有序Cram-Schmidt正交化在研究无取向硅钢质量中的应用

Apply Order Gram-Schmidt Orthogonalization to Research Quality of Disoriented silicon steel Sheet

无取向硅钢工艺是一个多变量系统,影响产品质量的素多,本工作从生产现场收集了147组数据,应用有序Gram-Schmidt正交化等多元分析方法处理数据,找出影响质量的主要因素,根据PRESS判据排除多噪音的因素,建立预报能力较强的模型,并从分析中找出提高质量的措施。

The production of disoriented silicon steel sheet is a multivariate system,in which there are many factors effecting the quality of the products. In this work the data sets included 147 samples are collected from the real production process and the Order Gram-Schmidt orthogonalization method is used to deal with the data in order to find the main factors effecting the quality of products. According to PRESS criteria,the vaiables with more noise are discarded and the model with high predictive ability is establised. The measure for improving the quality of products is determinded in the analysis.