Page 204 - Characterization and Properties of Petroleum Fractions - M.R. Riazi

T1: IML
QC: —/—
P2: KVU/KXT
P1: KVU/KXT
AT029-Manual-v7.cls
June 22, 2007
21:30
AT029-Manual
AT029-04
184 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
600
A SG is calculated from Eq. (4.76) and then distribution of SG
versus x cv can be obtained through Eq. (4.56). Once SG dis-
Experimental tribution is known, the reversed form of Eq. (4.7) should be
550 Method A used to estimate M distribution. In a similar approach if n 7+
is known distributions of M, T b , and SG can be determined by
Method B
Boiling Point, T b , K 500 this method gives the least accurate distribution since min-
assuming I o = 0.22 and B I = 3 and use of Eq. (4.7). Obviously
imum information is used to obtain the distributions. How-
ever, this method surprisingly well predicts boiling point dis-
tribution from specific gravity (as the only information avail-
450
able) for some crude oils as shown by Riazi et al. [40].
400
4.6 PSEUDOIZATION AND
LUMPING APPROACHES
350 Generally analytical data for reservoir fluids and crude oils are
0 0.2 0.4 0.6 0.8 1
available from C 1 to C 5 as pure components, group C 6 , and all
remaining and heavier components are grouped as a C 7+ frac-
tion as shown in Tables 1.2 and 4.1. As discussed earlier for
Cumulative Weight Fraction, x cw
wide C 7+ and other petroleum fractions assumption of a sin-
FIG. 4.24—Prediction of boiling point distribution in Exam- gle pseudocomponent leads to significant errors in the char-
ples 4.12 and 4.13.
acterization scheme. In such cases, distribution functions for
various characterization parameters are determined through
Methods A or B discussed in Section 4.5.4.6. Once the molar
because the system is gas condensate and value of B M is one. distribution is known through an equation such as Eq. (4.56),
For very heavy oils Method B predicts better prediction. As the mixture (i.e., C 7+ ) can be split into a number of pseudo-
shown in Fig. 4.24 for T b , Method B gives better prediction components with known x i , M i , T bi , and SG i . This technique is
mainly because information on at least one type of distribu- called pseudoization or splitting and is widely used to charac-
tion was available. terize hydrocarbon plus fractions, reservoir fluids, and wide
boiling range petroleum fractions [15, 17, 18, 23, 24, 26, 36].
Method C: M 7+ and SG 7+ are known—An alternative to In some other cases detailed analytical data on the compo-
method A when only M 7+ and SG 7+ are known is to predict M sition of a reservoir fluid are available for SCN groups such
distribution by assuming B M = 1 and a value for M o as steps as those shown in Table 4.2. Properties of these SCN groups
1–5 in method A. For every value of M, SG is estimated from are determined from methods discussed in Section 4.3. How-
Eq. (4.7) using coefficients given in Table 4.5 for SG. Then ever, when the numbers of SCN components are large (i.e.,
--`,```,`,``````,`,````,```,,-`-`,,`,,`,`,,`---
parameters SG o , A SG , and B SG are calculated. From these co- see Table 4.2) computational methods specially those related
efficients SG av is estimated and compared with experimental to phase equilibrium would be lengthy and cumbersome. In
value of SG 7+ . The initial guessed values for M o and B M are ad- such cases it is necessary to lump some of these components
justed until error parameter for calculated SG av is minimized. into single groups in order to reduce the number of com-
In this approach, refractive index is not needed. ponents in such a way that calculations can be performed
Method D: Distribution of only one property (i.e., N ci , smoothly and efficiently. This technique is called lumping or
M i , T bi , SG i , or N i ) is known—In this case distribution of grouping [24, 26]. In both approaches the mixture is expressed
only one parameter is known from experimental data. As an by a number of pseudocomponents with known mole frac-
example in Table 4.2, distribution of only M i for the waxy tions and characterization parameters which effectively de-
crude oil is originally known versus weight or mole fraction. scribe characteristics of the mixture. These two schemes are
In this case from values of M i , boiling point and specific grav- discussed in this section in conjunction with the generalized
ity are calculated through Eq. (4.7) and coefficients given in distribution model expressed by Eqs. (4.56) and (4.66).
Table 4.5 for T b and SG. Once distributions of T b , SG, and
M are known the distribution coefficients can be determined. 4.6.1 Splitting Scheme
Similarly if instead of M i , another distribution such as T bi ,
SG i ,or I i is known, Eq. (4.7) can be used in its reversed form Generally a C 7+ fraction is split into 3, 5, or 7 pseudocompo-
to determine M i distribution as well as other properties. nents. For light oils and gas condensate systems C 7+ is split
Method E: Only one bulk property (M 7+ , T b7+ , SG 7+ , into 3 components and for black oils it is split into 5 or 7 com-
or n 7+ ) is known—One bulk property is the minimum in- ponents. For very heavy oils the C 7+ may be split to even 10
formation that can be known for a mixture. In this case if components. But splitting into 3 for gas condensate and 5 for
M 7+ is known, parameter M o is fixed at 90 and B M = 1. Para- oils is very common. When the number of pseudocomponents
meter A M is calculated from Eq. (4.74). Once distribution of reaches ∞, behavior of defined mixture will be the same as
M is found, SG distribution can be estimated through use of continuous mixture expressed by a distribution model such
Eq. (4.7) and coefficients in Table 4.5 for SG. Similarly if only as Eq. (4.56). Two methods are presented here to generate
SG 7+ is known, assume SG o = 0.7 and B SG = 3. Coefficient the pseudocomponents. The first approach is based on the

Copyright ASTM International
Provided by IHS Markit under license with ASTM Licensee=International Dealers Demo/2222333001, User=Anggiansah, Erick
No reproduction or networking permitted without license from IHS Not for Resale, 08/26/2021 21:56:35 MDT