02638762/99/$10.00+0.

00
Institution of Chemical Engineers
Trans IChemE, Vol 77, Part A, May 1999

A SIMPLE THERMAL CONDUCTIVITY-TEMPERATURE

CORRELATION FOR UNDEFINED PETROLEUM AND
COAL LIQUID FRACTIONS
A. ABOUL-SEOUD and H. M. MOHARAM
Department of Chemical Engineering, Faculty of Engineering, Alexandria University, Egypt

simple correlation, requiring the speci c gravity as the only input, is found to predict
the thermal conductivity of unde ned petroleum and coal liquid fractions with
comparable or better accuracy than other methods. The correlation was derived from
64 data points for 25 unde ned petroleum fractions, and tested against 13 unde ned petroleum
fractions and 39 data points for 5 coal liquid fractions to give 2.93% and 3.99% AAD,
respectively.
Keywords: thermal conductivity; temperature effect; prediction; petroleum fractions; coal
liquid fractions

Teja and Tardieu3 proposed a method based on the

Effective Carbon Number (ECN) which can be calculated
from the thermal conductivity at a reduced temperature of
0.6.
Lakshimi and Prasad4 proposed a correlation for estimating the thermal conductivity of liquids. The correlation was
developed for pure liquids. However, it can be applied
to unde ned petroleum fractions. The correlation has the
following form:

INTRODUCTION
Thermal conductivities of substances are required in
correlations and design methods for heat transfer equipment
as well as non-isothermal mass transfer equipment and
reactors. Most of the procedures that were suggested for
the calculation of the thermal conductivity of hydrocarbon
mixtures (Reid et al.1 ) are only applicable to mixtures
of well-de ned components. Few theories, however, lead
to useful expressions for the calculation of the thermal
conductivity of liquids, and, consequently, empirical
methods are employed in engineering applications.
Riazi and Faghri2 developed a simpli ed equation for
the prediction of the thermal conductivity of liquid hydrocarbon systems as a function of the temperature with the
normal boiling point and the speci c gravity as input
parameters. The equation has the following form:
l

aT bb c c

d3

20 1

Tr

lr

l r P, T

lL

20 1 Tr
Tbr 2/3

l P, T / Tc

1/6

P 2/3
c MW

1/2

The nominator of the expression, at the right side of

equation (7), may be regarded as an approximation of the
thermal conductivity at the critical point. Using the simple
corresponding states theory, the thermal conductivity of
component x at a temperature T , and a pressure P, may be
found from the following equation:

where d is a constant.
Combining equation (2) and equation (3), the following
correlation can be derived,
1.11/MW1/2 3
3 20 1

f Pr , Tr

where f is the same function for a group of substances

obeying the corresponding states principle. For the reduced
thermal conductivity, l r , the following equation is used:

2/3

5
1

where T is in (K), and l is in (W m K ).

Among the most accurate procedures for the calculation
of the thermal conductivities of hydrocarbon mixtures
are those based on a corresponding states method. In
terms of pressure and temperature, the corresponding states
theory applied to thermal conductivity can be expressed
as follows:

where lL Tb is the thermal conductivity of the liquid at

the normal boiling point (W m 1 K 1 ).
To estimate lL at other temperatures, the Riedel equation,
as given in reference 1, may be used. The equation has the
form:
lL

0.00197 T /MW 0.5

1

1.11/MW 0.5

0.00005 T

1.3855

where a, b and c are functions of the temperature. From

reference 1, the work of Sato indicates that, at the normal
boiling point, the thermal conductivity is given by
l L Tb

0.0655

l x P, T

2/3

Tcx /Tco

1/6

Pcx /Pco

MWx /MWo
248

1/ 2

2/3

l o Po , To

SIMPLE THERMAL CONDUCTIVITY-TEMPERATURE CORRELATION

where

DEVELOPMENT OF THE CORRELATION

Po

PPco /Pcx

To

TTco /Tcx

lo

thermal conductivity of the reference substance

at the temperature To and pressure Po

Christensen and Fredenslund5 suggested a corresponding states model that has two contributions to the thermal
conductivity, one due to transport of transitional energy, l tr
and one to transport of internal energy, lint :
l

l tr

lint

For a component x, the following expression is used for

the transitional part:
Tc,x
Tc,0

l tr,x

1/6

Pc,x
Pc,0

ax
l 0 T0 , P0
a0

2/3

MWx
MW0

1/2

lint,0

Tc,x ax
Tc,0 a0

and

P0

Pc,x ax
Pc,0 a0

The subscript 0 is used for the reference substance, which

is methane.
Pedersen and Fredenslund6 used the following expression
for a derived from thermal conductivity data:
a

