SmartSignal-Shield-2-6-Fundamentals

SmartSignalShield2.6FundamentalsSmartSignalShieldFundamentalsJune2011IdentifyingProblemsHowdoyoufindaproblem?SmartSignaltellsyouwhereyoushouldlook.SmartSignalShieldFundamentalsJune2011TopicsHowtheprocessworksTechnologyDiagnosticsShieldmodelinglifecycleApplicationsandinfrastructureRolesandprocessesSmartSignalShieldFundamentalsJune2011MonitoringstrategyIncreaseRevenueAvailabilityReliabilityAvoidSurprisesReduceCostSmartSignalShieldFundamentalsJune2011BenefitsofusingSmartSignal•Preventissuesfrombecomingmoreserious•Improveequipmentavailabilityandreliability•ImproveefficiencyEquipment•Proactiveinsteadofreactive•Optimizemaintenanceactivities•SchedulerepairsbeforeseriousdamageoccursProcessesSmartSignalShieldFundamentalsJune2011UserexperienceEaseofuseEaseofmaintenanceFlexibilityAdaptabilityContinuousimprovementFastandeasyModelset-upMonitoringdataeasilyinterpretedAuto-adaptationApplychangesquicklyAccommodatesuniquedesignsTolerantofmultiplesensorlossesBaddatadoesnotdisruptmodelAllowsfornewinstrumentationFitsintoexistingworkprocessesNewsoftwarefeaturesImproveddiagnosticscapabilitiesSmartSignalShieldFundamentalsJune2011HOWTHEPROCESSWORKSSmartSignalShieldFundamentalsJune2011SimilarityBasedModelingTheSmartSignalprocessPrioritizedDiagnosticInformationSupplementaryDiagnosticsEquipmentSensorsPersonalizedEmpiricalModelsFaultpatterndiagnosticsSmartSignalShieldFundamentalsJune2011EarlywarningMultiplesensor,dynamicempiricalanalysisinrealtimeAxialpositionGaspressureΔSpeedBearingTempAlarm/TripAxialpositionSingle-sensorequipmentprotection:traditionalthresholds&rulesEarlystagesofdamageNormaloperationSmartSignalShieldFundamentalsJune2011EarlydetectionbasedonoperationalhistoryDeviationsfromHistoricPatternsWhereyourdataIs(currentvalue)Whereyourdatashouldbe(Model)SmartSignalShieldFundamentalsJune2011TECHNOLOGYINSIGHTSSmartSignalShieldFundamentalsJune2011ModelingtechniquesRegressionModelingEmpiricalModeling1stPrinciplesModelingNon-parametricParametricSBMNonlinearregressionARMAKalmanFilterSupportvectormachinesKernelregressionPhysics-basedDesign-basedPrincipleComponentAnalysisClassicalneuralnetworksSmartSignalShieldFundamentalsJune20111stprinciplesmodelingspecifiesentiresolution,irrespectiveofdataModelingaprojectilerangeairfrictionx=v02•sin(2θ)/gv0:Initialvelocityθ:Elevationangleg:AccelerationduetogravitySmartSignalShieldFundamentalsJune2011Parametricmodeling“fits”thesolutiontothedataWhatifparametricmodelingdoesn’tfit?SmartSignalShieldFundamentalsJune2011SimilarityBasedModeling(SBM)Non-parametricmodeling:Estimatefunctionisbasedpurelyondata,noequationsSmartSignalShieldFundamentalsJune2011SBMSummarized1.Builtonhistoricaldata,customizedtoaparticularpieceequipment2.Summarizesnormal,healthyequipmentbehavior3.Representsmanydifferentoperationalstatesandambientconditions4.Takesadvantagesensorrelationshipsinherentinequipmentandprocesses5.NotOEMspecific6.Predictswhatthedatashouldbereading,basedonthecurrentoperationanditshistoryforthistypeofoperationSmartSignalShieldFundamentalsJune20113-dimensionalsystemexample18z=f(x,y)zxySmartSignalShieldFundamentalsJune2011SBMestimategenerationBluedots=model“referencedata”:normal,healthyequipmentbehaviorforvariousoperationalstatesSmartSignalShieldFundamentalsJune2011SBMestimategenerationAnestimateisgeneratedforeachparameterInputvectoriscomparedtoreferencedataZEstimateX,YInputSmartSignalShieldFundamentalsJune2011SBMestimategenerationReal-timesimilaritycalculationsSmartSignalShieldFundamentalsJune2011“Real”vs.SBMSBMoutput3-dimensionalsystemSmartSignalShieldFundamentalsJune2011SimilaritymodelinginactionAswetrackthroughtheoperationalspaceoftheequipment……thesimilaritymodelbuildsestimatesfromexemplarsaccordingtoweightedcontributions.SmartSignalShieldFundamentalsJune2011GASTURBINEEXAMPLESmartSignalShieldFundamentalsJune2011GasFiredTurbine–ActualValuesSmartSignalShieldFundamentalsJune2011GasFiredTurbine–Estimates&PerformanceDeviationsSmartSignalShieldFundamentalsJune2011GasFiredTurbine–Residuals&DiagnosisIncident:CompressorFailureinprocessSmartSignalShieldFundamentalsJune2011HOWDIAGNOSTICSWORKSSmartSignalShieldFundamentalsJune2011DiagnosticIncidentsAnotificationofachangeinequipmentbehaviorthatconformstoafaultpattern.EquipmentisOperatingSignificantDeviationPersistenceMatchesaFaultPatternSmartSignalShieldFundamentalsJune2011EquipmentisoperatingEquipmentmustbeinoperationbeforebeingevaluatedforissuesEquipmentisinoperationEquipmentisnotinoperationOperatingModeThresholdSmartSignalShieldFundamentalsJune2011SignificantdeviationResidualdeviationsindicateun-modeledbehaviorResidual(Actual-Estimate)PositiveandNegativeResidualThresholdsSmartSignalShieldFundamentalsJune2011SignificantdeviationDeviationintheactualvalueActualSmartSignalShieldFundamentalsJune2011PersistenceIndicatesacontinuationofthedeviationLongPersistenceShortPersistenceIncidentsMarkersSmartSignalShieldFundamentalsJune2011FaultpatternsAsignaturesetofbehaviorspecifictoanequipmentfailuremode.BearingVibrationBearingTemperatureLubeOilOutTemperatureBearingProblemSealPressur