《计数型MSA.pptx》由会员分享，可在线阅读，更多相关《计数型MSA.pptx（36页珍藏版）》请在taowenge.com淘文阁网|工程机械CAD图纸|机械工程制图|CAD装配图下载|SolidWorks_CaTia_CAD_UG_PROE_设计图分享下载上搜索。
1、Measurement Systems Analysis (MSA) Attribute特性数据特性数据MSAThe Breakthrough Strategy And Measurement System Evaluation Measurement Systems Analysis will help us decide if we can adequately measure our output characteristic测量系统分析有助于确认是否能够胜任测量输出特性的要求 Then after weve improved the process, we will need to a
2、ssess it again and potentially need to improve it/在改善过程以后，我们需要再次评价测量系统，以及潜在的测量系统的改进1. Select Output Characteristic2. Define Performance Standards3. Validate Measurement System确认测量系统4. Establish Baseline Process Capability5. Define Performance Objectives6. Identify Variation Sources7. Screen Potentia
3、l Causes8. Discover Variable Relationships9. Establish Operating Tolerances Implement Improvements10.Validate Measurement System确认测量系统11.Determine Final Process Capability12.Implement Process Controls1Module ObjectivesBy the end of this module, the participant should be able to: Discuss the need for
4、 an Attribute Measurement System Analysis探讨属性数据测量系统分析所需条件 Describe the types of Attribute Measurement System Analyses and when to use them 叙述属性数据测量系统分析以及何时使用 Perform an Attribute Measurement System Analysis 实施属性数据测量系统分析2Typical Measurement SystemsAttribute MSA Examples- Data collection form- Survey-
5、 Customer satisfaction- Academic grades- Plug gages (go/no-go)- Thread gages (go/no-go)- Visual defect checkNote that often Attribute Measurement Systems involve human judgment Is it reliable for use in decision making?/注意属性数据MSA经常包含人员的判定判定可靠吗？3Purpose Of Attribute MSA Accuracy checks 准确性检查- Assess
6、standards against customers requirements 评价标准和顾客要求对比- Identify how well Measurement System conforms to a “known master” /识别测量系统符合”已知标准”的程度 Precision checks 精密性检查- To determine if inspectors (Appraisers) across all shifts, machines, lines, etc use the same criteria to evaluate items Reproducibility /
7、 确认检查员(评价者)所有班次、设备、生产线是否使用相同标准评价 再现性再现性- To quantify the ability of inspectors (Appraisers) or gages to accurately repeat their inspection decisions Repeatability / 量化检查员(评价者)或仪器准确重复检查结果的能力重复性重复性 To identify how well inspectors/gages measure a known master (possibly defined by the customer) to ensur
8、e no misclassification occurs /识别检查员/仪器测量”已知标准”(可能是顾客定义)确保没有错误分类的能力- How often operators decide to ship truly defective product 多久操作员会将不良品判为合格品？- How often operators do not ship truly acceptable product 多久操作员会将合格品判为不良品？ To determine areas where 确定以下范围- Training is needed 是否需要培训- Procedures or Contro
9、l Plans are lacking 是否缺乏程序或控制计划- Standards are not clearly defined 标准是否未定义- Gage adjustment or correlation is necessary 仪器调整或相互性是必须的4Attribute Data With Unordered Categories非顺序型的属性数据 Kappa Suppose that invoice quality is a key to the process Throughput. In other words, if an invoice is incomplete, t
10、he rework required impacts the quantity that can be processed in a day. Two Appraisers are asked to independently evaluate ten invoices randomly selected from different days. The results of the study are shown below:/假定发票质量是过程输出的关键，换句话说，如果一张发票不完整，再作业就会影响当天的产量。2个评价者独立评价随机在不同日期抽取的10张发票。结果如下所示：Invoice
11、#Appraiser 1 Appraiser 2 Agreement? 1BadBadY 2GoodBadN 3GoodGoodY 4BadBadY 5GoodGoodY 6GoodGoodY 7GoodBadN 8GoodBadN 9GoodGoodY 10GoodGoodY We could simply look at the percent of the time they agree as metric for between Appraiser agreement: 我们可以只考虑评价者判定一致的百分率作为衡量指标 But what would that not take into
12、 account? 但是还有什么没有考虑？_ % agree _ % disagree5Kappa compares the proportion of agreement between Appraisers after removing agreement by chance / Kappa 比较了评价者之间去除了的偶然性的一致性比率The proportion that the judges are in agreement is Pobserved /判断一致的比率是Pobserved The proportion expected to occur by chance is: 由于偶
