Professional Documents
Culture Documents
ESDU 73031 Convective Heat Transfer During Crossflow of Fluids Over Plain Tube
ESDU 73031 Convective Heat Transfer During Crossflow of Fluids Over Plain Tube
Endorsed by
The Institution of Mechanical Engineers
ESDU 73031
ESDU DATA ITEMS
Data Items provide validated information in engineering design and analysis for use by, or under the supervision
of, professionally qualified engineers. The data are founded on an evaluation of all the relevant information, both
published and unpublished, and are invariably supported by original work of ESDU staff engineers or consultants.
The whole process is subject to independent review for which crucial support is provided by industrial companies,
government research laboratories, universities and others from around the world through the participation of some
of their leading experts on ESDU Technical Committees. This process ensures that the results of much valuable
work (theoretical, experimental and operational), which may not be widely available or in a readily usable form, can
be communicated concisely and accurately to the engineering community.
We are constantly striving to develop new work and review data already issued. Any comments arising out of your
use of our data, or any suggestions for new topics or information that might lead to improvements, will help us to
provide a better service.
The work on this particular Data Item was monitored and guided by the following Working Party:
on behalf of the Heat Transfer Committee, which first met in 1966 and now has the following membership
Chairman
Mr E.C. Firman — Central Electricity Research Laboratories
Vice-Chairman
Dr G.F. Hewitt — HTFS, Atomic Energy Authority, Harwell
Members
Dr T.R. Bott — Birmingham University
Mr J.G. Cumming — Imperial Chemical Industries (Europa) Ltd
Prof. W.B. Hall — Manchester University
Mr E.A.D. Saunders — Whessoe Ltd
Mr M. Taylor — Foster Wheeler Ltd
Mr N.G. Worley — Babcock and Wilcox (Operations) Ltd.
The Steering Group has benefited from the participation of members from several engineering disciplines. In
particular, Mr E.C. Firman has been appointed to represent the interests of mechanical engineering as the nominee
of the Institution of Mechanical Engineers and Dr G.F. Hewitt has been appointed to represent the interests of
chemical engineering as the nominee of the Institution of Chemical Engineers.
The work on this Item was carried out in the Fluid Mechanics and Physical Properties Group of ESDU under the
supervision of Mr N. Thompson, Group Head. The members of staff who undertook the technical work involved in
the initial assessment of the available information and the construction and subsequent development of the Item
were
CONTENTS
Page
2. INTRODUCTION 3
6. EXAMPLE 7
8. TABLES 12
9. CALCULATION PROCEDURE 14
APPENDIX A 25
i
ESDU 73031
CONVECTIVE HEAT TRANSFER DURING CROSSFLOW OF FLUIDS OVER PLAIN
TUBE BANKS
Two sets of units are given below. The SI set is part of a completely coherent system of units. The British
units are not.
Units
SI British
h average external heat transfer coefficient for the bank W/m2K Btu/ft2h°F
T thermodynamic temperature K °R
Suffixes
m mean value
2
ESDU 73031
2. INTRODUCTION
For the calculation of heat transfer in a tube bank, data are required for the internal heat transfer between
the tube walls and the fluid flowing through the tubes and the external heat transfer due to the crossflow
of fluid over the tubes. Item Nos. 67016, 68006 and 68007 (References 2-4) deal with the internal heat
transfer; this Item covers the case of convective heat transfer for the crossflow of a fluid over tube banks
of circular cylinders of negligible thermal radiation and zero heat generation within the fluid.
The case of forced convection heat transfer into or out of the tubes is covered in detail but no data are
available from which to assess the effects of combined forced and free convection (see Section 5.1).
The data are presented in non-dimensional form and are valid for the crossflow of a single-phase Newtonian
fluid within the bounds of Table 8.2. This includes most normal liquids and gases. In applying the data to
the heating of liquids, the temperature of the tube outside surface at all points in contact with the liquid
must be less than the saturation temperature for the liquid concerned.
