Você está na página 1de 4

Casting Design Optimization driven by Simulation

Dr.B.Ravi,ProfessorofMechanicalEngineering,I.I.T.Bombay(b.ravi@iitb.ac.in) hemethodslayoutofacastingisanimportantactivity Further,takingtheaveragedifferenceinthepriceofa intoolingdevelopment.Itinvolvescriticaldecisions saleablecastingandscrapmetalasRs.10/kg,andassuming regardingpartorientationinmold,partingline,cores, averagerejectionsinafoundryas5%,theeconomicloss cavitylayout,feeders,feedaidsandgatingsystem.An causedbydefectivecastingsworksouttoRs.500pertonne ofproduction(inrealitythiscanbemuchhigher,with improperlayoutleadstoeitherpoorqualityorlowyield, 1 affectingmanufacturingcostsandproductivity . transport,warranty,andfailuresduringproductlife). Castingsimulationcanovercometheaboveproblems: virtualtrialsdonotinvolvewastageofmaterial,energyand Computersimulationprovidesaclearinsight labour,anddonotholdupregularproduction.However, regardingthelocationandextentofinternaldefects, mostofthesimulationprogramsavailabletodayarenot ensuringcastingsarerightfirsttimeandeverytime.It easytouse,takeasmuchtimeasrealtrials,andtheir however,requiresa3DCADmodelofthemethod accuracyisaffectedbymaterialpropertiesandboundary layout(withmoldcavities,cores,feeders,andgating conditionsspecifiedbyusers.Thebiggestproblemisthe channels),propersettingofboundaryconditionsfor preparationof3Dmodelofthemoldcavitywithcores, eachvirtualtrial,andcorrectinterpretationofresults. feedersandgatingforeveryiteration,whichrequiresCAD AutoCASTsoftwareintegratesandautomatesthe skillsandtakesconsiderabletimeforevensimpleparts. abovetasks,andprovidesanextremelyeasytouse Thisalsopreventsearlymanufacturabilityevaluationand graphicaluserinterfacesuitableforevenfirsttime improvementbyproductdesigners,whichcanbenefit computerusers.Themoldcavities,feedersandgating severaltimesmorethantoolingandprocesschanges. systemareautomaticallyoptimized,drivenbythe criteriaandconstraintsspecifiedbyuser.Thisreduces TheAutoCASTsoftwaredevelopedbyAdvancedReasoning thetotaltimeformethodsdesignandsimulationofa Technologies,MumbaiincollaborationwithI.I.T.Bombay typicalcastingtolessthanonehour. providesasingleintegrateduserfriendlyenvironmentfor castingmethodsdesign,solidmodeling,andsimulation2. Ithandlesbothferrousandnonferrousparts,andsandas wellasmetalmolds.Release10incorporatesmulticavity Methodsdesignisusuallycarriedoutmanuallyon2D moldlayout,automaticmodelingandoptimization,anda drawingsofthecastpart.Thentoolingisfabricated,trial costingmodeltocomparevariouslayouts(Fig.1). castingsareproducedinthefoundry,andinspected.If samplecastingscontaindefects(suchasshrinkageorgas porosity),thenthemethodslayoutismodifiedandthe processisrepeated.Eachsuchiterationcantakeupseveral days,affectingregularproduction.Afterafewiterations, thefoundrymayresorttoasafemethodsdesign(implying lowyield),orcontinuewithhighrejectionrates(implying highscraporrepaircost).Thisisespeciallytrueinthecase oflargecastings,wherethecostofatrialorrepaircanbe prohibitive. Assumingatypicalfoundrydevelops50newcastingsevery year,eachcastingrequiresatleast2trials,andtheaverage costofeachtrial(toolingmodification,melting&pouring, inspection,andeffectonregularproduction)asRs.20,000, theeconomiclossworksouttobetwomillion(20lakh) Fig.1.Castingmethodsdesignandsimulationsoftware. rupeesperyearperfoundry.

Fig.2.Partpropertycomputation. Fig.3.Partthicknessdistributionwithsensor. Fig.4.Methodsdesignanditsautomaticmodeling.

