A protein inspired RNA genetic algorithm for parameter estimation in hydrocracking of heavy oil

Abstract

Hydrocracking is a crucial process in refineries and suitable model is useful to understand and design hydrocracking processes. Simulating the procedure from RNA to protein, a protein inspired RNA genetic algorithm (PIRGA) is proposed to estimate the parameters of hydrocracking of heavy oil. In the PIRGA, each individual is represented by a RNA strand and a new fitness function combining traditional fitness value and individual ranking is employed to maintain population diversity. Furthermore conventional crossover operators are replaced by RNA-recoding operator and protein-folding operators to improve the searching ability. An adaptive mutation probability in the PIRGA makes the algorithm have more chance to jump out of local optima. Numerical experiments on seven benchmark functions indicate that the PIRGA outperforms other genetic algorithms on both convergence speed and accuracy greatly. 10 parameters are obtained by the PIRGA and the kinetic model for hydrocracking of heavy oil is established. Experimental results reveal that the predictive values are in good agreement with the experimental data with relative error less than 5%. The effectiveness and the robustness of the model are also validated by experiments.