US20160092660A1

US20160092660A1 - Characterization of Complex Hydrocarbon Mixtures for Process Simulation

Info

Publication number: US20160092660A1
Application number: US14/287,980
Authority: US
Inventors: Jorge M. Martinis; Charles C. Solvason
Original assignee: BRE GROUP Ltd
Current assignee: BRE GROUP Ltd; Bryan Research & Engineering
Priority date: 2013-10-04
Filing date: 2014-05-27
Publication date: 2016-03-31
Also published as: US20190267116A1; US10566079B2

Abstract

A process for simulating the composition and properties of hydrocarbon mixtures such as petroleum fractions with a reduced number of representative compounds that closely match the characteristics of the complete molecular composition of the mixture.

Description

CROSS REFERENCE TO RELATED APPLICATIONS

This application claims benefit of Provisional Application Ser. No. 61/886,756 filed Oct. 4, 2013, the content, Figures and disclosure of which are incorporated herein by reference in their entirety for all purposes.

BACKGROUND

1. Field of Invention
A process for simulating the composition and properties of hydrocarbon mixtures such as petroleum fractions. More specifically, a process for simulating a composition of hydrocarbon mixtures with reduced number of representative compounds that closely match the characteristics of the complete molecular composition of the mixture.
2. Background
Petroleum fractions are mixtures of a huge number of component molecules. This makes detailed analytical characterization on a molecular level extremely difficult, if not impossible, and so costly and time consuming as to be largely impractical. There is a recognized need to have a more detailed molecular characterization of these hydrocarbon mixtures than is available from conventional analysis. This need is especially acute in fundamental kinetic models are used for process simulation. Such models have been developed for various chemical processes, such as steam cracking, pyrolysis, steam reforming, hydrocracking, catalytic cracking, etc. These fundamental kinetic models are able to simulate the chemical kinetics over a wide range of process conditions and for a wide range of feedstock types, by accounting for the occurring chemical reactions as well as for the physical transport phenomena governing the process. (Steven P. Pyl, Kevin M. Van Geern, Marie-Francoise Reyniers, and Guy B. Marin; Molecular Reconstruction of Complex Hydrocarbon Mixtures: An Application of Principal Component Analysis; AlChE Journal; December 2010 Vol. 56, No. 12) and (US Published application 2009/0105966, Apr. 23, 2009).
The need for estimating detailed molecular composition of petroleum fractions is well recognized in the art and there have been numerous attempts to develop adequate solutions. Developments by Neurock et al. (1994)¹and Trauth et al. (1994)²have applied molecular reconstruction techniques to approximate individual components in petroleum fractions via stochastic optimization methods. The application of stochastic methods is, however, limited by the computational intensity imposed by the large sampling frequency needed to achieve statistical significance (Verstraete, 2004)³. For specific petroluem fractions such as naphthas and gasoils, Hudebine and Verstraete (2011)⁴have utilized a hybrid stochastic method that maximizes an entropy function with Lagrangian parameters associated with analytical constraints in order to reduce the computational burden. However, the method is strongly dependent on the initial set of molecules which must be chosen and built in relation to the type of petroleum fraction studied. The present invention provides another, more useable process for generating hydrocarbon mixture compositional information. ¹Neurock M., Nigam A., Trauth D. M., Klein M T. (1994) Molecular Representation of Complex Hydrocarbon Feedstocks through Efficient Characterization and Stochastic Algorithms, Chem. Eng. .5c L 49,24.4153-4177.²Trauth D. M., Stark S. M., Petti T. F., Neurock M. Klein M. T. (1994) Representation of the Molecular Structure of Petroleum Resid through Characterization and Monte Carlo Modeling, Energ. Fuel. 8. 3. 576-580.³Verstraele L I., Revellin N. Dulot H Hudebine D. (2004) Molecular reconstruction of vacuum gasoils, Prep. Am. Chem. Soc. Div. Fuel Chem. 49,1,20-21.⁴Hudebine D. and Verstraele L I., Dulot H.; Reconstruction of Petroleum Feedstocks by Entropy Maximization. Application to FCC Gasolines; Oil & Gas Sciebce abs Technology, Rev. IFP Energies novelles; 2011

SUMMARY

The present invention is a novel approach to the molecular reconstruction of complex hydrocarbon mixtures for use, for example, in commercial process simulation programs.

DESCRIPTION OF FIGURES

The Figures represent embodiments and aspects of the invention and are not intended to be limiting of the scope of the invention.

FIG. 1 is a graphical representation the results of a comparison of boiling point results from an embodiment of the invention with analytical results.

FIG. 2 is a graphical representation the results of a comparison of composition results from an embodiment of the invention with analytical results.

