A Hybrid Approach To Improve Reserve Estimates in Waterdrive Gas Reservoirs

摘要

The conventional reserve estimate methods for waterdrivernreservoirs are well established. Typically, the techniquesrnproposed by Havlena and Odeh and McEwen are used. Inrngeneral, the reserves estimated from the Havlena-Odeh andrnMcEwen procedures are not unique since the recordedrnhistorical production data is simultaneously solved for aquiferrnparameters and reserves. Normally, multiple combinations ofrnreserves and aquifer parameters match the field data equallyrnwell. Hence, a range of reserves rather than a unique value isrnfound. The upper and lower limits of the reserve range mayrnvary several folds.rnTo pick the best solution out of the multiple realizations,rnstatistical parameters are calculated for all the matches. Thernsolution with some minimum statistical indicator is consideredrnto be the best solution. Among the statistical indicatorsrnsuggested are the standard absolute relative error, normalizedrnsums of squares of differences, standard deviation, normalizedrnstandard deviation, curvature, and coefficient of variation.rnThe reserve estimate based on some minimum statisticalrnindicator may not be reliable. The best reserve estimaterncorresponding to the minimum standard absolute relative errorrnmay differ from those corresponding to the minimumrnnormalized standard deviation or the minimum coefficient ofrnvariation.rnIn this paper, a hybrid approach making use of availablernmethods is employed to narrow the reserve range estimates inrnwaterdrive gas reservoirs.