统计过程中的基于增强鲁棒辅助信息的记忆型控制图

发布时间：2020-11-08 10:10

　　 统计过程控制(SPC)是一系列通过统计分析来监控制造和非制造过程的方法。过程控制是用来提高产品和服务质量的连续的过程。波动是一个过程的重要部分,并且为了提高过程的质量,我们不能忽视这些波动。所有生产过程都会受波动的影响。这些波动可以分为两类:普遍原因引起的波动和特殊原因引起的波动。及时监测由特殊原因引起的波动对任何过程的执行都有重要作用。在检查产品是否符合他们所设计的要求时,控制图特别有用。控制图是最重要和常用的工具,用于识别由特殊原因引起的波动。通过使用控制图消除这些变化,可以控制和改进对任何制造,生产或工业过程的监控。Shewhart型控制图可以有效地控制或检测过程中的大量由特殊原因引起的波动,而指数加权移动平均值(EWMA)-累积和类型控制图(CUSUM)则在过程中由特殊原因引起的波动数量中等和少量时更有效。我们常常假设参数是已知的或通过IC采样被正确估计,并且数据没有异常值。因此,通过这些假设,可以利用均值和方差(或标准偏差)控制图完成位置和比例参数的监测。但实际上,这些假设并不正确,过程偶尔会有异常。此外,利用关于辅助变量的信息有助于提高估计器的精度,并因此提高制图结构。本文致力于研究一些改进的控制图表结构,其作为插件用于统计过程控制(SPC)工具包。文中提出的图标结构是基于一些辅助特性的信息,可用于设计位置和尺度参数。文中方法的性能表现通过一些有用的度量来评估,譬如平均运行链长(ARL),额外二次损失(EQL),相对平均运行链长(RARL),性能比较指数(PCI)。我们分别在正态、对数正态以及学生t分布(有和没有噪声污染)过程中利用简单随机抽样度量其表现能力。本文利用蒙特卡罗模拟比较了不同的控制图策略,并做了一些真实数据的分析,以突出其实际应用价值。
【学位单位】：大连理工大学
【学位级别】：博士
【学位年份】：2018
【中图分类】：O213
【文章目录】：
Preface
Abstract
摘要
List of Abbreviations and Acronyms
1 Introduction
    1.1 Definition of a process
    1.2 Statistical process control （SPC）
    1.3 Control charts
    1.4 Types of control charts
        1.4.1 Shewhart control charts
        1.4.2 EWMA control charts
        1.4.3 CUSUM control charts
    1.5 Performance measures
        1.5.1 Average run length （ARL）
        1.5.2 Extra quadratic loss （EQL）
        1.5.3 Relative average run length （RARL）
        1.5.4 Performance comparison index （PCI）
    1.6 Literature Review
        1.6.1 Improved EWMA control charts
        1.6.2 Improved CUSUM control charts
        1.6.3 Combined/mixed structure based control charts
        1.6.4 Auxiliary information based control charts
    1.7 Motivation and problem statement
    1.8 Objective of the study
    1.9 Organization of the thesis
2 On Auxiliary Information Based Improved EWMA Median Control Charts
    2.1 Median estimators
    2.2 Proposed EWMA structure
    2.3 Comparative analysis
        2.3.1 Comparison of control charts under uncontaminated environment
        2.3.2 Comparison of control charts under contaminated environments
    2.4 Illustrative example
        2.4.1 Simulated illustration
        2.4.2 Case study
    2.5 Concluding remarks
3 New Auxiliary Information Based Cumulative Sum Median Control Charts forLocation Monitoring
    3.1 Proposed CUSUM control charting structure
    3.2 Simulation procedure
    3.3 Comparative analysis
        3.3.1 Comparison of control charts in an uncontaminated scenario
        3.3.2 Comparison of control charts in a contaminated scenario
    3.4 Case study
    3.5 Concluding remarks
4 On a Class of Mixed EWMA-CUSUM Median Control Charts
    4.1 Proposed mixed EWMA-CUSUM median structure
    4.2 Simulation algorithm
    4.3 Results and discussion
        4.3.1 Comparison of control charts under uncontaminated environment
        4.3.2 Comparison of control charts under contaminated environment
    4.4 Case study
    4.5 Concluding remarks
5 New Interquartile Range EWMA Control Charts
    5.1 Quartiles, R and S estimators
    5.2 Design structures of EWMA R,S and proposed IQR control charts
    5.3 Simulation procedure
    5.4 Comparison and results discussion
        5.4.1 Comparison of control charts under uncontaminated environment
        5.4.2 Comparison of control charts under contaminated environment
    5.5 Case study
    5.6 Concluding remarks
6 New dual auxiliary information based EWMA control chart
    6.1 Usual, difference, and regression-type estimators
        6.1.1 Usual estimator under SRS
        6.1.2 Difference and Regression-type estimators
    6.2 Existing design structures
        6.2.1 EWMAC control chart structure
        6.2.2 EWMAD control chart structure
    6.3 Proposed control chart structures
    6.4 Results and discussion
    6.5 Real life example
    6.6 Concluding remarks
7 Summary, Conclusions and Future Recommendations
    7.1 Summary and conclusions
    7.2 Future work recommendations
References
Appendix A （Supplementary tables）
Research Projects and Publications during PhD Period
Acknowledgement
Author Information


本文编号：2874638