The MacKay River Project is located approximately 60 kilometers northwest of Fort McMurray, Alberta, Canada. The development consists of 25 well pairs and facilities to produce 30,000 bopd of bitumen from the McMurray Formation of the Athabasca Oil Sands Deposit over a 25 year life. An in-situ recovery process referred to as Steam Assisted Gravity Drainage (SAGD) is used. The SAGD process utilizes horizontal well pairs, drilled 5m vertically apart, to provide continuous steam injection into the upper well with continuous fluid production from the lower well. The initial phase of the MacKay River Project consists of two production pads that supply steam to the injection wells and accept return fluids from the production wells. Produced fluids are processed at the MacKay River central facility where the bitumen is shipped to market via an insulated pipeline. Solution gas is recovered for use in steam generation, while the water is treated and recycled. During the start-up at MacKay River, communication was initiated between each producer and injector by circulating steam in both of the wells. Once the reservoir between the two wells was sufficiently heated and bitumen mobility was evident, a pressure differential was applied between the wells for a short time after which the well pair was converted to SAGD mode. In SAGD mode the upper and lower wells became dedicated injection and production wells, respectively. In developing the start-up operating parameters for the MacKay River well pairs, a pseudo-compositional thermal reservoir simulator was used to perform sensitivities to determine the optimum operating strategy. The reservoir properties in the model were imported from a geostatistical model that was developed for the area. A discretized wellbore using the as-drilled trajectory was also implemented in the model. This allowed for the implicit modeling of the pressure and heat transfer dynamics in the wellbore, which are very significant in a SAGD operation. The main objective when developing this strategy was how to initiate communication between the producer and the injector in the optimal time without causing adverse affects on the long term SAGD performance. The variables that were investigated during the operating strategy development included steam circulation rate, steam circulation pressure, the magnitude and timing of pressure differential implementation between the injector and producer and the optimum timing of SAGD conversion. This start-up operating strategy was successfully implemented in the field from September to November 2002. Subsequently, the models are being calibrated to the field measured start-up data. Upon completion, these coupled wellbore reservoir models will be used to define the optimal ramp-up operating strategy for the MacKay River SAGD well pairs. This paper will detail the sensitivities conducted and the start-up prediction that were generated through the coupled wellbore reservoir simulator with a comparison to actual field data. Introduction Steam Assisted Gravity Drainage (SAGD) is an in-situ thermal recovery technique used to recover heavy oil or bitumen. This technology utilizes two parallel horizontal wells located near the bottom of the reservoir. Steam is injected continually into the upper injection well and rises, developing a steam chamber. The steam will flow to the interface of the chamber and condense, giving up the heat to the surrounding cold oil sands. The heated bitumen and condensate drain by gravity to the lower production well where the fluid is removed continuously. During this process the steam chamber grows upwards and laterally1. In the third dimension, an optimum SAGD process requires that the steam chamber, along the entire length of the well, be situated just above the production well with a minimum amount of steam production. The temperature along the production well can indicate the height of the chamber above the production well and is a function of the production well operating pressure.2

TX 75083-3836, U.S.A., fax 01-972-952-9435. AbstractIn a previous paper, a general theory for gas production during Steam-Assisted Gravity Drainage (SAGD) has been presented 1 . The main gases formed during "aquathermolysis" reactions, carbon dioxide and hydrogen sulphide, tend to be produced primarily via the produced water.The present paper reports a numerical analysis of the theory, and provides the results of a first simulation that describe the gas production history during SAGD. The numerical analysis includes hydrogen sulphide, carbon dioxide, and methane solution gas.Gases were included in the numerical analysis by using Kvalues derived from new theory. This theory accounts for the asymptotic behaviour of gases in solution as the critical point of water is approached.Simulation results corroborate the initial simple analysis, and agree with field observation. Minor differences are discussed.

World-wide experience indicates that thermal methods are the only effective methods for the in-situ recovery of heavy oils. Furthermore, the steady progress within horizontal well technology increases the applicability of thermal methods in con)unction with horizontal wells. The steam"assisted gravity drall1age (SAGD) process is the prominent example of the synergism of horizontal wells with steam injection.The major portion of the dollar per barrel operational cost for steam injection is currently attributed to the generation of steam and water handling requirements. In the future, more stringent environmental regulation might further prohibit the use of water for the recovery of heavy oils.aptha is presen~at the site as a diluent for pumping and pipe Itne transport ot the produced heavy oil. The injection of naptha vapour instead of water vapour combines an effective thermal process with the diluent mechanism of the naptha. We present the results of simple analytical calculations del~10nstrating the applicability of naptha-assisted gravit; dra1l1age (NAGD) compared to the SAGD process.Numerical simulations have been performed on a twodimensional cross section perpendicular to the horizontal well pai~" We discuss in detail the reservoir mechanisms present dunng NAGD. For the conditions studied naptha circulation rates are some 14 times the bitumen production rate. Evaluation of the process shows that it will be economic with naptha recycling and by minimising the naptha remaining in the reservoir.References at end of paper 475

Summary Stakeholders in in-situ oil-sands development take caprock-integrity issues seriously. The industry is faced with the challenge of determining an optimal operating pressure in the reservoir where, in general, the pressure should stay significantly low to ensure the caprock integrity while being significantly high for enhanced oil production and economics. This paper presents a comprehensive work program on the subject for a shallow oil-sands play. Caprock integrity considers the induced stress and deformation in a caprock during the thermal stimulation of an oil-sands reservoir. A minifrac-test program is undertaken to define the original in-situ stress state. Laboratory tests are carried out to measure the deformation and strength properties. Simulations are run to calculate the induced stresses and evaluate them against the mechanical strength. This paper describes some important quality-control issues for these activities. For the minifrac tests, multiple cycles and use of flowback are promoted for enhanced efficiency and accuracy. Laboratory tests are recommended on whole cores in a drained condition at a slow strain rate. Numerical simulations should use site-specific and laboratory-measured material properties. On the basis of the limited sensitivity analyses, the thermal expansion coefficient of the reservoir and Young's modulus of the caprock are found to significantly affect the caprock deformation and/or induced stresses.