SPCbilingual10547.pptx

《SPCbilingual10547.pptx》由会员分享，可在线阅读，更多相关《SPCbilingual10547.pptx（113页珍藏版）》请在淘文阁 - 分享文档赚钱的网站上搜索。

1、1统计过程控制统计过程控制工作坊工作坊 SPC Work Shopv精益管理顾问公司 -企业管理的好帮手 2基本统计概念 Basic SPC Concept3统计学统计学 (Statistics)v收集、整理、展示、分析、解析统计资料Collect,collate,display,analysis,explain of datav由样本推论母体/群体From sample to project the populationv能在不确定情况下作决策 Decision with uncertaintyv是一门科学方法、决策工具Scientific and decision tool基本统计概念 B

2、asic SPC Concept抽样推论4统计量 Statistical ValuevR 全距(range)v 算术平均数(arithmetic mean)vMd 中位数(median)vMo众数(mode)v2 方差/变异(variance)v 标准差(standard deviation)基本统计概念Basic SPC Concept5vR 全距(range)R=Xmax-Xminv 算术平均数(arithmetic mean)基本统计概念Basic SPC Concept6vMd 中位数(median)顺序数列中的中心项的数值The medial number in the series

3、vMo众数(mode)资料中出现最多的数值The number with highest frequency基本统计概念Basic SPC Concept7v2 方差/变异(variance)基本统计概念Basic SPC Concept8v 标准差(standard deviation)基本统计概念Basic SPC Concept9例：Case1,1,2,3,4,6,11R =10 =4Md =3Mo=1n 2=10.86 n-12=12.67n=3.29 n-1=3.56 基本统计概念Basic SPC Concept10数据的收集与整理Data Collection and Colla

4、tion分析分析测测试试行行动动ActionTestAnalysis数据数据Data结结Conclusion样本样本Sample群体群体PopulationSampling抽抽 样样11数据收集 Data Collection12数据整理 Data Collate-次数分配 Frequency Distribution13直方图(Histogram)14分布曲线 Distribution Curve15正态(常态)分布的曲线图Normal Distribution Curve16正态分布的性质Nature of Normal Distribution17数据 Data 直方图Histogram

5、 分布曲线 Distribution18直方图 Histogram 分布 Distribution正常型正常型Normal19直方图 Histogram 分布 Distribution偏向型偏向型 Bias20直方图Histogram 分布Distribution孤岛型孤岛型 Island21过程变差Process Variation22输入(材料)Input(Material)人 Man方法 Method过程(生产/装配)Process输出(产品)Output(Product)反馈(测量/检验)Feedback(Test/Insp.)测量系统MSA)材料Material过程变差 Proces

6、s Variation机器 Machine 23过程变差 Process Variationv输入材料 Material Input不同批次之间的差异 Lot to Lot Variation批次内的差异 Within Lot Variation随时间产生的差异 Variation in Time Series随环境而产生的差异 Variation due to Environment24过程变差 Process Variationv生产/装配 Production/Assembly设备及工装夹具的差异 Equipment and Tooling Variation随时间而产生的磨损，漂移等

7、Aging and Shift by time操作工之间的差异(如手工操作的过程)Variation between Operators(Manual Work)设置的差异 Different Setup环境的差异Different Environments25过程变差 Process Variationv输出产品 Output Product输出的产品随时间而产生的变化Product Variation by time series输出的产品随环境而产生的变化Product Variation due to environment26过程变差 Process Variationv反馈-测量

8、Feedback-Measurement由于测量用于过程中的所有组成部分，测量的变差会对过程的各个阶段产生影响Measurements are part of the process,variation in the measurement can affect the process in various stagesv偏倚 Biasv稳定性 Stabilityv重复性 Repeatabilityv再现性 Reproducibilityv分辨率 Resolution27持續改進過程循環的各個階段持續改進過程循環的各個階段2.維護過程維護過程 監控過程性能查找偏差的特殊原因並採取措施 3.改進

9、過程改進過程.改變過程從而更好地理解普通原因變差減少普通原因變差1.分析過程分析過程本過程應做些什麼?會出現什麼錯誤?本過程正在做什麼?達到統計控制狀態?確定能力計劃實施措施研究計劃措施實施研究計劃實施措施研究28过程变差 Process Variation范围范围范围范围每件产品的尺寸与别的不同Each product dimension has differences SPEC29过程变差 Process Variation范围范围范围但它们形成一个模型，若稳定，可以描述为一个分布The data are grouped to form the model.If stable,it can

