Initially, polymer flooding had not been considered as a viable enhanced-oil-recover (EOR) technology for Pelican Lake in northern Alberta, Canada, because of the high viscosity of the oil until it was considered in combination with horizontal wells. Polymer flooding generally has been applied in light- or medium-gravity oil, and, even today, standard industry screening criteria limit its use to viscosities up to 150 cp. Pelican Lake is the site of the first successful application of polymer flooding in much-higher-viscosity oil (1,000–2,500 cp).
Introduction
The Pelican Lake field, approximately 250 km north of Edmonton, Alberta, Canada (Fig. 1 above), was discovered in 1978 and started producing in 1980. With more than 6 billion bbl of oil originally in place (OOIP) and a primary recovery estimated at less than 7%, it presents a significant target for EOR. But it is also a challenging reservoir with high-viscosity oil in a thin formation.
Early History
The reservoir-depletion mechanism is solution-gas drive, but initial reservoir pressure was low and there is very little dissolved gas, so there is little energy in the reservoir. Because the oil is also viscous (from 600 to 80,000 cp), primary recovery is low, approximately 5 to 10% of OOIP. In addition, the reservoir is thin (an average thickness of 5 m). As a result, the first wells drilled in 1980–81 were not economic.
Horizontal Drilling in Pelican Lake
CS Resources drilled its first horizontal wells in the Winter pool in Saskatchewan and then turned to Pelican Lake in 1987. Horizontal drilling is well-adapted to Pelican Lake, provided that the well can be maintained in the pay zone. Because the reservoir is so thin, a horizontal well can increase the reservoir exposure tremendously. The production performances of the horizontal wells were markedly better than those of the vertical wells and seemed to correlate reasonably well with the length of the horizontal drain in the reservoir. In 1991, CS Resources drilled its first openhole lateral arm from a main horizontal drain. Then, in 1993, it went one step further and drilled two multilateral wells with a new tool, the lateral-tieback system. The use of multilaterals would greatly expand in the years to come.
Screening of EOR Methods for Pelican Lake
Despite the improvement in recovery and overall economics resulting from the use of horizontal and multilateral wells, it was clear that primary recovery would be limited to 5–10% of OOIP, and other options were considered to increase recovery.
Thermal EOR. Pelican Lake is not well-adapted to steam-injection methods because of its thickness, which causes severe heat losses to the over- and underburden. Thermal methods were tested early to increase recovery, but attempts were unsuccessful.
In 1993, Amoco tested a new EOR method—cyclic air injection—in the field using one horizontal injection well and a horizontal producer drilled on either side of it. The results were good, with the rates more than doubling, but the process was abandoned, apparently because of low oil prices.
Chemical EOR. Meanwhile, CS Resources decided to explore a form of chemical EOR: polymer flooding. Except for the oil viscosity, Pelican Lake’s reservoir conditions are ideal for polymer applications: low temperature, low water salinity, no bottom aquifer below most of the pool, and high permeability favorable for injectivity.
Polymer Flooding of Viscous Oil
In the early 1990s, very little thought had been given to using polymer flooding to improve the recovery of heavy oil. A few polymer pilots had been performed in high-viscosity oil, but those were poorly documented and apparently unsuccessful. The conventional wisdom at the time was that polymer flooding was limited to viscosities below 200 cp. That was also the criterion proposed in the most widely used EOR screening guide. This belief still remains widespread within the industry.
That belief was a result not only of practical field experience but also of the idea that the higher the oil viscosity, the more viscous a polymer solution had to be in order to make a meaningful contribution to recovery. But higher-viscosity polymer would result in lower injectivity; thus, there were some practical limitations to injecting viscous polymer—at least with vertical wells. Advances in horizontal drilling, though, had changed things: By using horizontal injection wells, injectivity could be improved significantly in certain cases.
CS Resources Polymer-Flood Pilot
Preliminary Studies. An evaluation using reservoir simulation was first performed to verify that polymer flooding had the potential to increase recovery in the field. The next step was laboratory measurements of polymer properties in cores. Given the mild reservoir conditions, a hydrolyzed polyacrylamide with a molecular weight of 13.6×106 Daltons was selected. Further reservoir-simulation work was performed with the laboratory data to evaluate the influence of various parameters such as polymer concentration and well spacing.
Pilot Description and Operations. The pilot consisted of three 1250-m-long horizontal wells—two production wells and one injection well drilled between them. The distance between production wells and the injection well was 150 m. All three wells were put on production until the pilot startup in order to improve injectivity. Injection lines and the liner of the injection well were coated with epoxy to prevent polymer contamination by iron.
