Page 136 - Characterization and Properties of Petroleum Fractions - M.R. Riazi

T1: IML
P2: IML/FFX
QC: IML/FFX
P1: IML/FFX
AT029-Manual-v7.cls
AT029-03
AT029-Manual
June 22, 2007
14:23
116 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
TABLE 3.16—Comparison of various methods of predicting pseudocritical properties and
acentric factor through enthalpy calculation of eight petroleum fractions [37].
Method of estimating input AAD, kJ/kg
parameters ∗a Liquid Vapor
Item T c , P c ω (437 data points) (273 data points)
1 Pseudocomp. Pseudocomp. 5.3 7.9
2 RD (80) LK 5.9 7.7
3 KL LK 5.8 7.4
4 Winn LK 9.9 12.8
Pseudocomp.: The pseudocomponent method by Eqs. (3.40)–(3.43) and (2.42) for T c , P c and ω; RD: Riazi–
a∗
Daubert [38] by Eqs. (2.63) and (2.64); LK: Lee–Kesler [39] by Eq. (2.103); KL: Kesler–Lee [40] by Eqs. (2.69)
and (2.70); Winn method [41] by Eqs. (2.94) and (2.95).
or by the pseudocomponent approach as discussed in Sec- The Twu method gives the highest error (AAD of 14.3%) fol-
tions 3.3.2–3.3.4. lowed by the Goossens with average deviation of 11.4%. The
For petroleum fractions, pseudocritical properties are not Twu and Goossens methods both underestimate the molecu-
directly measurable and therefore it is not possible to make lar weight of these heavy fractions. The Lee–Kesler method is
a direct evaluation of different methods with experimental more accurate for lighter fractions, while the API method is
data. However, these methods can be evaluated indirectly more accurate for heavier fractions. The pseudocomponent
through prediction of other measurable properties (i.e., en- method gives generally a consistent error for all fractions and
thalpy) through corresponding state correlations. These cor- the lowest AAD%. Errors generated by the API, Lee–Kesler,
relations are discussed in Chapters 6–8. Based on more than and the pseudocomponent methods are within the experi-
700 data points for enthalpies of eight petroleum fractions mental uncertainty in the measurement of molecular weight
over a wide range of temperature and pressure [1], different of petroleum fractions.
methods of estimation of pseudocritical temperature, pres-
sure (T pc ,P pc ), and acentric factor (ω) have been evaluated In summary, for light fractions (M < 300) methods recom-
and compared [37]. These petroleum fractions ranging from mended by the API for T c and P c (Eqs. 2.65 and 2.66) [2] or the
naphtha to gas oil all have molecular weights of less than 250. simple method of Riazi–Daubert (Eqs. 2.63 and 2.64) [38] are
Details of these enthalpy calculations are given in Chapter 7. suitable, while for heavier fractions the Lee–Kesler method
Summary of evaluation of different methods is given in Ta- (Eqs. 2.69 and 2.70) [40] may be used. The pseudocompo-
ble 3.16. As shown in Table 3.16, the methods of pseudocom- nent method may also be used for both T c and P c when the
ponent, Lee–Kesler, and Riazi–Daubert have nearly similar composition is available. For all fractions methods of calcu-
accuracy for estimating the critical properties of these light lation of acentric factor from the pseudocomponent or the
petroleum fractions. However, for heavier fractions as it is method of Lee–Kesler [39] presented by Eq. (2.105) may be
shown in Example 3.11, the methods of pseudocomponent used. Molecular weight can be estimated from the API method
provide more accurate results. [2] by Eq. (2.51) from the bulk properties; however, if the PNA
composition is available the method of pseudocomponent is
preferable especially for heavier fractions.
Example 3.11—Experimental data on molecular weight and
composition of five heavy petroleum fractions are given in
Table 3.17. In addition, normal boiling point, specific grav- 3.3.6 Estimation of Density, Specific Gravity,
ity, density, and refractive index at 20 C are also given [36]. Refractive Index, and Kinematic Viscosity
◦
Calculate the molecular weight of these fractions from the fol-
lowing five methods: (1) API method [2, 42] using Eq. (2.51), Density (d), specific gravity (SG), and refractive index (n) are
(2) Twu method [42] using Eqs. (2.89)–(2.92), (3) Goossens all bulk properties directly measurable for a petroleum mix-
method [43] using Eq. (2.55), (4) Lee–Kesler method [40] us- ture with relatively high accuracy. Kinematic viscosity at 37.8
◦
ing Eq. (2.54), and (5) the pseudocomponent method using or 98.9 C(ν 38(100) , ν 99(210) ) are usually reported for heavy frac-
Eqs. (3.40)–(3.43). Calculate the %AAD for each method. tions for which distillation data are not available. But, for
light fractions if kinematic viscosity is not available it should
be estimated through measurable properties. Methods of es-
timation of viscosity are discussed in Chapter 8; however, in
Solution—Methods 1, 2, and 4 require bulk properties of T b
this chapter kinematic viscosity at a reference of temperature
and SG, while the method of pseudocomponent requires T b
and the PNA composition as it is shown in Example 3.10. of 37.8 or 98.9 C (100 For210 F) is needed for estimation
◦
◦
◦
Method 2 requires T b and density at 20 C(d 20 ). Results of of viscosity gravity constant (VGC), a parameter required for
◦
calculations are given in Table 3.18. prediction of composition of petroleum fractions. Generally,
TABLE 3.17—Molecular weight and composition of five heavy petroleum fractions of Example 3.11 [36].
No. M T b , C SG d 20 , g/ml n 20 P% N% A%
◦
1 233 298.7 0.9119 0.9082 1.5016 34.1 45.9 20.0
2 267 344.7 0.9605 0.9568 1.5366 30.9 37.0 32.1
3 325 380.7 0.8883 0.8845 1.4919 58.4 28.9 12.7
4 403 425.7 0.9046 0.9001 1.5002 59.0 28.0 13.0
5 523 502.8 0.8760 0.8750 1.4865 78.4 13.3 8.3

--`,```,`,``````,`,````,```,,-`-`,,`,,`,`,,`---
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