质量管理(过程控制SPC)

1、第四章Statistical Process Control（统计过程控制）Case Study 实例研究实例研究v Mike was known as one of the best drivers in Flextronics Doumen Campus, One day at work, he was bragging about “How Perfectly Straight he could drive at all time, and no matter how the quality of roadway is ”, but his co-workers asked him to

2、 prove it, otherwise, Mike have to invite him for dinner.v迈克是伟创力斗门工业园众所周知的最好的司机之一, 有一天他吹牛说“我开车开得总是那么直,无论路面状况如何”,他的同事要求他示范表演, 否则迈克请吃饭.v How to prove it? Proving he was a straight driver was going to be a difficult task. Then someone suggested that Mikes truck was always leaking oil, so we can track t

3、he oil spots to determine if he was driving straightv怎么样证明呢? 要证明他开车很直是一件很难的任务, 有人建议“迈克的车总是漏油,我们可以以此来跟踪决定他是否开得直“v Are you sure Mikes co-worker will be treated or who would you Stand guarantee for?你肯定迈克的同事会被宴请吗? 或者你愿为谁做担保呢?The location of leaking oil 漏油位置漏油位置Looking at the road surface with Birds-eye

4、View after driving, did Mike drive straight? 看一下开车后的鸟瞰图看一下开车后的鸟瞰图, 迈克开车直吗迈克开车直吗? “Hi, Mike, when will you invite me for dinner?” Mikes co-worker talked to him.“嗨嗨,迈克迈克,你什么时候请我吃饭呢你什么时候请我吃饭呢?”You are just lucky! Mike answered. 迈克回答迈克回答:你只是运气好而已你只是运气好而已.If you were Mike, how do you justify for yourself

5、? Are you convinced?如果你是迈克如果你是迈克, 你会怎样为你辩护呢你会怎样为你辩护呢? 你服气吗你服气吗?The story is just beginning. 故事刚刚开始故事刚刚开始.Think about this: how does the oil fall ? 想一想想一想: 油是怎样滴下来的油是怎样滴下来的?If the truck was sitting still . . .Would the oil droplets always land in the same place? 如果卡车静止不动如果卡车静止不动, 油也会滴在同一个地方吗油也会滴在同一个地

6、方吗?Lets go on the study 我们继续研究吧我们继续研究吧!This is so-called natural variation or common cause 这就是所谓的这就是所谓的自然变自然变异或称之为普通原因异或称之为普通原因Think about the source of variation 想一下变异的来源想一下变异的来源Drop to Drop Variation油滴之间的变化油滴之间的变化Drop to Drop variation + Wind 油滴之间的变化加上风的作用油滴之间的变化加上风的作用Drop to Drop variation + Wind

7、 + the variation of Steering 油滴之间的变化加上风的作用油滴之间的变化加上风的作用, 以及掌舵方以及掌舵方向的变化向的变化Conclusion: 结论结论What variation is resulted from steering? Does it really affect the straight of driving? Up to now, are you trying to change your opinions or Mikes co-worker should be treated or not? 由掌舵引起的变化是什么呢由掌舵引起的变化是什么呢?

8、 它真的它真的影响开车的直线程度吗影响开车的直线程度吗? 到现在为止到现在为止, 你打算改变你原来的看法吗你打算改变你原来的看法吗? 或者或者说迈克的同事应不应该被款待说迈克的同事应不应该被款待?What is so-called special cause or non-random=assignable cause?所谓的特殊原因或称之为可查明原因是什么呢所谓的特殊原因或称之为可查明原因是什么呢?One more story 多一个故事多一个故事The factory scrap level is at a month low of 2% 公司的月报废率水平低于公司的月报废率水平低于2%M

9、anager presents an award to the team and hold Ceremony in the cafeteria, refreshments for all! 经理赠送了奖品给该团队并举行庆祝仪式为各位打气经理赠送了奖品给该团队并举行庆祝仪式为各位打气!“Everyone should be proud of what youve accomplished”. Manager addressed. 经理说经理说:每个人都应该为你们所取得的成就而自豪每个人都应该为你们所取得的成就而自豪Monthly Scrap Level Trend2.11.81.60123Jan

10、-03Feb-03Mar-03MonthScrap level %This story should be true 这应该是一个真实的故事这应该是一个真实的故事However, three consecutive months of scrap increases, then Manager was getting angry and took back the award, even deciding to fire someone who is in charge of this project directly. 然而然而, 报废率连续三个月递增报废率连续三个月递增! 经理很生气并将奖

11、经理很生气并将奖品收回品收回, 甚至要将直接负责人解雇甚至要将直接负责人解雇.“Actions must be taken” Manager required.经理要求经理要求:必须采取措施必须采取措施Monthly Scrap Level Trend2.11.81.61.92.22.30123Jan-03Feb-03Mar-03Apr-03May-03Jun-03MonthScrap level %Unfortunately, the scrap suddenly rises to a value of 2.3% in the next month, General manger decid