The startup operations did not go as planned; the viscosity target for the polymer solution (100–200 cp) could not be achieved. Further analysis revealed that the Quaternary aquifer water contained a significant quantity of dissolved iron (Fe II); this had not been observed in the laboratory because the water-sample bottle provided had not been sealed hermetically and, thus, iron had precipitated into Fe(OH)3. The oxidation reaction of Fe II into Fe III induced some polymer degradation. In order to resolve the issue, it was decided to precipitate the Fe II into Fe(OH)3 by aeration of the makeup water before preparing the polymer solution because Fe(OH)3 has no detrimental effect on polymer properties.
Lessons From the CS Resources Polymer-Flood Pilot. The first pilot was disappointing but still brought some important information. When PanCanadian decided in 2000 to evaluate the possibility of starting a new polymer pilot, the behavior of the first pilot was analyzed by use of reservoir simulations. A good history match was obtained by use of an increased skin in the injection well and low water relative permeability at residual oil saturation; further corefloods were conducted, and computed-tomography scans confirmed those assumptions.
Relative permeabilities for viscous oil are notoriously difficult to measure. In addition, although the influence of oil viscosity on relative permeability had long been debated, it now seems accepted that viscosity does indeed have an influence on relative permeability measurements. Water relative permeability at residual oil saturation has been shown to decrease with increasing oil viscosity. Given the uncertainty in the measurements of heavy-oil viscosity, reducing the relative permeability to obtain a match does not seem unreasonable.
The key practical consequence was that polymer viscosity did not need to be as high as originally thought to improve the oil/water mobility ratio. The other important conclusion was that the whole concept of polymer flood for Pelican Lake was still valid.
EnCana Polymer-Injection Pilot
Starting in 2003, EnCana conducted some workovers using polymer to improve conformance; conformance control consists of injecting a polymer or gel to solve water-channeling problems by plugging high-permeability intervals, or thief zones. The treatments seem to have been successful and were extended to a number of injection wells.
In January 2005 EnCana started injecting a much larger volume, and injection continued over several years. The results were positive, with increasing oil rate and decreasing water cut.
CNRL Polymer-Flood Pilot
CNRL initiated a polymer-flood pilot in the field in 2004. The laboratory work consisted of the selection of the polymer, the determination of its bulk properties, and the performance of corefloods to measure the adsorption and determine rheological parameters for reservoir simulations. The reservoir conditions for a polymer flood in Pelican Lake are mild; however, the experience of the CS Resources polymer pilot stressed the importance of injection-water quality.
Operations and Results. The pilot is composed of five 1400-m-long horizontal wells: three production wells and two injection wells in between them, with a spacing of 175 m between the wells. Polymer injection started in May 2005. The response occurred in February 2006 in the central production well and in April and September 2006 for the two other producers. The responses were excellent, with rates going from 18 to 232 BOPD in the first well, 9 to 364 BOPD in the central well, and 16 to 139 BOPD at the maximum in the last well.
Another striking feature is the slow and relatively moderate increase in the water cut for all three wells, especially compared with what was experienced in the waterflood pilot nearby. It must also be noted that the rates have remained high for almost 7 years.
Conclusions
The history of Pelican Lake reflects the progress in heavy-oil production technology during the past several decades and opens new perspectives for the future in areas where thermal recovery is inefficient. From the discovery up to 1987, the field was produced by vertical wells. The combination of thin reservoir, high oil viscosity, and no pressure support made production with vertical wells inefficient and uneconomical. Horizontal-well technology was the first main breakthrough, and development from the late 1980s until today has used long horizontal wells, sometimes with multilaterals.
Waterflood was implemented in the early 2000s and showed a substantial gain in recovery factor (from 5 to 10%). However, the adverse mobility ratio between the viscous oil (600 to 7,000 cp) and the water induced high water-cut and poor sweep efficiency in some patterns because of water channeling.
By the mid-2000s, polymer flood had been implemented and had proved highly successful, bringing the recovery factor up to 25% and higher while maintaining relatively low water cut. This second breakthrough in Pelican Lake exploitation is now implemented at the field scale. The important lessons from this experience are (1), for heavy oil in particular, polymer flood does not need to target a mobility ratio of unity to be efficient but should reach a compromise between mobility ratio and injectivity to optimize economics and (2) water quality is a key parameter.
This article, written by Special Publications Editor Adam Wilson, contains highlights of paper SPE 165234, “Pelican Lake Field: First Successful Application of Polymer Flooding in a Heavy-Oil Reservoir,” by Eric Delamaide, SPE, IFP Technologies; andAlain Zaitoun, SPE, Gérard Renard, SPE, and René Tabary, SPE, IFP Energies Nouvelles, prepared for the 2013 SPE Enhanced Oil Recovery Conference, Kuala Lumpur, 2–4 July. The paper was peer reviewed and published in the August 2014 SPE Reservoir Evaluation & Engineering journal, p. 340.