SPC过程和变差的种类

SPC-advanced training course在這些過程中，品質是倚重員工的技術。控制著員工的性向測驗、培訓及資格、員工素質的等評及以沒有差錯為宗旨以淢少員工的錯誤。以下為可行的控制方法：l SPCl IPCl 培訓過程和變差Labor Dominant ProcessSPC-advanced training course受控及不受控的工艺受控及不受控的工艺SPC-advanced training course變差的種類變差的種類受控變差受控變差是在過程內及一段時間內，有著平穩而連貫的變差模式。這又稱為普遍成因變差或隨機成因變差.不受控變差不受控變差是隨時間改變。變差模式的特性又稱為特別成因變差.變差事例：變差事例：考慮一製造過程現製造一系列獨立的零件，每一零件有著可量度的尺寸或特性。定期抽出某些零件並且量度。這些量度結果不同是因為物料、機器、操作員及方法相互牽引而形成變差.過程和變差SPC-advanced training course特殊原因造成不可预测特殊原因造成不可预测SPC-advanced training course变量变量的种类的种类基本統計及統計概念工艺过程可控制变量重要不重要输出变量不可控变量ABCD重要如B随A变化而变化则为非独立变量不重要Input factorOutput responseSPC-advanced training course變差變差2大主要變差的類別或來源1.受控變差-由普通原因引致的過程變差2.不受控變差-由特殊原因引致的過程變差 基本統計及統計概念=UCLCLLCLSPC-advanced training course變差變差變差的分類1.Within part/piece variation-同一样本内的变异2.Piece-to-piece variation-在某时间内，制造的不同批次产品样本的变异3.Time-to-time variation-不同时间制出产品的变异基本統計及統計概念Responset8:0013:003:00SPC-advanced training course普遍成因變差的示例普遍成因變差的示例當過程於現況中存在時，這經常在過程中發生並影響輸出。一些普遍成因的示例如下：物料的性質消耗(如硬度、光滑度)機器的特性(如年齡、運行溫度、公差)環境(如濕度、光線)用於量度過程輸入及輸出的量具因為這些成因會從過程運作中產生，唯一可減少普遍成因變差是透過重新設計過程/DOE/機器的投資而作出改善。在過程改善中，必須配合在過程改善中，必須配合管理行動管理行動過程和變差SPC-advanced training course特殊成因變差示例特殊成因變差示例特殊成因變差會導至不受控變差，它們是難以預知，並且由外間影響而非內在過程因素。透過簡單的統計技術，可把它們偵查出來。特殊成因示例：物料改變(如新供應商、不同的紙張)機器或過程故障(如模具磨損、郵務人員罷工)不同操作員(如新的受訓者)環境改變(如天氣時常很差)這些特殊成因變差可透過研究操作員或過程改善原因，從而中過程中刪除。TCM是其中一個以這為目標而有效的工具。過程和變差SPC-advanced training courseSPC-advanced training course特殊因素必須及早被發現，否則其出現會導至過程及輸出有不良的影響。當有害的特殊成因的原因被查明引証後，這便可從過程中被刪除，再發生的機會也很少。這行動必須小心研究有問題的程序，或為機器作定期維修。此外，在某些情況下，特殊成因也對過程有好處(如程序的改變)。再一次，必須查明改善的原因及，若可以，為過程作出永久性的零件。過程和變差消除特殊变因消除特殊变因SPC-advanced training course預防在過程中發生差劣品質預防在過程中發生差劣品質SPC在過程中建立品質在過程中建立品質(品質保証品質保証)傳統品質控制傳統品質控制檢查制成品檢查制成品偵測品質差劣的制成品偵測品質差劣的制成品SPC簡介传统品质和现代品质管理的差传统品质和现代品质管理的差别别SPC-advanced training courseSPC角度中的品質改善角度中的品質改善Upper Process LimitLower Process LimitxxUpper Spec LimitLower Spec LimitUpper Control LimitLower Control LimitCentral Limit Thorem:(for infinite size N)x=n(n:sample size)SPC-advanced training courseHow to adjust for Finite Population?x=nN:population sizen:sample sizeN-nN-1Otherwise,with n increase,xbecomes increasingly smaller.Correction factorThe correction factor is used only when:n/N5%SPC-advanced training courseExample:10 ohm resistors are manufactured in batches of 500,and 50 resistors aretested in each batch.If the mean resistence of the sample is less than9.9ohms,the batch is rejected.With =10ohms&=1ohms,what is the probability that the batch will be rejected?xSolution:N=500,n=50,X=9.9x=nN-nN-1=0.1414*0.902=0.1343xZ=X-=-0.745=10=19.9Z=-0.745Area=0.2719P =0.5-0.2719=0.2281=22.81%RejectSPC-advanced training courseY1-3.0-2.0-1.00.01.02.03.0Quantilesmaximum quartilemedianquartile minimum100.0%99.5%97.5%90.0%75.0%50.0%25.0%10.0%2.5%0.5%0.0%3.6964 2.6049 1.9503 1.2888 