項(xiàng)目運(yùn)作實(shí)例

（六西格瑪管理）項(xiàng)目運(yùn)作實(shí)例《6Sigma項(xiàng)目運(yùn)作實(shí)例》如何定義一個(gè)項(xiàng)目？項(xiàng)目定義是由冠軍來(lái)完成的。我們簡(jiǎn)單介紹以下項(xiàng)目是如何定義的。謝謝閱讀1確定主要商業(yè)問(wèn)題：a目標(biāo)b目的c可交付使用的2對(duì)與生產(chǎn)來(lái)說(shuō)：a循環(huán)時(shí)間b質(zhì)量/缺陷水平c耗費(fèi)3項(xiàng)目的選擇a選擇項(xiàng)目的工具a1宏觀圖a2Pareto圖分析a3魚(yú)骨圖a4因果矩陣圖b項(xiàng)目的標(biāo)準(zhǔn)（評(píng)估）b1減少缺陷的70%b2第一年節(jié)省$175Kb3項(xiàng)目完成周期為4個(gè)月b4最少的資金總額b5黑帶的第一個(gè)項(xiàng)目必須滿足培訓(xùn)目標(biāo)《6Sigma項(xiàng)目運(yùn)作實(shí)例》->《定義階段》->我們?cè)诙x階段做什么精品文檔放心下載------------------------------------------------------------------------感謝閱讀---------------------------感謝閱讀我們?cè)诙x階段需要做什么？1，完成項(xiàng)目陳述。2，完成項(xiàng)目預(yù)測(cè)節(jié)省金額。3，完成問(wèn)題陳述：3.1問(wèn)題是什么？3.2在哪里和什么時(shí)間發(fā)現(xiàn)的？3.3問(wèn)題將涉及哪些工序？3.4誰(shuí)將受到影響？3.5問(wèn)題的嚴(yán)重程度是什么？3.6你是如何得知這些的？4，繪制宏觀圖。5，描述項(xiàng)目的主線。6，完成目標(biāo)陳述。7，組成項(xiàng)目小組，列出小組成員。8，完成財(cái)務(wù)評(píng)估?！?Sigma項(xiàng)目運(yùn)作實(shí)例》->《定義階段》->如何進(jìn)行項(xiàng)目問(wèn)題陳述精品文檔放心下載------------------------------------------------------------------------精品文檔放心下載---------------------------感謝閱讀如何進(jìn)行問(wèn)題陳述？分六個(gè)方面進(jìn)行問(wèn)題陳述：1問(wèn)題是什么？2在哪里和什么時(shí)間發(fā)現(xiàn)的？3問(wèn)題將涉及哪些工序？4誰(shuí)將受到影響？5問(wèn)題的嚴(yán)重程度是什么？6你是如何得知這些的？《6Sigma項(xiàng)目運(yùn)作實(shí)例》->《定義階段》->如何繪制宏觀圖謝謝閱讀------------------------------------------------------------------------精品文檔放心下載---------------------------感謝閱讀如何繪制宏觀圖?繪制宏觀圖的順序：供應(yīng)商->輸入->工序->輸出->客戶感謝閱讀《6Sigma項(xiàng)目運(yùn)作實(shí)例》->《定義階段》->項(xiàng)目的目標(biāo)陳述要點(diǎn)感謝閱讀------------------------------------------------------------------------感謝閱讀---------------------------謝謝閱讀項(xiàng)目的目標(biāo)陳述要點(diǎn):1,目標(biāo)陳述2，計(jì)算方法3，全年節(jié)省額確定TeamMembers成員:1，小組成員要包括技術(shù)人員2，包括維修人員（如果需要）3，包括操作者4，小組人員不超過(guò)5人(特殊情況除外)。《6Sigma項(xiàng)目運(yùn)作實(shí)例》->《測(cè)量階段》->如何進(jìn)行項(xiàng)目描述精品文檔放心下載------------------------------------------------------------------------精品文檔放心下載---------------------------謝謝閱讀如何進(jìn)行項(xiàng)目描述:1，目標(biāo)陳述2，Metric圖3，月節(jié)省額如何繪制工藝流程圖:召集小組:流程圖繪制是集體努力的結(jié)果小組包括：流程負(fù)責(zé)人：項(xiàng)目結(jié)果的負(fù)責(zé)人工程部門(mén)-工藝，產(chǎn)品，設(shè)計(jì)及設(shè)備生產(chǎn)部門(mén)-操作員，各班次主管，培訓(xùn)員，操作班長(zhǎng)，維修技師精品文檔放心下載流程圖所需信息腦力風(fēng)暴觀察/經(jīng)歷操作手冊(cè)工程標(biāo)準(zhǔn)，工作指示六大方面（人，機(jī)，方法，測(cè)量，材料，環(huán)境）確定工藝范圍:范圍至觀重要越窄越好！大量工藝步驟可能表明項(xiàng)目定義不佳或問(wèn)題源于幾個(gè)項(xiàng)目問(wèn)題藏于問(wèn)題中若問(wèn)題可以由粗略分析解決，管理層會(huì)去做繪制可執(zhí)行的工藝圖你能確認(rèn)缺陷來(lái)源嗎？我們能有意識(shí)地改變輸入指標(biāo)變量嗎？有意識(shí)的改變輸入指標(biāo)變量能直接影響輸出結(jié)果嗎？工藝流程圖（PFD）:6Sigma工藝流程圖的要素：所有工藝步驟包括隱形工廠數(shù)據(jù)采集點(diǎn)所有設(shè)備/工具各步驟表明增值性（VA）和非增值性（NVA）控制標(biāo)準(zhǔn)文件用標(biāo)準(zhǔn)符號(hào)繪制工藝流程：在MicrosoftOfficeTM等軟件中可找到感謝閱讀工藝流程圖-程序:繪制工藝記載的工藝步驟包括所有檢查點(diǎn)，測(cè)量指標(biāo)和傳運(yùn)步驟確認(rèn)所有數(shù)據(jù)采集點(diǎn)標(biāo)示各工序標(biāo)準(zhǔn)控制文件各步驟標(biāo)明為增值性（VA）或非增值性（NVA）確認(rèn)各工藝步驟的X和Y標(biāo)明可能消除的NVA步驟加入并標(biāo)明“隱形工廠”工段標(biāo)明為VA或NVA，標(biāo)明可能消除的步驟標(biāo)明須指定控制文件的步驟加入DUP，RTY，COPQ，循環(huán)周期等估計(jì)值標(biāo)明須進(jìn)行量具和工藝能力研究的步驟通過(guò)直接或秘密觀察確認(rèn)準(zhǔn)確性文件記錄/確認(rèn):文件記錄的工藝流程首先繪制記錄下來(lái)的工藝加入并標(biāo)明隱形工廠步驟當(dāng)所有步驟展示出來(lái)后，流程圖就屬于實(shí)際工藝確認(rèn)流程圖的準(zhǔn)確性至關(guān)重要項(xiàng)目組必須花時(shí)間觀察工藝秘密進(jìn)行。觀察導(dǎo)致行為改變確認(rèn)實(shí)際工藝設(shè)置與記錄的設(shè)置相同跨班跨機(jī)器觀察工藝如何繪制工藝流程細(xì)圖:工藝流程細(xì)圖：6Sigma工藝流程圖要素：工藝或產(chǎn)品是輸出指標(biāo)Y和輸入指標(biāo)X標(biāo)準(zhǔn)上下限和標(biāo)準(zhǔn)控制文件所用設(shè)備/工具繪制工藝流程細(xì)圖工藝流程細(xì)圖必須依工藝流程圖而畫(huà)。更改其一應(yīng)在另一個(gè)中反映出來(lái)。謝謝閱讀應(yīng)使用最新的控制文件標(biāo)明所有隱形工廠步驟的輸入輸出指標(biāo)工藝流程細(xì)圖程序：1，從流程圖中列出工藝步驟2，加入下列內(nèi)容輸出指標(biāo)輸出指標(biāo)標(biāo)準(zhǔn)，若存在輸入指標(biāo)輸入指標(biāo)標(biāo)準(zhǔn)，若存在工藝能力或量具能力指標(biāo)所用設(shè)備3，標(biāo)明隱形工廠步驟4，標(biāo)明各步驟屬于增值性（VA）或非增值性（NVA）精品文檔放心下載5，標(biāo)明各步驟屬于可控性的（C）或噪音性的（N）6，確認(rèn)各設(shè)備的輸入指標(biāo)設(shè)置7，確認(rèn)流程圖準(zhǔn)確性8，必要時(shí)更改及更新流程標(biāo)準(zhǔn)限和工藝能力：工藝及產(chǎn)品標(biāo)準(zhǔn)加入X的工藝設(shè)置加入Y的標(biāo)準(zhǔn)限標(biāo)明未記錄的Y和可控的X測(cè)量系統(tǒng)加入量具重復(fù)性及復(fù)驗(yàn)性數(shù)據(jù)標(biāo)明須做測(cè)量系統(tǒng)分析的量具工藝能力展示RTY，DPU，CPK等的估計(jì)值標(biāo)明哪些工藝步驟數(shù)據(jù)陳舊或不完整而需做工藝能力分析更改及更新：更改記住：6Sigma的目標(biāo)之一是找出：Y=F（X）隨著對(duì)工藝的深入了解，更新工藝圖以反映新的信息更新項(xiàng)目最終成果之一是現(xiàn)有的工藝的流程圖更新工藝圖以反映任何工藝改變加入測(cè)量系統(tǒng)分析及工藝能力分析結(jié)果精簡(jiǎn)制造與5S:精簡(jiǎn)制造例似于日本的5S精簡(jiǎn)制造與5S:魚(yú)骨圖:魚(yú)骨圖一種系統(tǒng)確認(rèn)所有可能導(dǎo)致問(wèn)題(后果)產(chǎn)生的原因方法。