13、然性一致的比率 Pchance = (P Insp1 Good) (P Insp2 Good) + (P Insp1 Bad)(PInsp2 Bad) 1. What would the Kappa be if the judges agreed on every part?/如果每个 零件的判定都一致， Kappa是多少？2. How would you interpret a Kappa of zero?怎么解释Kappa为0？3. What would be the Kappa if the judges never agreed? 如果判定都不一 致Kappa又是多少？Definiti
14、on Of KappachancechanceobservedP1PPKappaLets walk through the calculation by hand.6How To Calculate Kappa By HandBuild contingency tableSum columns and rowsCalculate Pobs by adding diagonalCalculate PchanceCalculate KappaPchance = (PR1bad)(PR2bad)+(PR1good)(PR2good) = (.2)(.5)+(.8)(.5) = .5 Pobs - P
15、chanceKappa = 1 - Pchance=(.7-.5)/(1-.5)= 0.4.2.8.5.5 BadGoodAppraiser 1GoodBadAppraiser 2Pobs = .5 + .2 = .7Add3/10 = .32/10 = .20/10 = 05/10 = .5Add Add AddAddHow is this interpreted?7How To Interpret Kappa ResultsThe general rule for interpreting Kappa results are as follows: 解释Kappa的规则通常如下 0 Non
16、-random disagreement 非随机的不一致 0.9 Excellent Measurement System 极好的测量系统Remember, consider your Measurement System and how well the above criteria might apply.8Attribute Data Kappa Exercise1.Your company produces documents that are filled with alpha-numeric characters. If a document has one or more num
17、erals (0-9), it is defective. Your mission is to identify the defective documents./你公司的产品是字母与数字混合编排的文件。如果文件中有1个或多个数字(0-9)，就是不合格品。你的任务就是识别不合格的文件。2.In teams of two, appoint one person to be the data collector The other person is the inspector. The data collection sheet is located on the following page
18、. 2人一组，任命其中一个为数据收集员，另一个是检查员，数据收集表在后面一页中3.The document number is located under each of the 20 documents.The documents will be visible for three seconds before automatically advancing. The exercise will be finished in 60 seconds(three seconds/part).共有20个文件，每个显示3秒，然后自动进行下一个。练习在60s内完成(3s/部件).4.Each Appr
19、aiser will perform two trials 每个评价者重复2次5.Perform analysis in MINITAB per the next slides 进行MTB分析9SVQ O6Q NITQX N ZIZBM QTPZHJKM23! Example !Sample #Kappa ExerciseBegin Kappa exercise10Kappa Exercise Data Collection FormP a rt N o .T ria l 1 (P /F )T ria l 2 (P /F )A g r e e / D is a g r e e ?T ria l
20、 1 (P /F )T ria l 2 (P /F )A g r e e / D is a g r e e ?J u d g e 1 T ria l 1J u d g e 2 T ria l 2A g r e e / D is a g r e e ?1234567891 01 11 21 31 41 51 61 71 81 92 0C o m p a re J u d g e sJ u d g e 1J u d g e 2NOTE: Part count of 20 used only for demonstration. Sample size guidelines found on pag
21、e 13.11Kappa Analysis In MINITABPut data into MINITAB, each judges trial in a separate columnTo analyze, go to Stat Quality Tools Attribute Agreement Analysis123Enter Judges (Appraisers) and quantitySelect “Results” button and click last option412Attribute Gage R&R StudyAttribute Gage R&R Study for
22、J1T1, J1T2, J2T1, J2T2Within AppraiserAssessment AgreementAppraiser # Inspected # Matched Percent (%) 95.0% CI 1 20 19 95.0 ( 75.1, 99.9)2 20 18 90.0 ( 68.3, 98.8)# Matched: Appraiser agrees with him/herself across trials. Kappa Statistics Appraiser Response Kappa SE Kappa Z P(vs. 0) 1 b 0.8997 0.22
23、36 4.0238 0.000 g 0.8997 0.2236 4.0238 0.0002 b 0.8000 0.2236 3.5777 0.000 g 0.8000 0.2236 3.5777 0.000Between AppraisersAssessment Agreement# Inspected # Matched Percent (%) 95.0% CI 20 18 90.0 ( 68.3, 98.8)# Matched: All Appraisers assessments agree with each other.Kappa Statistics Response Kappa
24、SE Kappa Z P(vs. 0) b 0.8833 0.0913 9.6756 0.000g 0.8833 0.0913 9.6756 0.000MINITAB Output Session WindowSince Kappa 0.70 for both Appraisers, both agree well with themselves/ 2个评价者Kappa 0.70，一致度较好 Adequate repeatability/足够的重复性Since Kappa 0.70 for the Between Appraisers,/ 评价者间的Kappa 0.70 Adequate re
25、producibility/足够的再现性 13Guidelines For Kappa Studies Planning 策划- Sample Size样本大小100 samples with two trials per Appraiser / 100个样品，每个评价者重复2次If only 50-100 samples are available, do three trials per Appraiser/如果只有50-100个样品，每个评价者重复3次 50, understand that you may need very high Kappas to have adequate c