DETAILED DESCRIPTION

The invention, in broad aspect, is a process to provide molecular reconstruction of complex hydrocarbon mixtures, especially for use in commercial process simulators. In one aspect it is a process to simulate a petroleum fraction (or other chemical hydrocarbon mixture) in a manner that represents the actual molecular composition in a sufficiently simplified manner for convenient utility. In another embodiment it is a process as described above but that also includes a comprehensive database of identifiable compounds and compound groups and components and from which specific “compound species” (defined below) are selected to develop representation of a chemical mixture composition.
In another embodiment the invention is a process that expands a list of compound species to individual compounds that will be used in, for example, chemical process simulation. These three embodiments are described in more detail below. Petroleum mixtures contain thousands, perhaps millions of individual compounds. It is not feasible to analytically obtain a complete list of the molecules. Moreover, even if it were possible the huge number would generally be unmanageable for practical use such as product characterization or in chemical process simulations.
Several techniques exist to generate a reduced set of species to represent petroleum (hydrocarbon) fractions. At present, conventional commercial process simulators use pseudo-components based on component boiling points that carry no chemical information, thus making them unsuitable for the simulation of chemical reactors. Other methods use a set of molecules selected by stochastic techniques, where molecules are constructed in terms of moieties and their group contributions and then added to or rejected from the mixture until a good match with the oil fraction properties is achieved. In the latter case, a “group contribution” (addition of properties of individual moieties that may make up a molecule) method is needed to estimate the properties of each species. Consequently, different molecules are only as distinguishable as the group contribution method permits, depending on how detailed their contributions or moieties are.
In general, the available analysis for petroleum fractions will include distillation boiling range curve(s), chromatographic analysis (for chemical families such as paraffins, olefins, naphthenes and aromatics) and sometimes Mass Spectrometry analysis for more detailed analysis.
In one aspect the invention is a process that includes assembling, for a given hydrocarbon mixture, a listing of compounds and “compound species” (as is defined below), together with physical and chemical properties and calculating a molecular fraction composition (or equivalent) that mimics available analytical information (that includes, at least, boiling range analysis) for the mixture.
As used herein and in the claims “compound species” is a collection of molecules in chemical equilibrium which, for all practical purposes, can be treated as a pure species for the calculation of unit operations in a process simulation program that does not involve chemical reactions. In effect, a compound species is a mixture of fixed composition acting as a single component within a larger mixture that contains it. Development of compound species is a key to reducing the number of species for practical use.
“Comprehensive database” means a database (or other suitable data storage and retrieval means) containing a collection of identifiable compounds and compound species (represented by group contribution moieties) together with physical and chemical properties for each identified pure and compound species selected from the database when needed to match a given sample for which the composition and properties are to be calculated (as representative of the actual molecular composition). The database is accessible by computer means and is maintained in a tangible, not transitory form.
In general, the database will contain hydrocarbon species with carbon numbers within a range of interest. The specific pure and compound species selected depends on the intended purpose of the database. For petroleum refining operations, for example, the database will advantageously contain compounds up to about C₄₀. This range is adequate for most purposes since petroleum hydrocarbon species above C₄₀are generally asphaltenes for which detailed individual species identification is currently not possible. It is advantageous to structure the database so that when a compound species is selected, the corresponding pure species contained within it will automatically be excluded and conversely when a pure species is selected the compound species that includes it will be excluded. This aids in preventing confusing duplication of compound or compound species in any subsquent calculation of the representative composition of a given sample.
“Display” as the term is used herein in the context of the results of calculations means any suitable means of exhibiting the data, as for example, by printing or exhibiting on a computer screen or monitor. Additionally, it means the results of adapting the data in a manner that it can be displayed in a more complex system such as the systems utilized in chemical process simulation. An example is to have results of calculated data displayed in a word process system such as Microsoft Word™, or a spreadsheet such as Microsoft Excel™ or a more complete system such as Microsoft Visio™. The results can adapted to be integrated into a chemical process simulation and such as ProMax™ (a chemical process simulation system available from Bryan Research and Engineering of Bryan, Tex.) and displayed on a coupled Excel spreadsheet or Microsoft Visio™.
The following discussion of specific applications illustrates the first embodiment of the invention and the steps necessary for construction of a more comprehensive database of the second embodiment. Embodiments of the process of the invention were used to characterize a Jet A-1 fuel in terms of pure and compound species. The bulk of a Jet A-1 fuel is a kerosene oil fraction that boils within the range of 180-300 ° C. A basic analysis of a Jet A-1 sample is presented in Table 1.
This oil fraction is a complex mixture containing thousands of molecules. Currently, there are not analytical techniques capable of resolving all the components and even if it did exist, no property database would have a fraction of the property values required and no chemical process simulator could handle such a large number of species. Therefore, a reduced set of species that properly represent the entire population of molecules in the mixture is desirable.

TABLE 1

Jet A-1 sample analysis

	Analysis	Jet A-1

	Specific Gravity	0.81
	Molecular Weight	170
	Hydrogen Content, wt %	0.138
	Aromatics, vol %	20.0
	Olefins, vol %	<0.1
	ASTM D86, vol %/° C.
	IBP	174
	10%	205
	30%	218
	50%	232
	70%	245
	90%	258
	FBP	300

EXAMPLE 1

For this particular example, an extended version of the Joback's group contribution method was applied. The Joback method predicts eleven important and commonly used pure component thermodynamic properties from molecular structure only. (Joback K. G., Reid R. C., “Estimation of Pure-Component Properties from Group-Contributions”, Chem. Eng. Commun., 57, 233-243, 1987), the teachings of which are incorporated n herein by reference.
Table 2 presents the moieties and their group contributions for the boiling point (Tb) utilized by this method for hydrocarbon species. Thus, for instance, the extended Joback's group contribution method allows the estimation the boiling point of n-butane as:
Tb (n-butane)=2×23.58 (CH₃—)+2×22.88 (—CH₂—)=92.92 K
A similar procedure carrying additional functionality was applied to estimate other pure species properties such as critical temperature, critical pressure, critical volume, freezing point, Pitzer acentric factor, etc.

TABLE 2

Extended Joback's Group Contributions to Boiling Point

		Tb[K]
	Moiety	Contribution

	CH₃—	23.58
	—CH₂—	22.88
	>CH—	21.74
	>C<	18.25
	═C_DH₂	18.18
	—C_DH═	24.96
	>C_D═	24.14
	═C_D═	26.15
	—C_AH═	26.73
	>C_A═	31.01
	(C_A) > C_A═	31.01
	(C_N) > C_A═	31.01
	—C_NH₂—	27.15
	>C_NH—	21.78
	>C_N<	21.32
	(C_N) > C_NH—	21.32

	CA: carbon in aromatic ring
	CD: double-bond bearing carbon
	CN: carbon in cycloalkane ring

