Energy & Fuels 2003, 17, 689-693
689
Analysis of Quality of the Petroleum Fuels T. A. Albahri, M. R. Riazi,* and A. A. Alqattan Chemical Engineering Department, Kuwait University, P.O. Box 5969, Safat 13060, Kuwait Received October 27, 2002
In this work, analytical methods are developed to predict three basic characteristics that quantify the quality of petroleum fuels, such as gasoline and jet fuels. Simple equations have been proposed for the calculations of octane numbers, aniline points, and smoke points of petroleum fuels. Minimum input data for these equations are boiling point and specific gravity; however, when the composition of the mixture in addition to the boiling point is known, better, results can be obtained. Average deviations of about 4-7 for the octane number, 2.5 °C for aniline point, and 2 mm for smoke point were observed when evaluated with a wide range of data sets with fuels from around the world.
Introduction The quality of various petroleum fuels depends on their composition and types of hydrocarbons present in the mixture. The octane number is one of the characteristics of spark-ignition engine fuels, such as gasoline and jet fuel. This number indicates the antiknock characteristic of a fuel and strongly depends on the hydrocarbon type. Octane number of fuels can be improved by addition of oxygenates, such as TEL, MTBE, or TAME. The octane number of fractions without additives is usually referred to as the clear octane number. These additives are normally hazardous to the environment, and for this reason, methods to improve octane numbers through processes, such as alkylation, are used in refineries. Standards organizations and car manufacturers in many countries require a minimum octane number of 95. In Europe, a superstar gasoline should have a minimum octane number of 98. Two commonly used octane numbers are research octane number (RON) and motor octane number (MON). RON is measured under low speed conditions (ASTM D908), while the MON is measured under high speed conditions under the ASTM D 357 test method.1 The difference between RON and MON is called sensitivity, and fuels with lower sensitivities are desirable. Conducting such experimental tests on every motor fuel is prohibitive in both time and cost. There are very few predictive methods for calculating the octane number. Nelson methods are in the form of graphical correlations either in terms of boiling point and the Watson characterization factor for naphtha fractions or in terms of boiling point and paraffin weight percent.2 These methods have a limited range of application and are not suitable for computer applications; in addition, they produce high deviations in excess of 8 octane numbers. Analytical methods that couple regression analysis with * To whom correspondence should be addressed. Tel: (+965) 4817662. Fax: (+965) 4811772. E-mail:
[email protected]. (1) ASTM. Annual Book of Standards; American Society for Testing and Materials: Philadelphia, Pennsylvania, 1995. (2) Nelson, W. L. Octane Numbers of Naphthas. Oil Gas J. 1969, 67, 122.
results from gas chromatography3,4 or nuclear magnetic resonance5,6 are also prohibitive in both time and cost which makes them unattractive for quick online analysis. The aniline point (AP) is another characteristic of petroleum fractions that indicates the degree of aromaticity of hydrocarbon mixtures. The aniline point is defined as the lowest temperature at which equal volumes of aniline and the sample become completely soluble. As the amount of aromatics in a petroleum fraction increases, the aniline point decreases. Therefore, the aniline point is a parameter that is highly related to the hydrocarbon types in petroleum fractions. Naphthenic hydrocarbons and olefins exhibit values that are between aniline points of paraffins and aromatics. It can be measured by the ASTM D611 test method.1 The aniline point is a useful parameter in calculation of the heat of combustion, diesel index, and hydrogen content of petroleum fuels. For nonfuel products, such as solvents, the aniline point is usually specified to quantify their effectiveness for separation purposes. Linden used the method of characterization of Watson and Nelson to develop a simple correlation for prediction of aniline point in terms of boiling point and API gravity.7 However, the correlation was originally developed on the basis of only 37 samples in the following form:
AP ) 0.4(API)(Tb1/3) + 0.317(Tb) - 298
(1)
(3) Sasano Y. Measuring Research Octane Number of Gasoline by Gas Chromatograph. Jpn. Tokyo Koho 1997, JP09138613. (4) Van Leeuwen, J. A.; Jonker, R. J.; Gill, R. Octane Number Prediction Based on GasChromatographic Analysis and Nonlinear Regression Techniques. Chem. Intell. Lab. Syst. 1994, 25, 325. (5) Ramadhan, O. M.; Al-Hyali, E. A. S. New Experimental and Theoretical Relation to Determine the Research Octane Number (RON) of Authentic Aromatic Hydrocarbons that Could be Present in the Gasoline Fraction. Pet. Sci. Technol. 1999, 17, 623-635. (6) Meusinger, R.; Moros, R. Determination of octane numbers of gasoline compounds from their chemical structure by 13C NMR spectroscopy and neural networks. Fuel 2001, 80, 613-621. (7) Linden, H. R. The Relationship of Physical Properties and Ultimate Analysis of Liquid Hydrocarbon. Oil Gas J. 1949, 48, 60-65.