0.0006004 r2.043
MW 1.086
r

11

lint,0 and lint,x are given by the following expression

(Christensen and Fredenslund5 ):
lint

1.18653 g1 Cp1

f r

1.0053432 rt

2.5 R f r /MW
0.030182 r2r

Thermal conductivity data for petroleum fractions, reported

by Baltatu8 , were used to develop the present correlation.
The data set comprises 64 data points for 25 unde ned
petroleum fractions. From the analysis of these data, it
was observed that the thermal conductivity of petroleum
fractions are related to the square root of the quantity (c /T )
in a straight line relationship. This observation was used
to develop the following correlation:
l

12

0.029725 r3r
13

where g1 is the viscosity at the actual temperature and

at a pressure of 1 atm., R is the gas constant and Cp1 is the
ideal gas heat capacity at the temperature T, which can
be found from the procedure suggested by Kesler and
Lee7 .
The above methods are compared in terms of their
input data requirements in Table 1. The method proposed
by Lakshimi and Prasad4 requires the molecular weight
as the only input, whereas all other methods require a host
of other properties which may not be readily available for
unde ned fractions. In the present work, an attempt was
made to develop a simple estimation method requiring
the speci c gravity as the only input property.

2.540312 c /T

0.5

0.014.4485

1.
2.
3.
4.

Sato and Riedel

Riazi and Faghri
Lakshimi and Prasad
Corresponding States

RESULTS AND DISCUSSION

Table 2 shows the results of the comparison of the
prediction capabilities of equation (14) with the previously
published methods. The data of this table (25 petroleum
fractions with 64 data points), with which the proposed
correlation was developed, were published by Baltatu8 .
Each of the petroleum fractions contains hydrocarbons
with a normal boiling point within a narrow temperature
interval. The table shows that the present method, requiring the speci c gravity as the only input, is comparable
with most accurate and more complicated methods that
require more input parameters.
Generally, the developed correlation gives 5.62% AAD.
The corresponding AAD for the Sato-Riedel method
is 7.87%, for the Riazi and Faghri method, 18.31%, for
the Lakshimi and Prasad method 14.36%, and for the
corresponding states method [equation (10)] 8.74%.
The developed correlation was tested in Table 3 against
other methods for coal liquid fractions data published by
Baltatu9 . The test was made on 39 data points for 5 coal
liquid fractions. The present correlation gives 3.99% AAD
compared with 13.84% for the Sato-Riedel method, 27.27%
for the Riazi and Faghri method, 25.07% for the Lakshimi
and Prasad method and 6.64% for the corresponding states
method.
In Table 4, a comparison was made on the thermal
conductivity data of petroleum fractions published by
Baltatu9 . The data comprises 13 petroleum fractions. The
developed method gives 2.93% AAD compared with 8.54%
for the Sato-Riedel method, 17.58% for the Riazi and Faghri
method, 16.65% for the Lakshimi and Prasad method and
3.49% for the corresponding states method.

Input Data
Molecular weight, boiling point and critical temperature
Normal boiling point and speci c gravity
Molecular weight
Molecular weight, critical temperature and critical pressure

Trans IChemE, Vol 77, Part A, May 1999

where l is the thermal conductivity in (W m K ) and T

is in (K). The correlation uses the speci c gravity as the
only input parameter to predict the thermal conductivity
of petroleum fractions at different temperatures.
The values of the constants included in equation (14)
were determined for the studied unde ned fractions using
the Solver method (Excel 5.0, Microsoft).

Table 1. Input property requirements for liquid thermal conductivity estimation methods tested in the present work.
Method

14
1

10

lint,x

where
T0

249

No. of
constants
6
9
5

250

ABOUL-SEOUD and MOHARAM

Table 2. A comparison of the results of the developed correlation with those of previous methods for the petroleum fraction data from Baltatu8.
% Deviation

Sample
no.

