A statistical methodology for deriving reservoir properties from seismic data.pdf

《A statistical methodology for deriving reservoir properties from seismic data.pdf》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、GEOPHYSICS,VOL.60,NO.5(SEPTEMBER-OCTOBER1995);P.1437-1450,13FIGS.,5TABLES.AstatisticalmethodologyforderivingreservoirpropertiesfromseismicdataFrederiqueFournierandJean-FrancoisDerainABSTRACTmodelcomesfromamultivariateregressionbetweenthecanonicalseismicvariablesandthereservoir

2、prop-Theuseofseismicdatatobetterconstraintheerties,whenevertheyarepredictable.reservoirmodelbetweenwellshasbecomeanimpor-Inthecasestudy,wepredictedestimatesandassoci-tantgoalforseismicinterpretation.ateduncertaintiesonthelithofaciesthicknessescumu-Weproposeamethodologyforderiv

3、ingsoftgeo-latedoverthereservoirintervalfromtheseismicinfor-logicinformationfromseismicdataanddiscussitsmation.WecarriedoutaseismicfaciesidentificationandapplicationthroughacasestudyinoffshoreCongo.comparedthegeologicalpredictionresultsinthecasesofThemethodologycombinesseismic

4、faciesanalysisacalibrationonthewholedatasetandacalibrationdoneandstatisticalcalibrationtechniquesappliedtoseis-independentlyonthetraces(andwells)relatedtoeachmicattributescharacterizingthetracesatthereservoirlevel.Webuiltstatisticalrelationshipsbetweenseis-seismicfacies.Thelat

5、erapproachproducesasignificantmicattributesandreservoirpropertiesfromacalibra-improvementinthegeologicalestimationfromtheseis-tionpopulationconsistingofwellsandtheiradjacentmicinformation,mainlybecausethelargescalegeolog-traces.Thecorrelationstudiesarebasedonthecanon-icalvaria

6、tions(andassociatedseismicones)overtheicalcorrelationanalysistechnique,whilethestatisticalfieldcanbeaccountedfor.INTRODUCTIONgeologicalpropertiesofareservoirhasbeenanalyzedwithsyntheticandrealdata.AcomprehensivereviewofseismicSeismicdataareoftenusedasasourceofextrageologicalat

7、tributesisgiveninJusticeetal.(1985).informationbetweenwellssoabetterreservoirdescriptionBecauseofthehugeamountofdataandthenecessitytocanbeconstructed.Thisisparticularlytruesincethedevel-takeintoaccountseveralattributessimultaneously,ap-opmentof3-Dseismicdata,whichprovideanexce

8、llentarea1proachesusingstatisticalpatternrecognitiontechnique