In oil refineries, the residue fractions of crude oils need to be converted into more valuable products, such as gasoline and diesel. To obtain such products, the heaviest fractions are generally upgraded in residue conversion units. These petroleum residue conversion processes are based on the degradation of the largest molecules by thermal and/or catalytic cracking reactions at high temperature. Ebullated-bed hydroconversion units (sometimes referred to as residue hydrocracking units) are able to advantageously hydrogenate the high boiling components of the feed, such as vacuum residues, and crack them into smaller molecules.
An accurate prediction of process performances depends on the accuracy and reliability of the kinetic models. Classic kinetic models for complex hydrocarbon mixtures use a lumped kinetics strategy, grouping molecules into chemical families, based on global properties. However, such models assume that similar physical properties result in similar chemical reactivities. Moreover, for heavy hydrocarbon mixtures, the number of families and reactions turns out to be so vast that this lumping approach is no longer manageable. These limitations lead to the development of more detailed kinetic models containing molecule-based reaction pathways. However, these models expect a molecular description of the feed. Unfortunately, even the most advanced analytical techniques do not allow to identify the complete and quantitative molecular detail of heavy petroleum feed.
The present work focuses on the development of a novel two-step kinetic modeling strategy for heavy oil conversion processes. In this approach, both the feedstock composition and the process reactions are modeled at a molecular level. The composition modeling consists of generating a set of molecules whose properties are close to those of the process feedstock analyses. This set of molecules is generated by a two-step molecular reconstruction algorithm. In its first step, an equimolar set of molecules is built by assembling structural blocks in a stochastic manner. In the second step, the mole fractions of the molecules are adjusted by maximizing an information entropy criterion. Once the composition of the feedstock is represented, the conversion process is simulated by applying, event by event, its main reactions to the set of molecules by means of a kinetic Monte Carlo (kMC) method. The methodology has been applied to hydroconversion of Ural vacuum residue and both the feed and the predicted effluents were favorably compared to the experimental yield pattern.
Besides the prediction of the yield structure, the operability of the hydroconversion unit is also extremely important. Indeed, in order to maximize the conversion of the vacuum residue fraction, these hydroconversion units are operated at the highest possible severity. However, high severity operation is limited by the formation of sediments that lead to fouling of the down-stream equipment in the fractionation section. From their elemental analysis, these sediments are found to be a carbonaceous solid material, mainly formed by flocculation of asphaltenes
In this work, asphaltene fractions from various crude oils were also analyzed. Subsequently, the evolution of the asphaltenes was studied under hydroconversion conditions, as well as the sediments that were formed. During ebullated-bed hydroconversion of vacuum residues, the converted fractions (gasoline, middle distillate, VGO) are increasingly hydrogenated, while the unconverted VR, and especially the asphaltenes, become increasingly more aromatic with increasing severity. At high severity, the thermal conditions in the reactor therefore lead to the dealkylation of asphaltenes and hence to the formation of highly condensed polyaromatic hydrocarbons (PAH) that will finally aggregate and flocculate thereby forming sediments. This evolution is exemplified by reconstructing a model molecule from the available analyses that allows to represent the average asphaltene fraction at different conversion levels. Finally, the mitigation of sediment formation by co-processing a vacuum residue feedstock with relatively low amounts of heavy aromatic fractions (HCO) is illustrated. Ebullated-bed pilot testing demonstrated that sediments can be reduced by co-processing the vacuum residue with HCO from an FCC unit, and that HCO contents up to 15 vol% do not impact the conversion performances. By characterizing the effluents at various operating conditions, it was also shown that the HCO does not impact the asphaltene conversion mechanisms (since an identical average chemical composition and structure was obtained). Detailed analyses showed that this sediment reduction is unrelated to the composition and structure of the unconverted 540°C+ fraction. The stabilizing effect of the HCO can be attributed to the interactions between the unconverted asphaltenes and the tri- and tetra-aromatics in the vacuum gas oil that are introduced via the co-processed HCO.