Development of a Structure-Based Lumping Kinetic Model for Light Gas Oil Hydrodesulfurization

Abstract

10Adopting the petroleomics concept, significant effort has been made to create extensive databases of 11 molecular structures and chemical and physical properties of complex petroleum mixtures. By 12 collating the available information provided by petroleomics, this study develops a new 13 structure-based lumping kinetic model for hydrodesulfurization (HDS) of light gas oil. An advanced 14 experimentation system, analytical technique, and computer software tool are employed to generate 15 the necessary data. The model contains 16 structure-based lumps, each of which includes species 16 having similar structure and reactivity. The model allows for the tracking of changes in molecular 17 structure type of the input and output mixtures during HDS. Its prediction capability is validated over 18 a wide range of HDS operation temperatures (200°C-375°C). Petroleomics has emerged as a new field in petroleum technology. 17,18 In petroleomics, all petroleum 35 components are characterized and their properties and reactivity are correlated with the composition 36 data. The application of petroleomics promises a variety of opportunities for the efficient use of both 37 conventional and unconventional crude oils, the elimination of control and operation problems, and 38 the reduction of production costs. The advanced analytical technique Fourier Transform Ion 39 Cyclotron Resonance Mass Spectrometry (FT-ICR MS) has allowed for the characterization of 40 thousands of molecular structures 19 . New computer software tool KMT (Kinetic Modeler's Toolbox) 41 has stimulated the development of molecular-level kinetic models. 20-26 By combining a high 42 throughput experimentation (HTE) system, FT-ICR MS and KMT, Japan Petroleum Energy Center 43 has successfully developed a petro-informatics platform (PIP) 27 , which comprises large databases of 44 the elemental compositions, physical properties, and reactivity behaviors of different petroleum 45 fractions. Froment and co-workers developed single-event concept for building detail kinetic models 46 capable of describing the chemistry transformation of each molecule involved in the feedstocks. 28,29 47 Owing to inclusion of a large number of elementary reactions and chemical species, such 48 molecular-level kinetic models cannot be applied to the design, simulation, and optimization of 49 refinery reactors which are frequently performed via process simulators. 50 This study aims to develop a new structure-based lumping kinetic model for HDS of LGO (LGO-51 HDS). The chemical species involved in the feedstock and product mixtures are lumped in light of 52 types of molecular structures. To build the model, data are obtained from different sources: HTE, 53 FT-ICR MS, and KMT. The model is developed through a new approach that includes the following 54 steps: (1) analyzing elementary reaction rates to reduce unimportant reactions; (2) grouping reactant 55 species into lumps and defining the lump-based reaction network; (3) building a lump-based reaction 56 kinetic model; and (4) estimating ...

Model reduction in chemical dynamics: slow invariant manifolds, singular perturbations, thermodynamic estimates, and analysis of reaction graph

“…These combinations are often guessed on the basis of known kinetic properties and can be improved by iterative methods and observer theory. The standard examples are: (i) the lumped species are identified as the sums of species in selected groups (a very popular approach with many practical applications, for example, [9]); (ii) the lumped species are the numbers of links and structural fragments of various types and in different states (this approach has many applications, from petrochemitry [10] and modelling of intracellular networks [11] to the Internet dynamics [12]).…”

Prediction of Molecular Distribution and Temperature Profile of FCC Process through Molecular-Level Kinetic Modeling

“…The molecular-level kinetic model originated in the 1990s and has received wide attention. So far, it has been broadly applied to complex reaction systems, including petroleum, coal, and biomass [7][8][9][10][11][12][13][14][15][16] . For the FCC process, according to the molecular details presented in the reaction network, the kinetic model can be classified into two categories: mechanistic-level kinetic model and pathway-level kinetic model.…”

Molecular‐level kinetic modeling of heavy oil fluid catalytic cracking process based on hybrid structural unit and bond‐electron matrix

“…Recently, Zhou et al applied the BEM method to develop a molecular‐level kinetic model of a naphtha continuous catalytic reforming process 24 . Nguyen et al developed a kinetic model for light gas oil hydrodesulfurization using KMT 25 . Furthermore, the Green research group developed a reaction mechanism generator (RMG) 26 and used the software to develop various mechanistic level kinetic models 27,28 .…”