6SIGMA某工作案例-3

NOV-23-2000Page1C21C31C41C51C91K524.404.454.504.55NutMouldC/LDiameter_1BoxplotsofC/LDiambyNutMoul(meansareindicatedbysolidcircles)2）C/LMould(Cavity别)差异分析AC21-1C21-2C21-3C21-4C31-1C31-2C31-3C31-4C41-1C41-2C41-3C41-4C51-1C51-2C51-3C51-4C91-1C91-2C91-3C91-4K52-1K52-2K52-3K52-44.404.454.504.55NutCavityC/LDiameter_1BoxplotsofC/LDiambyNutCavi(meansareindicatedbysolidcircles)A特异点原因：1EAC/L破损对CKDMould进行Cavity别2-SampleTTestCavity别具体分析NOV-23-2000Page2Individual95%CIsForMeanBasedonPooledStDevLevelNMeanStDev----------+---------+---------+------C21-154.49400.0472(--*--)C21-254.51400.0055(--*-)C21-354.53000.0000(-*--)C21-454.52000.0071(-*--)C31-154.50600.0055(-*--)C31-254.51800.0045(--*--)C31-354.50800.0045(--*--)C31-454.51600.0055(-*--)C41-154.51600.0055(-*--)C41-254.51800.0045(--*--)C41-354.52800.0045(--*-)C41-454.52800.0045(--*-)C51-154.51800.0045(--*--)C51-254.52400.0055(--*-)C51-354.49800.0045(--*--)C51-454.51200.0045(--*--)C91-154.50600.0055(-*--)C91-254.51000.0000(--*-)C91-354.51000.0000(--*-)C91-454.50600.0055(-*--)K52-154.51200.0045(--*--)K52-254.52600.0089(--*--)K52-354.42200.0045(-*--)K52-454.45200.0045(--*--)----------+---------+---------+------PooledStDev=0.01084.4454.4804.515One-wayANOVA:C/LDiameter_1versusNutCavityAnalysisofVarianceforC/LDiamSourceDFSSMSFPNutCavi230.0667600.00290324.970.000Error960.0111600.000116Total1190.077920【续上】对CKDMould与其他全部Cavity别2-SampleTTestNOV-23-2000Page3ATwo-SampleT-TestandCI:K52,C**Two-sampleTforK52vsC**NMeanStDevSEMeanC**204.53220.01020.0023K52204.451910.009450.0021Difference=muK52-2-muK52-3Estimatefordifference:0.0802795%CIfordifference:(0.07396,0.08658)T-Testofdifference=0(vsnot=):T-Value=25.77P-Value=0.000DF=37假设：H0：K52=C**(精熙Mould)H1：K52≠C**P-Value0.05,所以丢弃H0。取H1（即K52≠C**）结论：CKD模具Cavity别与其他Mould外径尺寸有差异。对策：全部使用现地化部品。停止使用CKD资材【续上】NOV-23-2000Page4123456789101112白班白班白班夜班夜班夜班白班白班白班夜班夜班夜班白班白班白班夜班夜班夜班白班白班白班夜班夜班夜班白班白班白班夜班夜班夜班白班白班白班夜班夜班夜班白班白班白班夜班夜班夜班白班白班白班夜班夜班夜班白班白班白班夜班夜班夜班白班白班白班夜班夜班夜班白班白班白班夜班夜班夜班白班白班白班夜班夜班夜班05101505101505101511/2011/1811/17Line12345Multi-VariChartforDefectbyTime-DayShiftDayTimeA对Day，Line，Shift&Time进行多变量分析如下3）Line(Operator)AnalysisD–6Line有工程不良有异常NOV-23-2000Page514710131695%ConfidenceIntervalforMu1.01.52.02.595%ConfidenceIntervalforMedianVariable:DefectA-Squared:P-Value:MeanStDevVarianceSkewnessKurtosisNMinimum1stQuartileMedian3rdQuartileMaximum2.06752.31481.000014.5940.0002.325002.483926.169851.785895.054423600.00000.00002.00003.000016.00002.58252.67992.0000Anderson-DarlingNormalityTest95%ConfidenceIntervalforMu95%ConfidenceIntervalforSigma95%ConfidenceIntervalforMedianDescriptiveStatisticsAP-Value0.05对各个Factor进行Kruskal-WallisTestDefectDataAnalysisforLineNOV-23-2000Page6DefectversusTimeDefectversusLineTimeNMedianAveRankZLineNMedianAveRankZ1722.000187.80.661302.000170.5-0.552722.000183.00.222301.000102.7-4.283722.000191.61.013302.000187.10.364722.000167.6-1.184302.000140.5-2.205722.000172.5-0.735303.000209.01.57Overall360180.56305.000301.16.637303.000191.60.61H=2.75DF=4P=0.6018302.000172.3-0.45H=2.82DF=4P=0.588(adjustedforties)9301.000144.9-1.9610303.000202.41.2011303.000205.31.3612302.000138.7-2.30Overall360180.5H=76.02DF=11P=0.000H=78.04DF=11P=0.000(adjustedforties)DefectversusDayDefectversusShiftDayNMedianAveRankZShiftNMedianAveRankZ11/171202.000176.3-0.54白班1802.000189.81.6911/181202.000184.20.48夜班1802.000171.2-1.6911/201202.000181.00.06Overall360180.5Overall360180.5H=0.35DF=2P=0.841H=2.87DF=1P=0.090H=0.36DF=2P=0.837(adjustedforties)H=2.95DF=1P=0.086(adjustedforties)Line–DefectAnalysisANOV-23-2000Page7TimeShiftLineDay12345白班夜班12345678910111211/1711/1811/201.02.23.44.65.8DefectMainEffectsPlot-DataMeansforDefectDefectEffectbyLineAnalysisAFactoryP-ValueSignificant(P0.05)Time0.588NOShift0.086NOLine0.000YES(D-6line)Day0.837NONOV-23-2000Page8Chi-SquareTest:OK,NGExpectedcountsareprintedbelowobservedcountsOKNGTotalMachine123256532330923253.5555.45Machine222760452280522750.7554.25Machine320360722043220383.3948.61Machine422528422257022516.3153.69Total8890421289116Chi-Sq=0.000+0.108+0.004+1.578+0.027+11.259+0.006+2.546=15.528DF=3,P-Value=0.001p-value值小于0.05，因此可得出设备别不良发生率的概率有差异。3号机器的Chi-sq值最大。因此我们用2号机器和其它机器进行一下比较。A11/8日富钢业体4台Nut压入机一天不良统计表Machine1Machine2Machine3Machine4OK23256227602038322528NG534572424）MachineChi-SquareTestNOV-23-2000Page9Chi-SquareTest:OK,NGOKNGTotalMachine123256532330923261.4947.51Machine222760452280522758.5246.48Machine422528422257022524.0046.00Total6854414068684Chi-Sq=0.001+0.634+0.000+0.047+0.001+0.349=1.032DF=2,P-Value=0.597P-Value0.05Machine1,2,4无差异Chi-SquareTest:OK,NGOKNGTotalMachine320360722043220382.2549.75Machine1,2,4685441456868968521.75167.25Total8890421789121Chi-Sq=0.024+9.951+0.007+2.960=12.943DF=1,P-Value=0.000P-Value0.05Machine3与Machine1,2,4有差异，原因分析：Machine3Jig定位块高度不良，使冲头压入不完全。对策：取掉限位块，使冲头能完全压入。AMachine&Machine3比较Machine1,2&4比较【续上】NOV-23-2000Page10A-Phase小结1.NutDiameter不良严重。工程能力低下.改善方向：Nut业体多元化（NutDiameter&NutHeight作DOEFactor)2.CKDCarriageLowerDiameter与现低化部品差异大，（改善方向：该部品全部使用现地化）；3.Line别不良数量差异大，作业者作业方法需要改变；（改善方向：标准作业方法制定）4.外注设备差异大，且无A