计量型数据MSA.pptx

《计量型数据MSA.pptx》由会员分享，可在线阅读，更多相关《计量型数据MSA.pptx（47页珍藏版）》请在淘文阁 - 分享文档赚钱的网站上搜索。

1、019 MSA-1Effective Date: 2/2004Copyright 2000-2002 Sigma Breakthrough Technologies, Inc. Used with permission.Copyright 2004 Cummins, Inc. All Rights Reserved.计量性数据计量性数据MSA019 MSA-2Effective Date: 2/2004Copyright 2000-2002 Sigma Breakthrough Technologies, Inc. Used with permission.Copyright 2004 Cum

2、mins, Inc. All Rights Reserved.简单测量分析简单测量分析 续续 Simple Measurement Study - continued 使用使用， 计算计算Delta 的标准差的标准差Using Descriptive Statistics, calculate the standard deviation of Delta 这个标准差被称为这个标准差被称为This standard deviation is called the Standard Error for Measurement 其主要代表了什么，重复性还是再现性？其主要代表了什么，重复性还是再现性

3、？Which does this primarily represent, Repeatability or Reproducibility?Variable N Mean Median StDev Delta 10 -0.410 -0.400 0.387 019 MSA-3Effective Date: 2/2004Copyright 2000-2002 Sigma Breakthrough Technologies, Inc. Used with permission.Copyright 2004 Cummins, Inc. All Rights Reserved.简单测量分析简单测量分析

4、 续续 Simple Measurement Study - continued 使用上面的标准差，什么是测量分布的良好估计值Using the standard deviation above, what would be a good estimate of measurement spread?Variable N Mean Median StDev Delta 10 -0.410 -0.400 0.387 2.32=Precision 0.387*6=Precision *6=Precision 假设公差是5 （上限下限）Suppose tolerance (upper spec -

5、lower spec) is 5. 精确度和公差的比率如何？What would be the ratio between the precision and the tolerance?019 MSA-4Effective Date: 2/2004Copyright 2000-2002 Sigma Breakthrough Technologies, Inc. Used with permission.Copyright 2004 Cummins, Inc. All Rights Reserved.简单测量分析简单测量分析 续续 Simple Measurement Study - cont

6、inued这个比率被称为精确度和公差比率 . 这个例子中这个例子中P/T 比率是：比率是： This ratio is known as the Precision to Tolerance (P/T) Ratio For this example the P/T Ratio is:这个可以被解读为：“46的规格公差被测量系统误差所占用了”This is interpreted as: “46% of the specification tolerance is taken up by measurement error”Variable N Mean Median StDev Delta 1

7、0 -0.410 -0.400 0.387 P T/. PrecisionTolerance 46%2325019 MSA-5Effective Date: 2/2004Copyright 2000-2002 Sigma Breakthrough Technologies, Inc. Used with permission.Copyright 2004 Cummins, Inc. All Rights Reserved.测量能力指数测量能力指数- P/T Measurement Capability Index - P/T 精确性和公差比率精确性和公差比率 Precision to Tole

8、rance Ratio 描述有描述有是被测量系统误差占用了是被测量系统误差占用了Addresses what percent of the tolerance is taken up by measurement error 同时包含了重复性和再现性同时包含了重复性和再现性 Includes both repeatability and reproducibility-操作者操作者 x 单位单位x 试验次数试验次数 Operator x Unit x Trial experiment 最佳情况最佳情况 Best case: 10% ； 可以接受可以接受 Acceptable: 30%通常以百分

9、比表示通常以百分比表示Usually expressed as percent注意注意：6.0标准差代表了标准差代表了99.73% 的测量系统变差。记住在做测量系统和以前分析结果的计算的测量系统变差。记住在做测量系统和以前分析结果的计算时，要经常检查这个值。时，要经常检查这个值。 这里用的这里用的6.0 是是AIAG 的工业标准。的工业标准。Note: 6.0 standard deviations accounts for 99.73% of MS variation. Always check the value used when making comparisons between m

10、easurement systems or previous analysis calculations. The use of 6.0 is the AIAG industry standard. P TToleranceMS/ 6.0*019 MSA-6Effective Date: 2/2004Copyright 2000-2002 Sigma Breakthrough Technologies, Inc. Used with permission.Copyright 2004 Cummins, Inc. All Rights Reserved.精确度和公差比率精确度和公差比率Preci