0.6633-0.0095-0.6682-1.2673-1.9656-2.5343-3.4156MomentsMeanStd DevStd Error MeanUpper 95%MeanLower 95%MeanNSum Weights -0.00 0.99 0.01 0.02 -0.02 10000.00 10000.00Y2-3.0-2.0-1.00.01.02.03.0Quantilesmaximum quartilemedianquartile minimum100.0%99.5%97.5%90.0%75.0%50.0%25.0%10.0%2.5%0.5%0.0%3.7469 2.6341 1.9646 1.2951 0.6700-0.0229-0.6864-1.3064-1.9558-2.6102-3.5024MomentsMeanStd DevStd Error MeanUpper 95%MeanLower 95%MeanNSum Weights -0.01 1.01 0.01 0.01 -0.03 10000.00 10000.00Y3-4.0-3.0-2.0-1.00.01.02.03.0Quantilesmaximum quartilemedianquartile minimum100.0%99.5%97.5%90.0%75.0%50.0%25.0%10.0%2.5%0.5%0.0%3.7051 2.6042 1.9537 1.2906 0.6794 0.0053-0.6602-1.2795-1.9687-2.5748-3.9444MomentsMeanStd DevStd Error MeanUpper 95%MeanLower 95%MeanNSum Weights 0.01 1.00 0.01 0.03 -0.01 10000.00 10000.00Y4-3.0-2.0-1.00.01.02.03.04.0Quantilesmaximum quartilemedianquartile minimum100.0%99.5%97.5%90.0%75.0%50.0%25.0%10.0%2.5%0.5%0.0%3.8340 2.5585 1.9116 1.2613 0.6563-0.0344-0.6950-1.3100-1.9846-2.5801-3.6670MomentsMeanStd DevStd Error MeanUpper 95%MeanLower 95%MeanNSum Weights -0.02 1.00 0.01 -0.01 -0.04 10000.00 10000.00Y5-3.0-2.0-1.00.01.02.03.04.0Quantilesmaximum quartilemedianquartile minimum100.0%99.5%97.5%90.0%75.0%50.0%25.0%10.0%2.5%0.5%0.0%4.3333 2.7027 1.9827 1.2675 0.6733 0.0033-0.6597-1.2605-1.9436-2.5207-3.4887MomentsMeanStd DevStd Error MeanUpper 95%MeanLower 95%MeanNSum Weights 0.00 0.99 0.01 0.02 -0.01 10000.00 10000.00Mean-3-2-10123Quantilesmaximum quartilemedianquartile minimum100.0%99.5%97.5%90.0%75.0%50.0%25.0%10.0%2.5%0.5%0.0%2.4477 1.2913 0.9659 0.6467 0.3348-0.0117-0.3415-0.6404-0.9829-1.2946-1.9717MomentsMeanStd DevStd Error MeanUpper 95%MeanLower 95%MeanNSum Weights -0.01 0.50 0.01 0.00 -0.02 10000.00 10000.00(Y1+Y2+Y3+Y4+Y5)/5=MeanLook at the Sigma difference:Sigmax=Sigma/SQRT(5)SPC-advanced training coursey1y2y3y4y5y6y7y8y9y101-1.01-1.720.2971.044-0.590.352-1.21-0.370.486 1.3152-0.11-0.63-1.12-0.537 0.7480.02-0.29-0.02-1.15-2.131.16-0.210.347-0.653-1.56-0.711.056-1.291.686-0.1140.402 1.661-1.45-0.346-0.180.375-0.7-1.40.245-0.3351.224 0.497-1.10.997-0.58-0.2-0.91-1.240.356-1.94Min-1.01-1.72-1.45-0.653-1.56-0.71-1.21-1.4-1.15-2.1Max1.224 1.661 0.3471.0440.748 0.375 1.056-0.021.686 1.315Range2.2383.381.7981.6972.3091.082.266 1.373 2.839 3.412Mean0.333-0.08-0.60.101-0.43-0.03-0.41-0.860.324-0.63R=2.239Estimated Sigma=R/d2=0.961X3 Sigma=A2RSigma x=0.433Random Sampling for a Normal DistributionStdev0.9341.2620.8580.8470.8340.4460.8860.6251.0091.414SPC-advanced training courseSPC的步驟的步驟1.提供統計方法的教育。基礎的SPC方法、fractional factorial experiments、腦力激盪法及QFD/品質工具。.2.確定商業機會/推動者。3.把主要的生產特性與過程變差達到商業機會/推動者。4.Understand relationship of the process independent variables to process dependent variables and Product characteristics.5.確定最好的主要過程及主要產品參數。