Reservoir Simulation Study of a Thermal Horizontal Well Pilot in the Cold Lake Oil Sands

Abstract

Summary A reliable model was developed for evaluating and guiding the performance of horizontal wells in thermal operations to recover heavy oil in Cold Lake, Alta., Canada. The horizontal well model was developed with measured field data from the first thermal horizontal well pilot (HWPI) of Imperial Oil Resources Ltd. Introduction Excellent history matches of the pilot's bitumen and water production rates and bottomhole pressures (BHP's) and acceptable history matches of the observation-well temperatures and of the thermal energy of the produced fluid were achieved with a thermal reservoir simulator developed by Exxon Production Research Co. The 3D horizontal well model applied in the simulator had three phases and three components. It was interfaced with a 3D geological reservoir description model of the pilot area. The history-match study was valuable in understanding measured field data and the important physics affecting thermal horizontal well performance. Obtaining a reliable history match of the HWPI and developing simulator models to assess and guide horizontal well operations in Cold Lake was significantly facilitated by (1) the availability and use of a large number of diverse measured field data, (2) the incorporation of a detailed reservoir description model, and (3) strong emphasis on ensuring that measured field data supported major assumptions applied to history match the HWPI simulator model. Background The Imperial Oil Resources Cold Lake oil sands property is located 256 km [160 miles] northeast of Edmonton. It is the largest steam-injection project in Canada, and current bitumen production is 14 300 m3/d [90,000 B/D] from nearly 2,000 wells. The vast majority of the production comes from the use of cyclic steam stimulation (CSS). Currently, three horizontal well pilots are in operation at the site. The pioneering work1-3 of the horizontal well application to produce heavy oil (bitumen) led to drilling the first pilot in 1979. Gallant and Dawson4 described the objectives of the first two pilots. The third pilot will test the technical and economic feasibility of a horizontal well follow-up to CSS.5 The major thrust of this study was to develop simulator models to predict the performance of horizontal wells as a follow-up and as an alternative to CSS and to guide future operating strategies. This paper summarizes the history-match study of the first HWPI to develop a thermal horizontal well simulator model for heavy-oil operation in Cold Lake. The first HWPI in Cold Lake had a 245-m [804-ft] slotted liner positioned about 30 m [98 ft] below the top of the Clearwater formation. A vertical well steam injector was located almost directly above the horizontal well 45 m [148 ft] from its reservoir end (see Fig. 1). An operating strategy6 was applied to test gravity-drainage concepts.7 It consisted of a preheating phase followed by high-pressure steam injection into the vertical well (i.e., above the reservoir failure pressure) and continuous production from the horizontal well. A bottomhole pump was installed in the horizontal well during July 1982. The present study was performed with Exxon's fully implicit thermal version of the MARS simulator.8 The 3D reservoir simulator model (11×5×12) handles three phases and allows more than three components to be studied. It was interfaced with it 3D reservoir description model of HWPI. The reservoir description model was a computerized geologic model describing formation thickness, bitumen saturation, permeability, and the location of vertical flow barriers or "tight streaks." Reservoir Description Steam injection operations in Cold Lake occur in the Clearwater formation.9,10 The formation is an unconsolidated, clean, well-sorted, fine- to medium-grained sand. Initial average reservoir properties6 of the pilot are 33% porosity, 47-m [154-ft] net pay thickness, 1.3-darcy permeability, and 60% bitumen saturation. Initial bitumen viscosity was approximately 0.1 MPa·s [100,000 cp] at the ambient reservoir temperature of l3°C [55°F]. The ratio of vertical to horizontal permeability, kV/kH, was varied as a history-matching parameter. A kV/kH of 0.2 resulted in acceptable history matches of thermal energy production, breakthrough time, and total fluid production. Four major tight zones are present in the geological model: two in the top third and two in the bottom third of the Clearwater formation. Tight streaks with thicknesses greater than or equal to 1 m [3.3 ft] were included in the geologic model. An initial reservoir datum pressure of 3 MPa [435 psi] was applied at the horizontal well layer. Bitumen viscosities and reservoir fluid and rock properties used in the model were the same as those used in prior CSS studies.11,12 Simulation Model Description and Treatment of Field Data Simulation Grid Model. Fig. 2 shows the location of the HWPI in the Cold Lake Leming area and the boundaries of the reservoir area represented in the HWPI simulation model. Fig. 3 is the full-size areal grid system of the model with the locations of fracture blocks indicated. The Clearwater formation was gridded into. 12 vertical layers. Fig. 4 shows the average thicknesses of the blocks in each layer of the geological model. The vertical gridding was designed to allow accurate representation of the geological description and true locations of the vertical well perforations. Layer 6 in Zone 2 represents the vertical injector fracture layer, which results from steam injection into the reservoir: The horizontal well is located in Layer 10 of Zone 3. The model included a representation of Wells T-14 and T-15 in the neighboring T pad to allow investigation of interwell communication effects on the HWPI history match. The wellbore index formation applied to model the horizontal wellbore is consistent with that presented in Ref. 13. Pressure Drop in the Horizontal Well. In the model, pressure-drop effects were neglected in the horizontal wellbore because calculated pressure drops are small compared with expected pressure drawdowns in the reservoir. The calculated maximum pressure drop in the horizontal well was <3.45 kPa [<0.50 psi]. Calculations were based on the following data, which are consistent with field measurements: a maximum bitumen production rate of 50 m3/d [315 B/D] through the 193.6-mm [7.62-in.] -diameter slotted liner and a minimum bitumen temperature of 100°C [212°F] in the liner. p. 403-406 Simulation Grid Model. Fig. 2 shows the location of the HWPI in the Cold Lake Leming area and the boundaries of the reservoir area represented in the HWPI simulation model. Fig. 3 is the full-size areal grid system of the model with the locations of fracture blocks indicated. The Clearwater formation was gridded into. 12 vertical layers. Fig. 4 shows the average thicknesses of the blocks in each layer of the geological model. The vertical gridding was designed to allow accurate representation of the geological description and true locations of the vertical well perforations. Layer 6 in Zone 2 represents the vertical injector fracture layer, which results from steam injection into the reservoir: The horizontal well is located in Layer 10 of Zone 3. The model included a representation of Wells T-14 and T-15 in the neighboring T pad to allow investigation of interwell communication effects on the HWPI history match. The wellbore index formation applied to model the horizontal wellbore is consistent with that presented in Ref. 13. Pressure Drop in the Horizontal Well. In the model, pressure-drop effects were neglected in the horizontal wellbore because calculated pressure drops are small compared with expected pressure drawdowns in the reservoir. The calculated maximum pressure drop in the horizontal well was <3.45 kPa [<0.50 psi]. Calculations were based on the following data, which are consistent with field measurements: a maximum bitumen production rate of 50 m3/d [315 B/D] through the 193.6-mm [7.62-in.] -diameter slotted liner and a minimum bitumen temperature of 100°C [212°F] in the liner. p. 403-406