【作者】 俞磊；

【导师】 刘飞；

【作者基本信息】 江南大学 ， 控制理论与控制工程， 2008， 硕士

【摘要】 预防质量缺陷、追求品质卓越的理念深入质量管理和生产过程,预防原则成为现代质量管理的核心与精髓。统计过程控制SPC作为质量控制的重要工具和技术,对预防产品生产全过程的质量缺陷具有举足轻重的作用。自20世纪20年代Shewhart博士提出第一张控制图以来,经过几十年的发展和应用,SPC在工业生产和质量改进方面取得了显著的效果。但是,传统SPC的实施仅限于质量特性值统计独立的假设,而在现代高速、自动化的工业过程中,观测序列往往存在不同程度的自相关性,这使得SPC在监控时容易漏发报警;SPC控制图以“点出界”作为过程失控的判异准则,而对点的好坏程度没有加以量化说明,并且这样的判异准则仅依赖当前的测量数据进行,容易对控制限内的点排列不随机而导致的过程失控漏发报警。统计质量管理及SPC的实施,必须以大量的测量数据为基础。同时,测量本身也是一个过程。因此,有必要对产生测量数据的测量系统进行分析,即实施测量系统分析MSA,而MSA的理论基础和基本方法是SPC技术。针对SPC的局限性及其研究现状,本文在SPC与EPC的整合框架下研究自相关过程的质量控制问题;在过程异常模式的识别中引入模糊聚类分析加以实现。对MSA的稳定性分析及MSA推广应用也作了相应研究。本文的贡献主要在以下几个方面:1.研究统计过程监控与调整,提出了在状态空间模型下,自相关过程的统计监控方法,对EPC反馈调整进行深入分析,从而将EPC控制器的设计推广到了一般情况。2.在SPC与EPC整合框架下,通过引入反馈,将Taguchi思想推广到分析自相关过程的质量损失,并进一步将Taguchi信噪比作为特性指标衡量SPC与EPC整合以及不同监控方法的效果。3.提出了采用模糊聚类分析的方法识别过程异常模式,实现统计独立过程和自相关过程的质量控制。4.将EWMA控制图应用于MSA稳定性分析中,提高了检测过程均值中存在小波动的效果;将MSA推广到恒加荷试验中,研究MSA在压力试验机精度检测中的应用。最后,在本文的研究基础上,提出了对SPC改进与MSA应用研究的一些思考。更多还原

【Abstract】 The ideas of preventing defects in quality and seeking transcendence of quality have been embedded in quality management and production process. Prevent principle has been a core and kernel of modern quality management. SPC, as an important tool and technique, has a crucial part in whole production processes. Since Dr Shewhart invented the first control chart in 1920’s, SPC has acquired remarkable effects in industrial production and quality improvement by developments and applications in these decades.Quality data, which are analyzed by traditional SPC, must satisfy hypothesis of independence. However, observation sequences are sometimes autocorrelation processes at a certain extent in modern high speed and automatic industry, which easily makes SPC omit alarms. SPC control charts judge processes in or out of control by the rule of points between UCL and LCL or not. As a result, SPC doesn’t show how points are good or bad. And the judgment rule only depends on current measured data, which easily omits alarms of no random points between UCL and LCL.Statistical quality management and SPC are based on a mass of measured data. And meanwhile, measurement itself is a process. So it is necessary to analyze the measurement systems, which are used to measure process data. Moreover, the basis theory and methods of MSA are SPC techniques.For localizations and study actuality of SPC, Quality Control of autocorrelation process was studied based on integrated SPC and EPC, and Fuzzy Cluster Analysis was used to identify abnormal patterns of process. Stability analysis and applications of MSA were also studied in this thesis.The major contributions of this thesis are as follows:1. Statistical process monitoring and adjustment was studied, and statistical monitoring methods of autocorrelation processes were presented by means of state space model. EPC feedback adjustment was analyzed deeply, and the design methods of EPC controllers were generalized.2. Based on integrated SPC and EPC, Taguchi’s Quality Loss was used to analyze autocorrelation process via feedback adjustment. And Taguchi’s S/N was used as a property index to evaluate effects of Integrated SPC and EPC and different monitoring methods.3. Fuzzy Cluster Analysis was used to identify abnormal patterns of process and realize Quality Control of independence and autocorrelation processes.4. EWMA control chart was used to enhance effects of small fluctuation in MSA stability analysis. MSA was generalized to the test of invariable loading rate, and applications of MSA were studied in precision inspection of compression testing machine.In the end, perspectives of improvements of SPC and applications of MSA were given based on studies in this thesis.更多还原