10、 be descried as the distribution.30过程变差 Process Variation分布可以通用过以下因素来加以区分Distribution can be descried with various elements或这些因数的组合 Or the combination of these elements范围范围范围位置位置 Position分布宽度置分布宽度置 Spread形状形状 Form/Shape31过程变差 Process Variation预测Forecast范围 SPEC时间 Time如果只存在变差的普通原因，如果只存在变差的普通原因，随著时间的推移

11、，过程的输出随著时间的推移，过程的输出形成一个稳定的分布并可预测形成一个稳定的分布并可预测If there are only the common causes,and move in time series,the process outputs are predictable with stable distribution 目标直线Target Line32过程变差 Process Variation目标直线 Target Line预测 Forecast范围 Spec时间 Time如果存在变差的特殊原因，随著时间的推如果存在变差的特殊原因，随著时间的推移，过程的输出不稳定移，过程的输出不

12、稳定If the special cause is in the variation,in time series,the process become unstable？33过程控制和过程能力Process Control and Process Capability过程控制过程控制 Control Process 受控 In Control(消除了特殊原因)Eliminate the Special Cause时间 Time范围 SPEC不受制 Out Of Control(存在特殊原因)with Special Cause34过程变差 Process VariationvDistribu

13、tion can be described by:vLocation(typical value)位置(典型值)uSpread 分布宽度 uShape(pattern of variation)形状(变差的模式)LSLUSLTypical value典型值 SpreadSpread分布分布 35过程变差 Process Variationv过程的变差分两个类型 Two Categories特殊原因造成的变差 Special Causes普通原因造成的变差 Common Causes36过程变差 Process Variationv普通原因变差 Common Causes in Variatio

14、n影响过程中每个单位 affect each unit in the process是由所有不可分派的小变差源组成 Formed by various variation sources;and但遵循一个分布 follow a particular distribution通常需要采取系统措施来减小Use system corrective actions to reduce在控制图上表现为随机性random point in the control chartIf only common cause,the process output of a process is Predictable

15、如只有普遍原因，输出是可预计的过程。37过程变差 Process Variationv特殊原因变差 Special Causes in Variation间断的，偶然的，通常是不可预测的和不稳定的变差Discontinued,accidental,unpredictable and unstable非随机的图案 Non random pattern是由可分派的变差源造成该变差源可以被纠正Can be corrected,the cause is from the particular source38过程变差 Process Variationv工业经验建议为：Industrial Exper

16、ience局部措施 Local action通常用来消除变差的特殊原因Usually required to eliminate special causes of variation通常由与过程直接相关的人员实施 Usually taken by people close to the process大约可纠正15%的过程问题 Only 15%of process problem can be corrected39过程变差 Process Variationv工业经验建议为：Industrial Experience对系统采取措施 Actions on the system通常用来消除变差

17、的普通原因 Usually required to eliminate/or reduce common causes of variation,and be taken by management大约可纠正85%的过程问题 About 85%of process problem can be corrected40过程控制和过程能力Process Control and Process Capability过程能力 Process Capability规范上限受控且有能力符合规范(普通原因造成特变差)Control and capable to fit the spec(variation

18、due to Normal Cause)受控但没有能力符合规范 Control but without capability to fit the spec(普通原因造成变差太大)Large variation due to Normal Cause Spec时间 Time41过程控制和过程能力Process Control and Process Capability每个过程可以分类如下：Each process can be categorized as:Control or out of control?受控或不受控?Fit the customer requirements?是否有满足

19、客户要求?Meeting requirement满足要求In Control受控Not In Control 不受控Acceptance符合Case 1 第1类Case 3 第3类Not Acceptance不符合Case 2 第2类Case 4 第4类42过程控制和过程能力第1类(符合要求，受控)Case 1(Accept,In Control)是理想状况。为持续改进可能需要进一步减小变差Ideal Status,to reduce variation for continuous improvement.第2类(不符合要求，受控)Case 2(Not Accept,In Control)存