12、es to take action and a “special meeting” is called to solve this problem immediately. 不幸的是不幸的是, 报废率突然上升至报废率突然上升至2.3%, 总经理决定采取行动并总经理决定采取行动并立即召开特别会议以解决此问题立即召开特别会议以解决此问题.Later on, the manager is fired by general manager, also all responsible person are not sure what to do, but just loot at the trend o

13、f scrap level. So they do nothing.随后经理被总经理解雇随后经理被总经理解雇, 同样同样, 所有的负责人都不知道要做什么所有的负责人都不知道要做什么, 他们只是盯着报废他们只是盯着报废率的趋势率的趋势! 他们什么都没有做他们什么都没有做!The story is performing continuously 故事还在继续上演故事还在继续上演GM talked to manager: “Sorry, your soft management made this happen, you have to take the direct responsibility”

14、; manager left without any justification.总经理对经理说总经理对经理说:对不起对不起, 你松弛的管理导你松弛的管理导致此事的发生致此事的发生, 你不得不为此而负责你不得不为此而负责, 于是于是经理未做任何辩护离开了经理未做任何辩护离开了Monthly Scrap Level Trend2.11.81.61.92.22.31.81.70123Jan-03 Feb-03 Mar-03 Apr-03 May-03 Jun-03 Jul-03 Aug-03MonthScrap level %But it is very strange that “Things

15、are looking-up based on the data collected from the next two months!” , although we all knew that nothing had been done to change the system and process 很奇怪的是根据随后两个月的数据看上去事情有了好转很奇怪的是根据随后两个月的数据看上去事情有了好转, 尽管我们都知道我尽管我们都知道我们的系统和流程都未改变们的系统和流程都未改变How come?怎么会呢怎么会呢?To be continued 未完待续未完待续Question: Why the

16、 scrap level come down without any change to process?Someone asked General manager: How do you think about this? I think you have made wrong decision before! ” Hey, I made my decision based on data - How can I go wrong ? GM replied.有人问总经理有人问总经理: 你怎样看待这个问题呢你怎样看待这个问题呢? 我认为你做出了错误决定我认为你做出了错误决定!, 总经理回答总经

17、理回答:我根我根据数据做决定据数据做决定, 我怎样错的呢我怎样错的呢?Black Belt review the trend of scrap level, talked to General manager: Your decisions were made from observing high and low points as signals. When In reality, it was all noise. Look at the data, there was no significant change in the process., so you made wrong de

18、cision to blame our manager”Black Belt 对报废率的水平进行了审核后对总经理说对报废率的水平进行了审核后对总经理说: 你的错误决定来源于你将数据你的错误决定来源于你将数据的高低点作为信号的高低点作为信号! 事实上事实上, 它们都是干扰因素它们都是干扰因素, 流程中并没有显著的变化流程中并没有显著的变化, 因此你错因此你错怪了我们的经理怪了我们的经理“General manager: ?总经理无言以对总经理无言以对.What have you learned from the two lessons? Have you had good understandi

19、ng on Common cause and Special cause?Summary 总结总结UCLCLLCL统计小测试：统计小测试：F字母计数练习字母计数练习Imagine for one brief moment that each of the one hundred and forty-one words of this paragraph is a separate component from a first run of fourteen-inch flywheels. You are one of five inspectors performing the final i

20、nspection of these finished components which were produced on fairly small dial index machines that are not being controlled by the use of statistical techniques. As can be expected from an operation of this nature , there are a number of defectives components being made .Each word that contains an

21、f represents a defective component. How many of the defectives are you able to find? Check again and inspect for the present s off s. Write your final count in the bottom left hand corner of this page .This example should give you a fair idea of how reliable 100%inspection can be. 请用1分钟，彻底检查一次，看看字母“

22、F”出现的次数答案？结论：结论：100的检验不能保证的检验不能保证100的合格的合格所谓质量变异,指同一批的产品，即使所采取的原材料、生产工艺和操作方法相同，但其中每个产品的质量也不能丝毫不差，完全相同，他们之间或多或少总会有些差别，这种差别被称为变异。（产品实际达到的质量特征值与与规定的质量特征值之间发生的偏差：同样品牌型号价格买来的手机，为什么我的老出问题？） 一、质量变异一、质量变异（qualitative variation)及变异特征及变异特征变异产生原因导致变异的要素被称为5M1E人、机、料、法、环、测量MachineManpowerMaterialsMethodMeasureme

23、ntEnvironment随机变异（正常变异）：随机变异（正常变异）：由偶然性因素引起由偶然性因素引起偶然性因素繁多，如原材料性质的微小差异，模具的微小变形，操作方法的微小变化等。该类因素不易识别且难以清除，其引起的产品质量差异是随机的，但一般影响并不大，不会造成废品。系统变异是指有生产过程中的系统因素引起的变异。如混入了不同规格成分的原材料、设备过度消耗或者调整不准确、不同人员进行操作等。这类因素的特点是数目不多但对产品的影响却很大，可是数目不多但对产品的影响却很大，可能造成不合格产品。系统变异（异常变异）：系统变异（异常变异）：由系统因素引起由系统因素引起 战后经济遭受严重破坏的日本在战后