Tb
(K)

355.0

0.7559

359.3

0.6981

384.9

0.7499

388.9

0.7788

421.1

0.8067

460.3

0.8132

469.9

0.7887

8
9

489.9
492.6

0.7800
0.8072

10

498.4

0.9535

11

504.4

0.7990

12

524.6

0.8174

13

528.3

0.8478

14

531.2

0.7861

15

535.7

0.8031

16

543.9

0.8560

17

557.3

0.8555

18

566.4

0.8751

19

572.1

0.8383

20

586.4

0.8509

21

603.5

0.8504

22

610.7

0.8740

23

615.4

0.8555

24

616.4

0.8772

25

631.8

0.9236

% AAD

T
(K)

l
(Wm 1 K 1 )

273.15
333.30
295.30
338.30
293.30
383.30
295.60
326.70
360.00
295.60
338.90
388.30
298.30
398.30
296.30
398.30
288.30
295.60
338.30
338.90
303.30
355.00
388.30
273.15
333.30
298.90
328.30
368.90
394.40
293.30
453.30
273.15
333.30
305.60
341.10
373.90
393.30
273.15
333.30
298.90
332.20
357.10
378.00
295.60
326.70
360.00
295.30
443.50
295.30
473.30
305.60
347.80
377.20
273.15
333.30
273.15
398.30
298.90
332.20
378.90
412.20
298.90
332.20
378.90

0.13530
0.11860
0.11080
0.09830
0.11160
0.08960
0.12017
0.11315
0.09797
0.11819
0.10860
0.09036
0.11050
0.08960
0.11690
0.09620
0.12180
0.11785
0.10052
0.10878
0.12870
0.11755
0.11125
0.13100
0.11972
0.12446
0.11927
0.11136
0.10646
0.12100
0.10180
0.12760
0.12160
0.13329
0.12723
0.11978
0.11615
0.11220
0.10670
0.11993
0.10909
0.10564
0.09977
0.11437
0.11014
0.10341
0.12050
0.09564
0.11890
0.09240
0.12290
0.11667
0.11165
0.12900
0.13910
0.11600
0.11000
0.12149
0.11483
0.10668
0.09581
0.11705
0.10904
0.10158

Sato & Riedel1

Riazi & Faghri2

Lakshimi & Prasad 4

Corresponding
states6

Developed
correlation

3.69
3.47
22.30
24.19
18.69
20.37
9.23
9.07
17.02
8.66
9.44
18.75
12.27
16.20
5.92
7.78
1.67
3.15
13.47
4.75
3.43
1.51
1.05
5.00
4.55
5.20
5.00
4.22
3.87
2.35
9.45
3.89
7.07
12.63
12.88
11.92
11.96
6.48
3.79
4.52
0.65
0.52
3.34
0.29
0.30
1.70
5.42
3.33
5.42
5.33
10.69
10.57
10.02
12.72
24.32
2.89
11.87
10.32
8.62
7.08
0.99
8.28
4.98
3.33

6.04
7.71
8.66
8.63
11.94
10.50
4.18
3.14
9.72
9.51
9.26
17.61
20.66
23.31
15.01
14.90
13.50
16.51
27.45
17.66
7.87
11.42
13.46
9.29
8.90
12.58
12.65
13.55
14.17
17.39
11.96
14.06
9.49
4.79
4.29
5.42
5.48
31.22
28.28
20.12
27.05
27.44
31.67
27.49
28.08
31.73
21.02
28.24
23.96
32.78
19.98
21.28
23.22
18.71
4.29
32.22
24.55
23.57
27.09
31.85
43.31
30.15
36.72
43.14

3.75
5.39
21.06
25.32
15.57
19.78
6.44
6.46
14.83
3.63
3.77
12.36
4.60
5.51
2.32
3.15
7.33
5.54
2.31
5.57
11.39
12.19
13.48
12.99
14.56
14.33
15.25
16.11
16.77
10.93
23.82
13.57
18.50
22.09
23.54
23.84
24.55
2.22
7.58
13.83
10.77
12.12
10.73
9.65
11.24
11.22
15.37
20.58
15.38
24.29
21.01
22.87
23.82
20.78
33.76
12.65
26.66
19.85
20.01
21.15
17.97
17.17
16.11
17.52

1.33
9.53
1.17
5.49
8.87
1.56
6.68
2.08
4.73
13.80
8.56
10.78
17.47
10.16
2.22
4.57
2.55
3.01
2.51
1.86
26.34
25.56
23.33
5.34
8.20
3.82
5.01
6.79
8.32
3.88
11.59
13.01
17.76
17.40
18.89
19.52
20.62
17.83
12.28
2.89
7.25
6.12
8.05
10.08
8.77
9.85
4.32
8.51
1.60
9.20
8.46
10.09
10.34
1.94
19.91
0.60
1.09
1.14
2.12
2.14
3.12
13.28
12.89
11.05

