Three Dimensional SAGD Simulation of a Dipping Oil Sand Reservoir

Abstract A numerical simulation study was conducted for the application of the Steam Assisted Gravity Drainage (SAGD) technology developed at the Alberta Departmentof Energy Underground Test Facility (UTF) pilot to a dipping oil sandsreservoir in the Liaoning province of north east China. The Computer Modeling Group (CMG) STARS simulator was used in the study. The model was sensitized tokey reservoir and operating parameters, with specific focus on the differencesbetween the UTF site and the Liaohe (Shu-I) reservoir. Initially a two dimensional (2D) base case model was prepared using Liaohe Shu-l reservoir properties. A series of sensitivity runs was made to keyreservoir properties which differed significantly from the UTF site. Propertieswith a significant impact on performance include rock properties (heat capacityand thermal conductivity), oil viscosity as a function of temperature, porosity, and high temperature residual oil saturation. The use of high qualitysteam gave a better steam oil ratio than lower quality steam, consistent withthe concept that only the latent heat of the steam contributes to oil recoveryin the SAGD process. In all sensitivity runs, the ultimate sweep efficiencydown to the residual oil saturation is very high, above 90% the maindistinction between cases is the oil rate history, steam oil ratios, and theamount of recoverable oil in place (due to differences in porosity and residualoil saturation). All the UTF properties, except rock properties, lead toimproved performance compared to the base case using Shu-I reservoirproperties. Reservoirs with significant dip, such as the Shu-I necessitate the use of afull 3D model of the reservoir with" coupled wellbore-reservoir capability. Astepwise approach was taken from the 2D base case model, to coupled 2D models,3D models with standard source-sink well models (with no dip and with dip), andeventually to a 3D model with dip and the new discredited well model coupled tothe reservoir. With no dip, the discredited well model gave very similarresults to the source-sink model. In the presence oj dip, the source-sink wellcould not correctly model steam trap control in the production well. Thediscredited well allowed steam trap control as used in field operations, and itshowed that very similar performance is predicted with dip compared to no dipassuming that both the pressure drop in the wellbore, and heat transfer betweentubing and annulus, are small. With the base case Shu-l reservoir properties, and Wellbore completions havingminimal pressure drops and heat transfer effects, the expected performance for SAGD in the Shu-J reservoir is: CDOR 63 m3/d per well pair over 6 years SOR 3.8 (using 70% quality steam) over 6 years Recovery efficiency of 73% over 10 years INTRODUCTION Steam Assisted Gravity Drainage (SAGD) implementation, and associatedsimulation, has focused on the use of non-dipping horizontal wells. On a globalbasis, however, reservoirs may have sufficient dip that the SAGD wells must bedrilled along a line of dip in order to ensure that significant volumes of thereservoir are not left unaffected by the process, since reservoir volumes belowthe producing well will be unwept.