화학공학소재연구정보센터
학회 한국화학공학회
학술대회 2005년 봄 (04/22 ~ 04/23, 여수대학교)
권호 11권 1호, p.188
발표분야 공정시스템
제목 Canonical Variate Analysis based Variable Reconstruction and Sensor Fault Diagnosis
초록 Many multivariate statisitical process control (MSPC) techniques have been developed for detection, isolaiton, and diagnosis on the modern chemical processes which have high dimensionality, strong correlations, and severe dynamics. Recently, the canonical variate analysis (CVA) has been researched to analyze process dynamics and to monitor process abnormalities. Several researches showed that CVA is superior to PCA in process abnormality detection. This paper proposes new CVA based variable reconstruction algorithms and sensor fault identification strategy using them. Through comparison to conventional dynamic PCA, it is verified that the proposed can be expected to be superior to the previous approaches for sensor fault identificaiton.
저자 이창규1, 최상욱2, 이인범3
소속 1포항공과대, 2CPACT, 3Univ. of Newcastle
키워드 process monitoring; fault diagnosis; variable reconstruction
E-Mail
원문파일 초록 보기