摘要
在研究多向主元分析(MPCA——Multi-way Principal Component Analysis)理论的基础上,通过对间歇过程数据的分析研究移动窗口多向主元分析(MWMPCA——Moving Window Multi-way Principal Component Analysis)理论,并将该方法应用于TE过程进行故障检测与诊断.与MPCA方法比较,MWMPCA方法随采样的增加窗口长度不断改变,使窗口内有用的信息不断增加,所建模型更加准确,能提高监控系统的稳定性.通过对Q统计量、HotellingT2统计量的检测结果进行分析比较,证明MWMPCA理论在检测系统异常事件中能提高系统的准确性,使系统故障检测与诊断的性能得到改进.
In this paper,MWMPCA（Moving Window Multi-way Principal Component Analysis） was studied by analyzing MPCA（Multi-way Principal Component Analysis） for batch process,and this method was used for process fault detection and diagnosis.Based on this theory,as the samples raise,the window＇s length was raised too.Then the system stablility can be improved.The result of Q statistic and Hotelling T2 statistic were analysed and compared,demonstrating that the accuracy of fault detection and diagnosis and system characteristic can be improved.
出处
《沈阳化工大学学报》
CAS
2010年第2期170-174,共5页
Journal of Shenyang University of Chemical Technology
基金
国家自然科学基金资助项目(60774070)
辽宁省教育厅科学研究基金资助项目(20060669
20040041)
关键词
间歇过程
多向主元分析
移动窗口多向主元分析
故障检测与诊断
batch process
multi-way principal component analysis
moving window multi-way principal component analysis
fault detection and diagnosis