10.1021/ef020250w CCC: $25.00 © 2003 American Chemical Society Published on Web 04/16/2003
690
Energy & Fuels, Vol. 17, No. 3, 2003
Albahri et al.
where AP is the aniline point (°F) and Tb is the boiling point (°R). API is defined in terms of specific gravity at 60 °F (15.56 °C) S in the following form:
API )
141.5 - 131.5 S
(2)
Another method for determining the aniline point was proposed by Walsh and Mortimer,8 which is based on their gas chromatographic studies using the specific gravity and the normal paraffin carbon number as correlating parameters:
AP ) -204.9 - 1.498(C50) + 100.5(C501/3/S) (3) where AP is the aniline point in °C, C50 is the normal paraffin carbon number having a boiling point the same as the midboiling point of the sample, and S is the specific gravity at 15.5 °C. The third characteristic of fuels that was studied in this work is the smoke point. The smoke point is defined as the maximum flame height in millimeters (mm) at which fuel (kerosene or jet fuel) will burn without smoke in a standard wick-fed lamp. It can be measured by ASTM D 1322 or IP 57 standard test methods,1 and it is designated by SP. The smoke point is a very important characteristic of aviation turbine and kerosene fuels. The smoke point is used as an indication of the combustion performance, which is related to the hydrocarbon type composition of such fuels. A low smoke point fuel is not desirable. The higher the amount of aromatic content of fuel, the smokier the flame and the lower the smoke number. Paraffins have higher smoke numbers than those of aromatics and naphthenes. In the absence of experimental data, the smoke point may be estimated from correlations of the hydrocarbon composition and physical properties of the fuel. The Institute of Petroleum, IP,9 method relates the smoke point to the hydrocarbon composition of kerosene in the following form:
Figure 1. Research octane number of pure hydrocarbons from different families.
SP ) -255.26 + 2.040(AP) - 240.8 ln(S) + 7727.0(S/AP) (5)
One drawback in the above equation is that the aniline point is not always available, and the smoke point is easier to determine experimentally than the aniline point. Furthermore, the above correlations were developed for finished kerosene and jet fuels and may not be applicable for the full range of blend stocks used in these products. To correct the above equations for the ASTM D1322 test, 0.7 mm should be subtracted from the calculated value from the IP smoke point. Recently a set of data on crude assay and specification of petroleum products from around the world were collected and published in the Oil and Gas Journal Data Book.11 These data include boiling point, API gravity (or specific gravity), hydrocarbon type composition (paraffins, naphthenes, and aromatics), kinematic viscosity, research and motor octane numbers, and aniline, smoke, flash, pour, and cloud points for large number of various petroleum products. Although not all these specifications are given for a single fraction, the amount of information available for each characteristic is sufficient to develop new predictive methods or to reevaluate the existing correlations. Information on the composition type varies from one sample to another. In some cases, a detailed analysis of paraffins (P), iso-paraffins (IP), naphthenes (N), and aromatics (A) were given; however, more data on the PNA composition were available.11 Generally, the octane numbers are given for naphthas, gasolines, and jet fuels while smoke points are mainly given for jet fuels and kerosenes. Typical boiling ranges for light naphtha, gasoline, heavy naphtha, and kerosene are approximately 0-150, 0-180, 150-205, and 205-260 °C, respectively. The main purpose of this work is to use the available data bank to develop new and simple procedures for prediction of the RON, MON, AP, and SP values of various fuels with a minimum of information available on each fuel.
where SP is the IP smoke point (mm), AP is the aniline point (°C), and S is the specific gravity (60 °F/60 °F).