26、onfidence/ 0)1 -0.058824 0.235702 -0.24957 0.59852 0.035714 0.235702 0.15152 0.43983 -0.107692 0.235702 -0.45690 0.67614 0.000000 0.235702 0.00000 0.50005 0.437500 0.235702 1.85616 0.0317Overall 0.033058 0.132073 0.25030 0.4012Kendalls Coefficient of Concordance Coef Chi - Sq DF P0.670034 10.0505 5
27、0.0738MINITAB Output Session WindowYou still get the Kappa Statistics, but you wont need to use them here/还是能得到Kappa统计表，但是用不到它们First we need to look at the p-value. Remember, KCC 0, indicates association, if the p-value is small, we accept that KCC 0. The p-value tells you the probability that some
28、association was found purely by chance. This is saying that there is a 7.4% chance a non-zero KCC (some level of association) was found by chance alone. You decide based on the amount of risk youre willing to take. A value like 0.05is typical. / 首先要看p值，记住KCC0，表示有联系，如果p值很小，我们接受KCC0。P值表示纯粹由于偶然而发生联系的可能
29、性。这里表示有的纯粹由于偶然而发生非0 KCC的机会是7.4%(互相联系的水平)。你的决定取决于你愿意承担的风险。通常设为0.05The KCC is fairly low here, see criteria onnext slide这个KCC比较小，看下一页的标准20How To Interpret Kendalls Coefficient Of Concordance In general, KCC can vary from 0.0 to 1.0/ KCC值一般可以从0变到1 The higher the value, the higher degree association amo
30、ng the assessments made by the Appraisers /值越大，评价者的一致性越高 How close to 1.0 is needed?需要和1.0多接近？- Although very situational dependent, the general guidelines are as follows:虽然有很多不同状况，但是一般规则如下0-0.7 Low degree of association, Measurement System needs attention 低程度联系，需关注测量系统0.7-1.0 Generally acceptable 通
31、常可接收As stated on the prior page, the p-value for KCC should also be low, generally less than 0.05 This reduces your risk of getting an acceptable KCC just by random chance.正如前页所说，KCC的p值应该较小，通常比0.05小 减少由于偶然造成接收的KCC的风险21Guidelines For Kendalls Studies Planning 策划- Sample size 样本大小More is better As you
32、r sample size increases, your confidence intervals around your KCC decrease 越多越好 样本大小增大，和KCC有关的置信区间范围减小Collect as many samples as practically possible, 20 minimum is a guideline, 30 is best / 收集尽可能多的样品，至少20，最好 30 Perform at least two trials per Appraiser 每个评价者至少重复2次- Sample part selection 样品选择Parts
33、in the study should represent the full range of variation and thus utilize the full range of the rating scale 研究的样品能代表变差的整个范围，这样就能使用整个范围的分类标准 Execution 实施- Parts should be rated in random order independently (no comparisons) /部件应该以随机顺序评价- Study should be blind研究应该是盲测- Rating time should be similar t
34、o that “normally” used评价时间也和 正常使用近似22Guidelines For Kendalls Studies(Contd) Analysis 分析- Prior to reviewing the KCC value, check to see that the p-value is low(generally 0.05) If it is not, add more samples to the study or add another trial / 在看KCC值之前，检查p值是否较小(通常0.05) 如不是，增加样品或增加1次重复- Review the rep
35、eatability portion first (Within Appraiser), if an Appraisers KCC is very low, they may need improvement (see Improvement below) / 先检查重复性部分(评价者内),如果评价者的KCC很低，那就需要改进(看下面的改进) - For Appraisers that have acceptable repeatability, review the reproducibility portion (Between Appraiser) / 如重复性可接收，检查再现性部分(评
36、价者间)- If a “Gold Standard” is available (ratings of the samples known by some other means as being “correct”), compare each Appraiser to them for “calibration”/ 如果”标准”已知(已知的正确的等级)，通过比较每个评价者和已知值来校正Use the field in MINITAB, “Known Standard Attribute” /用MTB中“Known Standard Attribute”一栏 Improvement 改进-
37、If the Within Appraiser KCC scores are low, that Appraiser may need training. Do they understand the rating scale? Are the instructions clear to them? / 评价者内 KCC值较低，那么就需要培训，他们是否理解分类标准？是否清楚作业指导书？- If the Between Appraiser KCC scores are low, each Appraiser may have a differing definition of the ratin
