Adoption of Pareto Charts in the Analysis of Input Variables Interaction and Selection of Input-Output Pairings in a Refinery Fluid Catalytic Cracking Unit

Abstract

Princewill Nemibo Josiah, Ipeghan Jonathan Otaraku and Benson Osa Evbuomwan

This paper presents a study of the effects of input variables (main factors) and their couplings on key process vari- ables(responses) in a refinery fluid catalytic cracking (FCC)unit using pareto analysis. Five responses namely, Ris- er temperature(Trx),Regenerator temperature(Trg) ,flue gas oxygen concentration, (Od ),Gasoline yield(y2 ), Light gases yield(y3 )) and four main factors, namely gasoil feed rate(Fgr), regenerated catalyst flow rate(Frc),combustion air flow rate(Fa ), combustion air temperature(Ta )and were studied in a regular two-layer experimental design that generated 32 numerical experiments. A combination of Design Expert software and an in-mouse FCC unit simulator was used to con- duct the numerical experiments from which pareto plots were generated as a tool for response-factor analysis. Results from this study show that Frc and Fgr are the only significant main factors with respect to riser temperature response. While the significance is of the order Frc> Fgr, Frc produced negative effect while Fgr produced positive effect on riser temperature while the interaction factor (Fgr+Frc) is the only cross-coupling .Cross -coupling of variables is more signif- icant in the regenerator as two interaction factors (Fgr+Frc) and (Frc+Fa ) featured prominently and produced significant effects on regenerator temperature and flue gas oxygen concentration, respectively. The ranking of the effects on regen- erator temperature is Frc>Fa > Fgr while that on flue gas oxygen concentration is Frc>Fgr>Fa . Results further show that while Frc, Fgr and Fa produced positive effects on regenerator temperature, only Fa produced negative effects in flue gas oxygen concentration. Moreover, the four main factors in this study and their couplings did not dhow direct effects on gasoline yield and light gases yield respectively. The results are in tandem with FCC unit behaviour and thus assert the merit in the adopted tool.

PDF