11.91
10.18
1.57
2.69
2.15
9.28
3.52
3.16
5.85
0.06
0.82
12.15
6.96
11.98
0.24
2.49
3.38
0.38
9.07
0.69
0.56
0.29
0.17
6.15
8.18
4.87
5.83
5.59
4.94
0.93
6.27
4.52
10.43
13.14
14.48
13.77
13.61
13.87
7.11
1.27
4.93
4.02
6.65
8.22
6.25
7.14
0.33
0.37
2.53
0.93
2.72
4.72
4.91
0.19
16.87
10.10
6.11
1.39
1.10
1.03
7.24
8.30
9.59
9.25

7.87

18.31

14.36

8.74

5.62

Trans IChemE, Vol 77, Part A, May 1999

SIMPLE THERMAL CONDUCTIVITY-TEMPERATURE CORRELATION

251

Table 3. A comparison of the results of the developed correlation with those of previous methods for the coal liquid fraction data from Baltatu9.
% Deviation
Sample
no.

Tb
(K)

372.6

0.7701

410.4

0.8125

524.8

0.9761

650.4

1.0793

749.8

1.1733

T
(K)

l
(Wm 1 K 1 )

302.09
395.54
438.93
510.43
301.98
395.48
481.65
510.54
301.71
302.43
347.65
348.98
394.76
430.65
409.59
436.21
471.59
473.98
511.09
298.37
299.93
344.71
348.87
394.76
398.37
428.65
430.76
464.87
473.43
511.09
512.15
356.87
363.21
391.15
404.26
433.26
466.37
492.98
510.82

0.11672
0.09468
0.08728
0.07709
0.11985
0.10031
0.08544
0.08301
0.12930
0.12833
0.12402
0.12191
0.11634
0.11065
0.11743
0.11022
0.10389
0.10534
0.10204
0.12868
0.12556
0.12541
0.12326
0.11811
0.11975
0.11859
0.11418
0.11537
0.11221
0.11025
0.11056
0.13141
0.13030
0.13202
0.12823
0.12465
0.11522
0.11676
0.11368

% AAD

Sato & Riedel1

Riazi & Faghri2

Lakshimi & Prasad 4

Corresponding
states6

Developed
correlation

12.12
9.89
3.02
19.50
6.93
5.53
3.86
12.00
5.28
4.64
6.83
5.38
7.01
7.50
9.91
7.98
8.15
9.81
13.35
12.75
10.71
14.34
13.20
13.61
15.12
17.12
14.13
18.37
16.95
19.48
19.82
24.16
23.91
26.66
25.35
25.18
21.53
24.57
23.92

3.88
0.16
5.57
17.22
5.69
2.72
4.11
9.36
10.03
10.79
9.91
11.68
12.52
14.98
10.41
14.94
18.83
16.99
17.80
21.62
24.59
24.12
26.31
33.00
31.34
34.45
39.80
41.52
46.50
54.46
54.20
30.83
32.65
34.66
40.84
50.91
72.71
79.68
91.92

10.90
12.34
9.58
1.15
2.68
1.15
4.55
9.80
13.66
13.12
17.48
16.27
20.23
22.70
23.53
23.42
25.65
27.13
32.13
21.05
19.31
25.50
24.79
28.35
29.86
33.67
31.43
37.34
36.92
41.81
42.14
34.27
34.49
38.77
38.61
40.62
40.40
44.88
45.92

1.41
4.2
4.2
2.75
2.9
7.78
6.21
6.61
14.75
15.42
8.67
2.37
0.34
1.85
1.09
1.98
1.98
3.76
6.77
19.11
21.89
14.66
15.83
11.11
8.85
3.74
7.32
0.33
0.9
3.78
4.22
11.34
11.57
6.3
7.33
5.42
7.76
1.51
1.06

2.49
3.13
5.36
9.25
2.11
0.38
5.21
4.68
0.57
1.20
3.11
1.65
3.84
3.76
6.70
4.08
2.66
4.28
5.36
7.50
9.86
1.82
2.91
0.23
1.65
4.70
1.29
6.43
4.78
7.22
7.59
0.15
0.28
5.56
4.54
5.54
1.95
6.23
5.61

13.84

27.27

25.07

6.64

3.99

Table 4. A comparison of the results of the developed correlation with the results of previous methods for the petroleum fraction data from Baltatu9.
% Deviation
Sample
no.
1
2
3
4
5
6
7
8
9
10
11
12
13

Tb
(K)

T
(K)

l
(Wm 1 K 1 )

347.6
421.1
492.6
498.2
498.4
506.8
535.7
572.1
603.2
616.1
616.4
631.5
684.7

0.7310
0.8060
0.8070
0.9550
0.9550
0.8070
0.8030
0.8380
0.8570
0.8770
0.8770
0.9240
1.0010

333.30
295.60
295.60
303.30
303.30
295.30
305.60
443.30
305.30
341.10
298.90
295.00
398.30

0.10480
0.11820
0.11790
0.12900
0.12880
0.12220
0.13330
0.09600
0.12010
0.12270
0.12150
0.11700
0.11000

% AAD

Trans IChemE, Vol 77, Part A, May 1999

Sato & Riedel1

Riazi & Faghri2

Lakshimi & Prasad 4

Corresponding
states6

Developed
correlation

17.20
8.65
3.11
3.54
3.42
1.52
12.63
3.66
8.70
15.35
10.33
7.84
15.01

2.75
9.49
16.46
7.62
7.81
13.62
4.78
27.76
22.90
18.04
23.56
30.54
43.18

20.62
3.61
5.58
11.51
11.40
10.43
22.09
20.85
19.02
26.32
19.86
16.55
28.60

0.3
7.3
0.9
1.2
3.9
4.6
2.4
4.7
0.1
3.5
7.2
7.1
2.2

0.27
0.00
0.32
0.70
0.55
3.15
13.15
0.00
0.03
6.80
1.36
9.16
2.64

8.54

17.58

16.65

3.49

2.93

252

ABOUL-SEOUD and MOHARAM

Figure 1.

CONCLUSIONS

REFERENCES

The thermal conductivity of unde ned petroleum and

coal derived liquid fractions was successfully estimated
using a simple correlation needing the speci c gravity as
the only input. The developed correlation was derived from
64 data points for 25 unde ned petroleum fractions and
tested on 13 unde ned petroleum fractions and 5 coal liquid
fractions to give 2.93% and 3.99% AAD, respectively. In
comparison with other methods in the literature, a comparable or better accuracy was obtained.

1. Reid, R. C., Prausnitz, J. M. and Poling, B. E., 1987, The Properties of

Gases and Liquids, 4th ed., (McGraw-Hill, New York).
2. Riazi, M. R. and Faghri, A., 1985, Thermal conductivity of liquid
and vapor hydrocarbon system: Pentane and heavier at low pressures,
Ind Eng Chem Proc Des Dev, 24: 398401.
3. Teja, A. S. and Tardieu, G., 1988, Prediction of the thermal
conductivity of liquids and liquid mixtures including crude oil fractions,
Canada J Chem Eng, 66: 980986.
4. Lakshimi, D. S. and Prasad, D. H. L., 1992, A rapid estimation method
for thermal conductivity of pure liquids, Chem Eng J, 48: 211214.
5. Christensen, P. L. and Fredenslund, A. A., 1980, A corresponding states
model for the thermal conductivity of gases and liquids, Chem Eng Sci,
35: 871875.
6. Pedersen, K. S. and Fredenslund, A., 1987, An improved corresponding
state model for the prediction of oil and gas viscosities and thermal
conductivities, Chem Eng Sci, 42: 182186.
7. Kesler, M. G. and Lee, B. I., 1976, Improve prediction of enthalpy of
fractions, Hydrocarbon Processing, March, 153158.
8. Baltatu, M. E., 1984, Prediction of the transport properties of petroleum
fractions, paper presented at the 1984 Winter National Meeting AIChE,
Atlanta Georgia.
9. Baltatu, M. E., Ely, J. F., Hanley, J. M., Graboski, M. S., Perkins, R. A.
and Sloan, E. D., 1985, Thermal conductivity of coal-derived liquid
and petroleum fraction, Ind Eng Chem Proc Des Dev, 24: 325332.

NOMENCLATURE
a, b, c
Cp
d
MW
P
R
T

constants in equation (1)

ideal gas heat capacity
constants in equation (2)
molecular weight
pressure
the gas constant
temperature

Greek
a
c
g
l
r

parameter in equation (10)

speci c gravity
viscosity
Thermal conductivity
density

Subscript
b
boiling point
c
critical
int
internal
L
liquid
o
reference component
r
reduced property
tr
transitional
x
any component

ADDRESS
Correspondence concerning this paper should be addressed to
Dr H. Moharam, PO Box 172, El-Ibrahimia 21321, Alexandria, Egypt
(E-mail: moharam@alex.eun.eg).
The manuscript was received 6 April 1998 and accepted for publication
after revision 4 February 1999.

Trans IChemE, Vol 77, Part A, May 1999