A direct consequence of utilizing moieties from a group contribution method is that the population of distinguishable molecules in an oil fraction decreases significantly. For example, the following constructions shows how 2,3-dimethyl-pentane and 2,4-dimethyl-pentane become identical when described in terms of Joback's moieties:
2,3-dimethyl-pentane=2,4-dimethyl-pentane=4×(CH₃—)+3×(—CH₂—)
The general approach to construct and select a set of molecules that could represent the oil fraction mixture is by mathematical sampling. Monte-Carlo and other stochastic techniques typically generate molecules tied to a given probability density distribution (Gaussian, Etc.) following a defined set of building rules. On every attempt to construct a molecule, a moiety is tried and then accepted or rejected according to the building rule. Once constructed, every molecule is added to an equimolar mixture, pure and mixture properties are calculated and tested against the measured properties of the oil fraction. This method produces a different selection of molecules for every oil fraction characterized, offering no consistency when a single species base is required to represent several samples of oil fractions within the same boiling range.
A novel approach of embodiments of this invention is to take advantage of the fact that the number of indistinguishable molecules that can be represented for a given group contribution method is substantially less than the original population of molecules present in the oil fraction. Therefore, a set of building rules can be constructed so that all the possible indistinguishable molecules could be constructed. Thus, sampling, or selection of an optimum set is no longer required.
As an example, all possible alkanes within the kerosene approximate carbon range C₈-C₁₅are constructed. Based on the group contribution method, only (2) two structural characteristics are necessary to fully describe an alkane: carbon number (n) and number of side chains (s). Whereas the group contribution method provides a subset of four (4) moieties to construct any possible alkane: CH₃— (m₁), —CH₂— (m₂), >CH— (m₃) and >C< (m₄). After applying constraints to satisfy the conservation of bonding valences and atoms, the following relationships result:
n=m ₁ +m ₂ +m ₃ +m ₄ 1)
2n+2=3m ₁+2m ₂ +m ₃ 2)
s=m ₃+2m ₄ 3)
2+s=m ₁ 4)
With boundaries:
n=8, 2, . . . , 15 1)
0≦s≦|2(n−2)/3| 2)
0≦m ₄ ≦|s/2| (degree of freedom) 3)
In general, expressions (1)-(4) constitute a linear system of equations in the integer domain with a rank of three (3), leaving only one degree of freedom. The set of solutions for the moieties {m} given the structural characteristics {n,s}, constructs all the possible molecules that satisfy those structural characteristics.
A simple example is the construction of all the possible C₅alkanes: n=5 and s=0, 1, 2 as shown in Table 3. All possible solutions are found by counting over one degree of freedom (m₄) from 0 to |s/2|.

TABLE 3

C₅Alkanes Construction

		m₁	m₂	m₃	m₄
n	s	2 + s	n − 2 − 2s + m₄	s − 2m₄	m₄= 0 . . . \|s/2\|	Formula	Name

5	0	2	3	0	0	CH₃—CH₂—CH₂—CH₂—CH₃	n-pentane
5	1	3	1	1	0	(CH₃)₂CH—CH₂—CH₃	2-methyl-pentane
5	2	4	0	0	1	(CH₃)₄C	neopentane

A similar approach is also used for other hydrocarbon families such as alkenes, cycloalkanes, aromatics and naphtheno-aromatics. Further generalization requires additional structural characteristics such as number of aromatic rings, number of substituents, etc.
The iterative construction by a computer program of all possible distinguishable molecules within the kerosene carbon range (C₈-C₁₅) with a maximum number of side chains and ring substituents of four (4); results in the mixture presented in Table 4.

TABLE 4

Constructed molecules in the C₈-C₁₅range

		C₅	C₆	Aromatic
Hydrocarbon Family	Cores	rings	rings	rings	Molecules

Alkanes	—	—	—	—	123
Alkenes	—	—	—	—	389
Cycloalkanes	1	0	1	—	198
Cycloalkanes	1	0	2	—	40
Cycloalkanes	1	1	0	—	267
Cycloalkanes	1	1	1	—	65
Cycloalkanes	1	2	0	—	99
Cycloalkanes	2	0	2	—	24
Cycloalkanes	2	1	1	—	44
Cycloalkanes	2	2	0	—	69
Aromatics	1	—	—	1	112
Aromatics	1	—	—	2	25
Aromatics	2	—	—	2	11
Naphtheno-Aromatics	1	0	1	1	64
Naphtheno-Aromatics	1	1	0	1	99
Naphtheno-Aromatics	2	0	1	1	31
Naphtheno-Aromatics	2	1	0	1	58
					1718

Even though the number of molecules needed to characterize a kerosene sample has been reduced to a minimum by constructing only those molecules that the group contribution method of Joback is capable of distinguishing, a set of 1718 molecules is still too large for any practical application in commercial chemical process simulators. For process simulation purposes, the goal is to further reduce the number of species to less than a few hundred in order to match the average number of components that most compositional analyses can report.
A key to achieve a further reduction in the number of species is to introduce the concept of a compound species. A compound species is a collection of molecules in chemical equilibrium which for all practical purposes can be treated as a pure species for the calculation of any unit operation within a simulation that does not convey chemical reactions. In effect, a compound species is a mixture of fixed composition acting as a single component within a larger mixture that contains it.
As an example, suppose the smoke point of a kerosene sample is to be adjusted to the Jet A-1 fuel specification of 25 mm by passing it as a stream through a bed of alumina particles. It is known that in the presence of acid materials, hydrocarbons are transformed according to the rules of the carbenium ion chemistry and that in the particular case of alkenes, fast methyl- and hydride-transfer elementary steps occur leading to chemical equilibrium for all species with the same number of side chains for a given carbon number.
Based on this premise, 116 species are defined in terms of structural characteristics as shown in Table 5. Among the 116 species defined, there are 32 pure species and 84 compound species containing molecules with equal number of carbons, side chains and substituents in chemical equilibrium.
Furthermore, the set of 1718 molecules previously constructed to represent kerosene samples are mapped into 116 species. In the particular case of the kerosene sample shown in Table 1, the negligible content of olefins allowed the removal of alkene species from the mixture. As a result of mapping, 1010 molecules are accounted to characterize the kerosene sample in terms of 116 species.

