合成S~2图的设计及应用

发布时间：2018-01-19 03:03

本文关键词： 过程方差合成S~2图已知参数未知参数　出处：《浙江工商大学》2015年硕士论文　论文类型：学位论文

【摘要】：在生产过程中,有些生产因素如劣质的原材料或者工人操作不熟练并不会引起过程均值发生变化,但却会引起过程方差的变化,过程方差变大说明产品的质量不稳定,过程方差减小说明产品的质量提高。另外过程方差失控会使均值控制图的表现不同于预期,因此监测过程方差的变化是十分重要的。传统的休哈特S2图常被用于过程方差的监控,但存在一定缺陷,具体表现为对过程方差大漂移比较敏感,而对过程方差中小漂移不敏感。本文引入合成图的方法,将休哈特S2图与合格品链长图结合组成合成S2图来对方差进行监控,以改进休哈特s2图的表现,提高控制图监测过程方差变化的能力。本文主要研究内容如下：(1)本文将合成图的方法分别应用于对过程方差单、双侧漂移的监测。基于参数已知情况,设计了上单侧合成S2图、下单侧合成s2图以及双侧合成s2图分别应用于监测正态总体假定下过程方差变大、过程方差变小以及过程方差漂移方向不确定等三种情况,并与相应的休哈特s2图比较。结果显示本文设计的单、双侧合成s2图的监测效果均优于休哈特S2图。(2)控制图的设计通常假设受控参数己知,而实际应用中受控参数常常未知并需要由已获受控样本进行估计。本文分析了参数估计对三种合成s2图的影响,发现参数估计会增加伪警概率,大大降低合成s2图的监控能力,特别是在过程方差漂移量较小和用于估计受控参数的第1阶段样本组数较小之时。(3)因为参数估计大大影响三种合成S2图的表现,所以基于己知参数的控制图不再合适,因此本文基于估计参数设计了新的单、双侧合成s2图对过程方差进行监控,并与相应的的休哈特s2图比较,结果显示在参数未知时,新设计的单、双侧合成s2图的监测效果优于相应的单、双侧休哈特s2图。
[Abstract]:In the process of production, some production factors, such as inferior raw materials or unskilled operation of workers, will not cause changes in the mean value of the process, but will cause changes in the variance of the process. The increase of process variance indicates that the quality of the product is unstable, and the decrease of the process variance indicates the improvement of the quality of the product. In addition, the runaway process variance will make the performance of the mean control chart different from the expected. Therefore, it is very important to monitor the variation of process variance. The traditional Shewhart S2 diagram is often used to monitor the process variance, but it has some defects, which shows that it is sensitive to the large drift of process variance. However, it is not sensitive to the small drift of process variance. In this paper, the method of composite graph is introduced, which combines Shewhart S2 diagram with qualified product chain length graph to monitor the variance to improve the performance of Heinhart S2 diagram. The main contents of this paper are as follows: 1) in this paper, the method of composite graph is applied to the monitoring of single and bilateral drift of process variance, based on the known parameters. The S 2 map of upper side synthesis, the second side S 2 and the double side S 2 are designed to monitor the normal population under the assumption that the process variance becomes larger. The process variance becomes smaller and the direction of process variance drift is uncertain. The results are compared with the corresponding Hewhart S2 diagram. The results show that the design of this paper is single. The monitoring effect of bilateral s2 diagram is better than that of Hewhart S2 chart. 2) the design of the control chart usually assumes that the controlled parameters are known. In practical application, the controlled parameters are often unknown and need to be estimated by controlled samples. In this paper, the effect of parameter estimation on three kinds of composite S2 graphs is analyzed, and it is found that parameter estimation will increase false alarm probability. Greatly reduces the monitoring ability of the composite S2 graph. Especially when the variance drift of the process is small and the number of samples used in the first stage to estimate controlled parameters is small) the parameter estimation greatly affects the performance of the three kinds of composite S2 graphs. Therefore, the control chart based on known parameters is no longer suitable, so this paper designs a new single- and double-side composite S2 graph to monitor the process variance based on the estimated parameters, and compares it with the corresponding Shewhart S2 graph. The results show that when the parameters are unknown, the monitoring effect of the newly designed single and bilateral composite S2 charts is better than that of the corresponding single and bilateral Heinhart S2 diagrams.
【学位授予单位】：浙江工商大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：F273.2;F224

【相似文献】