Steam-Drive Correlation and Prediction

Introduction During the past 15 years, steam-injection processes have become an importantmeans of exploiting heavy oil reserves. Traditionally, these processes havebeen classified as either steam soaks or steam drives. With combinations, suchas presoaking drive wells and partially driving steam soaks, the distinction isnot always applocable. Furthermore, our experience auggests that oil/steamratios from most mature processes converge to a calue determined only byreservoir and steam properties and time. To date, the steam-soak process has proven the more attractive, partlybecause the immediate response allows an early evaluation of a reservoir andpartly because oil rates from initial soak cycles tend to be better than latercycles. Successful steam soaks are limited to reservoirs where natural recoverymechanisms (gravity drainage, pressure depeletion, and solition gas drive)areineffective because of the low oil mobilities. Successful steam drives requiregood conformance,a means ofstarting the process because high oil saturations can limit injectivityseverely and prevent effective initial reservoir heating, andsustainedhigh injectivity throughout the process life. Unlike steam soaks, steam drivesdo not respond until built-up oil banks and heat reach the production wells. Because peak production rates may not be observed for several years after thestart of injection, piloting is expensive and expansion to full scale issoemwhat hazardous. For those reasons, screening methods that predict ultimateoil/steam ratio are useful in planning new projects or in modifying existingones. In the past, steam injection has been applied to a wide spectrum ofreservoir conditions, many of which have proven unsuitable. In retrospect, wecan explain the varied response withh a simple mathematical model thatincorporates reservoir and steam properties in the prediction. This paperdescribes the model and compares prediction from it with laboratory and fieldresults.