1、统计流程控制简介 (Intro To SPC),连接控制图方法与流程改进方法Link Control Chart methods to the Process Improvement Methodology TM 讨论不同类型的偏差 Discuss different types of variation 介绍各种类型的控制图Introduce various types of Control Charts 讨论如何解释控制图Discuss the interpretation of Control Charts,目的 Objectives,连接控制图方法与流程改进方法Link Control
2、 Chart methods to the Process Improvement Methodology TM 讨论不同类型的偏差 Discuss different types of variation 介绍各种类型的控制图Introduce various types of Control Charts 讨论如何解释控制图Discuss the interpretation of Control Charts,目的 Objectives,小秘密- SPC 只是作为一个控制工具来使用 Myth - SPC Is Only Used As A Control Tool,测量 MEASURE,
3、分析 ANALYZE,改进 IMPROVE,控制 CONTROL,漏斗效应 The Funneling Effect,我们是否应该采取行动？ Should we take action?,每天我们都被数据淹没,而且不得不作出决定 Every day we are flooded by data and we are forced to make decisions,工厂产量下降 Plants Output Decreases By 4%美国贸易赤字增加亿US Trade Deficit Rises By $40Billion 某公司获利比上季度降低亿 Company Xs Earnings A
4、re Off $240Million From Previous Quarter,我们需要解释数据的方法 We Need Ways to Interpret Data,今天采集什么样的数据？ What Type Of Data Is Collected Today?,制造业 Manufacturing : _ 非制造业 Non-Manufacturing _,如何分析数据？ How Is It Analyzed?,制造业 Manufacturing : _ 非制造业 Non-Manufacturing _,得知数据好坏后该当如何？ What Happens If It Is Bad / Goo
5、d?,制造业 Manufacturing : _ 非制造业 Non-Manufacturing _ _,客户需求下限 Lower “Customer” Requirement,这一方法 THIS METHOD告诉你关于客户的需求 Tells you where you are in regards to customers needs不告诉你怎么满足用户需求及下一步怎么办It will NOT tell you how you got there or what to do next,客户需求上限 Upper “Customer” Requirement,我们管理数据的方式过去(历史来讲)的方
6、式 The Way We Manage Data Historically,不用管它,不会坏的 Leave It Alone.It Aint Broke,痛苦,受累 Pain & Suffering,痛苦,受累 Pain & Suffering,这一方法导致何种管理行为？ This method causes what type of management behavior?,客户需求下限 Lower “Customer” Requirement,客户需求上限 Upper “Customer” Requirement,我们管理数据的方式历史来讲的方式 The Way We Manage Dat
7、a Historically,不用管它,不会坏的 Leave It Alone.It Aint Broke,痛苦,受累 Pain & Suffering,痛苦,受累 Pain & Suffering,2,3,Scrap Level (%)废品率,1,1996,Celebration Time,工厂废品率为年度最低的The factory scrap level is at a year low of 2%经理给工厂颁奖Manager presents an award to the plant 在餐厅进行庆祝：每人都可分享免费皮萨饼和饮料Ceremony in the cafeteria: p
8、izza and refreshments for all! “每人都应为他们的成就骄傲” “Everyone should be proud of what theyve accomplished”.,Derived from Understanding Variation: The Key To Managing Chaos, Donald J. Wheeler, SPC Press. 1993.,年月APRIL 1996,J F M A,2,3,1,1996,经理希望能将发出去的奖收回来 Manager wants to take back award,废品率连续三个月持续增长Three
9、 consecutive months of scrap increases. 经理希望能将发出去的奖收回来Manager wishes he could take back the award 经理考虑要采取行动了Manager is thinking about taking action,Scrap Level (%)废品率,年月JUNE 1996,Derived from Understanding Variation: The Key To Managing Chaos, Donald J. Wheeler, SPC Press. 1993.,J F M A M J,2,3,1,19
10、96,No more “Nice Guy”不再充好人了,废品率上升到 Scrap rises to a value of 2.6% 经理决定采取行动 Manager decides to take action 召开一个“特别会议”来寻求一个永久性的解决方案A “special meeting” is called to solve this problem once and for all. 经理在长篇大论次品率多么重要后离开了雇员们不知道该干什么另外，他们有其他更重要的评估标准于是，他们什么也没做 After a sound lecture on the importance of scr
