Numerical Simulation and Optimization of the SAGD Process in Surmont Oil Sands Lease

Abstract

Abstract Optimization of Steam Assisted Gravity Drainage (SAGD) remains a major concern in Surmont leases (an Athabasca oil sands deposit located in northeastern Alberta, Canada) due to the extensive presence of top gas and top water zones over the bitumen. Observation well data has detected the pressure communication between the SAGD steam chamber and overlying thief zones. Maintaining the steam chamber pressure is very difficult due to these thief zone interactions. Previous numerical simulations, laboratory experiments and field production data have demonstrated that the overlying top water and gas thief zones have a detrimental effect on the SAGD process. Oil production and steam oil ratios tend to decrease as the depletion of top gas continues. Also, the heat loss to the overlying thief zone will be more significant when the top water zones are present. However, an optimal operating strategy for the full field scale SAGD process with both top gas and top water remains uncertain. In addition, a detailed investigation of the impact of top gas and water thief zones on SAGD performance provides the basis in calibration of geostatistical and flow models for commercial phase planning and forecasting. The objective of this paper is to construct numerical flow simulation of a Surmont pilot using a well-defined 3D geostatistical model to determine the impact of the top thief zones on bitumen recovery. The focus of the study will be on three horizontal well pairs plus 15 vertical observation wells at the McMurray formation. The stochastic geostatistical model is to build a representation of the McMurray geology that honors the deposition structure, facies proportions, reservoir characteristics and petrophysical properties. Structural tops are interpreted from well logs and porosity-permeability relationships established from quantitative log analysis and core-log calibration. The facies-based log-derived porosity, permeability, shale volume and water saturation are populated into a grid block by Sequential Gaussian Simulation (SGS) in the petrophysical modeling process. Then a static model is upscaled to coarse simulation grids, and a submodel for each single well pair is extracted for the purpose of history match in STARS™ simulator. Reasonable history match of oil and water rates has been achieved by calibrating this static model with the field production data. The steam chamber pressure and temperature profile from the numerical model has been conformed to the field data from the observation wells. Optimization of cumulative steam oil ratios (cSOR) by varying injection pressure and the steam trap control with the top thief zones has been investigated in great detail. In conclusion, SAGD performance is dominantly controlled by geological heterogeneity, completion and operation constrains, and steam chamber pressure variations. Finally, integrated optimization strategies have been developed and tested on a full field-based heterogeneous simulation model.