TABLE 5

Kerosene composition from pure and compound species definitions

Structural Characteristics:

	Carbon	Side	Double		Aromatic	Aromatic
	Number	Chains	Bonds	Cores	Rings	Substituents

Species	Min	Max	Min	Max	Min	Max	Min	Max	Min	Max	Min	Max

Octane	8	8	0	0	0	0	0	0	0	0	0	0
SBOctanes	8	8	1	1	0	0	0	0	0	0	0	0
DBOctanes	8	8	2	2	0	0	0	0	0	0	0	0
TBOctanes	8	8	3	3	0	0	0	0	0	0	0	0
MBOctanes	8	8	4	4	0	0	0	0	0	0	0	0
Nonane	9	9	0	0	0	0	0	0	0	0	0	0
SBNonanes	9	9	1	1	0	0	0	0	0	0	0	0
DBNonanes	9	9	2	2	0	0	0	0	0	0	0	0
TBNonanes	9	9	3	3	0	0	0	0	0	0	0	0
MBNonanes	9	9	4	5	0	0	0	0	0	0	0	0
Decane	10	10	0	0	0	0	0	0	0	0	0	0
SBDecanes	10	10	1	1	0	0	0	0	0	0	0	0
DBDecanes	10	10	2	2	0	0	0	0	0	0	0	0
TBDecanes	10	10	3	3	0	0	0	0	0	0	0	0
MBDecanes	10	10	4	6	0	0	0	0	0	0	0	0
Undecane	11	11	0	0	0	0	0	0	0	0	0	0
SBUndecanes	11	11	1	1	0	0	0	0	0	0	0	0
DBUndecanes	11	11	2	2	0	0	0	0	0	0	0	0
TBUndecanes	11	11	3	3	0	0	0	0	0	0	0	0
MBUndecanes	11	11	4	6	0	0	0	0	0	0	0	0
Dodecane	12	12	0	0	0	0	0	0	0	0	0	0
SBDodecane	12	12	1	1	0	0	0	0	0	0	0	0
DBDodecane	12	12	2	2	0	0	0	0	0	0	0	0
TBDodecane	12	12	3	3	0	0	0	0	0	0	0	0
MBDodecanes	12	12	4	7	0	0	0	0	0	0	0	0
Tridecane	13	13	0	0	0	0	0	0	0	0	0	0
SBTridecane	13	13	1	1	0	0	0	0	0	0	0	0
DBTridecane	13	13	2	2	0	0	0	0	0	0	0	0
TBTridecane	13	13	3	3	0	0	0	0	0	0	0	0
MBTridecane	13	13	4	8	0	0	0	0	0	0	0	0
Tetradecane	14	14	0	0	0	0	0	0	0	0	0	0
SBTetradecane	14	14	1	1	0	0	0	0	0	0	0	0
DBTetradecane	14	14	2	2	0	0	0	0	0	0	0	0
TBTetradecane	14	14	3	3	0	0	0	0	0	0	0	0
MBTetradecane	14	14	4	8	0	0	0	0	0	0	0	0
Pentadecane	16	15	0	0	0	0	0	0	0	0	0	0
SBPentadecane	15	15	1	1	0	0	0	0	0	0	0	0
DBPentadecane	15	16	2	2	0	0	0	0	0	0	0	0
TBPentadecane	15	15	3	3	0	0	0	0	0	0	0	0
MBPentadecane	16	15	4	9	0	0	0	0	0	0	0	0
Cyclopentane	5	5	0	3	0	0	1	1	0	0	0	0
Methylcyclopentane	6	6	0	3	0	0	1	1	0	0	0	0
Cyclohexane	6	6	0	3	0	0	1	1	0	0	0	0
C7Cyclopentanes	7	7	0	3	0	0	1	1	0	0	0	0
Methylcyclohexane	7	7	0	3	0	0	1	1	0	0	0	0
C8Cyclopentanes	8	8	0	3	0	0	1	1	0	0	0	0
C8Cyclohexanes	8	8	0	3	0	0	1	1	0	0	0	0
C9Cyclopentanes	9	9	0	3	0	0	1	1	0	0	0	0
C9Cyclohexanes	9	9	0	3	0	0	1	1	0	0	0	0
C10Cyclopentanes	10	10	0	3	0	0	1	1	0	0	0	0
C10Cyclohexanes	10	10	0	3	0	0	1	1	0	0	0	0
C11Cyclopentanes	11	11	0	3	0	0	1	1	0	0	0	0
C11Cyclohexanes	11	11	0	3	0	0	1	1	0	0	0	0
C12Cyclopentanes	12	12	0	3	0	0	1	1	0	0	0	0
C12Cyclohexanes	12	12	0	3	0	0	1	1	0	0	0	0
C13Cyclopentanes	13	13	0	3	0	0	1	1	0	0	0	0
C13Cyclohexanes	13	13	0	3	0	0	1	1	0	0	0	0
C14Cyclopentanes	14	14	0	3	0	0	1	1	0	0	0	0
C14Cyclohexanes	14	14	0	3	0	0	1	1	0	0	0	0
