摘要
采用常规控制图的基本假设是生产过程的输出变量服从正态分布。在实际生产中,虽然生产过程的某些质量特性(如形状与位置等几何质量特性)也呈现正态分布,但由于测量工具及这些特性自身的特殊性,测量数据通常呈现出以目标值为折叠点的折叠正态分布,这就无法应用常规的统计控制技术。为了解决这种情况下的质量控制问题,本文分析了形状与位置等几何质量特性测量过程中的特殊性,以及测量数据的分布规律,在此基础上,给出了评价工序能力的方法,构造了相应的控制图,并设计了相应的中文软件。实证结果表明:该软件既简单又实用,能适用于大部分形状和位置质量特性的质量控制问题。
Traditionally, control charts are applied for assuming that process observations follow normal distribution. Unfortunately, this assumption is frequently violated in practice. This is because most geometry errors of mechanical parts are distributed folded nornally by virtue of measurement tools and methods. They make conventional control chart invalid.In order to deal with this problem, this paper discusses the specific features in measuring geometrical performance and sets up the distribution function for this kind of measured data. On the basis of that, the method of evaluating the process performance is given, the related process control approach is presented and the software for feature of geometries performance analysis and control is introduced. The results show that the software is both simple and useful, and can be used to monitor production process being of quality characteristics of gemetry errors.
出处
《南京航空航天大学学报》
EI
CAS
CSCD
北大核心
1999年第4期480-483,共4页
Journal of Nanjing University of Aeronautics & Astronautics
基金
国家自然科学基金
国防科工委质量与可靠性中心A 类课题基金
关键词
折叠正态分布
统计控制
几何质量
geometry
quality
statistics
control
folded normal distribution