Introduction

Current refinery models devote little attention to composition-thermodynamic contexts and massive interest on plant logistics using relaxed unit models and lumped descriptions, where individual chemical species do not appear. Although satisfactory for management, this may occlude information for improvement of processes with important roles reserved in the future. Probably hydrotreating (HDT) and hydrocracking (HCC) figure among them.

This work proposes a methodology for HDT modeling, consisting of:

  • Mixed discrete-continuous Compositional Model (CM) for HDT feeds;

  • CM tuned to a given feed via reconciliation with raw oil data;

  • Network of HDT chemical reactions on the continua of reagent and product species, allowing influence of species size on kinetics;

  • Reactor model analytically solved;

  • Model parameters estimated with isothermal pilot data.

Compositional Model for HDT Feeds The CM is conceived as a mixture of N F=27 continuous families of descendant species from 27 chemical ancestors (Table 1). Ancestors and rules for generation of descendant, were proposed with principles accepted in petroleum models1,2,3,4,5. Descendants are created via lateral attachment of paraffinic chains with length defined by a continuous variable, X, distributed positively according to scaled and shifted gamma Probability Density Functions (PDF) (Eq. 1). A given species, belonging to family i, is thus represented by the pair (i, X); F0(i, X-Di) is its molar density function per 100g of oil; U(X) is the unitary step function and (.) is the gamma function. (çi 1 +) e F (i, X - D) =C i (.X -D)çi .exp(-e (X -D U)). (X -D) 0 i i (1) Ã(ç 1 +) i i i i i Family i has four parameters:. [Ci, Di, i, i]. Ci is its number of moles per 100g of oil; Di is its shift parameter; i, i are gamma PDF parameters. We assume that:

  • only the first 17 families exist in a given feed, other families are created by HDT;

  • only two pairs of PDF parameters, [ L, L], [ H, H] are assigned to the 17 families (Table 1), each with its own [Ci, Di].

To represent all families in the same species continuum, X is defined for the benzenic family (No.1).

Parameters Di are pre-assigned to each family such that species (i, X) has the same number of carbons than a benzenic descendant (1, X-Di). Thus, D2 = D4 = D10; D3 = D5 = D7 = D11 = 8; BLOCK 2 - - FORUM 11 419

A COMPOSITIONAL FRAMEWORK FOR FEED CHARACTERIZATION APPLIED IN THE MOELING OF DIESEL HYDROTREATING PROCESSES D13 = 2 etc.

This assures, for instance, that benzothiophenic and benzenic species, mapped with the same value of X, are correctly linked, as reactant and product, by a hydrogenating reaction. 1 (L) 2 (L) 3 (H) 4 (L) 5 (H) 6 (H) 7 (H) 8 (H) 9 (L) 10 (L) 11 (H) 12 (L) S S

This content is only available via PDF.
You can access this article if you purchase or spend a download.