11、ap, the manager leaves. Employees arent sure what to do. Besides, they have other metrics which have more importance. So they do nothing.,Scrap Level (%)废品率,年月 NOVEMBER 1996,Derived from Understanding Variation: The Key To Managing Chaos, Donald J. Wheeler, SPC Press. 1993.,J F M A M J J A S O N,经理看
12、到从去年开始废品率持续下降Manager has seen reduced scrap levels since the end of last year教训：“严格的管理会出成效！” The Learning : “A tough management style gets results!”,Manager concludes: “Tough Love Makes Things Happen”,2,3,1,1996,1997,Scrap Level (%)废品率,1997年6月 JUNE 1997,Derived from Understanding Variation: The Key
13、To Managing Chaos, Donald J. Wheeler, SPC Press. 1993.,J F M A M J J A S O N D J F M A M J,Derived from Understanding Variation: The Key To Managing Chaos, Donald J. Wheeler, SPC Press. 1993.,将数据置于统计流程控制图中 Putting The Data In A SPC Chart,2,3,1,1996,1997,Scrap Level (%)废品率,J F M A M J J A S O N D J F
14、 M A M J,UCL,LCL,统计流程控制图显示不同的解释,可为什么呢? SPC Tells A Different Story. But Why?,2,3,1,1996,1997,Scrap Level (%)废品率,J F M A M J J A S O N D J F M A M J,UCL,LCL,“人们已知的最佳方式之一是如不能使用控制图分析数据会： 增加成本, 浪费的努力和降低士气; “ - Donald J. Wheeler 博士 “Failure to use control charts to analyze data is one of the best ways kn
15、own to mankind to:increase costswaste effort and lower morale.”- Dr. Donald J. Wheeler,统计流程控制图显示不同的解释,可为什么呢? SPC Tells A Different Story, But Why?,S = 统计技术过去经常检查偏差 Statistical techniques used to examine process variation,C = 控制过程通过积极管理 Controlling the process through active management,P = 过程, 任何过程 P
16、rocess, ANY Process,现在我们管理数据的方法 - SPC The Way We Manage Data - Today SPC,控制图方法 Control Charts Method 它从哪里来的? Where Did It Come From?,19世纪20 年代- 西部电器 的 Walter Shewhart博士: 1920s - Western Electric / Dr. Walter Shewhart 惯于确认受控的& 未受控的偏差 Used to identify Controlled & Uncontrolled Variation 受控制的:普通原因或固有偏差
17、Controlled: Common Cause or Inherent Variation 未控制的:特殊起因或可指定的偏差Uncontrolled: Special Cause or Assignable Variation 在背景噪声中试图发现由特殊原因造成的偏差Tries to find the special cause variation in all of the background noise 使用控制图作为主要工具 Uses Control Charts as main tool,差异类型“共同性对特别” Types of Variation “Common vs. Spe
18、cial”,普遍原因 COMMON CAUSE 呈现在每个过程中 Is present in every process 是由过程中产生 的(我们做生意的方式)Is produced by the process itself (the way we do business) 可能被去除和或变小,但在过程上要求一个根本变化Can be removed and/or lessened but requires a fundamental change in the process,当只有共有原因偏差存在在过程中,过程是稳定, 可预测,并在控制下的. A process is Stable, Pr
19、edictable, and In-Control when only Common Cause Variation exists in the process,偏差类型“普通对特别“ Types of Variation “Common vs. Special”,不可预见的 Unpredictable 与普通偏差比较特别地大Typically large in comparison to Common Cause variation 由独特或一系列的干扰造成 Caused by unique disturbances or a series of them 可以由基本的过程控制和监视去除或变