11、sion to Tolerance RatioProduct ToleranceLSLUSL测量系统变差measurement system variationP/T = 20%P/T = 100%P/T = 200%019 MSA-7Effective Date: 2/2004Copyright 2000-2002 Sigma Breakthrough Technologies, Inc. Used with permission.Copyright 2004 Cummins, Inc. All Rights Reserved.测量能力指数测量能力指数 % GR&R Measurement

12、Capability Index - % GR&R 描述有多少描述有多少是由测量系统误差造成的是由测量系统误差造成的 Addresses what percent of the total variation is taken up by measurement error 同时包含了重复性和再现性同时包含了重复性和再现性 Includes both repeatability and reproducibility-操作者操作者 x 单位单位x 试验次数试验次数 Operator x Unit x Trial experiment 作为一个目标，我们期望作为一个目标，我们期望%R&R 30%

13、 As a target, look for %R&R 尽管这样，如果流程能力不够好时，尽管这样，如果流程能力不够好时， P/T率会带给你一个错误的保证率会带给你一个错误的保证Even then, if process capability (Cpk) is not adequate, the P/T ratio may give you a false sense of security019 MSA-16Effective Date: 2/2004Copyright 2000-2002 Sigma Breakthrough Technologies, Inc. Used with perm

14、ission.Copyright 2004 Cummins, Inc. All Rights Reserved.%R&R 的应用的应用Uses of %R&R 对黑带来讲，对黑带来讲， %R&R 是一个很好的指标是一个很好的指标The %R&R is the best measure for the Black Belt-对于整体流程变差而言，此估计可以用于评估测量系统对于整体流程变差而言，此估计可以用于评估测量系统的表现的表现This estimates how well the Measurement System performs with respect to the overall

15、process variation -%R&R是进行流程改进研究的一个很好的估计是进行流程改进研究的一个很好的估计%R&R is the best estimate when performing process improvement studiesCare must be taken to use samples representing full, but typical, process variation019 MSA-17Effective Date: 2/2004Copyright 2000-2002 Sigma Breakthrough Technologies, Inc. U

16、sed with permission.Copyright 2004 Cummins, Inc. All Rights Reserved. 通常需要通常需要23名操作员名操作员 Generally 2 to 3 operators 通常需要对通常需要对10个单位的对象进行测量个单位的对象进行测量Generally 10 units to measure 每一个单位被每一个操作者测量每一个单位被每一个操作者测量23次次Each unit is measured 2-3 times by each operatorGage R&R 分析的设定分析的设定Gage R&R Study Setup019

17、 MSA-18Effective Date: 2/2004Copyright 2000-2002 Sigma Breakthrough Technologies, Inc. Used with permission.Copyright 2004 Cummins, Inc. All Rights Reserved.研究样本的选择研究样本的选择Study Sample Selection 样本应从流程正常变差的范围中取得样本应从流程正常变差的范围中取得Samples that span the normal variation of the process should be pulled fro

18、m the process 如果生产一种平均厚度为如果生产一种平均厚度为1.000及及“sigma为为0.010”的材料，则样本厚度范围应为的材料，则样本厚度范围应为0.980 - 1.020 （95 的范围延伸）的范围延伸）If you produce a material with a mean thickness of 1.000” and a sigma of 0.010”, use samples that go from 0.980 - 1.020” thick (95% of the range)019 MSA-19Effective Date: 2/2004Copyright

19、2000-2002 Sigma Breakthrough Technologies, Inc. Used with permission.Copyright 2004 Cummins, Inc. All Rights Reserved.研究样本的选择研究样本的选择Study Sample SelectionBE CAREFUL!-如果同一流程生产不同厚度的材料，那就将他们分组并进行如果同一流程生产不同厚度的材料，那就将他们分组并进行R&R研究研究If you produce different thickness of material with the same process, sub-g

