




已阅读5页,还剩258页未读, 继续免费阅读
版权说明:本文档由用户提供并上传,收益归属内容提供方,若内容存在侵权,请进行举报或认领
文档简介
,QSM754SIXSIGMAAPPLICATIONSAGENDA,精品资料网,Day1Agenda,WelcomeandIntroductionsCourseStructureMeetingGuidelines/CourseAgenda/ReportOutCriteriaGroupExpectationsIntroductiontoSixSigmaApplicationsRedBeadExperimentIntroductiontoProbabilityDistributionsCommonProbabilityDistributionsandTheirUsesCorrelationAnalysis,精品资料网,Day2Agenda,TeamReportOutsonDay1MaterialCentralLimitTheoremProcessCapabilityMulti-VariAnalysisSampleSizeConsiderations,精品资料网,Day3Agenda,TeamReportOutsonDay2MaterialConfidenceIntervalsControlChartsHypothesisTestingANOVA(AnalysisofVariation)ContingencyTables,精品资料网,Day4Agenda,TeamReportOutsonPracticumApplicationDesignofExperimentsWrapUp-PositivesandDeltas,精品资料网,ClassGuidelines,Qthemean,themedianandthemode.,精品资料网,WHATISTHEMEAN?,精品资料网,WHATISTHEMEDIAN?,精品资料网,WHATISTHEMODE?,精品资料网,MEASURESOFCENTRALTENDENCY,SUMMARY,精品资料网,SOWHATSTHEREALDIFFERENCE?,精品资料网,SOWHATSTHEBOTTOMLINE?,精品资料网,COINTOSSPOPULATIONDISPERSION,精品资料网,WHATISTHERANGE?,精品资料网,WHATISTHEVARIANCE/STANDARDDEVIATION?,精品资料网,MEASURESOFDISPERSION,精品资料网,SAMPLEMEANANDVARIANCEEXAMPLE,精品资料网,SOWHATSTHEREALDIFFERENCE?,精品资料网,SOWHATSTHEBOTTOMLINE?,精品资料网,SOWHATISTHISSHIFTDF=(r-1)x(c-1).Inourcase,wehave3columns(c)and2rows(r)soourDF=(2-1)x(3-1)=1x2=2.Thesecondpieceofdataistherisk.Sincewearelookingfor.95(95%)confidence(andarisk=1-confidence)weknowtheariskwillbe.05.Inthec2table,wefindthatthecriticalvaluefora=.05and2DFtobe5.99.Therefore,ourc2CRIT=5.99Ourcalculatedc2valueisthesumoftheindividualcellc2values.Forourexamplethisis.04+.18+.27+.25+.81+1.20=2.75.Therefore,ourc2CALC=2.75.Wenowhaveallthepiecestoperformourtest.OurHo:isc2CALCc2CRIT.Isthistrue?Ourdatashows2.755.99,thereforewefailtorejectthenullhypothesisthatthereisnosignificantdifferencebetweenthevendorperformanceinthisarea.,精品资料网,ContingencyTableExercise,Wehaveapartwhichisexperiencinghighscrap.Yourteamthinksthatsinceitismanufacturedover3shiftsandon3differentmachines,thatthescrapcouldbecaused(Y=f(x)byanoffshiftworkmanshipissueormachinecapability.Verifywith95%confidencewhethereitherofthesehypothesisissupportedbythedata.,Constructacontingencytableofthedataandinterprettheresultsforeachdataset.,精品资料网,LearningObjectives,Understandhowtouseacontingencytabletosupportanimprovementproject.Understandtheenablingconditionsthatdeterminewhentouseacontingencytable.Understandhowtoconstructacontingencytable.Understandhowtointerpretacontingencytable.,精品资料网,DESIGNOFEXPERIMENTS(DOE)FUNDAMENTALS,精品资料网,LearningObjectives,Haveabroadunderstandingoftherolethatdesignofexperiments(DOE)playsinthesuccessfulcompletionofanimprovementproject.Understandhowtoconstructadesignofexperiments.Understandhowtoanalyzeadesignofexperiments.Understandhowtointerprettheresultsofadesignofexperiments.,精品资料网,WhydoweCare?,DesignofExperimentsisparticularlyusefulto:evaluateinteractionsbetween2ormoreKPIVsandtheirimpactononeormoreKPOVs.optimizevaluesforKPIVstodeterminetheoptimumoutputfromaprocess.,精品资料网,IMPROVEMENTROADMAPUsesofDesignofExperiments,精品资料网,KEYSTOSUCCESS,精品资料网,SoWhatIsaDesignofExperiment?