Page 166 - Characterization and Properties of Petroleum Fractions - M.R. Riazi

T1: IML
QC: IML/FFX
P1: IML/FFX
P2: IML/FFX
AT029-Manual-v7.cls
AT029-Manual
14:23
AT029-03
June 22, 2007
146 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
more general and applicable to various types of petroleum
curve. For prediction of each characteristic of a petroleum
fractions within the same boiling point range. The accuracy based on a minimum of three data points along the distillation
of a correlation for a specific group of fractions could be fraction, several methods are provided that have been in use
increased if the coefficients in the correlation are obtained in the petroleum industry. Limitations, advantages, and dis-
from the same group of fractions. In obtaining the constants, advantages of each method are discussed.
many equations may be convertible into linear forms and Basically two approaches are proposed in characterization
spreadsheets such as Lotus or Excel programs may be used to of petroleum fractions. One technique is based on the use
obtain the constants by means of least squared method. Non- of bulk properties (i.e., T b and SG) considering the whole
linear regression of correlations through these spreadsheets mixture as a single pseudocomponent. The second approach
is also possible. In analyzing the suitability of a correlation, called pseudocomponent technique considers the fraction as
2
the best criteria would be the R or correlation parameter a mixture of three pseudocomponents from the three fami-
defined by Eq. (2.136) in which values of above 0.99 indicate lies. This technique is particularly useful for heavy fractions.
an equation is capable of correlating data. Use of a larger data A third approach is also provided for wide boiling range frac-
bank and most recent published data in obtaining the nu- tions. However, since behavior of such fractions is similar to
merical constants would enhance accuracy and applicability that of crude oils the technique is mainly presented in the next
of the correlation. A fair way of evaluations and comparison chapter. Fractions are generally divided into light and heavy
of various correlations to estimate a certain property from fractions. For heavy fractions a minimum of three character-
the same input parameters would be through a data set not ization parameters best describe the mixture. Recommenda-
used in obtaining the coefficients in the correlations. In such tions on the use of various input parameters and advantages
evaluations AAD or %AAD can be used as the criterion to of different methods were discussed in Sections 3.8 and 3.9.
compare different methods. Average absolute deviation may For light fractions (M < 300, N C < 22) and products of atmo-
be used when the range of variation in the property is very spheric distillation unit, Eq. (2.38) for M, I, and d is quite ac-
large and small values are estimated. For example, in predic- curate. For such light fractions, T c , P c , and ω can be estimated --`,```,`,``````,`,````,```,,-`-`,,`,,`,`,,`---
tion of PNA composition, amount of aromatics varies from from Eqs. (2.65), (2.66), and (2.105), respectively. For pre-
1% in petroleum fractions to more than 90% in coal liquids. diction of the PNA composition for fractions with M < 200,
AAD of 2 (in terms of percentage) in estimating aromatic Eqs. (3.77) and (3.78) in terms of m and SG are suitable. For
content is quite reasonable. This error corresponds to 200% fractions with M > 200, Eqs. (3.71)–(3.74) in terms of R i and
in terms of %AD for fractions with 1% aromatic content. VGC are the most accurate relations; however, in cases that
Experience has shown that correlations that have fewer nu- viscosity data are not available, Eqs. (3.79) and (3.80) in terms
merical constants and are based on theoretical and physical of R i an CH are recommended. Special recommendations on
grounds with constants obtained from a wide range of data use of various correlations for estimation of different prop-
set are more general and have higher power of extrapolation. erties of petroleum fractions from their bulk properties have
been given in Section 2.10 and Table 2.16 in Chapter 2.
3.10 CONCLUSIONS AND
RECOMMENDATIONS 3.11 PROBLEMS

In this chapter various characterization methods for differ- 3.1. List four different types of analytical tools used for com-
ent petroleum fractions and mixtures have been presented. positional analysis of petroleum fractions.
This is perhaps one of the most important chapters in the 3.2. What are the advantages/disadvantages and the differ-
book. As the method selected for characterization of a frac- ences between GC, MS, GC–MS, GPC and HPLC instru-
tion would affect prediction of various properties discussed in ments?
the remaining part of the book. As it is discussed in Chapter 4, 3.3. A jet naphtha has the following ASTM D 86 distillation
characterization and estimation of properties of crude oils de- curve [1]:
pend on the characterization of petroleum fractions discussed
in this chapter. Through methods presented in this chapter vol% distilled 10 30 50 70 90
◦
one can estimate basic input data needed for estimation of ASTM D 86 temperature, C 151.1 156.1 160.6 165.0 171.7
thermodynamic and physical properties. These input param-
eters include critical properties, molecular weight, and acen- a. Calculate VABP, WABP, MABP, CABP, and MeABP for
tric factor. In addition methods of estimation of properties this fraction. Comment on your calculated MeABP.
related to the quality of a petroleum product such as distil- b. Estimate the specific gravity of this fraction and com-
lation curves, PNA composition, elemental composition, vis- pare with reported value of 0.8046.
cosity index, carbon residue, flash, pour, cloud, smoke, and c. Calculate the K W for this fraction and compare with
freezing points as well as octane and cetane numbers are pre- reported value of 11.48.
sented. Such methods can be used to determine the quality 3.4. A kerosene sample has the following ASTM D 86, TBP,
of a fuel or a petroleum product based on the minimum labo- and SG distribution along distillation curve [1]. Convert
ratory data available for a fraction. Methods of conversion of ASTM D 86 distillation curve to TBP by Riazi–Daubert
various types of distillation curves help to determine neces- and Daubert’s (API) methods. Draw actual TBP and pre-
sary information for process design on complete true boiling dicted TBP curves on a single graph in C. Calculate the
◦
point distillation curve when it is not available. In addition a average specific gravity of fraction form SG distribution
method is provided to determine complete distillation curve and compare with reported value of 0.8086.

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