20、roup them and perform the R&R study一个流程生产厚度为一个流程生产厚度为1.000”, 2.000”, and 4.000” 的材料的材料并且公差是并且公差是+/- 0.030，所有数值均由同一系统测量所得，所有数值均由同一系统测量所得A process produces 1.000”, 2.000”, and 4.000” materials all with a +/- 0.030” tolerance, all measured with the same system执行执行3次研究次研究 - 每个厚度一次每个厚度一次Perform 3 studies

21、 - one for each thickness-如果你将上述样本混合在一起，如果你将上述样本混合在一起， GR&R的值就被人为的的值就被人为的调低了调低了If you lumped the above samples together, the GR&R value will be artificially low019 MSA-20Effective Date: 2/2004Copyright 2000-2002 Sigma Breakthrough Technologies, Inc. Used with permission.Copyright 2004 Cummins, Inc.

22、All Rights Reserved.执行执行R&R 研究的步骤研究的步骤Procedure for Performing R&R Study 校准量具或者确认它们已经被校准过校准量具或者确认它们已经被校准过Calibrate the gage, or ensure that it has been calibrated 在在Minitab中建立数据矩阵，并确认它们已经被随机化中建立数据矩阵，并确认它们已经被随机化Set up your data matrix in Minitab. Be sure to randomize 让操作员以随机顺序测量样本让操作员以随机顺序测量样本Have op

23、erators measure the samples in random order 在在Minitab中分析数据来确定中分析数据来确定R&R的统计值的统计值Analyze data in Minitab to determine the statistics of R&R study-重复性重复性Repeatability-再现性再现性Reproducibility-上述之标准差上述之标准差Standard deviations of each of the above-% R&R-% P/T 分析结果并确定进一步行动分析结果并确定进一步行动Analyze results and dete

24、rmine follow-up action, if any019 MSA-21Effective Date: 2/2004Copyright 2000-2002 Sigma Breakthrough Technologies, Inc. Used with permission.Copyright 2004 Cummins, Inc. All Rights Reserved.样本大小的问题样本大小的问题Sample Size Issues-如果流程使用多个操作员，随机选择如果流程使用多个操作员，随机选择24个个If process uses multiple operators, choos

25、e 2-4 at random-如果流程只有一个操作员或者没有操作员，则分析时忽略操作员的影如果流程只有一个操作员或者没有操作员，则分析时忽略操作员的影 响（即忽略再现性的影响）响（即忽略再现性的影响）If process uses only one operator, or no operators, perform study withoutoperator effects (reproducibility effects ignored)-选择足够的样本，使得（样本数量）选择足够的样本，使得（样本数量）x （操作员数量）（操作员数量）15Select enough samples so

26、that (number of samples) X (number of operators) 15-如果不可行，选择测试的次数使得：如果不可行，选择测试的次数使得：If not practical or possible, choose number of trials so that:S- 样本数量样本数量O- 操作者数量操作者数量测试次数测试次数3 if S x O 测试次数测试次数3 到到 4if S x O 测试次数测试次数4 到到 5 if S x O 测试次数测试次数6 到到 8 if S x O Quality Tools Gage Study Gage R&R Study

27、(Crossed)019 MSA-24Effective Date: 2/2004Copyright 2000-2002 Sigma Breakthrough Technologies, Inc. Used with permission.Copyright 2004 Cummins, Inc. All Rights Reserved.Minitab 量具重复性和再现性研究量具重复性和再现性研究Minitab - Gage R&R Studies量具的选项：量具的选项：如果有公差，如果有公差，填入此处填入此处GAGE OptionsIf There Is A Tolerance For Thi

28、s Part -Put It Here研究的概要信息研究的概要信息General Info On Your Study放入放入6.0 (AIAG 标准）标准）Set to 6.0 (AIAG Standard)019 MSA-25Effective Date: 2/2004Copyright 2000-2002 Sigma Breakthrough Technologies, Inc. Used with permission.Copyright 2004 Cummins, Inc. All Rights Reserved.PercentPart-to-PartReprodRepeatGage

29、 R&R2001000% Contribution% Study Var% ToleranceSample Range0.100.050.00_R=0.0383UCL=0.1252LCL=0123Sample Mean1.000.750.50_X=0.8075UCL=0.8796LCL=0.7354123Part109876543211.000.750.50Operator3211.000.750.50PartAverage10 9 8 7 6 5 4 3 2 11.000.750.50Operator123Gage name:Date of study: Reported by:Tolera

