Increasing Oil Recovery From Heavy Oil Waterfloods

Abstract

Abstract A statistical study of 166 western Canadian waterfloods recovering heavy and medium gravity oils revealed new findings about best operating practices for heavy oil waterflooding. In classical light oil waterflooding, operators are advised to start waterflooding early and maintain the voidage replacement ratio (VRR) at 1. The study, however, produced surprising results for 2 parameters - among the 120 reservoir and operating parameters investigated - that ran counter to the recommended practices of classical light waterflooding. Delaying the start of waterflooding until a certain fraction of the original oil in place was recovered was found to be beneficial. Secondly, varying the VRR was shown to correlate with increased ultimate recovery - periods of under injection are needed, although a cumulative VRR of 1 should be maintained. Ultimate recovery was correlated with the primary recovery factor at the start of the waterflood. No trends appeared when the full set of 166 waterfloods was inspected. However, when the dataset is analyzed by ranges of API, a "sweet spot" of improved ultimate recovery was observed in a very narrow window of oil recovery factor prior to the start of waterflooding. Graphs of each category showed this "sweet spot" window where improved recovery occurred. These categories were API ranges; as well as ranges of permeability*height/viscosity (kh/µ); and pattern development. Also increases in ultimate recovery were observable when we examined graphs of ultimate recovery versus the fraction of injection volume that was underinjected - but again, only when the data was analyzed by the ranges. A certain period of injection when the VRR was less than 0.95 resulted in increased ultimate recoveries. However, it is important that this period of VRR < 0.95 be offset with periods of increased VRR so that the cumulative VRR cycles around 1.0. Again, each range manifested a narrow "sweet spot" for where this increase in ultimate recovery occurred. Introduction Waterflooding is becoming increasingly important in recovering heavy oil. In western Canada, 5201 million m3 of heavy oil is in place in Alberta and Saskatchewan, and more than 200 waterflood operations recover over 24% of that oil in place. Those western Canadian oil pools represent the largest source of data regarding recovery of oil as heavy as that found in many Alaskan oil pools. In order to optimize waterflooding strategy for the Alaskan oil pools, we examined the results of up to 50 years of waterflooding on 166 western Canadian waterfloods recovering oil of less than 30°API. Practices for waterflooding of conventional light oils were initially researched in the 1940's by Buckley and Leverett 1 and little has changed since Craig's2 work in the 1970's. Smith and Cobb3 cite that most of the sources refer to waterflooding oils of viscosity of less than 100 mPa·s. The major precepts of classical light oil waterflooding have been: start early; and maintain the voidage replacement ratio (VRR). Early start to waterflooding has been examined through the years and confirmed as the best practice for conventional oils by Gulick et al 4 and Ghozali 5. Maintaining even voidage replacement is so ingrained that Canadian producers must get permission from government regulators to deviate the VRR from 1.