您好,欢迎访问三七文档
TheDigitalTwinCompressingtime-to-valuefordigitalindustrialcompaniesTheIndustrialInternetholdsgreatpromiseforindustrialcompaniesworldwide.Asorganizationsprogresstowarddigitization,theirtime-to-valuecanbecompressedbyleveragingsuccessfulpatternsusedtodaybyconsumerInternettrailblazerslikeAmazon,Apple,andGoogle.Thispaperexaminestheconceptofdigitaltwins—dynamicdigitalrepresentationsthatenablecompaniestounderstand,predict,andoptimizetheperformanceoftheirmachinesandtheirbusiness.Digitaltwinsexistatthenexusofphysicalengineering,datascience,andmachinelearning,andtheirvaluetranslatesdirectlytomeasurablebusinessoutcomes—reducedassetdowntimeandmaintenancecosts,improvedplantandfactoryefficiency,reducedcycletimes,andincreasedmarketagility.ThepowerofoneConsiderthemeteoricgrowthofconsumerInternetbrandslikeAmazon,Apple,andGoogle.Amazongrewfromstartupto$89billioninannualrevenuein21years.Googlegrewfromstartupto$66billioninrevenuein17years.Appleincreasedannualrevenuefrom$8billionin2004to$183billionin2014.LeadingconsumerInternetbrandsbegantheirmarket-facingjourneysusingcoarse-graineddemographicmodelstoaddressbuyeraudiences.Overtime,astheyexpandedintonewmarkets,consumerInternetleadersalsosystematicallyenrichedbuyerdemographicdatawithpreferencedataderivedfromindividualbuyerbehaviors—productsfrequentlypurchasedincombination,responsestopromotionaloffers,viewsofcomparableproducts—tonameafew.Bycollectingandanalyzingdetailedbehavioraldata,consumerInternetbrandsevolvedfromcoarse-graineddemographicviews—whichmodeledlargepopulationsofconsumers—toamodelspecifictoeachconsumer.Furthersupplementedbyadvanceddatasciencethatpinpointsgroupaffinitiesandbehavioralcausalities,consumerInternetbrandsnowpreciselytargetbuyersbyaddressingeachasa“marketofone.”“Sofar,attentionhasbeenfocusedontheconsumerInternet,whereconnectedproductshavealreadytransformedthewaywelive:fromshoppingtomovingaroundtown,frommanagingourhomesystemstoorganizingaholiday.Butnowthespotlightisshiftingtoindustry,becausethisiswherethebiggesttransformationistakingplace—andwherethegreatestvaluewillbecreatedoverthecomingdecade.”Dr.MarcoAnnunziata,ChiefEconomist,GEDigitaltwinsrequiremassivecomputingscaleBycombiningdataandadvanceddatascience,consumerInternetleadersbuildadigitaltwin—adigitalmodelrepresentingdeepprofilecharacteristics—ofeverytargetconsumer.Andeverytimeaconsumerreadsaproductreview,viewsacomplementaryproductoffer,orpurchasesagift,thatbuyer’sdigitaltwingainsadditionalresolution.Thus,thedigitaltwinofeachtargetbuyeriscontinuouslyimprovedoveritslifetime.Tobuildandenrichthedigitalprofilesofhundredsofmillionsofindividuals,consumerInternetleadershavecreatedmassivelyscalablecloudinfrastructuresconsistingofmillionsofphysicalservers.Theyhardentheirinfrastructuresbyremovingallunnecessarysoftware,implementingadvancedsecurityprotocols,andevendeployinghardwarespecificallytunedtotheircloudtechnologystacks.Theskillandfinancialresourcesneededtooperatethesedistributed,massivelyscalablecomputeandstorageinfrastructuresarebeyondthereachofallbutasmallnumberofcompanies.DigitaltwinsfortheIndustrialInternetWhatcanIndustrialInternetcompanieslearnfromconsumerInternettrailblazers?Theycanlearntorequiresecure,hardenedplatforms;tofocusondataandintelligence;andtoapplydigitaltwinpatternstoobtainhigh-fidelitydigitalrepresentationoftheirsystems.Consideracombined-cyclepowerplantthatusesanaturalgas–poweredturbineasitsprimaryelectricitygenerator,andasteamturbinethatusesthebyproductofgas-firedproduction(i.e.,steam)asitssecondarygenerator.Turbinescomprisethousandsofcomponents,eachofwhichissubjecttofailurebasedonmetallurgy,load,environment,andotherfactors.Itisimpossibletopinpointthepropertimetotakeaturbineofflineformaintenanceusingtraditionalmeantimebetweenfailure(MTBF)estimates.Soplantoperatorshavenochoicebuttoscheduleexpensive—oftenunnecessary—maintenancecyclesbasedonhistoricaloperatingexperienceandthe“tribalknowledge”ofafewexpertemployees.Evenwithscheduledmaintenance,operatorshavenoabilitytopredictwhenturbineassemblieswillfail,exposingtheentireplanttotheriskofunscheduledoutages.Applydigitaltwinpatternstothisproblem,however,andtheoutcomeschangedramatically.Sensorsinstalledoncriticalturbineassembliestransmitreal-timedatatoasecurecomputingresource,buildingadigitaltwinofeachcomponentandassembly.Thedigitalmodelsareenrichedwithsituationaldatasuchassystemload,ambienttemperature,andairquality.Bybuildingadigitaltwin“modelofone”foreverycriticalturbineassembly,andcontinuouslyanalyzingeachmodelusingadvancedstatisticaltools,plantoperatorscanbringturbinesdownformaintenancepredictively,eliminatingthecostsofunnecessarydowntimeandmitigatingtherisksofunplannedoutages.Physicalscience+datascience+learningsystem=velocityUnliketheconsumerInternet,whichreliesonhumanbehaviorstobuilddigitaltwinmodels,industrialcompanieshaveamasseddecadesofphysicalscienceabouttheirproducts,equipmentassets,manufacturingprocesses,andoperatorcontrolsystems.Thisscienceformstheready-madescaffoldingofadigitaltwin,whichisthenbroughttolifewithdatacollectedbysensorson
本文标题:20180722智慧水务资料包04智慧水厂TheDigitalTwinCompressin
链接地址:https://www.777doc.com/doc-6529238 .html