謝謝閱讀構(gòu)造魚(yú)骨圖的方法：1.陳述問(wèn)題，并置于右邊的方框內(nèi)2.朝方框畫(huà)一水平箭頭。3.在箭頭上下寫(xiě)上傳統(tǒng)因素類型名稱*或你懷疑是的類型名稱。用感謝閱讀直線連到箭頭線上。4.在各主要的類型范圍內(nèi)，集思廣益并列出所有可能引起問(wèn)題發(fā)生的因子。謝謝閱讀5.進(jìn)一步優(yōu)化：對(duì)各種詳細(xì)列出的因子再列出其輸入變量。精品文檔放心下載*6m--man,machine,method,measurement,mothernature(environment)感謝閱讀(6M:人員，機(jī)器，測(cè)量方法，原材料，環(huán)境)定性測(cè)量系統(tǒng)研究:定性型量具R&R-術(shù)語(yǔ):檢驗(yàn)員分?jǐn)?shù)（%）-在定性型R&R檢驗(yàn)過(guò)程中，檢驗(yàn)員前后一致的比例謝謝閱讀定性數(shù)據(jù)--定性（合格/不合格）數(shù)據(jù)，可用來(lái)做記錄和分析謝謝閱讀定性型測(cè)量系統(tǒng)--把每個(gè)部件與標(biāo)準(zhǔn)進(jìn)行比較，從而決定部件是否符合標(biāo)準(zhǔn)的測(cè)量精品文檔放心下載系統(tǒng)。消費(fèi)者偏見(jiàn)--員工傾向把合格產(chǎn)品判為廢品%--在定性型R&R精品文檔放心下載也一致的比例。標(biāo)準(zhǔn)值--由一個(gè)高準(zhǔn)確度量具所測(cè)的平均值生產(chǎn)者偏差--員工傾向于把不合格（有缺陷的）產(chǎn)品判為合格感謝閱讀篩選--用檢驗(yàn)方法對(duì)產(chǎn)品進(jìn)行100%的評(píng)估篩選有效性--定性量具系統(tǒng)區(qū)別合格與不合格的能力使用定性型量具R&R的目的:工藝評(píng)估評(píng)估你的檢查標(biāo)準(zhǔn)或工作質(zhì)量標(biāo)準(zhǔn)與客戶要求的一致性確定所有班次，機(jī)器等的檢查人員是否使用相同標(biāo)準(zhǔn)來(lái)決定合格與不合格感謝閱讀量化檢查人員準(zhǔn)確重復(fù)其檢驗(yàn)結(jié)果的能力確定檢查人員與“已知標(biāo)準(zhǔn)”的一致性及傾向于消費(fèi)者偏差還是生產(chǎn)者偏差感謝閱讀工藝改進(jìn)發(fā)現(xiàn)是否需要培訓(xùn)，缺少工序或缺乏標(biāo)準(zhǔn)定性型量具R&R的方法:準(zhǔn)備從工藝中挑選30個(gè)部件，50%合格，50%次品可能的話，挑選近乎于合格和不合格樣本挑選檢查人員--受過(guò)完全培訓(xùn)的和有資格的實(shí)施要求每一個(gè)檢查人員隨機(jī)地檢查部件，決定合格與不合格并重復(fù)此檢查精品文檔放心下載評(píng)估將結(jié)果載入文件如果必要，采取適當(dāng)?shù)拇胧┱{(diào)整測(cè)量工藝重做R&R試驗(yàn)，核實(shí)調(diào)整后的有效性定性型量具R&R--結(jié)論:檢查員分?jǐn)?shù)如果大多數(shù)員工都是100%，則培訓(xùn)作用極為有限篩選有效分?jǐn)?shù)如果員工本身前后一致但是相互間不一致，則重新培訓(xùn)可幫助減少錯(cuò)誤。謝謝閱讀標(biāo)準(zhǔn)化分?jǐn)?shù)如果員工時(shí)常與標(biāo)準(zhǔn)不一致，則需要改變測(cè)量系統(tǒng)（或局部標(biāo)準(zhǔn)）感謝閱讀工藝能力分析:為何測(cè)量工藝能力?使我們根據(jù)數(shù)據(jù)分配資源?。ㄟ@可不常見(jiàn)！）缺陷率得以量化確認(rèn)可以改進(jìn)機(jī)會(huì)分析工藝能力可使組織預(yù)測(cè)其所有產(chǎn)品和服務(wù)的真實(shí)質(zhì)量水平感謝閱讀確認(rèn)工藝發(fā)生問(wèn)題的本質(zhì)-居中程度或分散度工藝能力研究連續(xù)數(shù)據(jù)離散數(shù)據(jù)1.確認(rèn)標(biāo)準(zhǔn)限1.確認(rèn)標(biāo)準(zhǔn)限2.收集數(shù)據(jù)2.收集數(shù)據(jù)3.確定短期偏差3.決定：短期還是長(zhǎng)期？4.計(jì)算工藝能力指標(biāo)：（通常是長(zhǎng)期）a.短期：4.計(jì)算工藝能力指標(biāo)：ⅠZU,ZLa.長(zhǎng)期：ⅡCPⅠPPMⅢCPKⅡSigma水平ZLTⅣSigma水平ZSTⅢPPKb.長(zhǎng)期：b.短期：ⅠSigma水平ZLTⅠSigma水平ZST精品文檔放心下載ⅡPPKⅡCPK工藝能力計(jì)算實(shí)例一位技師負(fù)責(zé)醫(yī)院設(shè)備的蒸汽殺菌過(guò)程。其中一個(gè)關(guān)鍵參數(shù)是控制“暴露”階段的謝謝閱讀溫度。設(shè)備室溫度和在最小飽和蒸汽濃度的周期時(shí)間決定殺菌程度在整個(gè)設(shè)備室維持前后謝謝閱讀一致的溫度范圍很重要。第一步：確認(rèn)標(biāo)準(zhǔn)這一階段常被忽視。我們?nèi)绾卧O(shè)定標(biāo)準(zhǔn)？設(shè)計(jì)部門(mén)-設(shè)計(jì)藍(lán)圖設(shè)計(jì)部門(mén)如何得到各項(xiàng)要求？工藝部門(mén)-標(biāo)準(zhǔn)由工藝以前能夠做到的或開(kāi)始使用時(shí)的能力定謝謝閱讀這想法有錯(cuò)嗎？客戶我們總是對(duì)客戶說(shuō)可以嗎？對(duì)上例而言：設(shè)備室目標(biāo)溫度是1250C±1.50C第二步：采集數(shù)據(jù)-合理編組應(yīng)采集數(shù)據(jù)獲得“短期”性能，如可能，“長(zhǎng)期”性能通過(guò)固定時(shí)間區(qū)間采集一系列快照型數(shù)據(jù)應(yīng)按合理編組采集快照數(shù)據(jù)什么是合理編組？從流程連續(xù)不斷產(chǎn)生的零件或產(chǎn)品中合理取樣以期捕獲最小工藝偏差的方法感謝閱讀組內(nèi)偏差反映一般偏差平均標(biāo)準(zhǔn)差（用一種均方差方法平均）是對(duì)工藝應(yīng)有能力的良好估計(jì)精品文檔放心下載第二步：采樣-例子例子：技師在暴露周期從控溫探針讀數(shù)中選取五個(gè)數(shù)據(jù)，并從連續(xù)七個(gè)殺菌運(yùn)轉(zhuǎn)周謝謝閱讀期采集數(shù)據(jù)，數(shù)據(jù)列在ChamberTemp2.mtw文件的桿ChambTemp欄中精品文檔放心下載第三步：確定短期偏差多數(shù)現(xiàn)有數(shù)據(jù)居于長(zhǎng)期和短期之間為了估計(jì)真實(shí)短期數(shù)據(jù)：小心設(shè)計(jì)工藝能力研究方法確保編組策略合理某些工藝無(wú)法研究短期數(shù)據(jù)如低產(chǎn)量和長(zhǎng)循環(huán)周期工藝采樣昂貴或難以取樣的工藝第三步：短期還是長(zhǎng)期？