Abstract
Reliable reservoir fluid saturation data areimportant to the over-all planning for prospectivesupplemental recovery projects. A trial-and-error graphical procedure, based on the fractional flowequation and using field production data, has beendeveloped to determine these values with reasonable accuracy.
Introduction
The preliminary analysis of a reservoir todetermine the feasibility of supplemental recoveryoperations requires a reliable estimate of reservoir fluid saturations, and a judgment must be made asto the degree of accuracy desired. An increase inthe complexity, data requirements, andsophistication of the various available solution methods doesnot necessarily imply that the results will be moreprecise. The decision as to which solution method should be used is sometimes difficult to make.This decision should consider, among other things, engineering manpower requirements for the various alternatives, the quality of existing data, and thesensitivity of the project economics to a reasonablevariation in fluid saturation levels.Average oil, water, and gas saturations for areservoir as a whole are usually determined by apressureTroducdonhistorymatchusingconventionalmaterial-balance methods. More sophisticatednumerical simulation techniques are generally used to determine reservoir saturation distributions.Simulation work can be time consuming andexpensive when efforts are made to determineaccurately the saturation distribution by attemptingto match the individual performances of a largenumber of wells.A trial-error graphical procedure has been developed as an alternative to the moresophisticated methods. Because of its easy use, the procedure may be considered as an efficientand economical application of engineering effort forplanning purposes. An application of this techniqueis illustrated for the Ten Section field, KernCounty, Calif.
RESERVOIR DESCRIPTION
The Ten Section field, located near Bakersfield, Calif., was discovered in 1936; developmentprogressed continuously until 1942, at which time 126 wells had been completed within the originallyproductive area of about 2,200 acres (Fig. 1). TheTen Section structure is a gently dipping (3 deg. to 7 deg.)doubly plunging anticline. The producing measuresare the upper Miocene Stevens turbidite sands, which have been subclassified as Zones 1, 2, and 3.The Zone 2 reservoir includes the intermediateseries of these sands that are bounded above by athin continuous shale varying in thickness from 5to 10 ft and are separated from the deeper measuresby an extensive thin shale. The sand intervals ofthe upper Stevens turbidite sequence are composedof individually graded and rythmically beddeddeposits. Core observations indicate that the cyclesaverage 2 to 3 ft in thickness and that the sandscomposing each cycle display the usual turbiditevariation in grain size (grading from coarse at thebase to fine at the top). Both sands and shaleswithin Zone 2 are discontinuous and few units arecorrelative across the field.Original oil in place within the Zone 2 volume of170,000 net acre-ft is estimated to have been 91million STB. About 35 million STB of oil (38.5percent) have been produced to date. An activepartial water drive is currently maintaining reservoirpressure in most portions of the field and anultimate primary recovery efficiency of 41.5 percentof the stock tank oil originally in place should berealized. Table I lists the pertinent reservoirproperties.About 58,000 acre-ft of the crestal portion of thereservoir are currently productive (Fig. 2). Part ofthis volume has been swept by natural water influx, as evidenced by the high producing water cuts inmany of me wells. Reservoir heterogeneitiesaccount for the lack of an easily discernible commonwater level in Zone 2.
SPEJ
P. 477^
Ensemble Machine Learning Assisted Reservoir Characterization Using Field Production Data–An Offshore Field Case Study
