摘要
针对砂处理过程生产数据的特点,运用改进的PCA建模方法,建立了多变量统计过程控制模型,用SPE统计量来检测建模数据中的个别异常样本;通过SPE图和T2图两个控制图能监控出砂生产过程的波动情况。
Based on the analysis of characteristics of sand preparation process, an improved PCA modeling method has been proposed, and a multivariate statistical process （MSPC） model to detect individual abnormity sample in modeling data has been set up according to the rules ofthe variation squared prediction error （SPE） statistics. The rules of the variation SPE statistics and T2 statistics and variance increasing of process variables can manifest on the SPE chart and T2 chart, which have been used to monitor the variations of the real process.
出处
《铸造》
EI
CAS
CSCD
北大核心
2005年第8期797-799,共3页
Foundry
关键词
多变量统计过程控制
波动
监控
multivariate statistical process control
variation
monitoring
