Parameter estimation in naphtha pyrolysis based on chaos quantum particle swarm optimization algorithm

Parameter estimation is the key step to improve the precision of a mechanistic model, which is fundamental for simulation, control and optimization of industrial processes. A novel method based on nonlinear optimization has been developed to estimate the initial selectivities of the first-order primary reaction for the naphtha decomposition. The proposed approach is to minimize the discrepancy between the model simulated outputs and the industrial measured values, based on the naphtha feed characteristics and operating condition. The chaos quantum particle swarm optimization (CQPSO) algorithm is proposed and employed since the problem is strongly nonlinear and high dimensional. By introducing the chaos-mutation operator with quantum-states-updating strategy, a good balance between exploration and exploitation is maintained throughout the entire searching, which is demonstrated by numerical experiment. The proposed algorithm is proved to be effective by estimating 10 parameters in the reaction model for naphtha pyrolysis.