Technical Development
SP ) 1.605(X) - 0.0112(X2) - 8.7 X ) 100/[0.0061(P) + 0.03392(N) + 0.13518(A)] (4) where SP is the IP smoke point (mm) and P, N, and A are the paraffin, naphthene, and aromatic contents (wt %). One disadvantage of the above correlation is that it requires knowledge of the composition of the kerosene, which is not always available. It is easier to measure the smoke point experimentally than the composition of kerosene in terms of paraffin, naphthenes, and aromatics. Jenkins and Walsh10 developed the following correlation for prediction of smoke point based on the specific gravity and aniline point of the fuel:
(8) Walsh, R. P.; Mortimer, J. V. New Way to Test Product Quality. Hydrocarbon Process. 1971, 50, 153-158. (9) Modern Petroleum Technology, 3rd ed.; The Institute of Petroleum: London 1962; p 609. (10) Jenkins, G. I.; Walsh, R. P. Quick Measure of Jet Fuel Properties. Hydrocarbon Process. 1968, 47, 161-164.
The clear octane numbers of some pure hydrocarbons generally found in naphtha and gasoline are given in the API(11) Oil and Gas Journal Data Book, 2000 ed.; PennWell: Tulsa, Oklahoma, 2000.
Quality of Petroleum Fuels
Energy & Fuels, Vol. 17, No. 3, 2003 691
Table 1. Coefficients for Equation 6 for Estimation of RON hydrocarbon family
a
b
c
d
e
n-paraffins 92.809 -70.97 -53 20 10 iso-paraffins 2-methylpentanes 95.927 -157.53 561 -600 200 3-methyl-pentanes 92.069 57.63 -65 0 0 2,2-dimethyl-pentanes 109.38 -38.83 -26 0 0 2,3-dimethyl-pentanes 97.652 -20.8 58 -200 100 naphthenes -77.536 471.59 -418 100 0 aromatics 119 144.8 -12 0 0
TDB.12 For the same carbon number or boiling point, the octane numbers of n-alkanes (n-paraffins), 2-methylalkanes (iso-paraffins), n-alkylcyclopentanes (naphthenes), and nalkylbenzenes (aromatics) vary significantly. As shown in Figure 1, the octane number of aromatics is generally higher than the n-paraffins, iso-paraffins, and naphthenes. Even different types of iso-paraffins have different octane numbers. However, most petroleum fractions are olefin free, and the main hydrocarbon families present in a fuel are n-paraffins (NP), iso-paraffins (IP), naphthenes (N), and aromatics (A). The clear research octane numbers of these homologous hydrocarbon families have been correlated to the normal boiling point through the following relation:
RON ) a + b(T) + c(T)2 + d(T)3 + e(T)4
Figure 2. Calculated vs experimental values of RON with known PNA.
(6)
where T ) Tb/100 in which Tb is the normal boiling point (°C). The coefficients a-e are given in Table 1. To calculate the RON of a fuel it is assumed that the mixture contains four model compounds from n-paraffin, iso-paraffin, naphthene, and aromatic families. Then the RON of the mixture is calculated as
RON ) xNP(RON)NP + xIP(RON)IP + xN(RON)N + xA(RON)A (7) where xNP, xIP, xN, and xA are volume fractions of n-paraffin, iso-paraffin, naphthene, and aromatic groups, respectively. (RON)NP is the clear RON of a hydrocarbon from a n-paraffin family whose boiling point is the same as the mid-boiling point or ASTM D86 temperature at 50% point of the fraction and can be calculated from eq 6. Similarly, values of (RON)N and (RON)A are calculated from eq 6 using the boiling point of the mixture. To calculate (RON)IP, eq 6 should be used for four different iso-paraffin families given in Table 1, and an average value (also shown in Figure 1) is used for (RON)IP. This is to account for the large differences in octane number for the various iso-paraffins in the gasoline fraction. When the amount of iso-paraffins is not reported, xNP and xIP are taken equal to half of the volume fraction of paraffins. When experimental data on the composition are not available, it can be estimated through methods proposed by Riazi et al.13-16 These correlations require refractive index, density, and molecular weight, all of which can be estimated from the mid-boiling point and (12) Daubert, T. E., Danner, R. P., Eds. API Technical Data BookPetroleum Refining, 6th ed.; American Petroleum Institute (API): Washington, DC, 1997. (13) Riazi, M. R.; Daubert, T. E. Prediction of the Composition of Petroleum Fractions, Industrial and Engineering Chemistry. Process Des. Dev. 1980, 19, 289-294. (14) Riazi, M. R.; Daubert, T. E. Prediction of Molecular Type Analysis of PetroleumFractions and Coal Liquids. Ind. Eng. Chem., Process Des. Dev. 1986, 25, 1009-1015. (15) Riazi, M. R.; Roomi, Y. A. Compositional Analysis of Petroleum Fractions. Presented at the Division of Petroleum Chemistry, 222nd National Meeting of the American Chemical Society, Chicago, IL, August 28-31, 2001. (16) Riazi, M. R.; Roomi, Y. Use of the Refractive Index in the Estimation of Thermophysical Properties of Hydrocarbons and Their Mixtures. Ind. Eng. Chem. Res. 2001, 40, 1975-1984.
