Six Sigma 统计学基础及控制图.pptVIP

* * * * * * * * * * * * * * * * * * * * * * * * * * * EWMA 图的结构 考虑到利率因素,至小应该收集50个数据点. 将第一个EWMA点设置为目标值. 对每一个数据点和l的不同值,确定误差，et,误差的平方,及下一个EWMA点. 将误差的平方加总 EWMA 图的结构 选择l的一个值,使平方和为最小. 在图上绘出数据点 计算出控制范围,并在图上绘出. 观察图表,确定过程是否处于控制. (没有点位于控制范围之外) 如果不是这样的话, 找出可解释的原因并重做一遍. EWMA 图和信号 打开工作表CRANKSH.MTW 选择 StatControl ChartsEWMA 在独立栏中输入: AtoBDist, 子集大小为: 5 在注释中,输入 “Crnksht 0.2” ,然后点击 OK 注意“EWMA的权重”为2.0 结果: 第五个点和第六个点位于控制范围之外. 重做练习,将EWMA的权重变为4.0,改变注释. 当EWMA的权重为6.0时,重复这一过程. EWMA 图和信号 结果: 第五个点和第六个点位于控制范围之外 以4.0为权重的EWMA 图 结果: 第五个点位于控制范围之外 以6.0为权重的EWMA 图 结果: 没有点超出控制范围 教学反馈/下周发表课题 1.为何同样的SQL及Cp却代表不同的良品率？（详见SLIDE-67） 2.描述关于样本分布的1.5个SIGMA的移动 3.描述一般原因和特殊原因的可能来源及其在控制图上的体现 4.描述控制图的制定步骤（如何设计？考虑哪些因素？为什么？ 5.解释X-R控制图及其所代表的可能的变动来源 * * * * * * * * * * * * * * * Cp and Cpk are measures of the natural spread and centering of the population to specified limits. Cp is an index of the spread of the population in relation to the spread of the spec limits. Cpk is an index of the centering of the population in reference to the spec limits. In the denominator, we use the term ‘Normal Variation of the Process’ and write it as +/- 3 sigma. This is due to 99.7% of the population falling within +/-3 sigma of the mean. We use this as the ‘estimate’ of the total population. If the Spec limit range is equal to +/- 3 sigma, what is Cp? Want Cp to be 2 or better - why? Because then we have only consumed 1/2 of the spec range with our process, permitting shift in the mean without impacting our customer - maybe! Why maybe? It depends on where the process is centered. Good Analogy to use: car in garage example where the garage walls are the specification limits and the car width is the process width. , LSL USL +3 s -3 s LCL UCL m UCL, LCL and USL, LSL are the same LSL USL +6 s -6 s LCL UCL m Example s 6 12 s C p = , C p = 2 , C p = 1 , C p = 2 C p = 1 * * Cp and Cpk are measures of the natural spread and centering of the population to specified limits. Cp is an index of the spread of the pop