统计过程控制简介.pptVIP

The Individuals chart violates rule 1 on the first point, below lower limit. The process also looks like it is creeping up and about to exceed the upper limit. Ranges are stratified, only three values. Stability and normality should be checked prior to capability analysis. Process Capability: if specifications for the Shaft OD were given, we can use Minitab to determine Process Capability. In this case, since we have Individuals(X) data and no clear idea on subgroup size, then use of the Six Sigma Process Report (which requires subgroup size 1) is not correct. The tool to use in this case is the Capability Analysis (Normal) or Capability SixPack(Normal)—per Minitab ver 12. Assume specification is: 0.250 +/- 0.005 for Shaft OD Result = although showing “lack of CONTROL,” the process is very CAPABLE (so far) Minitab Commands: Stat Quality Tools Capability Analysis (Normal) Single Column: Shaft_OD Subgroup Size: 1 Lower Spec: .245 Upper Spec: .255 Use the relationship: ZST = 3CP and ZLT = 3PP 离散数据: 缺陷个数5 I-MR 缺陷个数=5 关注缺陷的比例 样本总量是常数 np chart 样本总量变化 p chart 关注缺陷的个数 Other terminology may also be used. A defect is also known as a nonconformance. A defective is known as a nonconforming. What is the appropriate distribution to use and chart to generate is based on the following factors: Is it defects or defective that you are investigating in the data you are collecting (e.g. Is it pass/fail for the part or process or is it several sub-components or sub-processes passed/failed for the part or process)? Is it a constant lot size or variable lot size? 离散数据: 缺陷个数5 I-MR 缺陷个数=5 关注缺陷的比例 样本总量是常数 np chart 样本总量变化 p chart 关注缺陷的个数 C bar = DPU = number of defects detected/subgroup A subgroup may contain one or more “physical” units. You need C bar up around 5 - If it is less than 5, your distribution will be skewed because it will be truncated by 0. Increase your subgroup size if you need to increase C bar or collapse existing catego