24、经济遭受严重破坏的日本在19501950年通过休哈特早年通过休哈特早期的一个同事戴明期的一个同事戴明(W. Ed- wards Deming)(W. Ed- wards Deming)博士，将博士，将SPCSPC的的概念引入日本。从概念引入日本。从1950195019801980年，经过年，经过3030年的努力，日本年的努力，日本跃居世界质量与生产率的领先地位。美国著名质量管理专跃居世界质量与生产率的领先地位。美国著名质量管理专家伯格家伯格(Roger W. Berger)(Roger W. Berger)教授指出，日本成功的基石之教授指出，日本成功的基石之一就是一就是SPCSPC。 美美国贝

25、尔实验室休哈特博士（国贝尔实验室休哈特博士（W. A. W. A. ShewhartShewhart）于）于19241924年年发发明控制明控制图图，开启开启了了统统计计品管的新品管的新时时代代。SPCSPC兴兴起的背景：起的背景：起源起源 1940 1940s s 二次世界大战期间，二次世界大战期间，美国军工产品使用抽样方案和控美国军工产品使用抽样方案和控制图以保证军工产品的质量。制图以保证军工产品的质量。16SPCSPC兴兴起的背景：起的背景：日本日本1950s 质量管理大师戴明博士在日本工业产品生产过程中全面推行SPC。 日本JUSE（科学家协会）设置“戴明”奖，奖励那些有效实施统计技术

26、的企业。 石川磬提出“QC七工具”，帮助生产现场人员分析和改进质量问题，并推动广泛应用。1970s 有效地推行 “QCC圈”和应用统计技术使日本经济的快速发展，成为高品质产品的代名词。1980s 美国等其他国家紧随日本的步伐，开始推行“QC小 组”和统计技术的应用。 美国汽车工业已大规模推行了美国汽车工业已大规模推行了SPCSPC，如福特汽车公司，通用汽车公司，克莱斯勒，如福特汽车公司，通用汽车公司，克莱斯勒汽车公司等，上述美国三大汽车公司在汽车公司等，上述美国三大汽车公司在ISO9000ISO9000的基础上还联合制定了的基础上还联合制定了QS9000QS9000标标准，编制了准，编制了SP

27、CSPC手册。在与汽车有关的行业中，颇为流行手册。在与汽车有关的行业中，颇为流行。 MOTOROLA 公司颁布“QC挑战”，通过SPC的实施改进过程能力，并提出追求“66”目标。1987 ISO9000标准建立并颁布实施，明确要求实施统计技术。17质量控制中的数据质量控制中的数据 算术平均数（arithmetic mean） 标准差（standard deviation） )(xnxfxniii1)()()(111)(2222112xxxxxxnnxxfnniii质量控制中的数据质量控制中的数据R 极差（Range ） RXmaxXmin群体样本结论数据抽 样分析测试行动符号有的不一样，如：符

28、号有的不一样，如：总体标准差总体标准差 与样与样本标注差本标注差 S二、正态分布曲线二、正态分布曲线（p408参数表）参数表）正态分布中，任一点出现在 1内的概率为 P(-X +) = 68.26% 2内的概率为 P(-2X +2) = 95.45% 3内的概率为 P(-3X USLPPM USLPPM USLPPM LSLPpkPPLPPUPpCpmCpkCPLCPUCpStDev (Overall)StDev (Within)Sample NMeanLSLTargetUSL310.97308.71 2.27 51.41 51.32 0.10 0.00 0.00 0.001.141.531.

29、141.33 *1.291.731.291.514.994384.40236100102.9 80.0 *120.0Exp. Overall PerformanceExp. Within PerformanceObserved PerformanceOverall CapabilityPotential (Within) CapabilityProcess DataWithinOverall过程控制图过程控制图能力分析X-bar-sSometimes, the sub-group size is larger and we can easily compute the process Sigm

30、a ( standard deviation), then X-bar s can be used instead of the range in variables control charts, then the X-bar & s Control Chart limits can be calculated as follow:Std Dev X barsBUCLsCLsBLCL43 sAXUCLXCLsAXLCL33 4c/s I-MR ChartIn some circumstances, only individual values can be taken. Contro

31、l charts can still be created.They are called Individuals & Moving Range charts, or “I MR Chart”Because the charted X values are actual process points, specifications may be placed on the chart.计量型控制图：计量型控制图：X（I）-MR图图XX啤酒厂是驰名多年的啤酒制造厂，平均每季度产啤酒300万吨。计划部的同事正在依据以往的销售数据制定2014年的生产计划。2013年平均每季度产量为357.492吨。从数据看2014年应该把季度产量定位360吨。过程控制图过程控制图20062007200820092010201120122013318.687275.766282.757297.379326.103268.907300.556347.23260.148274.934277.198394.235309.231304.528254.327548.537356.279232.999258.315300.098320.57338.934375.984279.144301.917289.714302.723289.209293.335296.068330