C15Cyclopentanes	16	15	0	3	0	0	1	1	0	0	0	0
C15Cyclohexanes	15	15	0	3	0	0	1	1	0	0	0	0
C10Decalins	10	10	0	3	0	0	1	1	0	0	0	0
C11Decalins	11	11	0	3	0	0	1	1	0	0	0	0
C12Decalins	12	12	0	3	0	0	1	1	0	0	0	0
C13Decalins	13	13	0	3	0	0	1	1	0	0	0	0
C14Decalins	14	14	0	3	0	0	1	1	0	0	0	0
C15Decalins	15	15	0	3	0	0	1	1	0	0	0	0
C8Cyclopentanes(Di−)	8	8	0	3	0	0	1	1	0	0	0	0
C9Cyclopentanes(Di−)	9	9	0	3	0	0	1	1	0	0	0	0
C10Cyclopentanes(Di+)	10	10	0	3	0	0	1	1	0	0	0	0
C11Cyclopentanes(Di−)	11	11	0	3	0	0	1	1	0	0	0	0
C12Cyclopentanes(Di−)	12	12	0	3	0	0	1	1	0	0	0	0
C13Cyclopentanes(Di−)	13	13	0	3	0	0	1	1	0	0	0	0
C14Cyclopentanes(Di+)	14	14	0	3	0	0	1	1	0	0	0	0
C15Cyclopentanes(Di−)	15	15	0	3	0	0	1	1	0	0	0	0
C9CyclohexaneIndanes	9	9	0	3	0	0	1	1	0	0	0	0
C10CyclohexaneIndanes	10	10	0	3	0	0	1	1	0	0	0	0
C11CyclohexaneIndanes	11	11	0	3	0	0	1	1	0	0	0	0
C12CyclohexaneIndanes	12	12	0	3	0	0	1	1	0	0	0	0
C13CyclohexaneIndanes	13	13	0	3	0	0	1	1	0	0	0	0
C14CyclohexaneIndanes	14	14	0	3	0	0	1	1	0	0	0	0
C15CyclohexaneIndanes	15	15	0	3	0	0	1	1	0	0	0	0
C10TetrahydroAromatics(Mono)	10	10	0	3	0	0	1	1	1	1	0	4
C11TetrahydroAromatics(Mono)	11	11	0	3	0	0	1	1	1	1	0	4
C12TetrahydroAromatics(Mono)	12	12	0	3	0	0	1	1	1	1	0	4
C13TetrahydroAromatics(Mono)	13	13	0	3	0	0	1	1	1	1	0	4
C14TetrahydroAromatics(Mono)	14	14	0	3	0	0	1	1	1	1	0	4
C15TetrahydroAromatics(Mono)	16	15	0	3	0	0	1	1	1	1	0	4
C10Indanes	10	10	0	3	0	0	1	1	1	1	0	4
C11Indanes	11	11	0	3	0	0	1	1	1	1	0	4
C12Indanes	12	12	0	3	0	0	1	1	1	1	0	4
C13Indanes	13	13	0	3	0	0	1	1	1	1	0	4
C14Indanes	14	14	0	3	0	0	1	1	1	1	0	4
C15Indanes	15	16	0	3	0	0	1	1	1	1	0	4
C8Aromatics	8	8	0	3	0	0	1	1	1	1	0	2
C9Aromatics	9	9	0	3	0	0	1	1	1	1	0	3
C10Aromatics(Mono)	10	10	0	3	0	0	1	1	1	1	0	4
C11Aromatics(Mono)	11	11	0	3	0	0	1	1	1	1	0	4
C12Aromatics(Mono)	12	12	0	3	0	0	1	1	1	1	0	4
C13Aromatics(Mono)	13	13	0	3	0	0	1	1	1	1	0	4
C14Aromatics(Mono)	14	14	0	3	0	0	1	1	1	1	0	4
C15Aromatics(Mono)	15	16	0	3	0	0	1	1	1	1	0	4
C10Aromatics(Di−)	10	10	0	3	0	0	1	1	2	2	0	4
C11Aromatics(Di−)	11	11	0	3	0	0	1	1	2	2	0	4
C12Aromatics(Di−)	12	12	0	3	0	0	1	1	2	2	0	4
C13Aromatics(Di+)	13	13	0	3	0	0	1	1	2	2	0	4
C14Aromatics(Di−)	14	14	0	3	0	0	1	1	2	2	0	4
C15Aromatics(Di−)	16	15	0	3	0	0	1	1	2	2	0	4
C12Biphenyls	12	12	0	3	0	0	2	2	2	2	0	4
C13Biphenyls	13	13	0	3	0	0	2	2	2	2	0	4
C14Biphenyls	14	14	0	3	0	0	2	2	2	2	0	4
C15Biphenyls	16	15	0	3	0	0	2	2	2	2	0	4
C12Hexahydro-biphenyls	12	12	0	3	0	0	2	2	1	1	0	4
C13Hexahydro-biphenyls	13	13	0	3	0	0	2	2	1	1	0	4
C14Hexahydro-biphenyls	14	14	0	3	0	0	2	2	1	1	0	4
C15Hexahydro-biphenyls	16	15	0	3	0	0	2	2	1	1	0	4