,whereamathematicalreasoningcanbehad,itsasgreatafollytomakeuseofanyother,astogropeforathinginthedark,whenyouhaveacandlestandingbyyou.Arbuthnot,AdesignofexperimentintroducespurposefulchangesinKPIVs,sothatwecanmethodicallyobservethecorrespondingresponseintheassociatedKPOVs.,精品资料网,DesignofExperiments,FullFactorial,VariablesInput,Controllable(KPIV)Input,Uncontrollable(Noise)Output,Controllable(KPOV),HowdoyouknowhowmuchasuspectedKPIVactuallyinfluencesaKPOV?Youtestit!,精品资料网,DesignofExperiments,Terminology,Mathematicalobjectsaresometimesaspeculiarasthemostexoticbeastorbird,andthetimespentinexaminingthemmaybewellemployed.H.Steinhaus,MainEffects-Factors(KPIV)whichdirectlyimpactoutputInteractions-Multiplefactorswhichtogetherhavemoreimpactonprocessoutputthananyfactorindividually.Factors-IndividualKeyProcessInputVariables(KPIV)Levels-MultipleconditionswhichafactorissetatforexperimentalpurposesAliasing-Degreetowhichanoutputcannotbeclearlyassociatedwithaninputconditionduetotestdesign.Resolution-Degreeofaliasinginanexperimentaldesign,精品资料网,DOEChoices,Aconfusingarray.,FullFactorialTaguchiL16HalfFraction2leveldesigns3leveldesignsscreeningdesignsResponsesurfacedesignsetc.,Forthepurposesofthistrainingwewillteachonlyfullfactorial(2k)designs.Thiswillenableyoutogetabasicunderstandingofapplicationandusethetool.Inaddition,thevastmajorityofproblemscommonlyencounteredinimprovementprojectscanbeaddressedwiththisdesign.Ifyouhaveanyquestiononwhetherthedesignisadequate,consultastatisticalexpert.,Mumble,Mumble,blackbelt,Mumble,statisticsstuff.,精品资料网,TheYatesAlgorithmDeterminingthenumberofTreatments,Oneaspectwhichiscriticaltothedesignisthattheybe“balanced”.AbalanceddesignhasanequalnumberoflevelsrepresentedforeachKPIV.Wecanconfirmthisinthedesignontherightbyaddingupthenumberof+and-marksineachcolumn.Weseethatineachcase,theyequal4+and4-values,thereforethedesignisbalanced.,Yatesalgorithmisaquickandeasyway(honest,trustme)toensurethatwegetabalanceddesignwheneverwearebuildingafullfactorialDOE.Noticethatthenumberoftreatments(uniquetestmixesofKPIVs)isequalto23or8.Noticethatinthe“Afactor”column,wehave4+inarowandthen4-inarow.Thisisequaltoagroupof22or4.Alsonoticethatthegroupinginthenextcolumnis21or2+valuesand2-valuesrepeateduntilall8treatmentsareaccountedfor.Repeatthispatternfortheremainingfactors.,精品资料网,TheYatesAlgorithmSettinguptheAlgorithmforInteractions,Nowwecanaddthecolumnsthatreflecttheinteractions.RememberthattheinteractionsarethemainreasonweuseaDOEoverasimplehypothesistest.TheDOEisthebesttooltostudy“mix”typesofproblems.,Youcanseefromtheexampleabovewehaveaddedadditionalcolumnsforeachofthewaysthatwecan“mix”the3factorswhichareunderstudy.Theseareourinteractions.Thesignthatgoesintothevarioustreatmentboxesfortheseinteractionsisthealgebraicproductofthemaineffectstreatments.Forexample,treatment7forinteractionABis(-x-=+),soweputaplusinthebox.So,inthesecalculations,thefollowingapply:minus(-)timesminus(-)=plus(+)plus(+)timesplus(+)=plus(+)minus(-)timesplus(+)=minus(-)plus(+)timesminus(-)=minus(-),精品资料网,YatesAlgorithmExercise,Weworkforamajor“Donut&Coffee”chain.Wehavebeentaskedtodeterminewhatarethemostsignificantfactorsinmaking“themostdeliciouscoffeeintheworld”.Inourworkwehaveidentifiedthreefactorsweconsidertobesignificant.Thesefactorsarecoffeebrand(maxwellhousevschockfullonuts),water(springvstap)andcoffeeamount(#ofscoops).,UsetheYatesalgorithmtodesigntheexperiment.,精品资料网,Selectthefactors(KPIVs)tobeinvestigatedanddefinetheoutputtobemeasured(KPOV).Determinethe2levelsforeachfactor.Ensurethatthelevelsareaswidelyspreadapartastheprocessandcircumstanceallow.DrawupthedesignusingtheYatesalgorithm.,So,HowdoIConductaDOE?,精品资料网,Determinehowmanyreplicationsorrepetitionsyouwanttodo.Areplicationisacompletenewrunofatreatmentandarepetitionismorethanonesamplerunaspartofasingletreatmentrun.Randomizetheorderofthetreatmentsandruneach.Placethedataforeachtreatmentinacolumntotherightofyourmatrix.,So,HowdoIConductaDOE?,精品资料网,Calculatetheaverageoutputforeachtreatment.Placetheaverageforeachtreatmentafterthesign(+or-)ineachcell.,AnalysisofaDOE,精品资料网,Addupthevaluesineachcolumnandputtheresultundertheappropriatecolumn.Thisisthetotalestimatedeffectofthefactororcombinationoffactors.Dividethetotalestimatedeffectofeachcolumnby1/2thetotalnumberoftreatments.Thisistheaverageestimatedeffect.,AnalysisofaDOE,精品资料网,Theseaveragesrepresenttheaveragedifferencebetweenthefactorlevelsrepresentedbythecolumn.So,inthecaseoffactor“A”,theaveragedifferenceintheresultoutputbetweenthe+levelandthe-levelis6.75.Wecannowdeterminethefactors(orcombinationoffactors)whichhavethegreatestimpactontheoutputbylookingforthemagnitudeoftherespectiveaverages(i.e.,ignorethesign).,AnalysisofaDOE,Thismeansthattheimpactisinthefollowingorder:A(6.75)AB(5.25)BC(3.25)B(2.25)AC(1.75)ABC(1.25)C(-0.25),精品资料网,AnalysisofaDOE,精品资料网,ConfidenceIntervalforDOEresults,精品资料网,ConfidenceIntervalforDOEresults,精品资料网,IMPROVEMENTPHASEVitalFewVariablesEstablishOperatingTolerances,HowaboutanotherwayoflookingataDOE?,精品资料网,Itlookslikethelanesareingoodconditiontoday,Mark.TimhasbroughtthreedifferentbowlingballswithhimbutIdontthinkhewillneedthemalltoday.Youknowheseemstohaveimprovedhisgameeversincehestartedbowlingwiththatwristband.,HowdoIknowwhatworksforme.Laneconditions?Balltype?Wristband?,精品资料网,HowdoIsetuptheExperiment?,FactorAFactorBFactorC
温馨提示
- 1. 本站所有资源如无特殊说明,都需要本地电脑安装OFFICE2007和PDF阅读器。图纸软件为CAD,CAXA,PROE,UG,SolidWorks等.压缩文件请下载最新的WinRAR软件解压。
- 2. 本站的文档不包含任何第三方提供的附件图纸等,如果需要附件,请联系上传者。文件的所有权益归上传用户所有。
- 3. 本站RAR压缩包中若带图纸,网页内容里面会有图纸预览,若没有图纸预览就没有图纸。
- 4. 未经权益所有人同意不得将文件中的内容挪作商业或盈利用途。
- 5. 人人文库网仅提供信息存储空间,仅对用户上传内容的表现方式做保护处理,对用户上传分享的文档内容本身不做任何修改或编辑,并不能对任何下载内容负责。
- 6. 下载文件中如有侵权或不适当内容,请与我们联系,我们立即纠正。
- 7. 本站不保证下载资源的准确性、安全性和完整性, 同时也不承担用户因使用这些下载资源对自己和他人造成任何形式的伤害或损失。
最新文档
- 高中中秋节课件
- 四川中考英语真题模拟试卷含参考答案 5份
- 非银金融行业市场前景及投资研究报告:寿险公司负债成本改善
- 高一必修二《离骚》课件
- 夫妻离婚协议书:涉及借款清偿及房产分割的详细条款
- 环评技术咨询与项目可行性研究合同
- 品牌季度广告代理服务合同
- 大学实验室验收标准制定方案
- 企业人才流失原因分析和预防措施
- 提高营销团队的执行效率
- 环境保护工程质量保证措施
- 2025外研版英语八年级上册多元化教学计划
- 新团员培训第一课:青年你为什么要入团
- 公司6s管理制度
- 消防系统施工方案
- 台湾问题演讲稿
- 基本建设会计制度
- 银行员工消保知识培训
- 2025年防范电信网络诈骗知识竞赛题库及答案
- 2025年华能重庆珞璜发电有限责任公司招聘笔试参考题库含答案解析
- 《机器视觉技术及其应用》课件-模块1项目1 机器视觉技术简介
评论
0/150
提交评论