20、在过大的普通原因变差 with large variation from normal cause 短期内，进行100%检测以保护客户不受影响In short term,go for 100%inspection to reduce the effect for customer 必须进行持续改善以减少普通原因的变差 Must go for continuous improvement to eliminate the effect from common cause 43过程控制和过程能力Process Control and Process Capability第3类类(符合要求，不受控符

21、合要求，不受控)Case 3.Accept but OUT of controlv有相对较小的普通原因及特殊原因变差With less normal and special causes v如果存在的特殊原因已经明确但消除具影响可能不大经济，客户可能接受这种过程状况If the special causes are known but the cost may be high to eliminate them,customer may accept this process status.44过程控制和过程能力第4 4类类 (不符合要求，不受控不符合要求，不受控)Case 4(Not Acc

22、ept,OUT of Control)v存在过大的普通原因及特殊原因的变差Larger Normal and special causes in the variation v需要进行100%检测以保护客户利益 必须采取紧急措施使过程稳定，并减小变差Must go for contingency/containment actions to stable the process and reduce variations,100%inspection to protect customer 45过程控制和过程能力v判断一个过程是否满足规格要求：Judge the process can fit

23、 the spec requirementsCapability Index 能力指数能力指数-Cpk(For stable Process only 只用於穩定过程)Performance Index 性能指数性能指数 Ppkv判断一个 过程是受控还是不受控用：Judge the process is in control or out of control控制图控制图 Control Charts46控制图 Control Charts 为判断制程变差的为判断制程变差的普通原因或特殊原普通原因或特殊原因因的主要统计分析工具的主要统计分析工具To judge the process vari

24、ation with COMMON or SPECIAL cause as the major statistical analysis tool47基本概念 Basic Conceptv控制图中包括三条线 3 lines in Control Chart控制上限(Upper Control Limit;UCL)中心线(Center Line;CL)控制下限(Lower Control Limit;LCL)UCLCLLCL48控制图的种类 Type of Control Charts v计量型数据的控制图 Variable Control ChartsXbar-RXbar-SX-MRv计数型数

25、据的控制图 Attribute Control Chartspnpcu49控制圖的選用程序控制圖的選用程序注:本圖假設測量系統已經過評價並且是適用的50控制图的使用策划Planning for the Control Charts作控制图需要按以下步骤：Follow the necessary steps:v计划 Planningv资源 Resourcesv评估和改进 Evaluation and Improvement51控制图的使用策划Preparation for the use of Control Charts建立适于采取措施的环境-公司文化 Set the right enviro

26、nment to take actions Company Culture确定过程 Define the process确定待管理的特性(如:特殊特性)，考虑Determine characteristics(e.g.Special Characteristics)that need to be managed Shall Consider:v客户的要求 Customer Requirementsv当前及潜在的问题区 Existing and potential problem areav特性间的相互关系 Interrelation between characteristics确定测量系统

27、Define the Measurement System MSA使不必要的变差最小化 Reduce the variations in those avoidable area52控制图的使用策划Preparation for the use of Control Charts过程过程 ProcessProcess输出输出 OutputOutput输入输入 InputInput使用什么方式？使用什么方式？(材料材料/设备设备)What?What?(Equipment/Raw Material(Equipment/Raw Material)谁进行谁进行?(能力能力/技能技能/培训培训)Who?

28、Who?(Competence,ability,training)(Competence,ability,training)如何做？如何做？(方法方法/程序程序/技术技术)How?How?(Method/Procedure/Technique)(Method/Procedure/Technique)使用的关键准则是什么？使用的关键准则是什么？(测量测量/评估评估)Key measurement?Key measurement?(Measuring/Evaluation)(Measuring/Evaluation)53Xbar-R Chart 图Average and Range Chart 平

29、均数平均数-全距全距控制图控制图工业界最常使用的计量值控制图The most common chart for the industry54建立Xbar-R控制图的步骤Set up the steps for X-R Charts1.收集数据收集数据 Collect Data v选择子组大小，频率和数据Select the subgroup size,frequency and numbers of subgroup 子组大小 Subgroup size:v一个子组内该单元之间出现变差机会小(如45件连续于单一过程的产品)variation among units within the sub