一個(gè)指導(dǎo)思想：如果允許80%的輸入指標(biāo)在其自然范圍內(nèi)浮動(dòng)，數(shù)據(jù)就是長(zhǎng)期的短期及長(zhǎng)期：組內(nèi)及組間平均標(biāo)準(zhǔn)差與總標(biāo)準(zhǔn)差對(duì)各組方差取平均值可得到組內(nèi)標(biāo)準(zhǔn)差的平均值總標(biāo)準(zhǔn)差由所有數(shù)據(jù)算出，不計(jì)編組平均標(biāo)準(zhǔn)差不計(jì)組間偏差，而總標(biāo)準(zhǔn)差計(jì)入組間偏差平均標(biāo)準(zhǔn)差是對(duì)組內(nèi)標(biāo)準(zhǔn)差的最佳估計(jì)長(zhǎng)期和短期指導(dǎo)思想短期數(shù)據(jù)在有限的周期或間隔采集數(shù)據(jù)在有限的機(jī)器和員工中采集差不多總是連續(xù)變量長(zhǎng)期數(shù)據(jù)在很多的周期，間隔，機(jī)器和員工中采集可以是離散或連續(xù)數(shù)據(jù)離散數(shù)據(jù)幾乎都是長(zhǎng)期性的第四步：計(jì)算ZU和ZL:Z-分?jǐn)?shù)提供統(tǒng)計(jì)數(shù)據(jù)以便用共同語(yǔ)言交流提供一個(gè)與標(biāo)準(zhǔn)上下限相關(guān)的工藝性能指標(biāo)第四步：計(jì)算CP例子工藝平均值為325標(biāo)準(zhǔn)差為15標(biāo)準(zhǔn)上限為380，下限為270CP是多少？若平均值為355而標(biāo)準(zhǔn)差不變CP又是多少？Cp與工藝應(yīng)有能力Cp是工藝應(yīng)有能力的良好指標(biāo)工藝應(yīng)有能力--一個(gè)工藝觀察到的最好的短期性能機(jī)會(huì)--工藝長(zhǎng)期性能與工藝應(yīng)有能力間的差距Sigma項(xiàng)目--致力與把長(zhǎng)期性能與工藝應(yīng)有能力的差距縮短感謝閱讀定量測(cè)量系統(tǒng)研究:定性型量具R&R--模型測(cè)量系統(tǒng)μ總和=μ工藝+Δμ測(cè)量系統(tǒng)偏離度：觀察值=實(shí)際真實(shí)值+測(cè)量偏移通過(guò)“校準(zhǔn)計(jì)劃”Δ測(cè)量偏移來(lái)評(píng)估真實(shí)值測(cè)量值（準(zhǔn)確度）測(cè)量系統(tǒng)σ2總合=σ2工藝+σ2測(cè)量系統(tǒng)偏離度：觀察的偏差=工藝的偏差+測(cè)量的偏差通過(guò)“校準(zhǔn)計(jì)劃”來(lái)評(píng)估真實(shí)值測(cè)量值（準(zhǔn)確度）測(cè)量系統(tǒng)的指標(biāo):量具R&R結(jié)果->量具偏差（σmeasurementsystem）感謝閱讀真實(shí)值精確度（量具偏差）觀察值測(cè)量系統(tǒng)的精確度（P）：精確度包括重復(fù)性和復(fù)制性測(cè)量系統(tǒng)的指標(biāo)-PT:精確度與公差之比--P/T代表量具偏差占公差的部分此部分通常用百分?jǐn)?shù)來(lái)表示最好的情形P/T<10%--可接受的P/T<30%謝謝閱讀測(cè)量系統(tǒng)的測(cè)量方法--P/TV:精確度與總偏差之比代表量具偏差占據(jù)總偏差的部分此部分通常用百分率來(lái)表示最好情形<10%量具可接受條件<30%測(cè)量系統(tǒng)的指標(biāo)--分辨指數(shù):分辨指數(shù)是測(cè)量系統(tǒng)從工藝數(shù)據(jù)中可辨認(rèn)的不同讀數(shù)的數(shù)量感謝閱讀分辨指數(shù)是一個(gè)分辨率指標(biāo)分辨指數(shù)是重復(fù)性和復(fù)制性的函數(shù)最好情形：>4，可接受的：3-4P/T和P/TV的用處:P/T（%公差）最常用于測(cè)量系統(tǒng)的精確度評(píng)估將量具的精確度與公差要求進(jìn)行對(duì)比如果量具用來(lái)對(duì)生產(chǎn)樣品進(jìn)行分類P/T還可以P/SV（%R&R）--6Sigma首選測(cè)量量具與量具研究偏差相比其性能如何最適合進(jìn)行工藝改進(jìn)的評(píng)估使用時(shí)應(yīng)小心。量具研究偏差并不一定代表真實(shí)的工藝偏差感謝閱讀P/TV（%R&R）--6Sigma首選測(cè)量量具與工藝偏差相比其性能如何使用時(shí)應(yīng)小心。量具研究偏差并不一定代表真實(shí)的工藝偏差謝謝閱讀當(dāng)量具樣本中的偏差代表真實(shí)工藝偏差時(shí)，P/TV等于P/SV謝謝閱讀定量型量具R&R--使用方法說(shuō)明:1，校準(zhǔn)量具或確認(rèn)最近校準(zhǔn)仍然有效2，收集10個(gè)代表工藝偏差全部范圍的樣本3，從每日使用這種測(cè)量方法的員工中選出檢驗(yàn)員4，運(yùn)用Clac>MakePatternedData>準(zhǔn)備量具研究數(shù)據(jù)表感謝閱讀5，讓員工測(cè)量所有無(wú)標(biāo)識(shí)，隨機(jī)次序的樣本6，分別讓另外其他員工測(cè)量所有無(wú)標(biāo)識(shí)，隨機(jī)次序的樣本精品文檔放心下載7，重復(fù)第五步及第六步循環(huán)三次。也盡量打亂員工次序8，用Minitab作下列兩個(gè)分析Stat>QualityTools>GageR&RStudy(Crossed)精品文檔放心下載Stat>QualityTools>GageRunChart精品文檔放心下載9，對(duì)測(cè)量系統(tǒng)能力研究結(jié)果進(jìn)行分析10，確定適當(dāng)?shù)暮罄m(xù)措施定量型量具R&R--Minitab實(shí)例:一個(gè)黑帶想對(duì)冶金工藝使用的溫度表進(jìn)行量具研究，他嚴(yán)格按前面一頁(yè)的方法進(jìn)行感謝閱讀實(shí)驗(yàn)，并將數(shù)據(jù)輸進(jìn)了R&Rexample.xls中。謝謝閱讀運(yùn)用Minitab分析數(shù)據(jù)并評(píng)估量具能力Stat>QualityTools>GageR&RStudy(Crossed)...感謝閱讀Minitab量具R&R研究--選項(xiàng)輸入該工藝公差和偏差，如果你想要Minitab幫你計(jì)算P/T和P/TV的話。精品文檔放心下載Minitab默認(rèn)計(jì)算P/SV量具R&R結(jié)果--ANOVA表P值是變化源在統(tǒng)計(jì)上對(duì)總偏差影響是否不顯著的概率在這個(gè)例子中，部件和員工均為顯著的偏差源另外，你能用Minitab的計(jì)算器計(jì)算總的平方和嗎？這個(gè)值代表什么意思？感謝閱讀《6Sigma項(xiàng)目運(yùn)作實(shí)例》->《分析階段》->失效模式及后果分析謝謝閱讀------------------------------------------------------------------------謝謝閱讀---------------------------精品文檔放心下載失效模式及后果分析:FailureModesandEffectsAnalysis(FMEA)感謝閱讀Background:FailureModesandEffectsAnalysis(FMEA)精品文檔放心下載Firstdevelopedinthe1950’sAppropriatedbyNASAinthe1960’sforthespaceprogram精品文檔放心下載FordMotorCompanywasthefirstNorthAmericancompanytowidely精品文檔放心下載implementtheuseofFMEAsTypesofFMEASystem–Top-level,earlystageanalysisofcomplexsystems感謝閱讀Design–Systems,subsystems,parts&componentsearlyindesign謝謝閱讀stageProcess–Focusesonprocessflow,sequence,equipment,tooling,精品文檔放心下載gauges,inputs,outputs,setpoints,etc感謝閱讀Who?When?WhoconstructstheFMEA?TheBlackBeltistheteamleader.謝謝閱讀TheprocessownerinheritsthefinishedFMEA.精品文檔放心下載Usetheprocessmapping,C&Ematrixteam.謝謝閱讀Mayneedtoaddarepfromquality,asupplier,reliability感謝閱讀WhenshouldtheFMEAbeconstructed?謝謝閱讀Aftertheprocessmap&theC&Ematrix感謝閱讀Beforeorafterthecontrolplan,dependingonthematurity感謝閱讀oftheprocessWhy?Warmupexercise:Youhave60secondstodocument:感謝閱讀Whatwouldyouwanttoknowabouta“defect”?