20、小 Can be removed/lessened by basic process control and monitoring 我们认为如果过程中有特别原因偏差,它们就是失控和不稳定的. A process exhibiting Special Cause variation is said to be Out-of-Control and Unstable,练习 Exercise,当它与你的项目有关系时, 确认某种“普通原因” 和“特别原因” 偏差可能的形式As it relates to your project, identify some possible forms of “co
21、mmon cause” and “special cause” variation普便原因 Common Cause特殊原因 Special Cause,Minitab 控制图 Control Charts,Minitab 控制图练习 Control Charts Exercise,我们用一些随机的数据 Lets use some Random data 从您的生意中， 我们使用一些代表性的数值和正态偏差创造25 行任意正常数据，Create 25 rows of random normal data using some representative values for Mean and
22、Std Dev from your business 绘制单独图 Plot an Individuals chart 注意监视时间和价值被绘制在Y 轴 Note that monitoring over time and the value is plotted in the Y axis,数据结构随时间变化DATA PLOTTED OVER TIME,MONITORED CHARACTERISTIC,UCL,Center Line,LCL,UCL = Upper Control Limit / LCL = Lower Control Limit,Plotted Data,主要部分-控制图 K
23、ey Component - Control Charts,我们创造出一套标准测试，帮助识别在过程中的特别起因事件A set of standard tests have been created to help identify SPECIAL CAUSE events in our processes 我们的过程中当测试(或规则)被打破,我们称之为“失控” We use the phrase “Out of Control” when a test (or rule) has been broken,关于测试我们建议The tests we suggest: MINITAB 测试Minit
24、ab tests: 全部测试 All Tests ( 测试1-8Test 01 through 08 ) 样品规则Pattern rule: 如果你看到一个样品，过程已经失控 If you see a pattern, the process is out of control,这意味着“异常” 事发生过 This means something “unusual” has happened 去检查它! Go check it out!,过程控制测试 Process Control Tests,时间TIME,我们测量的项目 The Item We Are Measuring,应用测试 Appl
25、ying the Tests,1 Sigma,2 Sigma,3 Sigma,1 Sigma,2 Sigma,3 Sigma,60-75%,90-98%,99-99.9%,% of Data Points,UCL,LCL,时间 TIME,我们测量的项目The Item We Are Measuring,标准偏差的规则 Rules of Standard Deviation 数据应该在哪? “Where should the data lie?”,Minitab 测试Tests,Test #1,Test #2,过程控制测试 Process Control Tests,我们建议使用。全部测试 We
26、 suggest using.all tests.,失控意味什么? What does Out-of-Control mean?,查出控制的缺陷 Detecting Lack of Control,如果您确定您的过程是“失控” 你应该做什么 ？What should you do if you determine that your process is “Out of Control?”,查出控制的缺陷 Detecting Lack of Control,UCL,LCL,1 Sigma,2 Sigma,3 Sigma,1 Sigma,2 Sigma,3 Sigma,测试1， 距中心线的一个比
27、3个西格玛还多的点 Test # 1: One point more than 3 sigmas from center line.,如果打破测试1时，数据将在哪里下降？Where Will Data Fall If It Breaks Test #1?,什么是这个测试“观察“ 数据点断裂的机会? Whats The Chance of “Seeing” A Data Point Break This Test?,程序控制测试 Process Control Tests,UCL,LCL,1 Sigma,2 Sigma,3 Sigma,1 Sigma,2 Sigma,3 Sigma,如果打破测试
28、7时，数据将在哪里下降？ Where Will Data Fall If It Breaks Test #7?,测试7：在1西格码中心线内的一行上的15个点（任意一边） Test # 7: Fifteen points in a row within 1 sigma of center line (either side),过程控制测试 Process Control Tests,UCL,LCL,1 Sigma,2 Sigma,3 Sigma,1 Sigma,2 Sigma,3 Sigma,测试8：从中心线超过1 个西格码 (任意一边)一行上的八个点。Test # 8: Eight point
