A Monitoring Method Based on Modified Dynamic Factor Analysis and Its Application
Aimed at the dynamic characteristic of the process, the autocorrelation and the cross-correlation among process variables, a monitoring method based on Dynamic Factor Analysis was investigated which was widely concerned in academia and industry. However, the traditional Dynamic Factor Analysis utilizes the same time lags currently among all variables, but the computing work is complex and the auto-correlation isnt adequately exemplified. The paper proposes a new monitoring method based on modified dynamic factor analysis that determines different time lags according to different variables by correlation analysis, and constructs several statistics as the monitoring indices to monitor the operation condition of the dynamic process. The improved monitoring method is applied in the Tennessee-Eastman process and compared with DFA, the validity and superiority are proved.
Modified Dynamic Factor Analysis correlation analysis time-lags EM algorithm dynamic process monitoring
Xueyan Yin Fei Liu
Institute of Automation, Jiangnan University, 214122, wuxi, China
国际会议
无锡
英文
442-449
2010-09-17（万方平台首次上网日期，不代表论文的发表时间）