Improved two-level monitoring system for plant-wide processes

摘要: Abstract In modern industries, the plant-wide process has become more and popular, which always consists of various operation units, equipments, workshops even factories. Therefore, it is difficult to monitor those processes, monitoring complexity also much higher. While traditional multivariate statistical analysis provides satisfactory results within a single part, e.g. unit, may fail catch detailed cross-information among different parts process. this paper, an improved two-level system formulated for processes. first level, latent variable information extracted by principal component (PCA) model, based on global matrix generated combining variables from order characterize cross-data process, efficient support vector data description (SVDD) method employed modeling relationships matrices. Based Tennessee Eastman (TE) enhanced performance obtained system. Compared PCA strategy, new useful describe accurate can be obtained.