30、group are small(e.g.4 to 5 consecutively produced pieces by single production stream)每个子组内的变差是普通原因造成每个子组内的变差是普通原因造成 Variation within each subgroup reflects COMMON causev代表短时间内零件间变产 Represent the piece-to-piece variation within short period of timev不正常变差=过程发生变化 unusual variation between subgroups=cha

31、nges in process55建立Xbar-R控制图的步骤Set up the steps for X-R Chartsv选择子组大小，频率和数据Select the subgroup size,frequency and numbers of subgroup 子组频率 Subgroup frequency:v检查过程经过一段时间的变化，如:换班、温升趋势、材料变更 To detect the change of process over time,e.g.changes due to work shift difference,warm-up trend,material lots56

32、子组数大小 Number of Subgroups:v确保变差的主要原因有机会出现(25子组包含100个单值)Major sources of variation have had an opportunity to appear(25 or more subgroups containing 100 or more readings)v可利用现有数据加速研究进度Can use existing data to speed up the process建立Xbar-R控制图的步骤 Set up the steps for X-R Charts571.建立控制图及記錄原始數據Set up con

33、trol chart and raw data gathering建立Xbar-R控制图的步骤 Set up the steps for X-R Charts582.计算每个子组的X平均值和极差R Calculate the Average Range and Process average建立Xbar-R控制图的步骤 Set up the steps for X-R Charts59UCLCLLCL3.子组数足夠時(25子组包含100个单值),計算及繪成控制線 試驗控制限When number of subgroup is adequate(25 or more subgroups cont

34、aining 100 or more readings),calculate and draw the control limit trial control limits建立建立Xbar-R控制图的步骤控制图的步骤 Set up the steps for X-R Charts60建立Xbar-R控制图的步骤 Set up the steps for X-R Charts4.过程控制解释 Explain the Process Controlv分析极差图(R图)上的数据点 Analyse the R chart first识别并标注特殊原因(R图)Identify and mark the

35、special causes糾正及預防再發生 Correct and prevent it form recurringv消除受已識別已識別的特殊原因影響的子組Exclude all subgroups affected by the identified special cause v重新计算控制极限(R图)Recalculate the R control limits 61當极差受控 過程分佈是穩定的,分析平均值图(Xbar图)When the ranges are in statistical control,process spread isstable,thenv看过程的位置是否改

36、变 x 图 Check if the process location is changing overtime x chartv为继续控制延长控制限 Extend the Control Limits for Process Control建立Xbar-R控制图的步骤 Set up the steps for X-R Charts62控制图-过程的声音Control Charts-the voice of the processv控制图可以给我们提供出出现了哪种类型的变差的线索，供我们采取相应的措施The control chart can provide us the hints of w

37、hat types of variations,for the right corrective actions63控制图上的信号解释 Explain The Signals on the Control Chartsv有很多信号规则适用于所有的控制图(Xbar图和R图)，主要最常见的规则 有以下3种：Many types of signals and rules in the XR charts,mainly 3 types:规则1：超出控制限的点 Rule#1:Points Exceed the control limitsUCLLCL64控制图上的信号解释Explain The Sign

38、als on the Control Charts规则2：连续7点在中心线一侧Rules#2:continuous 7 points in one side of the centre linerunUCLLCL65控制图上的信号解释 Explain The Signals on the Control Charts规则3：连续7点升或下降 Rule#3:continuous 7 points in descending or raisingtrendUCLLCL66过制图上的信号解释 Explain The Signals on the Control Charts多于2/3的点落在图中1/

39、3以外。More than 2/3 of points outside the 1/3 zone 67过制图上的信号解释 Explain The Signals on the Control Charts呈现有规律变化。Regular Cyclical Trend68改变控制时限 Change In Control Limit将这些的控制限画到控制图上，并作为不断进行的过程制的基础。Plot these new control limits on the chart as the basis forongoing process control 只要过程的均值和极差保持受控 可将控制限延长以于

40、以后的时期 As long as the process remains in control for both and range ongoing limits can be extended for additional periods.69圖形圖形 Patterns極差圖極差圖 Range Chart均值圖均值圖 Average Chart超出控制限的點 Points beyond the control limits控制限繪錯 分佈寬度增加(零件間變化)測量系統未有足夠分辨或已變化Mis-plotted of control limits Increase or decrease of

