微流控芯片注塑成型質(zhì)量缺陷的成因及預(yù)測(cè).pdf

1、大連理工大學(xué)碩士學(xué)位論文微流控芯片注塑成型質(zhì)量缺陷的成因及預(yù)測(cè)姓名：程康申請(qǐng)學(xué)位級(jí)別：碩士專業(yè)：機(jī)械制造及其自動(dòng)化指導(dǎo)教師：宋滿倉20091201微流控芯片注塑成型質(zhì)量缺陷的成因及預(yù)測(cè)CauseandPredictionoftheDefectsintheProcessofInjectionM01dingforMicro—fluidicChipAbstractwiththematu打tyofmicrom01dingtechnology，m

2、icrop】asticpa九sandthepartswithmicrostrucnlrearewidelyusedin珈anyfieldsAsanewbranchofinjectionmoldingtechn0109y，microinjectionm01dingisonthestageofrapiddeVe】opmentSch01arsathomeandabroadhavealreadystudieda10tonitandhadsome

3、achjeVementsHoweVer，themoldingmechanismanddefecttheoriesstillneedtobe￡JrtherinvestigatedinordertobuildprocessspecjljcationformicroinjectionmoldingThemicro—nuidicchipwhichisthetypicalplasticwithmicro—stmcmreinthe6e】dsofmi

4、crojnjectionm01ding，waschosenasthesubjectTheinjectionmoldingprocessofmicro—nuidicchipcontainsthecouplingofmacroscopicalandmicroinjectionm01dingAsaresull，therearebothconventjona】defectsandmicrodefectsinmicrofluidicchipThe

5、poss佑1e。eorⅣenti硼a1defectswere6rstlystudjedan扣thfmechanismofdefectswereanalysed，andthenco盯ectionsuggestionswereadVancedfromtheaspectsofprocessdesignThen，thedifkrencesbetweentraditiona】andmicroinjectionmoldjngwerediscusse

6、dtoexplaintheoriginsofthedefectsandtointrodl】cethestandardofqualjty—estimationformicro—flujdicchip，whichwastheopeningwidthofthecha加e1AmoldwasdesignedandmanufactllredaccordingtothestructuralcharactersofmicrofluidicchipARe

7、rtheexpenments，thechipswascutatgjVenpositioninofdertoobserveandphotographbymicroscopeThecontourofchannelwaspickedupthroughsoRwareMatlabARerthat，thewidthofcharulelwasmeasuredbyusingimageprocessingt001sofsoRwareMatlabThere

8、lationshipbetweenprocessingparametersandchipqualitywasanalyzedbybothsjmu】ationandBPNeuralNetworkThrou曲soRwareM01dflow，theinjectionprocesswassimulatedwiththegiVenparamelers，andthepressureatthecrosssectionofthechannelswasg

9、otinIhesimulationresultsByanalyzing，theconclusionwasthatthebiggerthepressurewas，thebetterthequa】ityofthechanne】wasButthismethodwasonlythepredictoneTheBPNeura】NetworkmethodwasonthebasisofMatlabwhichbuildedlheBPNeuralNetwo

10、rkandtrainedjtwithtakingprocessparameters(mou】dtemperature，meltstemperature，injectionpressure，injectionrateandhoklingpressure)astheinputandthewidthofthechannelastheoutputuntilthefittingprecisionreachedtherequirementThepr