Considered economic requirement for maximizing propylene yield of fluidized catalytic pyrolysis process in two-stage riser

as well as complex characteristics of the FCC process such as strong nonlinearity

coupling multivariable and unavailability in online measurement of product yields

a self-optimizing control strategy on a basis of multi-objective optimization was proposed. First

a multi-objective optimization framework for maximizing propylene yield while minimizing dry gas output was created from steady state model and operational constraints of the process

and solved for optimal operation condition with a complete and uniform Pareto distribution by standardized normal constraint method. Secondly

a self-optimizing scheme of cascade controls was generated from relationships between the optimal operation condition and the active constraints. Product yield that could not be measured online were estimated by an unscented Kalman filter transformation. Compared to the tracking control on temperature setpoints at the riser outlet

the self-optimizing control method could spontaneously adjust operating condition under circumstances of interference and reduce the disadvantageous impact of noise factors to optimizing process operation.