Abstract—

Multi-criterion optimization is achieved for the performance of a reactor unit used in reforming oil fractions on the basis of a mathematical model created earlier. The model considers changes in the temperature and molar flow rate of a reaction mixture during chemical reactions, the composition of refined feedstocks, and the quality and ratio of the circulation of hydrogen-rich gas. The criteria for multi-criterion optimization are the octane number, the yield of the target product (reformate), and the combined and individual contents of aromatic hydrocarbons and benzene. Optimization is achieved by means of Pareto approximation using a genetic algorithm. The control parameters are the temperatures at the inlet to the reactors. Results from optimization show a reduction of 56 to 45 wt % in the total content of aromatic hydrocarbons, and the research octane number (RON) is lowered from 92.7 to 90.7. Three-objective optimization that ensures a 4 to 3.08 wt % reduction in the content of benzene and a 92.7 to 91.8 drop in the research octane number is also considered. The drop in the octane number is in both cases acceptable, considering the formulated requirements for the ultimate content of benzene and aromatic hydrocarbons in the reformate.

中文翻译：

---

基于遗传算法的催化重整装置多条件优化

摘要—

在先前创建的数学模型的基础上，针对重整油馏分的反应堆装置的性能实现了多标准优化。该模型考虑了化学反应过程中反应混合物的温度和摩尔流量的变化，精制原料的组成以及富氢气体循环的质量和比例。多标准优化的标准是辛烷值，目标产物（重整产物）的产率以及芳烃和苯的总和和单独含量。使用遗传算法通过帕累托逼近来实现优化。控制参数是反应器入口处的温度。优化结果表明，芳烃的总含量减少了56至45 wt％，研究辛烷值（RON）从92.7降低到90.7。还考虑了三目标优化，该优化可确保苯含量减少4至3.08 wt％，并且研究辛烷值下降92.7至91.8。考虑到对重整产品中苯和芳烃最终含量的规定要求，辛烷值的下降在两种情况下都是可以接受的。