ESDU Copyright material. For current status contact ESDU.
Section 3 presents data averaged for a complete bank of tubes and Section 4 considers the variations in heat
transfer from row to row. Section 5 considers briefly a number of other factors that may influence convective
heat transfer. The methods presented here are expected to yield heat transfer coefficients accurate to within
about ± 15 per cent over most of the Reynolds number range but Appendix A and Table 8.2 discuss the
derivation and accuracy of the data in greater detail. A summary of the calculation procedure is given in
Section 9.
In general terms, when free convection effects can be ignored, the Nusselt number can be related to the
flow parameters, fluid properties and tube bank geometry by the expression
Nu = f ( Re b, Pr b, Pr w, X t, X l, N, L ⁄ D, θ ) .
Nu = Nu r × F 1 × F 2 × F 3 , (3.1)
where the term Nur (see Section 3.2 and Figures 2 and 3) is defined as the average Nusselt number based
on the following reference conditions,
(ii) tube bank having 10 tube rows with at least 6 tubes in each row,
The terms F1 , F2 and F3 are factors to account for the effect on Nu of departures from the reference
conditions given in the following Table. (No information is available for L/D < 5 ).
3
ESDU 73031
Thus the mean Nusselt number for a particular tube bank is found by evaluating the various terms on the
right-hand side of Equation (3.1) as described in the following Sections. The mean heat transfer coefficient
is then given by
ESDU Copyright material. For current status contact ESDU.
h = Nuλ b ⁄ D
·
Q = h∆TπDL (number of tubes)
where ∆T is the appropriate mean temperature difference between the tube wall and the bulk of the
crossflow fluid, evaluated by such methods as given in Reference 1 for example.
When large temperature changes occur in the fluid flowing over the bank it is more accurate to assess the
heat transfer by splitting the bank into sections containing a few tube rows where the use of mean
temperatures does not lead to significant errors in the temperature driving force or physical properties.
Information relating to the heat transfer for individual rows is discussed in Section 4.
The variations Nu r of with Re b and Pr b covering gases and liquids are given by Figures 2a–2c and
Figure 4 for various in-line tube banks and Figures 3 and 4 for staggered tube banks. Figure 1 gives a guide
to enable the appropriate curve in Figure 2 to be chosen for in-line tube arrangements. Values of Nu r given
by Figures 2–4 are represented by the equation
m 0.34
Nu r = aRe b Pr b (3.2)
where Re b and Pr b are based on bulk fluid conditions. Values of the parameters a and m are given in
Table 8.1 and represent a best fit to the available data over the quoted ranges of Reynolds number.
Because Re b is based on the maximum velocity between tubes, Vmax , which in turn is dependent on the
transverse tube spacing, the effect of Xt on the variation of Nu with Re b for a given Pr b is minimised.
However, an effect of tube row spacing on the parameters in Equation (3.2) remains for in-line tube banks
with small longitudinal pitching. The grouping of the data in Table 8.1 for various tube row spacings,
together with the dependency of the parameters on Re b , is discussed in Appendix A in relation to the
uncertainty of the data.
4
ESDU 73031
Note that in the evaluation of Nur for banks of tubes inclined to the axis of the duct, values of Vmax or
Gmax used in the calculation of Reb should be estimated assuming that the tubes are not inclined. The
factor F3 (Section 3.5) makes allowance for the effect on Nu of inclined tubes.
The values of F1 to be used in Equation (3.1) for both staggered and in-line tube banks are given in Figure 5.
This factor takes into account the effect on Nur of variable fluid properties due to the difference in
temperature between the bulk fluid and the tube wall. It is given by the empirical equation
Pr 0.26
F1 = ---------b- (3.3)
Pr w
ESDU Copyright material. For current status contact ESDU.
which was derived from an analysis of data for both staggered and in-line tube banks, for values of Pr b
up to about 600, and can be applied in practice for Re b > 10 .
