Fluid catalytic cracking optimisation using factorial design and genetic algorithm techniques

Authors

  • José F. Cuadros,

    Corresponding author
    1. LOPCA/UNICAMP-Laboratory of Optimisation, Design and Advanced Control, Department of Chemical Process, School of Chemical Engineering, University of Campinas, UNICAMP—13083-970, Campinas, Sao Paulo, Brazil
    • LOPCA/UNICAMP-Laboratory of Optimisation, Design and Advanced Control, Department of Chemical Process, School of Chemical Engineering, University of Campinas, UNICAMP—13083-970, Campinas, Sao Paulo, Brazil
    Search for more papers by this author
  • Delba C. Melo,

    1. LOPCA/UNICAMP-Laboratory of Optimisation, Design and Advanced Control, Department of Chemical Process, School of Chemical Engineering, University of Campinas, UNICAMP—13083-970, Campinas, Sao Paulo, Brazil
    Search for more papers by this author
  • Rubens Maciel Filho,

    1. LOPCA/UNICAMP-Laboratory of Optimisation, Design and Advanced Control, Department of Chemical Process, School of Chemical Engineering, University of Campinas, UNICAMP—13083-970, Campinas, Sao Paulo, Brazil
    Search for more papers by this author
  • Maria R. Wolf Maciel

    1. LDPS/UNICAMP, Laboratory of Separation Process Development, Department of Chemical Process, School of Chemical Engineering, University of Campinas, UNICAMP—13083-970, Campinas, Sao Paulo, Brazil
    Search for more papers by this author

Abstract

Statistical techniques coupled with genetic algorithm (GA) were used to identify optimal values of key operational variables in fluid catalytic cracking (FCC) process. A Kellog Orthoflow F fluid catalytic cracking process model was considered. It is known as a highly nonlinear process with a large number of variables with strong interactions among them. A reduced process model was obtained through factorial design technique to be used as a process function in the optimisation work giving as result the operational conditions that maximise conversion without infringing operational restrictions with savings in computational burden and time. An increase of 8.71% in process conversion was achieved applying GA as optimisation technique. © 2012 Canadian Society for Chemical Engineering

Abstract

Les techniques statistiques couplées à l'algorithme génétique ont été utilisées pour identifier les valeurs optimales des principales variables opérationnelles du procédé de craquage à catalyseur fluide. Un modèle du procédé de craquage à catalyseur fluide Kellog Orthoflow F a été utilisé. Ce procédé, renommé pour être fortement non-linéaire, possède un grand nombre de variables ayant de fortes interactions entre elles. Un modèle réduit de ce procédé, obtenu via l'élaboration d'un plan factoriel, a été utilisé pour déterminer conditions optimales d'opération qui maximisent la conversion tout en respectant les contraintes d'exploitation. De plus, la complexité et le temps de calcul ont été avantageusement réduits. Une augmentation de 8.71% de la conversion a été réalisée grâce à l'application de l'algorithme génétique comme technique d'optimisation. Copyright © 2012 Canadian Society for Chemical Engineering

Ancillary