Dual-objective optimization for petroleum molecular reconstruction based on property and composition similarities

Petroleum molecular reconstruction method can be used to calculate molecular composition from limited analytical data, which is the basis of the molecular-level process modeling of petroleum refining. However, due to the problem of multiple solutions, it is difficult to obtain accurate and stable molecular compositional models. Therefore, based on the traditional bulk property constraints, composition similarity was proposed and a new petroleum molecular reconstruction method was proposed. In this article, diesel is taken as the study object, and the method was applied to construct various diesel compositional models. The results showed that the models' bulk properties and molecular fraction distribution were consistent with the experimental data. Finally, the method was applied to construct diesel compositional models of an actual refinery. With the reference oil compositional model as the molecular reference data, the problem of accurate construction of compositional models in a long-term production process without detailed characterization data was solved.