Structural Characteristics:

	5-Carbon	6-Carbon	Cycloalkane	Pure
	Rings	Rings	Substituents	Species	Composition

Species	Min	Max	Min	Max	Min	Max	Count	molar

Octane	0	0	0	0	0	0	1	0.0000
SBOctanes	0	0	0	0	0	0	1	0.0000
DBOctanes	0	0	0	0	0	0	2	0.0000
TBOctanes	0	0	0	0	0	0	2	0.0000
MBOctanes	0	0	0	0	0	0	1	0.0000
Nonane	0	0	0	0	0	0	1	0.0000
SBNonanes	0	0	0	0	0	0	1	0.0000
DBNonanes	0	0	0	0	0	0	2	0.0000
TBNonanes	0	0	0	0	0	0	2	0.0000
MBNonanes	0	0	0	0	0	0	2	0.0000
Decane	0	0	0	0	0	0	1	0.0000
SBDecanes	0	0	0	0	0	0	1	0.0000
DBDecanes	0	0	0	0	0	0	2	0.0000
TBDecanes	0	0	0	0	0	0	2	0.0000
MBDecanes	0	0	0	0	0	0	3	0.0000
Undecane	0	0	0	0	0	0	1	0.0000
SBUndecanes	0	0	0	0	0	0	1	0.0000
DBUndecanes	0	0	0	0	0	0	2	0.0000
TBUndecanes	0	0	0	0	0	0	2	0.0000
MBUndecanes	0	0	0	0	0	0	3	0.0000
Dodecane	0	0	0	0	0	0	1	0.0810
SBDodecane	0	0	0	0	0	0	1	0.0000
DBDodecane	0	0	0	0	0	0	2	0.0000
TBDodecane	0	0	0	0	0	0	2	0.0000
MBDodecanes	0	0	0	0	0	0	3	0.0000
Tridecane	0	0	0	0	0	0	1	0.2280
SBTridecane	0	0	0	0	0	0	1	0.0000
DBTridecane	0	0	0	0	0	0	2	0.0000
TBTridecane	0	0	0	0	0	0	2	0.0000
MBTridecane	0	0	0	0	0	0	3	0.0000
Tetradecane	0	0	0	0	0	0	1	0.0194
SBTetradecane	0	0	0	0	0	0	1	0.0000
DBTetradecane	0	0	0	0	0	0	2	0.0000
TBTetradecane	0	0	0	0	0	0	2	0.0000
MBTetradecane	0	0	0	0	0	0	3	0.0000
Pentadecane	0	0	0	0	0	0	1	0.0000
SBPentadecane	0	0	0	0	0	0	1	0.0000
DBPentadecane	0	0	0	0	0	0	2	0.0000
TBPentadecane	0	0	0	0	0	0	2	0.0000
MBPentadecane	0	0	0	0	0	0	3	0.0000
Cyclopentane	1	1	0	0	0	0	1	0.0000
Methylcyclopentane	1	1	0	0	0	1	1	0.0000
Cyclohexane	0	0	1	1	0	0	1	0.0000
C7Cyclopentanes	1	1	0	0	0	2	3	0.0000
Methylcyclohexane	0	0	1	1	0	1	1	0.0000
C8Cyclopentanes	1	1	0	0	0	3	6	0.0000
C8Cyclohexanes	0	0	1	1	0	2	3	0.0000
C9Cyclopentanes	1	1	0	0	0	4	12	0.0000
C9Cyclohexanes	0	0	1	1	0	3	6	0.0000
C10Cyclopentanes	1	1	0	0	0	4	17	0.0000
C10Cyclohexanes	0	0	1	1	0	4	12	0.0000
C11Cyclopentanes	1	1	0	0	0	4	25	0.2050
C11Cyclohexanes	0	0	1	1	0	4	17	0.0000
C12Cyclopentanes	1	1	0	0	0	4	33	0.0000
C12Cyclohexanes	0	0	1	1	0	4	25	0.0000
C13Cyclopentanes	1	1	0	0	0	4	40	0.0000
C13Cyclohexanes	0	0	1	1	0	4	33	0.0000
C14Cyclopentanes	1	1	0	0	0	4	45	0.0000
C14Cyclohexanes	0	0	1	1	0	4	40	0.0000
C15Cyclopentanes	1	1	0	0	0	4	48	0.0000
C15Cyclohexanes	0	0	1	1	0	4	45	0.0000
C10Decalins	0	0	2	2	0	4	1	0.0000
C11Decalins	0	0	2	2	0	4	1	0.0000
C12Decalins	0	0	2	2	0	4	3	0.0000
C13Decalins	0	0	2	2	0	4	6	0.0000
C14Decalins	0	0	2	2	0	4	12	0.0000
C15Decalins	0	0	2	2	0	4	17	0.0000
C8Cyclopentanes(Di−)	2	2	0	0	0	4	1	0.0000
C9Cyclopentanes(Di−)	2	2	0	0	0	4	1	0.0000
C10Cyclopentanes(Di+)	2	2	0	0	0	4	3	0.0000
C11Cyclopentanes(Di−)	2	2	0	0	0	4	6	0.0000
C12Cyclopentanes(Di−)	2	2	0	0	0	4	12	0.0000
C13Cyclopentanes(Di−)	2	2	0	0	0	4	17	0.0000
C14Cyclopentanes(Di+)	2	2	0	0	0	4	26	0.0000
C15Cyclopentanes(Di−)	2	2	0	0	0	4	33	0.0000
C9CyclohexaneIndanes	1	1	1	1	0	4	1	0.0000
C10CyclohexaneIndanes	1	1	1	1	0	4	1	0.0000
C11CyclohexaneIndanes	1	1	1	1	0	4	3	0.0000
C12CyclohexaneIndanes	1	1	1	1	0	4	6	0.0000
C13CyclohexaneIndanes	1	1	1	1	0	4	12	0.2496
C14CyclohexaneIndanes	1	1	1	1	0	4	17	0.0000
C15CyclohexaneIndanes	1	1	1	1	0	4	26	0.0000
C10TetrahydroAromatics(Mono)	0	0	1	1	0	4	1	0.0000
C11TetrahydroAromatics(Mono)	0	0	1	1	0	4	2	0.0000
C12TetrahydroAromatics(Mono)	0	0	1	1	0	4	5	0.0000
C13TetrahydroAromatics(Mono)	0	0	1	1	0	4	10	0.0000
C14TetrahydroAromatics(Mono)	0	0	1	1	0	4	19	0.0000
C15TetrahydroAromatics(Mono)	0	0	1	1	0	4	27	0.0000
C10Indanes	1	1	0	0	0	4	2	0.0000
C11Indanes	1	1	0	0	0	4	6	0.0000
C12Indanes	1	1	0	0	0	4	10	0.0273
C13Indanes	1	1	0	0	0	4	18	0.0000
C14Indanes	1	1	0	0	0	4	26	0.0000
C15Indanes	1	1	0	0	0	4	37	0.0000
C8Aromatics	0	0	0	0	0	0	2	0.0000
C9Aromatics	0	0	0	0	0	0	4	0.0000
C10Aromatics(Mono)	0	0	0	0	0	0	7	0.0000
C11Aromatics(Mono)	0	0	0	0	0	0	10	0.0000
C12Aromatics(Mono)	0	0	0	0	0	0	14	0.0000
C13Aromatics(Mono)	0	0	0	0	0	0	18	0.1897
C14Aromatics(Mono)	0	0	0	0	0	0	21	0.0000
C15Aromatics(Mono)	0	0	0	0	0	0	23	0.0000
C10Aromatics(Di−)	0	0	0	0	0	0	1	0.0000
C11Aromatics(Di−)	0	0	0	0	0	0	1	0.0000
C12Aromatics(Di−)	0	0	0	0	0	0	2	0.0000
C13Aromatics(Di+)	0	0	0	0	0	0	4	0.0000
C14Aromatics(Di−)	0	0	0	0	0	0	7	0.0000
C15Aromatics(Di−)	0	0	0	0	0	0	10	0.0000
C12Biphenyls	0	0	0	0	0	0	1	0.0000
C13Biphenyls	0	0	0	0	0	0	2	0.0000
C14Biphenyls	0	0	0	0	0	0	3	0.0000
C15Biphenyls	0	0	0	0	0	0	5	0.0000
C12Hexahydro-biphenyls	0	0	1	1	0	4	1	0.0000
C13Hexahydro-biphenyls	0	0	1	1	0	4	4	0.0000
C14Hexahydro-biphenyls	0	0	1	1	0	4	7	0.0000
C15Hexahydro-biphenyls	0	0	1	1	0	4	15	0.0000