Figure 3. Calculated vs experimental values of RON with known NP-IP-N-A composition. specific gravity of a petroleum fractions using methods proposed by the API12 and Riazi and Daubert.17,18 Once the clear RON is known, the clear MON can be estimated from the following correlation derived from the equation proposed by Jenkins19 for olefin free fuels.
MON ) 22.5 + 0.83(RON) - 20.0(S)
(8)
where S is the specific gravity of the fuel at 15.5 °C. A summary of the results for the calculations of RON and MON using both experimental and predicted composition is given in Table 2. To develop a predictive method for the aniline point of petroleum fractions, we use parameters that quantify the amount of aromatics is a hydrocarbon mixture. It has been shown that parameters Ri (or m) and S are suitable for the prediction of the aromatic content of petroleum fractions, (17) Riazi, M. R.; Daubert, T. E. Simplify Property Predictions. Hydrocarbon Process. 1980, 59 (3), 115-116. (18) Riazi, M. R.; Daubert, T. E. Characterization Parameters for Petroleum Fractions. Ind. Eng. Chem. Res. 1987, 26, 755-759. (Corrections, p 1268.) (19) Jenkins, G. I. Calculation of the Motor Octane Number from the Research OctaneNumber. J. Inst. Pet. 1968, 54 (529), 14-18.
692
Energy & Fuels, Vol. 17, No. 3, 2003
Albahri et al.
Table 2. Summary of Results for Predictions of RONs and MONs RON
item 1 2 3 4 5
no. of data points 19 97 100 18 39
API gravity range
methoda NP-IP-N-A P-N-A pred P-N-A Nelson 1 Nelson 2
51-87 44-88 44-88 58-78 47-74
RON range 26-74 30-80 30-80 45-73 18-72
AADb 4 6 7 5 6
MON
MD
AAD
MD
11 19 22 19 14
5c
12 7
1d
NP-IP-N-A refers to use of eqs 6 and 7 with complete PINA composition from experimental data, P-N-A refers to use of eqs 6 and 7 (xNP ) xIP ) xP/2) with PNA composition from experimental data, and pred P-N-A refers to use of eqs 6 and 7 (xNP ) xIP ) xP/2) with predicted PNA composition from boiling point and specific gravity. Nelson 1 refers to the graphical method developed by Nelson2 using the average boiling point and the Watson characterization factor. Nelson 2 refers to the graphical method developed by Nelson2 using the average boiling point and paraffincontent. b AAD ) average absolute deviation; MD ) maximum deviation. c MON calculated from predicted RON and eq 8. d MON is calculated by eq 8 and experimental RON. Table 3. Summary of Results for Prediction of Aniline Pointsa SP (°C) item 1 3 3
method
equation
AAD
MD
proposed model Linden5 Walsh-Mortimer6
10 1 3
2.5 6.5 4.6
7 28 13
a Number of data points: 300. API gravity range: 14-56. Boiling point range: 115-545 °C. Aniline point range: 45-107 °C.
Table 4. Summary of Results for Prediction of Smoke Pointsa SP, mm item 1 2 3
method
equation
AAD
MD
proposed model IP7 Jenkins and Walsh8
11 4 5
2 5.6 8.8
7.8 13.8 43
a Number of data points: 136. API gravity range: 26-52. SP range: 12-37 mm.
where Ri is defined as
Ri ) n - d/2
Figure 4. Calculated vs experimental values of RON using predicted PNA composition.
