Kinetics modelling of ethylene cracking furnace based on SQP-CPSO algorithm

Abstract

The Kumar molecular reaction kinetics model (proposed by Kumar and Kunzru) is widely used to establish an ethylene cracking model because of its accuracy and simplicity. Parameters of the primary reaction equation are determined definitely by different oil properties and compositions. These parameters include changeable selectivity coefficients and minor-changeable kinetics parameters, which should be optimized in terms of different oils in an actual process. The paper proposes a new optimization algorithm based on a hybrid sequential quadratic programming–chaos particle swarm optimization (SQP-CPSO) to realize two optimal targets: elemental balances of carbon and hydrogen atoms, and minimum product yield errors. The SQP algorithm is to accelerate the local search in the CPSO global optimum process. The SQP-CPSO algorithm first acquires the optimized selectivity coefficients of non-linear kinetics modelling, and then if the elemental balance is not reached, the kinetics parameters are optimized to satisfy both optimal targets. The experimental results exceed the existing methods in its preferable realization of two targets. Moreover, this method has been proven effective and applicable with several kinds of naphtha with different characteristics from a real ethylene plant. The optimized kinetics model is helpful in predicting product yields, and optimizing and controlling operational conditions in an ethylene product system.