin (2007), a horizontal fracture was placed at the top of the reservoir and a pressure-dependent transmissibility modifier was applied to the top reservoir layer. These fractures will then open or close, depending on the reservoir pressure.…”

“…The estuarine zone has poorer quality (i.e., lower permeability) than the deltaic one. In current operations, closely spaced multilateral J-wells are drilled from a central pad facility (Koci and Mohiddin 2007). Shell is now shifting toward an inverted seven-spot pattern in which each pattern will cover 3.4 ha within the Bluesky formation (Shell Recovery Process 2009).…”