38、g scale A standardized definition can improve this situation /评价者间 KCC 值较低，那么每个评价者对零件分类的定义不同标准定义可以改变这种状况- If improvements are made, the study should be repeated to confirm improvements have worked /如果已进行改进，还需要再次研究以确认改进有效23Objectives ReviewThe participant should be able to: Discuss the need for an At
39、tribute Measurement System Analysis探讨属性数据测量系统分析所需条件 Describe the types of Attribute Measurement System Analyses and when to use them 叙述属性数据测量系统分析以及何时使用 Perform an Attribute Measurement System Analysis 实施属性数据测量系统分析24AppendixCopyright 2001-2005Six Sigma Academy International, LLCAll rights reserved; f
40、or use only in compliance with SSA license. Attribute MSA MethodStep 1:Select a minimum of 30 parts from the process 至少选择30个部件- 50% of the parts in your study should have defects/ 50%有缺陷- 50% of the parts should be defect-free/ 50%没有缺陷- If possible select border line (or marginal) good and bad sampl
41、es / 尽可能选择界限附近的好的和不好的样品Step 2:Identify the appraisers which should be qualified / 选择有资格的评价者Step 3:Have each appraiser, independently and in random order, assess these parts and determine whether or not they pass or fail / 每个评价者随机评价部件，确认是否通过或不通过Step 4:Enter the data into Excel or MINITAB to report th
42、e effectiveness of the attribute measurement system / 输入Excel或MTB 报告属性数据MSA的有效性Step 5:Document the results. Implement appropriate actions to fix the inspection process if necessary / 结果文件化，如有必要，进行相应的改善Step 6:Re-run the study to verify the fix / 再运行MSA，验证改善有效Note: A 30-piece sample will yield an esti
43、mate of appraiser efficiency and capability which has a fair amount of uncertainty. Typically a larger sample is not needed because the appraisal process is obviously ineffective. The Excel spreadsheet can handle up to 100 samplesCopyright 2001-200526Attribute MSA Supplemental FilesCopyright 2001-20
44、05Six Sigma Academy International, LLCAll rights reserved; for use only in compliance with SSA license. Attribute MSA - Excel Method Allows for R&R analysis within and between appraisers 可以评价者内和评价者间的重复性和再现性研究 Test for effectiveness against standard 检验和标准值比较的有效性 Limited to nominal data at two levels
45、仅限于2水平的nominal 数据Copyright 2001-200528DATE:1/4/2001Attribute Legend5 (used in computations)NAME:Acme Employee1 PassPRODUCT:Widgets2 FailBUSINESS:Earth ProductsKnown PopulationSample #AttributeTry #1Try #2Try #1Try #2Try #1Try #21PassPassPassPassPassPassPass2PassPassPassPassPassPassPass3PassPassPassP
46、assPassPassPass4PassPassPassPassPassFailPass5FailFailFailFailFailPassFail6FailPassPassPassPassPassPass7PassPassPassPassPassPassPass8PassPassPassPassPassPassPass9FailFailFailFailFailFailFail10PassPassPassPassPassPassPass11PassPassPassPassPassPassPass12PassPassPassPassPassPassPass13PassPassPassPassPas
47、sPassPass14PassPassPassPassPassFailPass15FailFailFailFailFailPassFail16PassPassPassPassPassPassPass17PassPassPassPassPassPassPass18PassPassPassPassPassPassPass19FailFailFailFailFailFailFail20PassPassPassPassPassPassPass21PassPassPassPassPassPassPass22PassFailFailPassPassPassPass23PassPassPassPassPas
48、sPassPass24PassPassPassPassPassFailPass25FailFailFailFailFailFailFail26PassPassPassPassPassPassPass27PassPassPassPassPassPassPass28PassPassPassPassPassPassPass29FailFailFailFailFailFailFail30PassPassPassPassPassPassPassOperator #1Operator #2Operator #3Attribute MSA ExampleOpen file: MSA-Attribute.xl
49、sCopyright 2001-200529Scoring Example 100% is target for all scores - 57.14%SCREEN % EFFECTIVE SCORE vs. ATTRIBUTE - 42.86% APPRAISER SCORE - % SCORE VS. ATTRIBUTE - Copyright 2001-200530Statistical ReportStatistical Report - Attribute Gage R&R StudyDATE: 1/4/01NAME: Acme EmployeePRODUCT: WidgetsBUS
50、INESS: Earth Products% Appraiser1%Score vs Attribute2SourceOperator #1Operator #2Operator #3Operator #1Operator #2Operator #3Total Inspected303030303030# Matched303025282924False Negative (operator rejected good product)100False Positive (operator accepted bad product)111Mixed00595% UCL100.0%100.0%9