感謝閱讀Fortheprocess:FMEAimprovesthereliabilityoftheprocess精品文檔放心下載AnFMEAidentifiesproblemsbeforetheyoccur感謝閱讀FMEAservesasarecordofimprovement&knowledge謝謝閱讀Forthefuture:FMEAhelpsevaluatetheriskofprocesschanges謝謝閱讀FMEAidentifiesareasforotherstudies–謝謝閱讀multi-vari,ANOVA,DOE6sProcessFMEA--Terminology精品文檔放心下載FMEA:Asystematicanalysisofaprocessusedtoidentifypotential精品文檔放心下載failuresandtopreventtheiroccurrence感謝閱讀PotentialFailuremode:Themannerinwhichtheprocesscould精品文檔放心下載potentiallyfailtomeettheprocessrequirements.謝謝閱讀PotentialFailureEffect:Theresultsofthefailuremodeonthe精品文檔放心下載customer.Severity:Anassessmentoftheseriousnessofafailuremode.精品文檔放心下載Severityappliestotheeffectsonly.感謝閱讀Cause:Howthefailurecouldoccur,describedintermsofsomething謝謝閱讀thatcanbecorrectedorcontrolled.感謝閱讀Occurrence:Thelikelihoodthataspecificfailuremodeisprojected精品文檔放心下載tooccur.Detection:Theeffectivenessofcurrentprocesscontrolstoidentify精品文檔放心下載thefailuremode(orthefailureeffect)priortooccurring,priorto精品文檔放心下載releasetoproduction,orpriortoshipmenttothecustomer.精品文檔放心下載RPN--RiskPriorityNumber:TheproductofSeverity,Occurrence精品文檔放心下載&DetectionFMEAExamplesPlatingExampleAnaerospaceplatingcompanywasshippingproducttoits感謝閱讀customerswithnickelplatingthatwastoothin.Partswerefailing精品文檔放心下載corrosiontestingatthecustomer.感謝閱讀ShippingExampleTheshippingdepartmentofanelectronicscompanyisunableto謝謝閱讀shipanassemblywithoutitsclamshellprotectivepackaging.This謝謝閱讀causesoccasionallateshipmentstothecustomer.感謝閱讀Inthefollowingexamples,asinglelinefromtheFMEAisusedasan感謝閱讀illustrationforeachoftheaboveexamples.精品文檔放心下載圖形技術(shù)分析:GraphicalMethodsProcessVariationNoisevariationfromdiscreteinputs感謝閱讀Differentoperators,machines,setups感謝閱讀Differentdays,shiftsDifferentbatches,mixtures,rawmaterials感謝閱讀Noisevariationfromcontinuousinputs精品文檔放心下載Ambienttemperature,humidity,pressure精品文檔放心下載Wear,drift,erosion,chemicaldepletion精品文檔放心下載),...,,(21kProcessxxxfy=),...,,(21kNoisennnf+精品文檔放心下載IntentionalUnwantedTheequationjustmeansthatanyoutputis謝謝閱讀determinedbytheintentionalprocesssettings謝謝閱讀andtheunwantednoisevariation.謝謝閱讀CommonClassificationofNoiseVariables謝謝閱讀Positional(withinpartvariation)謝謝閱讀Variationwithinasingleproductionunit精品文檔放心下載Thicknessvariationacrossaplatedpart感謝閱讀Variationacrossaunitcontainingmanyparts謝謝閱讀Variationacrossasemiconductorwaferwithmanydie感謝閱讀Variationbypositioninabatchprocess謝謝閱讀Cavity-to-cavityvariationsinaninjectionmoldingoperation謝謝閱讀Cyclical(part-to-partvariation)謝謝閱讀Variationbetweenconsecutiveproductionunits謝謝閱讀Batch-to-batchaveragedifferences–consecutivebatches精品文檔放心下載Temporal(time-to-timevariation)謝謝閱讀Shift-to-shift,Day-to-Day,Setup-to-setup精品文檔放心下載VariationnotaccountedforbyPositionalorCyclical感謝閱讀2222TemporalCyclicalPositionalNoiseσσσ++=感謝閱讀GraphicalAnalysis–ExampleInjectionmoldingisusedtomakeatypeofsocket,fourpiecesatatime,one感謝閱讀pieceperslot.Measurementsofthesocketsconsistofthicknessvaluesin感謝閱讀excessof5.00millimeters.Thegaugesmeasureinhundredthsofa感謝閱讀millimeter.Thespecificationis11±6.謝謝閱讀Fourtimesadaythesupervisorwouldgotothepressandgatherupthe精品文檔放心下載partsproducedbyfiveconsecutivecyclesofthepress.Sinceeachcycle感謝閱讀producedfourparts,hewouldhave20partstomeasureeverytwohours.感謝閱讀Thesupervisorkepttrackofthecycleandthecavityfromwhicheachpart精品文檔放心下載cameandwrotehistwentymeasurementsinanarraylikethis:Thesupervisorcollectedsamplesfourtimesadayforfivedays(20samples精品文檔放心下載total,20partspersample).CalculatetheprocesscapabilityanduseaMulti-Vari謝謝閱讀charttohelpdeterminesourcesofvariation.感謝閱讀ABCDES11819201921S21316141313S31011131013S41112131313Exercise:DetermineCapability感謝閱讀UsingMinitab,analyzetheThickdata精品文檔放心下載inSocketData.mtwforprocesscapability感謝閱讀Remember,thespecificationsare:11±6感謝閱讀Whatistheshort-termprocesscapability?精品文檔放心下載Whatisthelong-termprocesscapability?感謝閱讀Arethesegoodorbadvalues?Remember,onegoalofSixSigmaisto精品文檔放心下載reducevariation,whichwillincrease感謝閱讀capability.Itisalwaysimportantto感謝閱讀understandtheprocesscapability.