重油加氢处理分子层次模型构建与原料适应性考察

复杂体系反应动力学模型在石油加工过程工艺设计、操作条件优化等方面具有重要的应用价值。传统集总模型由于组成划分强烈依赖于建模所用原料，导致其原料适应性差。相比于集总模型，分子层次模型以分子作为最小构筑单元，模型的预测能力及原料适应性大为提高。笔者提出了分子层次模型构建策略，完成了重油加氢处理分子层次模型的构建，并将构建完成的重油加氢处理分子层次模型应用于不同重油原料，以考察其产物性质预测能力与原料适应性。结果表明：不同重油原料的组成模型宏观性质与实验数据一致，分子分布与高分辨质谱数据一致；反应动力学模型计算得到的产物宏观性质与分子分布符合实验结果，硫化物与氮化物的转化规律与实验规律一致。因此，分子层次模型具有一定的预测能力与原料适应性，相比于集总模型具有更高的工业应用价值。

Construction of Molecular Level Model for Heavy Oil Hydrotreating and Investigation of Feedstock Adaptability

The kinetic model of a complex system is of important application value in process design and operation condition optimization during petroleum refining. The traditional lumped model has poor feedstock adaptability because its composition division strongly depends on the feedstocks used for modeling. Compared with the lumped model, the molecular-level model takes molecules as the minimum building unit, which can greatly improve model prediction ability and feedstock adaptability. In this paper, the construction strategy of molecular-level model was proposed; the molecular-level model for heavy oil hydrotreating model was constructed and further applied to different heavy oil feedstocks to investigate its product prediction ability and feedstock adaptability. The results show that the bulk properties of the compositional models of different heavy oil feedstocks are consistent with the experimental data, and the molecular distribution is consistent with the high resolution mass spectrometry data. In addition, the bulk properties and molecular distribution of the products calculated by the reaction kinetics model are in line with the experimental results, and the transformation regularities of sulfides and nitrides are consistent with the experimental regularities. Therefore, it can be found that the molecular-level model has certain prediction ability and feedstock adaptability, and is of higher industrial application value than the lumped model.