Computer-aided Methods Design


Themaininputisthe3DCADmodelofanascastpart: withoutdrilledholes,andwithdraft,shrinkageand machiningallowance(Fig.2).Themodelfilecanbeobtained fromtheOEMfirm,orcreatedbyalocalCADagency. Variousdisplayoptionssuchaspan,zoom,rotate, transparencyandmeasure,areprovidedtoviewand understandthepartmodel.Thecastmetalandprocessare selectedfromadatabase.Thicknessmapisgenerated.Part manufacturability(compatibilitywiththeselectedprocess) iscomputedandpictoriallydisplayed(Fig.3). Methodsdesigninvolvescores,feedersandgatingsystem. Holesinthepartmodelareautomaticallyidentifiedfor coredesign,orpluggediftheyaredrilled.Evenintricate holescanbeidentifiedbyspecifyingtheiropenings.The printlengthiscomputedbasedoncorediameterand length(theusercanchangetheirvaluesifrequired),and theentirecoremodelisautomaticallycreated.The programsuggeststhenumberofcavitiesdependingon themoldsize(selectedfromacustomizablelibrary), consideringbothcavitycavityandcavitywallgaps.Then thepartmodelisautomaticallyduplicatedinthecorrect locationsasperthedesiredcavitylayout(Fig.4). Tofacilitatefeederlocation,aquicksolidificationanalysisis carriedoutthatidentifiesfeedingzones.Theuserselectsa suitableconnectionpointclosethehottestzone,andthe sizeofthefeederiscomputedusingmodulusprinciple (solidificationtimeoffeederslightlymorethanthatofthe feedingzone).Standardfeedershapesincludecylindrical, oval,sphericalbottom,cruciform,etc.Othershapescanbe imported.Thefeedermodelisautomaticallycreated;the usercanchangeitsdimensionsorapplyfeedaidssuchas insulatingsleevesandexothermiccovers.Chills,padding andfinscanalsobecreated.Morefeedersorfeederswith multipleneckscanbecreatedbyspecifyingtheirpositions. Thegatingchannelsarealsocreatedsemiautomatically. First,theuserindicatesgatepositionsonthepartorfeeder model.Thenthespruepositionisdecided,andconnected tothegatesthroughrunners.Runnerextensionsarealso automaticallycreated.Eitherhorizontalorverticalgating systemcanbedesignedandmodifiedwithinminutes.The programsuggestsasuitablefillingtime(whichcanbe changedbyuser),accordinglycomputesthedimensionsof thegatingchannels,andcreatestheirsolidmodel.

Fig.5.Meltjetpathandmoldfilling. Fig.6.Castingsolidificationsimulation. Fig.7.Feedmetalpaths(temperaturegradients).

Automatic Optimization
Themoldcavitylayout,feeders,andgatingmodelsare automaticallyoptimizedwithinminutesbasedonquality requirementsandotherconstraints3.Formoldcavity layout,theprimarycriterionistheweightratioofcast metaltomoldmaterial.Ahighratiosuchas1:2(cavities tooclosetoeachother)canreducetheheattransferrate andleadtoshrinkageporositydefects.Alowratiosuchas 1:8(cavitiestoofarfromeachother)impliespoor utilizationofmoldmaterialandreducedproductivity.The programtriesoutvariouscombinationsofmoldsizesand numberofcavitiestofindthecombinationthatisclosestto thedesiredvalueofmetaltomoldratio. Thegatingoptimizationisdrivenbytheidealmoldfilling time,whichdependsoncastmetal,castingweightand minimumwallthickness.Fastfillingleadstoturbulence relateddefects(molderosion,airaspirationandinclusions). Slowfillingmaycausedefectsrelatedtopremature solidification(coldshutsandmisruns).Tooptimizethe gatingdesign,moldfillingissimulatedandtotalfilltimeis computed(Fig.5).Alayerbylayerfillingalgorithmtakes intoaccounttheinstantaneousvelocitythroughthegates (consideringbackpressure),andthelocalcrosssectionof themoldcavity.Thisgivesafairlyaccurateestimationof fillingtime,whilebeingcomputationallyfast.Ifthe differencebetweentheidealandsimulatedfillingtimeis morethanaspecifiedlimit,theprogramautomatically changesthegatingdesign,createsitssolidmodel,and verifiesthefillingbysimulation. Thefeederoptimizationisdrivenbycastingquality,defined asthepercentageofcastingvolumefreefromshrinkage porosity.Theuserindicatesatargetquality.Theprogram automaticallychangesthefeederdimensions,createsits solidmodel,carriesoutsolidificationsimulation(Fig.6),and estimatesthecastingquality.Thesolidificationsimulation employstheVectorElementMethod,whichcomputes temperaturegradients(feedmetalpaths)insidethe casting,andfollowstheminreversetoidentifythelocation andextentofshrinkageporosity(Fig.7).Thishasbeen foundtobemuchfasterthanFiniteElementorVolume Method,andusuallymoreaccuratetoo.Feederdesign iterationsarecarriedoutuntilthedesiredqualityis achieved,orthenumberofiterationsexceedsasetlimit. Theusercanaccepttheresults,orcanmodifythefeeder designinteractively.

