基于知识的智能加工理论及其应用基础研究135

南京航空航天大学博士学位论文基于知识的智能加工理论及其应用基础研究姓名：王宗荣申请学位级别：博士专业：机械制造及其自动化指导教师：王珉;左敦稳20070301i12BPCnP35iiiiiAbstractMetalcuttingisthemostcommonprocessmannerinmechanicalmanufacturing.Inmetalmachiningprocess,theskilledmachiningtechnicians,throughtowatchingofthesurfacebeingmachined,hearingvoiceofmachiningprocess,thesmellingodorincuttingprocessandtoolwear,candeterminetheworkingstateofmachine,sothattheyadoptappropriatecontingencymeasurestoensurethedimensionaltolerancemachinedparts,surfaceroughness,shapeandpositiontolerancesandothertechnicalrequirementsfortheuseofaproducttomeetperformance.Alongwitheconomicglobalization,thecompetitionintensifiesandmachinery,themanufacturingenterprisesurgentlyneedtoimprovethequalityofproductsandreduceprocessingcosts,sothattheyneedlargenumberofskilledmachiningtechnicians.Butinfact,theskilledmachiningtechniciansareshortofextremely.Enterprisesareurgentlyrequiredtoreplaceskilledmachiningtechnicianscompletelyorpartiallybycomputer,sothatintelligentandautomaticlevelinprocessingisimproved.Inordertoresolvetheaboveproblems,inthispaper,thefirsttimeartificialintelligencetechnologyisintroducedintothefieldsofmetalcutting,thearchitectureofintelligentmachiningsystemisstudied,theobservablityandcontrollabilityoftheintelligentmachiningsystemandstabilityofthesystemareexploredforresolvingpracticalissues.Itismadeknowninengineeringpracticesandcomputersimulationsthatthemanufacturingautomationlevelsareincreased,machiningcostsarereducedandthequalityofpartsimprovedbythestudy.1Afteranalysisoftheintelligentstructureaboutthehumanbody,theintelligentstructuralmodelofskilledmachiningtechniciansisputforward.Theanywaysareanalyzedtohandlecomplex,uncertain,processingsysteminstabilityandthevarioussituationandbreakdownbyskilledmachiningtechnicians.Itisresearchedthathowtousecomputerstosimulatetheskilledmachiningtechnicianstoreplacethebraininpart.2Throughthesummaryandconcludeofconcept,composeandtechnologyofintelligentmachiningsystem,theblackboardmodelbasedonintelligentmulti-levelandmulti-processingsystem’sarchitecture,modelofhierarchicalcontrolsystemandtheBPneuralnetwork,generalizedCnPintelligentintegratedsystemforprocessingandmulti-statespacehierarchicalstructureofitsthreeintelligentstructureofthecontrolsystemarchitecturearepost.Eachsystem’sarchitectureisdescribedivqualitativelyandquantitatively.3Thestatespacetheoryinmoderncontroltheoryisappliedtointelligentmachiningsystem.Theobservablity,controllabilityandstabilityofintelligentmachiningsystemtheory,anddesignoftheintelligentmachiningsystemarestudiedandsoon.Thedesignisfocusedonoperationalcontrollevel,thecalibrationlevel,andthetasksadaptlevel.4Thetheoryandstructureoftheintelligencemachiningsystemareusedtosystemsbasedoncomputervision.Theautomaticidentificationofcrackandburnsimageongrindingsurfaceoftitaniumpartsissolved,andthattechnologyareappliedtomanufacturingofAeronauticandAstronauticstructuralpartssothatthedifficultproblemsinproductionareresolved.5Imitatingtheskilledmachiningtechniciansfromintelligencemachiningsystembasedoncomputervision,thevoicesignalsfrommechanicalmachiningshopareletinputsignalstostudycuttingtoolwear.Forexampleturning,theturningvoiceismademonitoringsignalsofcuttingtoolwear.First,usingthewavelettransfermethod,thecuttingsoundsignalisseparatedintodifferentfrequencybands,andthepercentagesofeachband'senergyareextractedtobethecharacteristicstomonitorthetoolcondition.Then,athree-layerBack-Propagationartificialneuralnetworkisbuiltuptorecognizethecuttingtoolwear.TheoptimizedBPneuralnetworkstructureofnetworksaredeterminedfromcontrast.Takingadvantageofthesignalsamplesimulationoftheneuralnetworktrainingandtesting,andachievedgoodresultsintheidentification.Keywords:intelligentmachiningsystem,intelligentmonitoringandsupervise,intelligentprocessingofinformation,intelligentdimension-making,intelligentcontrol,computervision,computerhearing1AdvancedManufacturingTechnologyAMTFlexibleManufacturingSystem,FMSFlexibleManufacturingCell,FMCMaterialRequirementsPlanning,MRPManufacturingResourcePlanning,MRPIIJustInTime,JITComputerAidedManufactur-ing,CAMManagementInformationSystem,MISP.K.WrightD.A.BourneManufacturingIntelligence2IntelligentManufacturingTechnology,IMTIntelligentManufacturingSystem,IMSNewYorkP.K.WrightCarnegie-MellonD.A.BourneManufacturingIntelligenceSMEA.KusiakInt.J.ofIntelligenceManufacturingCIRPProductLifecyclesandEnvironmentNSF3NISTAMRFIMSAMRFBirminghamAlcanKittsGreenFutureGenerationManufacturingSystem,FGMSH.YoshikawaIMSVersionIMSIMSIMCIMIMS4IMSIMTIDIPPIMID/IPP/IMIMC51.2HollowingoutCIMSIntelligentMachining,IMgIMTIMSIMSIMTIMSFMS6ChryssolourisIntelligentMachining7ComputerAidedProcessPlanningCAPPCADCAMCIMS89IwatakMoriwakiMoriwakiKannatey-AsibuDornfeldMurakamiGikenchip-55ALanDornfeldShiraishi10SandvikCoromantTM-BU-1001CincinnatimilacronSMARTSENSORS11BP1.312132.114152.2162.12.3172.2-[101]-18192.3[101]202.4[101]--812551-1231212.5[101]22-2324252.4262728292.530312.6323.1ComputerAided3.23334X%X%X%3512312312Bayes3637n…...21m…...21k…...21T10T2T8T3T5T6T4T7T9T138ab3.2.2.339123n403.3BPDecen