41、 spread of distribution(piece-to-piece variation)Change in Measurement system or lack of discrimination 控制限繪錯 過程已變化 測量系統己變化Mis-plotted of control limits Process has shifted Change in Measurement system 70圖形圖形 Patterns極差圖極差圖 Range Chart均值圖均值圖 Average ChartRun 輸出值的分佈寬度增加：無規律的成因，如設備異常或夾具鬆動 測量系統改變 Great

42、er spread of output value due to:irregular cause such as equipment malfunction or loose fixturing shift in process elements such as new,less uniform raw material lots Change in Measurement system 過程均值已變化 測量系統逐漸改變Process average has changed Change in Measurement system such as drift 71 圖形圖形 極差圖極差圖 均值

43、圖均值圖 明顯非隨機圖形 顯著多於2/3以上的點落在中線外：控制限繪錯 過程或抽樣方法會被分層：子組系統化包括兩個或以上的過程流過程均值變化 數據被編輯 顯著少於2/3以上的點落在中線外：過程或抽樣方法形成連續的分組中包括多個具不同變化的過程流 顯著多於2/3以上的點落在中線外：控制限繪錯 過程或抽樣方法會被分層：子組系統化包括兩個或以上的過程流過程均值變化 數據被編輯 顯著少於2/3以上的點落在中線外：過程或抽樣方法形成連續的分組中包括多個具不同變化的過程流 如果存在幾個過程流，應分別識別和追蹤 72改变控制时限 Change In Control Limit估计过程的标准偏差用现在的子组容

44、量计算：Estimate the process standard deviation.Using the existing subgroup size calculate:=R/dR/d2 2式中 R为子组极差的均值(在极差受控时期)，d2 随样本容量变化的常数：Where R is the average of the subgroup ranges(for periods with the ranges in control)and d2 is a constant varying by sample size:n2345678910d21.131.692.062.332.532.702

45、.852.973.0873改变控制时限 Change In Control Limit按照新的子组容量查表得到系数d2、D3、D4和A2，计算新的极差和制限：Using the tabled factors for d2,D3,D4 and A2 based on the new subgroup size,calculate the new range and control limits:=R Rnewnewd d2 2UCLUCLR R=D D4 4 R RnewnewLCLLCLR R=D D3 3 R RnewnewUCLUCLX X=X+AX+A2 2 R RnewnewUCLUC

46、LX X=X-AX-A2 2 R Rnewnew74改变控制时限 Change In Control Limit将这些的控制限画到控制图上，并作为不断进行的过程控制的基础。Plot the new control limits on the chart as the basis for ongoing process control.只要过程的均值和极差保持受控 可将控制图的目的是合理、经济的控制状态As long as the process remains in control for both averages and rangesOngoing limits can be extend

47、ed for additional periods75有关控制的最后概念用于进一步的考虑 1.在一个生产过程中永远无法达到一种完美的控制状态。A perfect state of control is never attainable in a production process.2.过程控制图的目的是合理、经济的控制状态。The goal of the process control charts is reasonable and economical state of control.3.如果一张控制图上从来不出现失控点，我们将严肃地查问该操作是否应画图。If a chart neve

48、r went out of control we would seriously question whether that operation should be charted.4.对于车间目的来说，一个受控的过程即是仅有很少百分比的点失控并且对失控点采取过适当的措施。For shop purposes a controlled process is considered to be one where only a small percentage of the points go out of control and where out-of-control pointsare fol

49、lowed by proper action.76请完成请完成 Xbar-R控制图的制作练习控制图的制作练习Please Complete Control Chart Drawing练习 Exercise77控制图-过程的声音v控制图可以区分出普通原因变差和特殊原因变差Control Chart can distinguish the normal causes and the special causes特殊原因变差要求立即采取措施Special Causes needs immediate actions减少普通原因变差需要改变产品或过程的设计To Reduce the normal ca

50、uses need product or process redesign78控制图-过程的声音v错误的措施 Wrong Actions试图通过持续调整过程参数来固定住普通原因变差，称为过度调整，结果会导致更大的过程变差造成客户满意度下降Continue to adjust the process parameters to fix the normal process variations-over adjustments only cause larger variations and customer unsatisfied试图通过改变设计来减小特殊原因变差可能解决不了问题，会造成时间和