30、nce:Misc:Components of VariationR Chart by OperatorXbar Chart by OperatorResponse by PartResponse by Operator Operator * Part InteractionGage R&R (ANOVA) for ResponseMinitab 输出输出Minitab - OutputExpected ToBe Out of Control%R&R& P/T RatiosGives us the 1st look atdiscriminationGraphical ViewOf Results

31、019 MSA-26Effective Date: 2/2004Copyright 2000-2002 Sigma Breakthrough Technologies, Inc. Used with permission.Copyright 2004 Cummins, Inc. All Rights Reserved.Minitab 输出输出Minitab - Output细节将在第二周讨论细节将在第二周讨论 We will discuss in detail during week 2.Two-Way ANOVA Table With Interaction Source DF SS MS

32、F PPart 9 2.05871 0.228745 39.7178 0.000Operator 2 0.04800 0.024000 4.1672 0.033Part * Operator 18 0.10367 0.005759 4.4588 0.000Repeatability 30 0.03875 0.001292Total 59 2.24913019 MSA-27Effective Date: 2/2004Copyright 2000-2002 Sigma Breakthrough Technologies, Inc. Used with permission.Copyright 20

33、04 Cummins, Inc. All Rights Reserved.Minitab 输出输出Minitab - Output从实务的角度看从实务的角度看 我们目前不需要这里的数据我们目前不需要这里的数据From the practical side we dont need anything from here, yet.从统计的角度看从统计的角度看 这些原始的统计数据会被套入计算这些原始的统计数据会被套入计算%R&R and %P/T 的公式的公式From the stat side . These are the raw stats that will be going intoth

34、e %R&R and %P/T formulasGage R&R %ContributionSource VarComp (of VarComp)Total Gage R&R 0.0044375 10.67 Repeatability 0.0012917 3.10 Reproducibility 0.0031458 7.56 Operator 0.0009120 2.19 Operator*Part 0.0022338 5.37Part-To-Part 0.0371644 89.33Total Variation 0.0416019 100.00019 MSA-28Effective Date

35、: 2/2004Copyright 2000-2002 Sigma Breakthrough Technologies, Inc. Used with permission.Copyright 2004 Cummins, Inc. All Rights Reserved.Minitab 输出输出Minitab - Output红色的圆圈重要的比率红色的圆圈重要的比率 Red Circles = Key Ratios%P/T % R&R Study Var %Study Var %ToleranceSource StdDev (SD) (6 * SD) (%SV) (SV/Toler)Total

36、 Gage R&R 0.066615 0.39969 32.66 79.94 Repeatability 0.035940 0.21564 17.62 43.13 Reproducibility 0.056088 0.33653 27.50 67.31 Operator 0.030200 0.18120 14.81 36.24 Operator*Part 0.047263 0.28358 23.17 56.72Part-To-Part 0.192781 1.15668 94.52 231.34Total Variation 0.203965 1.22379 100.00 244.76Numbe

37、r of Distinct Categories = 4019 MSA-29Effective Date: 2/2004Copyright 2000-2002 Sigma Breakthrough Technologies, Inc. Used with permission.Copyright 2004 Cummins, Inc. All Rights Reserved.我们该做怎样的决定？我们该做怎样的决定？What decision do we make?Minitab 输出输出 Minitab - Output %P/T 最佳：最佳：10，可接受：，可接受：30%Best case:

38、10% Acceptable: 30% R&R以以%R&R 30%为目标为目标As a target, look for %R&R Number of Distinct Categories = 4019 MSA-30Effective Date: 2/2004Copyright 2000-2002 Sigma Breakthrough Technologies, Inc. Used with permission.Copyright 2004 Cummins, Inc. All Rights Reserved.测量系统测量系统 要问的问题要问的问题Measurement Systems -

39、Questions to Ask 是否选择了正确的测量方法？该测量系统是否和关键的输入或输出相关联是否选择了正确的测量方法？该测量系统是否和关键的输入或输出相关联 Have you picked the right measurement system? Is it associated with either critical inputs or outputs? 精确性，准确性及趋势图的特征如何？精确性，准确性及趋势图的特征如何？ What do the precision, accuracy, and trend charts look like? 变差的来源有哪些，由那些测量误差？变差