Once the reduction in the number of species to 116 without a significant sacrifice in accuracy is achieved, a problem remains in estimating the composition of the kerosene sample (Table 1) in terms of the mixture components (Table 5). Such inference is accomplished by determining the composition (X_i) that maximizes the Shannon entropy of the mixture (S) (Equation 1) and satisfies the constraints imposed by the application of mixing rules (Equation 2) in the calculation of the properties of the mixture to match the average properties of the sample (P_j).
$\begin{matrix} \max_{x} {S = - \sum_{i} X_{i} \cdot Ln (X_{i})} & (1) \\ {\overline{P}}_{j} = \sum_{j} X_{i} \cdot P_{i, j} & (2) \end{matrix}$
The estimated composition for the kerosene sample in Table 1 is reported in the last column of Table 5. The reconstruction of the kerosene sample by applying the mixing rules (Equation 2) is shown in Table 6.

TABLE 6

Kerosene sample reconstruction

	Analysis	Measured	Reconstructed

Specific Gravity	0.81	0.86
Molecular Weight	170	177
Hydrogen Content, wt %	0.138	0.138
Aromatics, vol %	20.0	20.4
Olefins, vol %	<0.1	0.0
ASTM D86, vol %/° C.
IBP	174	177
10%	205	206
30%	218	223
50%	232	224
70%	245	232
90%	258	262
FBP	300	264

The reconstructed sample shows a good agreement with the measured values. The decreasing accuracy of the estimated boiling temperatures with the proximity of the final boiling point are the consequence of increasing deviations with carbon number in the estimation of the boiling temperatures of pure species by Joback.
In cases where a detailed composition of the sample is required in a simulation (e.g. chemical reactor unit operation), the molar fraction (Z_j) of every pure species (j) in a compound species (i) is calculated from the molar fraction (X_i) of the compound species (i) by breaking it down back into their components in equilibrium as formulated in equation (3). The only information required to execute this calculation is the Gibbs free energy of formation of the pure species at the reference temperature (Equation 4). In the present example, it the equation is used to calculate the composition of the kerosene sample in terms of 1010 pure species.
$\begin{matrix} X_{i} = Z_{j} {1 + \sum_{k \neq j} K_{k, j}} & (3) \\ K_{k, j} = \exp {- \frac{G_{k} - G_{j}}{RT}} & (4) \end{matrix}$

EXAMPLE 2

Extended Application

The previous example illustrated the most basic application of the complex mixture modeling algorithm for the characterization of petroleum fractions. Within a commercial chemical process simulator, the algorithm enables the application of detailed kinetics and fundamental reactor models while keeping the number of species manageable. Moreover, it permits the creation of interfaces to go from petroleum fractions towards mixtures of pure and compound species to feed reactor models and then back to oil fractions.
The solution of equations (1) and (2) yields a composition that maximizes the Shannon Entropy of the mixture. However, this solution is not unique as any other composition that satisfies the constraints in equation (2) might also be valid and even approach closer to the real composition of the sample.
Consequently, in order to evaluate the performance of the algorithm of the process of the invention in estimating the compositions of the oil fractions, estimated compositions for a database of 50 naphtha samples were compared against experimental measurements. The database contains not only properties for every oil fraction such as specific gravity, molecular weight and boiling point curve, but also measured compositions of every sample from gas chromatography.
In this example, a species base comprising 235 molecules matching those previously identified in a gas chromatographic analysis was used. Furthermore, this base is grouped into a mixture of 68 species (47 compound species) that describes a naphtha cut. The results are presented in FIG. 2 in the form of parity plot for the overall composition in terms of hydrocarbon families.
This example illustrates the capability of the process of the invention to accurately model the composition of naphtha fractions and provide suitable composition information for further process simulation processes.
In another, second, aspect the invention includes assembling a compilation of hydrocarbon compounds and compound species (as defined below) that includes all compounds and compound species that will be useful for a desired purpose (such as petroleum refining), selecting from the assembly the compounds and compound species and their physical and chemical properties that may be included in a given hydrocarbon mixture, and calculating a composition that is consistent with available analytical information for the mixture.
Applicants developed an alternative, deterministic approach to molecular reconstruction to generate the new compilation. The approach replaces the stochastic methods with means to select from molecules from a molecular database that contains all chemical constituents and their isomers within a set of constraints tailored to capture petroleum hydrocarbon fractions of interest. Rather than build the database when needed to model a sample, applicants prepare a comprehensive database a priori, from which appropriate compound species and their respective physical and chemical properties, may be selected when needed to match a given sample for which the composition and properties are to be calculated (as representative of the actual molecular composition).
In general the database will contain hydrocarbon species with carbon numbers likely to be of interest. Carbon species up to about C₄₀is adequate for most purposes since petroleum hydrocarbon species above C₄₀are generally alphaltenes for which detailed individual species identification is impossible and of little process interest.
The method used by Applicants is dependent on two competing factors, the completeness of the database and the size of the database. Unfortunately, most gasoil reconstructions require representations of up to 40 atoms in a hydrogen suppressed molecular isomer which requires the creation of a database in size beyond the scope of the current state-of-the-art graph generate-and-test algorithms, such as those described by Pieroncely et al. (2012)(Julio E Peironcely, Miguel Rojas-Chertó, Davide Fichera4, Theo Reijmers, Leon Coulier, Jean-Loup Faulon and Thomas Hankemeier; OMG: Open Molecule Generator; Journal of Cheminformatics 2012, 4:21; http://www.jcheminf.com/content/4/1/21).
A higher carbon number hydrocarbon is even more difficult to create. As a result a hybrid, two stage method consisting of colored canonical graphs is used. Colored canonical graphs are representations of molecules consisting of extended Joback groups as the colors and the connectivity between all of the extended Joback groups in each molecule as the canonical graphs. Colored graphs are chosen to maximize computational speed during database construction and CPS run-time.
In the third embodiment the representative mixture as calculated in the first two embodiments is expanded to the individual compounds in the mixture. In effect it is a reversal of the determination of the compound species used in the first two embodiments. This is desirable when the mixture is to be used in chemical reaction simulation since the simulation will be far more accurate with specific compounds. While the consolidation of compound species is highly desirable to provide workable number of component, for chemical reaction simulations it cannot be presumed that the groups (compound species) will behave alike. Thus the calculated representative mixture is expanded by searching for all molecules in the colored graph database represented by each compound species. The resulting expanded hydrocarbon representative mixture the composition is then easily calculated by multiplying the percentage of the compound species in the unexpanded mixture by the equilibrium mixture percentage of the individual compounds.
While the invention has been particularly shown and described in particular embodiments above, those skilled in the art will understand that changes in form and detail may be made without departing from the spirit and scope of the invention.

