PCA y Lógica Fuzzy para la detección de eventos anormales de operación, en una planta de Craqueo Catalítico Fluidizado FCC

In the present article we apply and adjust some techniques and methodologies for detecting faults in a Fluidized Catalytic Cracking refinery process (FCC) UOP model, initially performing a reliability analysis that defines the level of criticality of each equipment and identify potential fault modes and their effect on the operation of the process, with the information provided by this analysis, case studies are set as requirements for a monitoring, detection and classification of anomalous situations, which may lead to process a fault condition. To study cases, conditions are simulated with a model of the FCC unit operation and faults are detected with a tool based on PCA (Principal Component Analysis) and Fuzzy logic, and finally the tool is adjusted using historical process data in the presence of the fault.