基于SWV方法的偏态过程能力分析

过程能力分析中标准的过程能力指数（PCI）计算方法是基于过程输出服从正态分布的假设建立的。但在实际生产中存在大量不服从正态分布过程输出，有时过程的分布形式未知。在此情况下，简单使用标准PCI会导致对过程能力的错误判断。当过程输出是偏态分布时，通常可通过确定分布函数，使用Clements方法确定PCI。但抽样数据的不足将造成对分布函数的估计不准确，结果存在较大偏差。基于赋权方差（WV）方法的PCI弥补了标准PCI的不足，但它也存在缺陷。本文对此进行了修正，并针对WVPCI的不足，基于比例赋权方差（SWV）方法构建了新的偏态PCI。通过仿真研究，比较了基于wV和SWV方法的PCI。仿真结果显示，基于SWV的PCI能够更准确地反映过程输出的能力。

Process Capability Analysis in Skew Distribution Based on SWV Methods

The calculating methods of standard capability indices in process capability assessment are based on the hypothesis that the distributions of process output are normal.There are a lot of practice cases in which the output isn't normal.(Because) the forms of these distributions are unknown,the results of evaluation may be misleading to a judgment about the processes.When the distribution is on the skew,the Clements method is used after estimating the form of distribution by a lot of data from process output.With the not truly known distribution,there is big gap in the assessment result.The paper analyzes the bug of the WV-based PCI.And based on Scaled Weighted Variance(SWV) method,some process capability indices formulas,which can be used in skew distribution,are established in the paper.The simulation study was done for comparison between the WV-based and SWV-based PCI.The results show that the SWV-based PCIs are better.