精品文檔放心下載PreparingDataforMarginalPlotby“Slot”謝謝閱讀Marginalplotsrequirebothvariablestobedefinednumerically感謝閱讀Weneedtoconvert“Slot”toanumericcolumnfirst精品文檔放心下載Step1:Convert“Slot”Manip>Code>TexttoNumericManip>Code>TexttoNumericMulti-VariAnalysis–Defined謝謝閱讀AgraphicalanalysistoolUseslogicalsub-groupingAnalyzestheeffectsofdiscreteX’soncontinuousY’s感謝閱讀Acapabilityandprocessanalysistool謝謝閱讀Datacollectedforarelativelyshorttime謝謝閱讀Datacanestimatecapability,stability,andy=f(x)’s謝謝閱讀Majorfocus:studyuncontrollednoisevariationfirst謝謝閱讀Variationinnoisevariablesproduceschronicandacute感謝閱讀meanshifts,changesinvariability,andinstability謝謝閱讀Noisevariationmustbereducedoreliminatedinorderto感謝閱讀leveragetheimportantcontrollablevariablessystematically精品文檔放心下載Multi-varianalysisisaveryusefultool謝謝閱讀forgraphicallyidentifyingsourcesof感謝閱讀variation,especiallynoisevariation.Later精品文檔放心下載thisweek,wewillbestudyingcorrelation&感謝閱讀regression(ananalysisoftheeffectof感謝閱讀continuousX’soncontinuousY’s),analysis謝謝閱讀ofvariance(ANOVA)andtheGeneralLinear精品文檔放心下載Model(GLM),bothnumericalanalysesof感謝閱讀variancedata.Multi-varianalyseswillhelpidentifythe感謝閱讀variationsourceswiththepurposeofreducing感謝閱讀oreliminatingthem.AMulti-VariPlan1.Clearlystatetheobjective感謝閱讀2.ListtheX’sandY’stobestudied感謝閱讀3.Ensuremeasurementsystemcapability精品文檔放心下載4.Describethesamplingplan5.Describethedatacollection&storageplan(who,what,when,etc.)精品文檔放心下載6.Describetheprocedureandsettingsusedtoruntheprocess精品文檔放心下載7.Assembleandtraintheteam.Defineresponsibilities精品文檔放心下載8.Collectthedata9.Analyzethedata10.Verifytheresults11.Drawconclusions.Reportresults.Makerecommendations精品文檔放心下載InjectionMoldingExample1.Clearlystatetheobjective感謝閱讀Determinetheprocesscapabilityoftheinjectionmoldingprocess謝謝閱讀Determinethemajorsourcesofnoisevariation感謝閱讀2.ListtheX’sandY’stobestudied謝謝閱讀Output:ThicknessInputs:Cavity(slot),cycle,sample精品文檔放心下載3.Ensuremeasurementsystemcapability精品文檔放心下載AnMSAwasconductedandthesystemwasfoundcapable感謝閱讀4.DescribethesamplingplanOnesamplefromeachslot,fiveconsecutiveruns,fourtimesa感謝閱讀dayforfivedays.5.Describethedatacollection&storageplan(who,what,when,where,謝謝閱讀etc.)Thesupervisorcollectedthedataandentereditinaworksheet精品文檔放心下載6.Describetheprocedureandsettingsusedtoruntheprocess精品文檔放心下載Standard,constantprocesssettings.謝謝閱讀7.Assembleandtraintheteam.Defineresponsibilities.感謝閱讀Forasmallproject,thesupervisordidallthework謝謝閱讀8.Collectthedata.ThedataareinMinitabworksheetSocketData.mtw感謝閱讀9.AnalyzethedataAnalysisisonthefollowingslides感謝閱讀中心限理論:CentralLimitTheoremQ:WhyAreSoManyDistributionsNormal?感謝閱讀Whyissomethingthiscomplicatedsocommon?Sciencehasshownusthatvariablesthat精品文檔放心下載varyrandomlyaredistributednormally.So感謝閱讀anormaldistributionisactuallyarandom精品文檔放心下載distribution.Anotherreasonwhysomedistributions感謝閱讀arenormallydistributedisbecause精品文檔放心下載measurementsareactuallyaveragesover謝謝閱讀timeofmanysub-measurements.The感謝閱讀singlemeasurementthatwethinkweare感謝閱讀makingisactuallytheaverage(orsum)of精品文檔放心下載manymeasurements.TheCentralLimit謝謝閱讀Theorem,discussedinthefollowingslides,感謝閱讀providesanexplanationofwhyaveragesof精品文檔放心下載non-normaldataappearnormal.精品文檔放心下載DiceDemonstration(IntegerDistribution)謝謝閱讀Whatdoesaprobabilitydistribution謝謝閱讀fromasingledielooklike?Whatisthemean?Whatisthestandarddeviation?感謝閱讀ConstructadatasetinMinitab感謝閱讀SelectCalc>RandomData>Integer…fromthemain謝謝閱讀menuGenerate1,000rowsofdatainC1:Min=1,Max=6謝謝閱讀UseMinitab’sGraphicalSummaryroutineforanalysis謝謝閱讀Stat>BasicStatistics>DisplayDescriptiveStatistics…感謝閱讀MinitabOutput(Typical)Theprobabilitydistributionofthe感謝閱讀possibleoutcomesoftherollofasingledie謝謝閱讀isobviouslynon-normal.Aperfectdistributionwouldhavehad精品文檔放心下載allsixbarsexactlyequal,butevenwith謝謝閱讀10,000datapoints,thereisstillsome謝謝閱讀differencesinthehistogram.Ifabetter謝謝閱讀estimateisrequired,adifferentdataset精品文檔放心下載couldbeconstructedwithexactlyequal感謝閱讀countsofeachpossibleoutcome.Tryit謝謝閱讀andseeifthenumbersareanydifferent.