1、Subject:StratigraphicdivisionsedimentaryfaciesanalysisbasedonloggingcurvesSpecialty:TechnologyofComputerApplicationName:ShenKailang(Signature)Name:ZhangQunhui(Signature)ABSTRACTWiththematurityofloggingcomprehensiveinterp

2、retationithasbecomeatendencytomakefulluseofdifferentloggingcurves.Thispapertakesloggingcurveastheresearchobjectdiscussrealizesomekeymethodsinthestratigraphicdivisionlogfaciesanalysis.Primaryresearchachievementsinnovation

3、pointsofthispaperaredescribedasfollows:1.TakeadvantageoftheanalyticmethodbasedonobjectientedthroughanalyzingWISdocumentswhichcontainloggingcurvesparsesteloggingcurvesdataaccurately.Inviewoftheinfluencefactsofloggingdataa

4、doptfixedwindowlengthcrelationcontrastmethodfdepthcrectionofcurvethenuseSavitzkyGolaysmoothingfiltertoremovenoiseswhichisaimedateliminatingnoisescausedbynonstratigraphicfactswhichprovideascientificdatasupptfstratigraphic

5、divisionlogfaciesanalysis.2.Accdingtothesmallsamplepropertiesofceoilfieldsthesimilarityofsameregionalgeologicalacteristicsresearchloglayeringmethodbasedonsupptvectmachine(SVM).Onaccountofinconvenienceincalculationstratig

6、raphicincontinuityoftraditionalmulticlassificationalgithmsimproveaonetoonemulticlassificationmethodsothatenhancelearningpredictionspeedofthesample.Aseverystratumhassomecontinuitypointsweapplyfilterfunctionstepfunctiontor

7、ecognizestratigraphicinterfacefurtherpromotingtheaccuracyofstratigraphicdivision.3.Combinewiththesimilarityoflogfacieslocatedatthesameregionsamelayerstudysupptvectmachine(SVM)algithmbasedonloggingcurvesthenuseitflogfacie

8、sanalysis.Sinceloggingcurvesreflectinglogfaciesismethereiscrelationbetweencurvedataweemployprincipalcomponenttoanalyzeextractprincipalcomponentoflogfacieswhichnotonlyreducethesampledimensionbutalsoimprovecomputationaleff