质量控制中的统计相关分析与建模

介绍了统计相关分析方法及其在连续过程产品质量相关因素分析和质量指标的建模与估计中的应用。描述了多元统计分析中的主元素分析和部分最小二乘两种方法，并将这些方法用于连续过程的质量控制，进行炼油过程油品质量相关因素的分析和质量指标的统计建模与预测，对不同条件下的模型预测结果进行了分析比较。结果表明，统计分析是一种收集、整理、分析数据的有效方法，模型预测精度很高，它将为连续过程质量控制开辟新的途径。

Statistical correlation analysis and modelling in quality control

The methods of multivariate statistical analysis and their application in quality control of continuous process are introduced. An overview of the methods such as principal component analysis (PCA) and linear partial least squares (PLS) is presented. The statistical correlation analysis between process data and quality variables in the chemical process environment is discussed. A prediction model of quality variables is constructed using the PLS method. The prediction results of models in different conditions are compared. The results show these methods are very effective in data analysis and modelling, ever widening applications can be found in quality control of continous process.