謝謝閱讀SamplingaNon-normalDistribution–Exercise謝謝閱讀Eachpersonintheclassistotossasinglediesixteen感謝閱讀timesandrecordthedata.Calculatethemeanandstandarddeviationofeach謝謝閱讀sampleofsixteenRecordthemeansandstandarddeviationsfromeach謝謝閱讀personintheclassinaMinitabworksheet精品文檔放心下載UseMinitab’sGraphicalSummaryroutineforanalysis精品文檔放心下載Stat>BasicStatistics>DisplayDescriptiveStatistics…謝謝閱讀Alternately,asampleofsixteenthrows感謝閱讀ofthedicecanbesimulatedinMinitabas謝謝閱讀follows:Select:Calc>RandomData>Integer…from謝謝閱讀themainmenuGenerate16rowsofdatainC1:Min=1,Max謝謝閱讀=6AnalyzetheSampleDataWhatisthemeanofthesampleaverages?精品文檔放心下載Mean≈3.5Whatisthestandarddeviationofthesampleaverages?精品文檔放心下載Sigma≈0.4Isthedistributionnormal?Whatisthep-value?Whatistherelationshipbetweentheaverageofthe精品文檔放心下載samplemeansandthepopulationaverage?精品文檔放心下載Whatistherelationshipbetweenthesigmaofthe精品文檔放心下載averagesandthesigmaoftheindividuals?精品文檔放心下載TheCentralLimitTheoremFormalDefinition:Ifrandomsamplesofnmeasurementsarerepeatedly精品文檔放心下載drawnfromapopulationwithafinitemeanμμμμandastandard謝謝閱讀deviationσσσσ,then,whennislarge,therelativefrequency精品文檔放心下載histogramforthesamplemeans(calculatedfromthe感謝閱讀repeatedsamples)willbeapproximatelynormalwitha感謝閱讀meanμμμμandastandarddeviationequaltothepopulation感謝閱讀standarddeviation,σσσσ,dividedbythesquarerootofn.精品文檔放心下載(Note:Theapproximationbecomesmorepreciseasn謝謝閱讀increases.)CentralLimitTheorem–Exercise感謝閱讀FromaMinitabanalysisoftheuniformlydistributed感謝閱讀data:Foranexercise,verifythattheCentralLimitTheoremis感謝閱讀validforthisuniformdataVariableNMeanStDevn=1(Individuals)10000-0.003310.57918謝謝閱讀n=2(Means)100000.002590.40613謝謝閱讀n=5(Means)10000-0.001130.25953謝謝閱讀n=30(Means)10000-0.002370.10559精品文檔放心下載相關(guān)性及簡(jiǎn)單線性回歸:Regression&CorrelationIntroductionUsedforquantitativevariables(X’sandY’s)謝謝閱讀Forreview:WhatisthefocusofSixSigma?感謝閱讀Q.Whatdoesthisequationrepresent?謝謝閱讀A.Amathematicalmodelofaprocess謝謝閱讀PurposeofRegression:topredictYfromasettingofx謝謝閱讀Examples:Distance=f(acceleration,initialvelocity,time)謝謝閱讀Productyield=f(concentrationsofreactants)謝謝閱讀Hardness=f(alloy,annealtemperature)謝謝閱讀)(xfY=Remember,thefocusofSixSigmaisto精品文檔放心下載determinethedefiningequationofthe感謝閱讀process.Itistoidentifytheimportantinput精品文檔放心下載variables,determinetherelationshiptothe感謝閱讀outputs,determinetheoptimumvaluesofthe精品文檔放心下載criticalinputsandthencontroltheinputsat感謝閱讀theoptimumsettings.Todothis,theBlackBeltmustknowthe感謝閱讀relationshipbetweentheinputsandthe精品文檔放心下載outputs.Thismodulediscusseslinear感謝閱讀modelingtechniquesforidentifyingthe謝謝閱讀relationshipbetweencontinuousvariable精品文檔放心下載inputsandcontinuousvariableoutputs.精品文檔放心下載ASimpleLinearModelLinearequationsrequirecontinuousinput感謝閱讀andoutputvariables.Oneotherassumptionis感謝閱讀thattheindependentvariable(input)isknown感謝閱讀andfixedandthatallofthevariationisinthe精品文檔放心下載dependentvariable(output).Thisisnot精品文檔放心下載usuallythecase,butoftentheinputsare精品文檔放心下載settingsondialsorgaugesorsoftwarethat精品文檔放心下載seemsfixedandinvariable.Manytimesthe精品文檔放心下載variationintheoutputisafunctionofthe精品文檔放心下載inabilityoftheinputcontrollertoholdthe謝謝閱讀inputatthesamevalue.CollectingData(y&x)–AFewThoughts感謝閱讀Pg8?March01,BreakthroughManagementGroup.Unpublishedproprietaryworkavaila精品文檔放心下載bleonlyunderlicense.Allrightsreserved.March16,2001感謝閱讀Makesuretheprocesssettingscoverthelikelyproduction謝謝閱讀range(butnottoofar).Toogreatarangepointsoutsidethenormalrangemay精品文檔放心下載havetoogreataneffectonthemodel.精品文檔放心下載ToosmallarangeErrortermmaydominatethefit.精品文檔放心下載Takeseveralreplicatesateachinputsetting(x).感謝閱讀Replicaterunshelpincreasethemodelaccuracy.感謝閱讀Randomizerunswheneverpractical.謝謝閱讀Runorderisoftensignificantfactor.精品文檔放心下載Theoutput(y)atdifferentinputs(x抯)isnotalways精品文檔放心下載independentofprevioussettings.