29、s in a row more than 1 sigma from the center-line (either side).,如果打破测试8时，数据将在哪里下降？ Where Will Data Fall If It Breaks Test #8?,过程控制测试 Process Control Tests,在控制下还是失控？In Control or Out of Control ? _,如果在控制以外, 打破了什么规则或表现出什么条件?If out of Control, which rule(s) is broken or condition(s) is present? _,在控制下
30、或失控？ In Control or Out of Control ? _,如果在控制以外, 打破了什么规则或表现出什么条件? If out of Control, which rule(s) is broken or condition(s) is present? _,在控制下或失控？ In Control or Out of Control ? _,如果在控制以外, 打破了什么规则或表现出什么条件? If out of Control, which rule(s) is broken or condition(s) is present? _,在控制下或失控？ In Control or
31、 Out of Control ? _,如果在控制以外, 打破了什么规则或表现出什么条件? If out of Control, which rule(s) is broken or condition(s) is present? _,违反规则并发现模式合适吗?Is it okay to break rules and find patterns?,规则(测试) 和模式 Rules (tests) and Patterns,因此, 根据现在你所知道的, 如果你的过程在控制下，在控制上限和下限之间百分之多少数据点将会下降? Therefore, based on what you know s
32、o far, what percent of data points should fall between the upper control limit (UCL) and lower control limit (LCL) if your process is in-control?,控制极限 Control Limits,如果点落在上限之外或控制下限之下，是否意味着我们为顾客做了一个缺陷产品?If a point falls beyond the upper or lower control limit does this mean we are making a defect for
33、 the customer?,控制极限对规格限制 Control Limits vs. Specification Limits,UCL,LCL,TIME,控制极限对规格限制 Control Limits vs. Specification Limits,过程控制极限是在过程中数据自己计算出来的Process Control Limits are calculated based on data from the process itself 他们根据+/- 3s (99.73%我们期望过程偏差落在这些极限之间) They are based on +/- 3s (99.73% of the
34、process variation is expected to fall between these limits) 产品规格极限不是在控制图上发现的Product Specification Limits ARE NOT found on the control chart 很重要一点是要了解程序控制与顾客要求如何吻合. Understanding how the process matches up against customer requirements IS important to know,确定过程执行如何满足顾客期望, 需要进行过程能力研究。 To determine how
35、 the process performs to Customer Expectations, a Process Capability Study is required.,把规格限制放在在控制图上Putting specification limits on a Control Chart把控制上限和控制下限当做规格限制Treating UCL and as a specification limit,2个控制图的大错误 TWO BIG CONTROL CHART ERRORS,控制极限对规格限制 Control Limits vs. Specification Limits,当你把任意上
36、下限作为监视工具时，他就不再是个控制图 LCL When you do either of these the control chart becomes just an inspection tool - its no longer a control chart控制上限和控制下限并不直接与客户缺陷有联系! UCL / LCL are not directly tied to customer defects !,控制图 取样对100%监视 Control Charts - Sampling vs. 100% Inspection,控制图采样对了解过程是一个简单并且有效的方法Control
37、Chart sampling is a simple and effective method for process understanding 如果采取消除特别起因行动 (稳定过程) 并且能力被证明, 100% 监视可能被取消(但是要知道客户的特殊检查计划) If actions are taken to eliminate special causes (stabilize the process) and capability is proven, 100% Inspection can be eliminated (but beware of customer specified i
38、nspection plans),对控制图有效使用的要求 Requirements for Effective Use of Control Charts,管理必须建立并且支持促进适当的行动，且支持对控制图上信息的收集Management MUST establish and support an environment that promotes proper action and support to the information collected on the control charts 控制图只在改善能够给公司或客户带来好处的过程中运用实施Control Charts are i
