Investigating chaos in an industrial fluid catalytic cracking unit

A fluid catalytic cracking (FCC) unit is known to contribute approximately 40% of revenue in a typical petroleum refinery. The FCC unit in one of the Southeast Asian refineries is not performing up to expectations, and wide fluctuations in the riser temperature are observed. This undesired occurrence has an adverse effect on the refinery´s profit. Our study aspires to identify the root cause of this problem by implementing a set of statistical tools that utilizes routine operating data to characterize the dynamics of the riser temperature. Results show that the riser temperature data are nonlinear, chaotic and/or contaminated with correlated noise. Implications are that linear controllers are inadequate for controlling the nonlinear/chaotic FCC unit, thus resulting in wide fluctuations. Further investigation of chaotic behavior by developing and using a dynamic model of the FCC unit is in progress.