Estimation parameters of hydrocracking model with NSGA-ii (Non-dominated Sorting Genetic Algorithm) by using discrete kinetic lumping model

• Published in: Fuel (Guildford) Vol. 200; pp. 333 - 344
• Main Authors: Li, Guoqing, Cai, Chuxuan
• Format: Journal Article
• Language: English
• Published: Kidlington Elsevier Ltd 15.07.2017; Elsevier BV
• Subjects: Catalytic cracking; Classifying; Genetic algorithms; Hydrocracking; Kinetics; Lump; Lumping; Mathematical models; Nsga-ii (Non-dominated Sorting Genetic Algorithm); Parameter estimation; Perturbation methods; Process parameters; Sorting algorithms; Stangeland model; Lump; Parameter estimation; Stangeland model; Nsga-ii (Non-dominated Sorting Genetic Algorithm)
• ISSN: 0016-2361; 1873-7153
• DOI: 10.1016/j.fuel.2017.03.078

[Display omitted] •NSGA-ii is proposed to free model parameter estimation from providing weight factors.•Scopes of parameters are narrowed and one of them is found to be calculated directly.•Effects of parameters perturbations are shown by graphs for final selection. Process of parameter estimation of hydrocracking modeling has not received enough attention which is actually indispensable. In this work, NSGA-ii was introduced to estimate parameters for avoiding unifying temperature and yield deviations by giving weight factors which would be a challenge and obstacle for calculating ideal parameters. Besides, by studying the nature of parameters, one of six parameters in Stangeland model is found can be calculated directly for the first time and scopes of others are also discussed with some of them reduced which would help to find more reliable parameters. Moreover, effects of parameter perturbation on model results are displayed which would provide a guidance for adjusting and choosing better parameters as the final step. Comparison between proposed method and past method showing that it could help provide parameters with good universality and forecast ability.