40、的来源有哪些，由那些测量误差？ What are the sources of variation and what is the measurement error? 应该如何改善测量系统？应该如何改善测量系统？ What needs to be done to improve this system? 是否将结果告知了适当人员？是否将结果告知了适当人员？Have we informed the right people of our results? 谁负责这个测量系统谁负责这个测量系统 ？ Who owns this measurement system? 谁负责解决问题？谁负责解决问题

41、？Who owns trouble shooting? 系统中是否包括了控制计划？系统中是否包括了控制计划？ Does this system have a control plan in place? 培训的频率怎么样？这个频率是否足够？培训的频率怎么样？这个频率是否足够？ Whats the training frequency? Is that frequent enough? 是否与类似的系统一致？是否与类似的系统一致？ Do identical systems match?019 MSA-31Effective Date: 2/2004Copyright 2000-2002 Sigm

42、a Breakthrough Technologies, Inc. Used with permission.Copyright 2004 Cummins, Inc. All Rights Reserved.Metric 总结？总结？Metric Summary%P/T 相对于相对于客户的要求客户的要求，我的测量系统是否足够好我的测量系统是否足够好Is my measurement system good enough to measure to customer requirements?%R&R我的测量系统是否可以发现我的测量系统是否可以发现流程中流程中的变化？的变化？ Is my mea

43、surement systems good enough to detect changes in the process?清晰的分类清晰的分类 Distinct Categories针对可接受的不同的量具，我的测领系统是否有足够针对可接受的不同的量具，我的测领系统是否有足够的清晰的分类？的清晰的分类？Does my measurement system have a sufficient number of distinct categories to be an acceptable variable gage？019 MSA-32Effective Date: 2/2004Copyrig

44、ht 2000-2002 Sigma Breakthrough Technologies, Inc. Used with permission.Copyright 2004 Cummins, Inc. All Rights Reserved.Radom order builtEffective Date: 2/2004Copyright 2000-2002 Sigma Breakthrough Technologies, Inc. Used with permission.Copyright 2004 Cummins, Inc. All Rights Reserved.Measurement

45、Systems AnalysisSupplemental Material MSA Macro019 MSA-34Effective Date: 2/2004Copyright 2000-2002 Sigma Breakthrough Technologies, Inc. Used with permission.Copyright 2004 Cummins, Inc. All Rights Reserved.Create Data Matrix for MSA Different methods to create the matrix for collecting MSA data:-ma

46、nually - not efficient, error prone, hard to randomize-repeatedly use Minitab patterned data - -use Minitab DOE “Create Factorial Design” -GRRsetup - custom Minitab macro Example MSA: -plan an MSA study for 3 Operators, 10 Samples per trial, 2 Trials -We want to create a Minitab file with the follow

47、ing columnsPart or Sample, Operator, Trial, Measurement019 MSA-35Effective Date: 2/2004Copyright 2000-2002 Sigma Breakthrough Technologies, Inc. Used with permission.Copyright 2004 Cummins, Inc. All Rights Reserved.GR&R Setup: Factorial Design ApproachPrepare your Measurement System “Experiment”-Let

48、s use Minitabs experiment generators. -Choose . -Click the “General full factorial design” radio button. -Leave the “Number of factors” at two (2). 019 MSA-36Effective Date: 2/2004Copyright 2000-2002 Sigma Breakthrough Technologies, Inc. Used with permission.Copyright 2004 Cummins, Inc. All Rights R

49、eserved.GR&R Setup: Factorial Design ApproachAfter clicking on Designs, fill in -the Factors (Operator, Part) and -the Number of Levels for Operator = 3 (three measurement operators)for Part = 10 (ten parts)-choose two “replicates” -Click OK. 019 MSA-37Effective Date: 2/2004Copyright 2000-2002 Sigma

50、 Breakthrough Technologies, Inc. Used with permission.Copyright 2004 Cummins, Inc. All Rights Reserved.GR&R Setup: Factorial Design ApproachClick on the Factors button019 MSA-38Effective Date: 2/2004Copyright 2000-2002 Sigma Breakthrough Technologies, Inc. Used with permission.Copyright 2004 Cummins