M5-Petrophysical-Modeling岩石物理模型(Schlumberger-petre

Module5PropertyModelingPetrelWorkflowTools3DGridConstructionStructuralGriddingStratigraphicModelingPillarGriddingWellLogUpscaleFacies&PetrophysicalModelingMakecontacts&VolumeCalculationIntrotoPetrelInterfaceWorkflowEditorPropertyModelingMakeHorizonsZones&Layering3DGridConstruction:StructuralModelingStudio3DGridConstructionStructuralFrameworkFaultModelingIntroductionSurfacesandDataeditPetrophysicalModelingObjectivesPetrophysicalmodelingtechniques–Deterministictechniques–StochastictechniquesUsecommonsettings–Setfilters–UsemanyrealizationsUsezonesettings:DefinezonesUsestochasticalgorithms–SequentialGaussiansimulation–GaussianRandomFunctionsimulation–QCresultsPropertyCalculator.AlgorithmsCoveredintheCourseStochasticmethodDeterministicmethodPixel-basedtechniqueInterpolationGaussianRandomFunctionSimulationalgorithmMovingAveragealgorithmItisfasterthanSequentialGaussianSimulation(SGS)andgivesbettervariogramreproduction.Ithasafastcollocatedco-simulationoptionusinganinteractivecorrelation-coefficientasslidebar.Basedontheinput,itgivesanaveragevalueandcalculatestheweigthsaccordingtothedistancefromwells.CommonandZoneSettingsTwomainmodelingsettingsbuttonsareavailable(CommonandZonesettings).TheseworkinexactlythesamewayasinFaciesmodeling.DeterministicModelingusingtheMovingAverageAlgorithmNote:Iftherangeoftheinputdataislarge,itwillcreatebulleyes.Movingaverageisafastdeterministicmodelingalgorithm.Itcalculatesanaveragebasedoninputdata(upscaledcells)andweightsitaccordingtodistanceawayfrominputdata.1.Setanupscaledproperty:(U)assuffix.2.ChoosetheZonetomodelandunlockit.3.SetMovingaverageasthemethodforthezone.4.Variogramtab(2methods):•SpecifyOrientation,Major/minorratio,andVerticalrange.•ClickGetavariogramfromDataAnalysis.5.SpecifytheOutputdatarange(max/min).GaussianRandomFunctionSimulation(GRFS)GRFSisthein-housestochasticmodelingalgorithmforPetrophysicalmodeling.ItisaverygoodalternativetoSequentialGaussianSimulation(developedbyGSLIB)GRFSadvantages:BasedontheideaofConditionalSimulation=Kriging+Unconditionalsimulation.Parallelized,soitismuchfasterthansequentialalgorithms.GivesabetterVariogramreproductionthanSGS.Hasacollocatedco-krigingoption,allowingfastupdatesoftheresultswhenchangingthecorrelationcoefficient(usingtheinteractiveslidebar).StochasticModelingusingGRFSConditionedtoaFaciesModel1.Setanupscaledproperty:(U)assuffix.2.Choosethezonetomodelandunlockit.5.Variogramtab(2methods):•SpecifyRange,Nugget,andTypemanually•ClickGetavariogramfromDataAnalysis.GaussianRandomFunctionSimulationisamuchusedstochasticmodelingalgorithm.Ithonorswelldata,inputdistributions,variograms,andtrends.Itcreatesmultipleequallyprobableoutputs.4.SetGaussianrandomfunctionsimulationasthemethodforthezone.3.ClicktheFaciesbuttonandchoosethefaciesmodelifconditioningisnecessary.StochasticModelingusingGRFS:Distribution7.SettheMinimumandMaximumvalueoftheOutputdatarange(absoluteorrelative%).8.DefinetheshapeoftheDistributionusingFromupscaledlogswhenavailable.ABivariatedistributionusesaSecondarypropertyonwhichtobasethedistribution.UsedifupscaledlogsUsedifno/fewupscaledlogsFromdistributionfunction6.ManuallyspecifytheSeednumber.ThesesettingsareavailableontheDistributiontab.StochasticModelingusingGRFS:UsingaSecondaryVariable1.Definethesecondaryvariable.A.Property:Continuousseismicattributevolumethatcorrelateswiththepropertytobemodeled.B.HorizontalsurfaceC.Verticalfunction2.Choosethemethod.A.Localvaryingmean:Normallytransformsthedatabeforeaddingtheresidual.B.Collocatedco-kriging:Usesacorrelationcoefficienttocalculatethecontributionofthesecondaryvariable.Steerthesimulationusingthespatialdistributionofasecondaryvariable,withacorrelationcoefficient(Collocatedco-kriging).QualityCheckingtheModeledPropertyFilterthehistogram.1.UseZonefilter.2.FilteronotherpropertyvaluesbyclickingtheFilterbuttonandgoingtopropertyfilterinthesettingsforthePropertiesfolder.QCresultsinahistogram.GotothesettingsforthePropertyandclicktheHistogramtab.Verifythatthehistogramfollowsthedistributionfrom:RawlogsUpscaledcells3DgridPropertyCalculator:OverviewThePropertycalculatorcanbeusedtomodifyexistingpropertiesortomakenewproperties.PropertyCalculator:UsingtheCalculator1.Right-clickthePropertiesfolderandselectCalculator.2.Enteraformulausingthekeypadsoftheinterface.Useexistingpropertiesfromthedatalist.4.Chooseapropertytemplate.3.(Optional)Browseforafiletobeused.5.RunyourfileorclickENTERtorunanyformula.Note:ReuseanyformulabyselectingitintheHistoryarchive.EXERCISEPetrophysicalModelingExtra:SequentialGaussianSimulation(SGS):ResultsSGSisakrig-basedstochasticmethod:Needsavariogram.Honorsthedistributionoftheinputdata(upscaledlogs).PetrelwillautomaticallydoaNormalscoretransformationbeforethesimulationandwillback-transformthedata.Theoutputisdistributedinablurrymanner.Extra:PetrophysicalModelingMethods:Overview(1)DeterministicEstimationInterpolationKrigingInterpolationKrigingKrigingbyGSLIBClosestFunctionalMovingaverageHonorswelldata,inputdistributions,variograms,andtrends.Itcanworkinrealcoordinatesandisfas