Equation (3.3) is applicable to both gases and liquids and within the scatter of the available data appears
to be valid for both heating and cooling cases.
Values of the factor F 2 = Nu N ⁄ Nu N = 10 to be used in Equation (3.1) are given in Figures 6a and 6b
for in-line and staggered tube banks respectively. This factor accounts for the effect on the average Nusselt
number of the number of tube rows (in the direction of the flow) being greater or less than 10.
Values of F2 were derived by summing and averaging the mean data from measurements carried out on
individual tubes in a bank (see Section A1.3 of Appendix A). The method by which the row to row variation
of Nun can be obtained is discussed in Section 4.
2 3
F2 = C0 + C1 ⁄ N – C2 ⁄ N + C3 ⁄ N (3.4)
are valid for 4 ≤ N ≤ 20 with the following values of C0, C1, C 2 and C 3 . (For N > 20 , the value of F2 for
N = 20 can be assumed.)
Arrangement ( Xl – 1 ) ⁄ ( Xt – 1 ) Re b range C0 C1 C2 C3
3
In-line > 0.2 > 2 × 10 0.990 0.873 9.60 18.6
5
ESDU 73031
3.5 Calculation of Tube Inclination Factor F 3 (Figure 7)
When the tube bank is inclined to the oncoming flow (see Figure 7) the heat transfer is decreased by a factor
F3 where
Nu θ
F 3 = --------------------------
Nu θ = 90°
Values of the factor F3 are given in Figure 7 and have been derived as described in Section A1.2 of
Appendix A. For values of θ between 90° and 20° Figure 7 is represented by the equation
0.6
F 3 = ( sin θ ) (3.5)
ESDU Copyright material. For current status contact ESDU.
and this can be assumed to apply to all tube banks for values of Re b between about 102 and 106.
For values of θ less than 10° it is recommended that the problem be regarded as one of internal flow (in
the space along and between tubes) using data from References 2-4 and an equivalent hydraulic diameter
derived from the cross-sectional area between the tubes in the axial direction.
The row to row variation in heat transfer depends in general on the spacing between tube rows and the
Reynolds number. In most cases the heat transfer from tubes in the first row is considerably less than that
from succeeding rows. Turbulence created by the first few rows of tubes causes an increase in the heat
transfer coefficient on inner rows although the heat transfer becomes more or less constant after the first 3
or 4 rows. An exception to this is for in-line banks at low level Reynolds numbers. In this case the first row
contributes a greater proportion to the overall heat transfer because of a shielding effect on subsequent
rows, particularly where the longitudinal spacing between tubes is small compared with the lateral spacing.
For most cases the heat transfer for a particular row may be obtained in a similar way as for a bank
(Equation (3.1)) except that the factor F2 is replaced by a new factor F4 , relating the performance of a
particular row (the nth) to the average for a 10 row bank. Thus
Nu n = F 4 Nu N = 10
= Nu r × F 1 × F 3 × F 4 (4.1)
where Nun is the mean Nusselt number for the nth row and F4 is a function of n . Values of the factor F4
are given in Figures 8a and 8b for in-line and staggered tube banks respectively.
3
For in-line tube banks the data of Figure 8a for Re b > 2 × 10 are applicable for arrangements where
( X l – 1 ) ⁄ ( X t – 1 ) > 0.2 . The data of Figure 8a for Reynolds numbers between 102 and 103 may be
tentatively applied for values of ( X l – 1 ) ⁄ ( X t – 1 ) in the range 0.5 to 2. For more detail of the special
case when ( X l – 1 ) ⁄ ( X t – 1 ) is very small reference should be made to Derivation 22.
2
The data for staggered tube banks were derived for Reb > 10 and, within the scatter of data between the
various sources, can be assumed independent of Re b and tube spacing.
The accuracy of the data presented in Figures 8a and 8b is discussed in Section A1.3 of Appendix A.
