- SEMINARIO - Fault detection and isolation via causality detection methods

DISERTANTE: Prof. Celso Jose Munaro, de la Universidad Federal de Espiritu Santo, Brasil.
RESUMEN: Disturbances that propagate in industrial plants can cause many control loops to operate outside of security levels and modes of better economic profits. A method to indicate the source of disturbances directly from operating data is presented. The control loops affected by disturbances are clustered based on principal component analysis and Hotelling’s T2 statistic and the causal relationships between them are detected via causality detection methods. The methodology was applied to data from industrial plants and the proposed algorithm correctly indicated the sources of disturbances related to the fault.