精品文檔放心下載Agoodspreadinthedataisrequiredfora精品文檔放心下載goodmodel.Considertwoexamples:謝謝閱讀Allofthedataiscollectedatthenormal感謝閱讀processsettings.Inthiscase,regressionwill謝謝閱讀trytofitalinearmodeltoacombinationof精品文檔放心下載randomprocessvariationandrandom精品文檔放心下載measurementvariation.Theresultswillbeof感謝閱讀novalue.Thesecondcaseiswhenmostofthedata謝謝閱讀isclusteredaroundthestandardsettings謝謝閱讀exceptforacoupleofpointsattheextreme謝謝閱讀ranges.Inthiscase,theextremepoints精品文檔放心下載controlthefitofthemodel.Ifoneofthe感謝閱讀extremepointsisaflyer,thenthemodelwill感謝閱讀beinerrorduetotheflyer.TheidealcaseisfortheBlackBeltto精品文檔放心下載collectarangeofdatathroughouttheprocess精品文檔放心下載space.置信區(qū)間:ConfidenceIntervalsApopulationisthesetofallmeasurementsofinteresttotheexperimenter精品文檔放心下載Asampleisasubsetofmeasurementsselectedfromthepopulation感謝閱讀Aninferenceisastatementaboutapopulationparameterbasedon感謝閱讀informationcontainedinasample感謝閱讀TwotypesofinferenceEstimationApollhasbeendevisedtodeterminethepublic’sreactiontoa精品文檔放心下載newpoliticalscandal.Thepurposeistoestimatethereaction感謝閱讀ofallAmericansbypollingarepresentativesample精品文檔放心下載HypothesistestingAvaccineforLymediseasehasbeendevelopedbuttherate精品文檔放心下載ofnegativesideeffectsis1.45%.Anewvaccinehasbeen謝謝閱讀developedanditisdesiredtoknowiftherateofnegativeside精品文檔放心下載effectsislowerthan1.45%.Theotherbranchofstatisticsis感謝閱讀descriptive.Itspurposeismerelyto感謝閱讀describeasetofmeasurements.感謝閱讀Inferentialstatisticsisusedtoguesswhat謝謝閱讀Godknowsaboutapopulationfromasample.謝謝閱讀Withininferentialstatistics,therearetwo感謝閱讀types:estimationandhypothesistesting.感謝閱讀Estimationistryingtoguessthepopulation精品文檔放心下載statisticsfromasample.Hypothesistesting感謝閱讀concernsevaluatingasamplestatisticand感謝閱讀comparingittosomehypotheticalvalue.精品文檔放心下載EstimatesandtheCLTWhatisthebestestimateofthepopulationmeanusingsampledata?謝謝閱讀Thesamplemean!Howgoodofanestimateisthesamplemean?謝謝閱讀Whatfactorsinfluencetheaccuracyoftheestimateofthemean精品文檔放心下載fromsampledata?Recallthat:Thevariationinthedistributionofsamplemeansisafunctionofthe感謝閱讀varianceofthePopulationandthesamplesize!感謝閱讀nPopX/σσ=WhatAboutSmallSamples?Ifthepopulationstandarddeviationisknown(italmostneveris)use謝謝閱讀thepreviousformulaforsmallsamples,too感謝閱讀Ifthepopulationsigmaisunknown(itusuallyis):精品文檔放心下載Theestimateforstandarddeviation(s)isused精品文檔放心下載Thet-distributionisusedinsteadofthenormal(Z)distribution精品文檔放心下載Q:Whatisat-distribution?Thet-distributionisafamilyofbell-shaped(normal-like)謝謝閱讀distributionsthataredependentonsamplesize感謝閱讀Thesmallerthesamplesizen,thewiderandflatterthe精品文檔放心下載distributionnstXμnstXnn1,2/1,2/+≤≤ααThet-distributionisthegeneralcasefor精品文檔放心下載anysamplewherethepopulationstandard感謝閱讀deviationisunknown.However,withlarge精品文檔放心下載samples,thet-andz-distributionsarenearly精品文檔放心下載identical,soeithercanbeused.謝謝閱讀YoucanverifythisinMinitabby謝謝閱讀generatingalargesampleofnormaldataand感謝閱讀thenanalyzingitwithboththez-andt-感謝閱讀distributionroutines.ProportionsandBinomialExperiments謝謝閱讀Pg35.April01,BreakthroughManagementGroup.Unpublishedproprietaryworkavail精品文檔放心下載ableonlyunderlicense.Allrightsreserved.April3,2001感謝閱讀Proportiondataisusuallytheresultofabinomial-type謝謝閱讀experimentBinomialexperiments(orBernoullitrials)arethosethat精品文檔放心下載haveonlyoneoftwooutcomes,eithera“success”ora感謝閱讀“failure”Theprobabilityofthistypeofexperimentisdescribedbya謝謝閱讀binomialdistribution,acomplicateddistribution謝謝閱讀Inmanycasesthenormaldistributioncanbeusedto感謝閱讀approximatethebinomialdistribution感謝閱讀Whennxp>5andnx(1-p)>5μ=nxpandσ2=nxpx(1-p)Binomialdistributionsarediscussedin謝謝閱讀almosteverystatisticstextbook.Calculations謝謝閱讀withthemisnotnecessarilydifficult,butitis感謝閱讀tediousifitmustbedonemanually.Minitab精品文檔放心下載hasroutines,however,thatgreatlysimplifies謝謝閱讀thecalculations.Ifthebinomialapproximationappliesand謝謝閱讀thedatacanbeestimatedwithanormal感謝閱讀distributionotherstatisticaltestsandcontrol感謝閱讀chartscanbeusedthatwouldnotbeavailable謝謝閱讀otherwise.Trytoconstructyourexperimentssuch精品文檔放心下載thatthebinomialapproximationisvalid.