Fig.8.Costanalysisandmethodsreportgeneration. Finally,thecostofthecastingiscomputedintermsof amortizedtooling,castmetal,othermaterials(mold,core, etc.),energy,andlabour.Variouscostratesandparameters canbesetbytheuser.Thisenablescomparingdifferent castinglayoutsintermsoftoolingandmanufacturingcost. Adetailedmethodsdesignreportalongwithanimageof theentirecastingisautomaticallygenerated,whichcanbe printedorstoredforfuturereference(Fig.8). Themetaldatabasecoversallmajoralloys(aluminum, copper,castiron,ductileiron,steel,andpreciousmetals) andprocesses(sand,shell,investment,diecasting).Itcan becustomizedtoanynewmetalprocesscombination. ThesoftwarehasbeendevelopedforstandardWindowsXP computers,andperformswellonevenportablecomputers. Thegraphicalinterfaceisdesignedtominimizethelearning andoperationtime,andtheuserisgentlyguidedthrough forgottenorwrongsteps.Eventhosewithoutanyprior exposuretocomputersareabletousethesoftwareaftera singledayoftraining.Allstepsstartingfrompartmodel importingtomold,core,feederandgatingsystemdesign, simulationandoptimizationarecompletedwithinonehour fortypicalcastings. Directbenefitsincludeatleast50%reductionincasting developmenttimeandporositydefects.Otherbenefits includeyieldimprovement,fasterquotation,handlingmore complexpartsandknowledgereuseforfutureprojects. ContinuousinteractionoftheR&Dteamwithlocalindustry hasmadeitpossibletoincorporatehundredsofuseful improvementsoverthelast20years.

TodayAutoCASTisthemostwidelyusedcastingsoftwarein Indiawith50licenses(foundries,engineeringandR&D institutes,andconsultants)coveringallmajorcastmetals andprocesses.Manyothershaveusedthesoftwarefor benchmarking.Simulationconsultantsareavailableacross thecountrytoprovidelocaltechnicalsupport,ensuringa smoothtransitiontocomputeraidedmethoding.

Summary and Future


Castingsimulationcanminimizethewastageofresources requiredfortrialproduction.Inaddition,theoptimization ofqualityandyieldimplieshighervalueadditionandlower productioncost,improvingthemargins.Forwidespread application,simulationprogramsmustbefast,reliable,and easytouse.Thishasbeenachievedbyintegratingmethods design,solidmodeling,simulationandoptimizationina singlesoftwareprogram,andautomatingmanytasksthat otherwiserequirescientificknowledgeandcomputational skills.Inmanybenchmarkingexercisesandsimulation clinics(Fig.9),thesoftwarehasconsistentlyprovenits reliabilityinpredictinginternaldefects(ex.shrinkage porosity)withinminutes,oftenbyseniorengineerswhoare firsttimecomputerusers.Thesimulationcostsarea fractionofthecostsoffoundrytrials,whileprovidingbetter andfasterinsightforcastingoptimization.Anetworkof localtechnicalsupportcentresandsimulationconsultants acrossthecountryensuresthatevenSMEfoundriesin remoteareascannowtakeadvantageofthetechnology. Thegoalofcastingsrightfirsttime,everytime,inthe shortesttime,iswithinthereachofeveryfoundry.


Fig.9.CastingsimulationtrainingandclinicatMumbai.

References

1. 2. 3.

B Ravi, Metal Casting: ComputerAided Design and Analysis, PHI India,NewDelhi,20052008,ISBN8120327268. AdvancedReasoningTechnologies,AutoCASTSoftwareUserManual andCaseStudies,http://www.autocast.co.in,2009. B Ravi, Casting Simulation and Optimisation: Benefits, Bottlenecks, andBestPractices,IndianFoundryJournal,54(1),Jan2008.

Você também pode gostar