39、mplemented ONLY on Key Processes on which improvement will bring benefit to the organization and/or the customer 数据从能力测量系统中收集的过程是有效的 Data collected from the process is validated through the use of a CAPABLE measurement system,控制图程序的最大问题 Biggest Failure Of Control Chart Programs系统被控制图淹没，没有对数据做任何行动 Fl
40、ood the system with control charts,then dont take action on the data,控制图与控制极限 Control Charts & Control Limits,初步控制极限演算 Initial Control Limits Calculation 当第一次建立或学习过程 When first setting up or studying a process,控制极限重算 Control Limits Re-calculation有一个已知的过程变动,并且它的作用被确认对老的控制极限相反 There is a known change
41、to the process and its effect has been verified against the old control limits.,两种一般数据 Two General Kinds of Data,属性 ATTRIBUTE - 离散, 可计的数据 Discrete, Counted DataEx: 1, 2, 3, 4 etcGood / BadMachine 1 , 2 , 3 .变量VARIABLES - 连续, 可测量的数据 Continuous, Measured DataEx: Weight = 10.2 LbsThickness = 11.211 inc
42、hes,螺钉扭矩在每个装配线传输的左前角离开 Bolt torque on the front left corner of every transmission coming off the assembly line每个螺钉离开装配线传输的平均扭矩Average bolt torque of every bolt for each transmission coming off the assembly line (3) 每个发动机所缺的螺钉数 # of missing bolts per engine(4) 每个销售合同的排字数 # of typos per sales contract
43、每月生产缺陷发动机的数目Number of engines with defects in monthly production每月生产缺陷发动机的的百分比数% of defective engines in monthly production根据应收帐款, 收回它的时间 Per accounts receivable, amount of time it takes to close it 每100个发动机的缺陷数 Number of engines with defects per 100 built,练习：什么类型的数据？ Exercise: What Type of Data?,Di
44、screte,Variable,What Type Of Data ?,Data Collected In Groups or Individuals ?,CountingSpecific Defects orDefective Items ?,GROUPS (Averages) (n1),INDIVIDUAL VALUES (n=1),X-Bar R X-Bar S,Individuals Moving Range,Specific Types Of “Defects”,Defective Items,Is The Probability OfA Defect Low ?,If You Kn
45、ow How Many Are Bad, Do You Know How Many Are Good?,Poisson Distribution,Binomial Distribution,Individuals Moving Range,NO,YES,YES,Area of Opportunity Constant In Each Sample Size ?,YES,NO,c Chart,u Chart,ConstantSample Size ?,np Chart,NO,YES,p Chart,选择正确控制图 Choosing the Correct Control Chart,NOTE:
46、X-Bar S is appropriatefor subgroup sizes (n) of 10,控制图的主要类型 Major Types of Control Charts,变量图 Variables Charts I-MR (个体 individuals) X-Bar (平均 average)特性图 Attribute Charts NP (有缺陷的数字 Number defective) P (有缺陷的比率 Proportion defective) C (过失数量 Number of defects) U (每个单位的过失数量Number of defects / unit),I-
47、MR 个别-移动范围 Individuals - Moving Range,不同偏差的控制图 Different Variables Control Charts,X-Bar-R 平均-范围图 Average Range Chart,NP 不一致的单位数量 Number ofnon-conforming units,不同属性的控制图：缺陷 Different Attribute Control Charts: Defectives,P 不一致的单位数量比例Proportion of non-conforming units,U 单位缺陷数量 Number of defects per “unit”,不同属性的控制图：缺陷 Different Attribute Control Charts : Defects,C 缺陷数量 Number of defects,练习: 什么类型的控制图? Exercise: What Type of Control Chart?,