精品文檔放心下載Ageneralruleofthumb:forthenormal謝謝閱讀approximationtoapply,haveasamplesizeof謝謝閱讀atleast30andlargeenoughtoguaranteesat感謝閱讀least5successes.假設(shè)測(cè)試:IntroductiontoHypothesisTesting謝謝閱讀ABrightIdeaNotes:Pg511Nov2000?April01,BreakthroughManagementGroup.Unpublishedproprietaryw精品文檔放心下載orkavailableonlyunderlicense.Allrightsreserved.謝謝閱讀Alightbulbcompanyistryingtoproduceabrighterlightbulbforthe謝謝閱讀sameenergy.Itishopedthatachangeinthefilamentcoating謝謝閱讀processwillproduceabrighterlight.感謝閱讀Theengineercollectedthelasttenlightbulbsmadebeforethe感謝閱讀processchangeandthefirsttenafterthechange.Themeanlight精品文檔放心下載outputoftheoldprocessbulbsis1251lumensandthenewprocess精品文檔放心下載is1273lumens.Doestheincreaseof22inthemeansofthetwogroupsrepresenta精品文檔放心下載realimprovement?Couldthedifferencebetweenthesetwogroupshavehappenedby感謝閱讀randomchance?Shouldtheengineerswitchtothenewprocess?感謝閱讀Thesekindsofproblemsarevery精品文檔放心下載familiartoengineers.Anengineeris謝謝閱讀givenatasktoimproveaprocessor精品文檔放心下載product.Afterachangeintheprocess,感謝閱讀theengineerisleftwiththeproblemof感謝閱讀determiningwhethertheprocesschange感謝閱讀hasmadeasignificantimprovementor精品文檔放心下載not.Thoughengineersoftenusemore精品文檔放心下載advancedtechniquestodeterminethe謝謝閱讀improvedsettings(DOE,forexample,to感謝閱讀bediscussedlater),ahypothesistestis精品文檔放心下載oftenusedtoverifytheexperiment感謝閱讀results.Theprocessmaybeasfollows:?Identifytheproblem.?Designandrunanexperimentto謝謝閱讀findanimprovedcondition.?Analyzethedataanddeterminethe感謝閱讀improvedoperatingpoint.?Verifytheeffectivenessofthe精品文檔放心下載improvementwithahypothesistest.謝謝閱讀AFewIlluminatingDetailsQCdatawereavailableforlightbulbsproducedinthesamefactory.精品文檔放心下載Allofthebulbshadbeenproducedusingthestandardfilament精品文檔放心下載coatingprocess.Thedatawascomprisedoftheaveragesof10精品文檔放心下載samplesfromconsecutivebatchesoflightbulbs.Theengineer精品文檔放心下載calculatedthedifferencesbetweenconsecutivegroupsandrecorded謝謝閱讀itinaMinitabworksheetincolumnQCData.感謝閱讀Consideringthenewdata,thequestionnowbecomes:精品文檔放心下載“Howoftenhasthemeanbrightnessofagroup謝謝閱讀oflightbulbsbeen22lumensbrighterthanthe謝謝閱讀groupproducedimmediatelybefore?”感謝閱讀OpenMinitabworksheetFilament.mtwforthedata精品文檔放心下載IntheMinitabdatasheet,column感謝閱讀‘OldData’isthefirstsetof10fromthe精品文檔放心下載oldprocess.‘NewData’isthesetofdatafromthenew感謝閱讀process.‘PlantData’isthe210averagesof10謝謝閱讀consecutivemeasurementsfromtheold謝謝閱讀processinsampleorder.‘Diff’isthemagnitudeofthedifference精品文檔放心下載betweenconsecutivegroups.WhatIsaHypothesisTest?Ahypothesistestissimplycomparingrealitytoan感謝閱讀assumptionandasking,“Aretheythesame?”精品文檔放心下載OrAhypothesistestistestingwhetherrealdatafitsamodel謝謝閱讀OrAhypothesistestiscomparingastatistictoahypothesis感謝閱讀均值測(cè)試:MeansTests1.Statethepracticalproblem2.Statethenullhypothesis3.Statethealternatehypothesis4.Testtheassumptionsofthedata*5.Calculatetheappropriateteststatistic(orcalculatep-value)6.Lookupthecriticalvaluefromtheappropriatedistribution(orsetalpha)7.Ifthecalculatedstatistic謝謝閱讀meetsthedecisionrulecriteria(orifp-value<α)thenrejectH08.FormulatethestatisticalconclusionintoapracticalsolutionThisisthegeneralrecipeforhypothesis謝謝閱讀testing.Thetestsdifferintheappropriate謝謝閱讀statisticsandappropriatedistributions.精品文檔放心下載Themeanstestsrecipeisthesameaswas精品文檔放心下載whatlearnedforvariancetesting.謝謝閱讀Example–OneMeanVs.TargetBackground–Acompanyauditsitsstockoftennisballsbytestingthebounce感謝閱讀heightof10randomlyselectedballs.Theaveragebounceheightofthe謝謝閱讀sampleis19.8in.Thehistoricaldataisμ=20.1andσ=0.5.Hasstorage精品文檔放心下載degradedthebounceofthetennisballs?Useα=0.05.精品文檔放心下載1.Statethepracticalproblem:精品文檔放心下載Isthebounceheightofthestoredpopulationlessthanthehistorical謝謝閱讀value?2.StatethenullhypothesisHo:μstored=20.13.Statethealternatehypothesis精品文檔放心下載HA:μstored<20.14.Testassumptions:normalityofthedata謝謝閱讀p-value=0-.432–Dataisnormal感謝閱讀Ave