Claims

1. A method of characterizing a hydrocarbon mixture comprising:

a) providing compilation of identified chemical pure or compound species marked with identifiers and including chemical and physical properties, wherein compound species comprise a compilation of related molecules in chemical equilibrium and wherein the compound species are computed by group contribution methods;

b) providing average analytical properties of a hydrocarbon mixture;

c) selecting pure and compound species for a specific hydrocarbon mixture of interest;

d) determining the composition of each pure or compound species in the mixture that maximizes the Shannon entropy of the specific mixture and satisfies the constraints imposed by the application of mixing rules;

e calculating properties of the hydrocarbon mixture that match the average analytical properties of the hydrocarbon mixture, and;

f) displaying the results.

2. The method of claim 1 wherein the mixture that maximizes the Shannon entropy is calculated by determining the composition (X_i) of each compound or compound species that maximizes the Shannon entropy of the mixture, S, by a model that satisfies the equation

\max_{x} {S = - \sum_{i} X_{i} \cdot Ln (X_{i})}

and satisfies the constraints imposed by the application of mixing rules by the equation

{\overline{P}}_{j} = \sum_{j} X_{i} \cdot P_{i, j}

and calculating of the properties of the mixture to match the average properties of the sample, P_j;

3. The method of claim 1 wherein the wherein the compound species are computed by a combination of group contribution and molecular connectivity methods.

4. The method of claim 1 wherein the hydrocarbon mixture comprises hydrocarbons having carbon number of 1 to 40.

5. The method of claim 1 wherein the method is integrated into a chemical process simulation computerized program.

6. The method of claim 1 wherein the results, f), are displayed in a spreadsheet computer program.

7. The method of claim 5 wherein the chemical process computerized program comprises Computer-Aided Process Engineering, CAPE-Open, process simulation modules.

8. The method of claim 5 wherein the chemical process computerized program is ProMax™.

9. The method of claim 1 wherein the results are displayed by a component of Microsoft Visio™.

10. A method for modeling a petroleum fraction comprising;

a) providing a database containing a compilation of listed molecular components and/or compound species of compounds together with their respective chemical and physical properties wherein each compound or compound species is tagged to allow it to be accessed by computer means;

b) selecting appropriate compounds and compound species from the database to be included in a second database specific to the petroleum fraction to be modeled;

b) providing average analytical properties of a hydrocarbon mixture;

c) determining the composition of each pure or compound species in the mixture that maximizes the Shannon entropy of the specific mixture and satisfies the constraints imposed by the application of mixing rules and placing the result in the second database;

e calculating properties of the hydrocarbon mixture that match the average analytical properties of the hydrocarbon mixture; and,

f) displaying the results.

11. The method of claim 10 wherein the mixture that maximizes the Shannon entropy is calculated by determining the composition (X_i) of each compound or compound species that maximizes the Shannon entropy of the mixture, S, by a model that satisfies the equation

\max_{x} {S = - \sum_{i} X_{i} \cdot Ln (X_{i})}

{\overline{P}}_{j} = \sum_{j} X_{i} \cdot P_{i, j}

12. The method of claim 10 wherein the wherein the compound species are computed by a combination of group contribution and molecular connectivity methods.

13. The method of claim 10 wherein the hydrocarbon mixture comprises hydrocarbons having carbon number of 1 to 40.

14. The method of claim 10 wherein the method is integrated into a chemical process simulation computerized program.

15. The method of claim 10 wherein the results, f), are displayed in a spreadsheet computer program.

16. The method of claim 14 wherein the chemical process computerized program comprises Computer-Aided Process Engineering, CAPE-Open, process simulation modules.

17. A method of representing the compounds in a hydrocarbon mixture by expanding the results of the calculated mixture of claim 10 by selecting all molecules in the first database represented by each compound species in the mixture and calculating the composition by multiplying the percentage of the compound species in the unexpanded mixture by the equilibrium mixture percentage of the individual compounds.