(9)
in which n is the refractive index and d is the liquid density (g/cm3), both being at 20 °C. These two parameters can be estimated from the boiling point and specific gravity through available relations.17 On the basis of a data bank consisting of 300 data points on aniline point of petroleum fractions from crude oils from around the world,11 the following relation is proposed to predict the aniline point:
AP ) - 9805.269(Ri) + 711.858(S) + 9778.7
(10)
where AP is in °C. This equation has been evaluated and compared with eqs 1 and 3. The results of the evaluations are summarized in Table 3. On the basis of a data bank consisting of 136 data points on the smoke points of petroleum fractions from crude oils from around the world from the Oil and Gas Journal Data Book,11 the following relation is proposed to predict the smoke point.
SP ) 0.839(API) + 0.0182634(Tb) - 22.97
(11)
where SP is the smoke point (mm), API is the API specific gravity, and Tb is the average boiling point (K). The evaluations of the proposed equation and other predictive methods are shown in Table 4. The graphical presentations for the evaluations of the proposed methods are shown in Figures 2-8 for octane numbers and aniline and smoke points.
Conclusion and Future Outlook In conclusion, the results presented in Tables 2-4 show that the proposed methods are capable of estimat-
Figure 5. Calculated RON vs experimental RON from method Nelson 1.
ing octane numbers and aniline and smoke points of petroleum fuels. Proposed methods are evaluated with a wide range of data points from oils from around the world with accuracies greater than those of similar existing methods. Proposed methods require a minimum of information on boiling point and specific gravity;
Quality of Petroleum Fuels
Figure 6. Calculated RON vs experimental RON from method Nelson 2.
Energy & Fuels, Vol. 17, No. 3, 2003 693
Figure 8. Calculated vs experimental smoke point.
eters for calculation of the octane number on the basis of the developed group contribution method.20,21 Nomenclature
Figure 7. Calculated vs experimental aniline point.
however, when additional data are available from experimental measurements, the proposed methods predict RONs and aniline points with higher degrees of accuracy. Further work on developing a new procedure on the basis of detailed compositional data and the octane number of pure components using the structural group contribution method is underway. The next phase of this study is to develop mixing rules and interaction param(20) Albahri, T. A. Structural Group Contribution Method for Predicting the Octane Numberof Pure Hydrocarbon Liquids. Ind. Eng. Chem. Res. 2003, 42, 657-662. (21) Albahri, T. A. Structural Group Contribution Method for Predicting the Octane Number of Pure Hydrocarbons and their Mixtures. Presented at the Division of Petroleum Chemistry, 224th National Meeting of the American Chemical Society, Boston, MA, Fall, 2002. See: Prepr.-Div. Fuel Chem. 2002, 47, 531-532.
A ) aromatic content, wt % AP ) aniline point API ) American Petroleum Institute Gravity at 60 °F (eq 2) a-e ) constants B1 ) RVP blending number for cut i C50 ) normal paraffin carbon number at which 50 percent of the same sample was eluted d ) liquid density at 20 °C, g/ml MON ) motor octane number MTBE ) methyl tert-butyl ether N ) naphthene content, wt % n ) sodium D-line refractive index at 20 °C and 1 atm P ) paraffin content, wt % Ri ) the refractivity intercept RON ) research octane number RVP ) Reid vapor pressure, psia S ) specific gravity (60 °F/60 °F) SP ) smoke point in mm T ) Tb/100 TAME ) tert-amyl methyl ether Tb ) normal boiling point of hydrocarbon TEL ) tetra-ethyl lead xNP, xIP, xN, and xA ) volume fractions of n-paraffins, isoparaffins, naphthenes, and aromatics groups, respectively
Acknowledgment. This paper was orally presented at the 224th National Meeting of the American Chemical Society, Boston, MA, August 2002. The summary version of this work appeared in the ACS preprints.22 EF020250W (22) Albahri, T. A.; Riazi, M. R.; Qattan, A. R. Octane Number and Aniline Point of Petroleum Fuels. Presented at the Division of Petroleum Chemistry, 224th National Meeting of the American Chemical Society, Boston, MA, Fall, 2002. See: Prepr.-Div. Fuel Chem. 2002, 47, 710-711.
