Genetic algorithm model development for prediction of main products in thermal cracking of naphtha: Comparison with kinetic modeling

A novel model for prediction of the main product yields of thermal cracking of naphtha was proposed using genetic algorithms (GAs). Researchers mainly use statistical approach and artificial neural network as empirical models for modeling the thermal crackers; however, in the present study, a new application of genetic algorithm in the modeling is proposed. A polynomial model is linked with a genetic algorithm optimization procedure for a more accurate estimation of the set of model parameters compare to a Design of Experiment (DoE) derived model. A semi-mechanistic kinetic model, as a conventional method, based on free radical chain reactions was also developed using experimental results. This semi-mechanistic kinetic model contains 96 reactions. The performance of mathematical method and GA tuned model was compared with each other and the ANOVA calculation was carried out. Moreover, GA tuned model significantly improved the accuracy of the fit. This model can be used to recommend operating conditions for further experimental works in naphtha thermal cracking especially in the kinetic studies.

► We model thermal cracking of naphtha with first principle modeling, kinetic model.
► Ninety six reactions are optimized for their activation energies and pre-exponential factors.
► Design of experiment model is linked with GA optimization to improve the accuracy.
► GA tuned, kinetic, and DoE models are compared and their advantages were emphasized.