基于线性轮廓数据的测量系统能力分析

发布时间：2018-09-04 14:00

【摘要】：测量系统能力分析是质量控制及提升过程中十分重要的质量工具,其在现代质量管理活动中的使用频率也越来越高,进行测量系统能力分析,其目的是对测量系统能力是否足够的问题进行判断,并最终确定该测量系统是否可接受。不管是在生产实践还是在科学研究中,测量系统能力分析都得到了广泛的应用,它是在进行过程能力分析、实验设计等研究前必不可少的一个步骤,也是企业用于提高产品和服务质量的一种重要工具。目前,该领域内的大部分文献是关于一元测量系统能力分析的,而且其研究也相对比较成熟。但是,在现代工业中,有些产品的质量需要多个特性来反映,这种情况下的测量值是一组相关的多维数据,单变量测量系统分析方法在此环境下一般是不适用的,因此越来越多的企业开始关注多元测量系统能力分析,这方面的科学研究也逐渐增多。然而,随着产品工艺及性能的日益提升,产品质量的测量数据类型已经不仅仅局限于单变量和多变量两种形式,很多产品质量的测量结果可以被描述成为一条轮廓线,在这种情况下,现有的一元和多元的方法在解决测量系统问题时是不够简洁的甚至是不适用的。基于此,本文提出了针对测量结果被描述成为简单线性轮廓线时的测量系统能力分析方法,在该研究中,测量误差主要表现在线性轮廓线的截距和斜率上,因此,可以将简单线性轮廓线的测量系统能力分析简化为二维数据的分析,在该方法中,多元方差分析被用于评估测量系统误差的方差分量矩阵和其他变异源的方差分量矩阵。在指标评价方面,本文选取了测量系统能力分析中的常用指标,精度公差比和重复性与再现性百分比,并将其做了一定程度的转化,使其满足用于评估线性测量系统的能力。本文在对所提方法进行了充分整理的基础上,选取了制造业的一个案例来进行分析,通过这个案例来验证该方法的可行性,文章的最后对全文进行了系统的总结并提出了今后科研的一些希望与思路。
[Abstract]:Measurement system capability analysis is a very important quality tool in the process of quality control and promotion, and its frequency of use in modern quality management activities is also increasing. The purpose is to determine the adequacy of the measurement system and to determine whether the measurement system is acceptable or not. Capability analysis of measurement systems has been widely used in both production practice and scientific research. It is an essential step before the process capability analysis, experimental design and so on. It is also an important tool for enterprises to improve the quality of products and services. At present, most of the literature in this field is about the capability analysis of monometric measurement system, and its research is relatively mature. However, in modern industry, the quality of some products needs more than one characteristic to reflect, in this case the measurement value is a set of related multidimensional data, the single variable measurement system analysis method is generally not applicable in this environment. Therefore, more and more enterprises begin to pay attention to the capability analysis of multivariate measurement system, and the scientific research in this field is increasing gradually. However, with the increasing improvement of product technology and performance, the measurement data types of product quality are not limited to two forms, single variable and multivariate, many of the measurement results of product quality can be described as a contour line. In this case, the existing unitary and multivariate methods are not simple enough or even suitable for solving measurement system problems. Based on this, this paper presents a method for analyzing the capability of measurement system when the measurement results are described as simple linear contours. In this study, the measurement errors are mainly reflected in the intercept and slope of the linear contour, so, In this method, the multivariate variance analysis is used to evaluate the variance component matrix of measurement system error and the variance component matrix of other sources of variation. In the aspect of index evaluation, the common indexes, accuracy tolerance ratio, repeatability and reproducibility percentage are selected in this paper, and they are transformed to a certain extent to satisfy the ability of linear measurement system. Based on the analysis of the proposed method, a case of manufacturing industry is selected to analyze the feasibility of the method. At the end of the paper, a systematic summary of the full text is given and some hopes and ideas for future scientific research are put forward.
【学位授予单位】：天津大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TB9

【相似文献】