过程能力要点详解.ppt

过程能力分析 那么 . . . 什么是 “好” 和 “不怎么好”? 红 (坏的) 黄 (还行) 绿 (好的) 为什么这么做? 为了回答 . . . 总体上讲, 过程 表现如何? 练习: 缺陷计算 alueriohng Long term is composed of many short terms put together. You can see above that the short terms may have different centering and different variability. The purpose of this slide is to emphasize the differences between an “entitlement” study and a “performance” study. Emphasize that the entitlement study should be taken over a short enough period of time to only involve 1 “set-up”, 1 lot of material, 1 set of people, basically minimal sources of variability with no time for a shift in the mean to occur whereas performance needs to have multiple set-ups and all sources of variability acting onthe process. Participant Notes: Remember, Six Sigma is a short term goal. GE will report short term sigma scores and long term DPMO. Look up 3.4 DPMO or 3.4E-06 on the normal table and you will see it has a Z value of 4.5 NOT 6.0. When we have long term data we calculate DPMO and convert it to the Z value. Then we add 1.5 to that number to get a short term Z value. Hence, GE talks about 6? (short term Z score) being equivalent to 3.4 DPMO in the long term. Be consistent...If you have short term datayou cannot add 1.5 to the Z score! Short term data does not capture total range of variation of a design or process. A reasonable estimate of the long term standard deviation can be obtained by multiplying the short term standard deviation by 1.3. This is the method applied to continuous data. The key to all of this is : 1. Consistency in applying the 1.5 shift -If you gather short term data (Capability or Entitlement Study), DO NOT apply a 1.5 ? shift -If you gather long term data (Performance Study or Historical data), add the shift 2. Report short term ? values and long term DPM 3. Minitab or the normal table is useful for both Variables and Attribute type of data 46 The Z-value is a means of comparing different processes and different CTQs.