6
ESDU 73031
5. OTHER FACTORS INFLUENCING CALCULATION OF HEAT TRANSFER
Two factors should be distinguished. Firstly, that due to the distribution of buoyancy forces arising from
temperature variations in the fluid close to the tube surface; no experimental data exist for this effect with
turbulent crossflow in tube banks, although it is known that in flow through vertical pipes such forces can
significantly affect turbulence production and hence heat transfer. Secondly, in the case of a crossflow heat
exchanger in which the tubes are horizontal and the flow vertical, there will of course be a “chimney effect”
so that the change in average density of the fluid must be accounted for in, for example, the determination
of pressure drop across the heat exchanger.
For all cases of liquid heating, and for most cases of gas heating, thermal radiation will be negligible
compared with forced convection. In cases where large surface to surface temperature differences occur,
or where surface to gas temperature differences are very large and the gas has significant absorption
characteristics (e.g. CO2 and H2O), a special investigation will be required (see, for example, Reference 1).
No account has been taken in this Item of the effect on the heat transfer of vibration of the tubes or of sound
waves propagated through the fluid. Frequencies in certain narrow band ranges can have a significant effect,
usually to increase the heat transfer, so that if this effect is ignored performance will err on the safe side as
far as surface area calculations are concerned.
6. EXAMPLE
A rectangular duct, 0.95 m by 5 m is used to carry an airflow of 12 000 kg/h across a simple plain-tube
heat exchanger consisting of 5 rows of tubes containing 20 tubes in each row each having an outside
diameter of 2.5 cm. Steam at atmospheric pressure is condensed within the tubes which have an effective
length of 5 m and are arranged on a 4.5 cm equilateral triangular pitch. What will be the rise in temperature
of the air across the tube bank due to forced convection heat transfer assuming the tube wall temperature
to be 100°C and the air inlet temperature to be 20°C?
(i) T w, m = 100°C .
The mean bulk temperature of the air flow across the bank is not known but a value of Tb, m = 30°
will be assumed as a first approximation.
(ii) For air the following physical properties apply at Tb, m and Tw, m .
3
c pb = c pw = 1.013 × 10 J/kg K ,
–6 –6
µ b = 19 × 10 kg/m s , µ w = 22 × 10 kg ⁄ m s ,
–3 –3
λ b = 26.4 × 10 W ⁄ m K , λ w = 31.6 × 10 W⁄ m K.
7
ESDU 73031
3 –6 –3
(iii) Pr b = 1.013 × 10 × 19 × 10 ⁄ ( 26.4 × 10 ) = 0.729 .
3 –6 –3
Pr w = 1.013 × 10 × 22 × 10 ⁄ ( 31.6 × 10 ) = 0.705 .
Pr b ⁄ Pr w = 1.034 .
Pr b = 1 Pr b = 1
= Wc pb ( T 2 – T 1 )
where T1 and T2 are the inlet and outlet temperatures of the airflow in °C and ∆T is here taken
as the log-mean temperature difference between the tube wall and the airflow over the tubes.
3
30.1 × ∆T × π × 0.025 × 5 × 100 = ( 12 000 ⁄ 3600 ) × 1.013 × 10 ( T 2 – 20 ) .
Thus,
∆T = 2.856T 2 – 57.12 . (6.1)
8
ESDU 73031
This equation can be solved by trial and error:
References
Derivations
*
If an arithmetic mean temperature difference had been used a value of T 2 = 43.8°C would have been obtained.
9
ESDU 73031
10. BERGELIN, O.P. A study of three tube arrangements in unbaffled tubular heat
DAVIS, E.S. exchangers. Trans. Am. Soc. Mech. Engrs, Vol. 71, pp. 369-374, 1949.
HULL, H.L.
11. BERGELIN, O.P. Heat transfer and fluid friction during viscous flow across banks of
BROWN, G.A. tubes - III: a study of tube spacing and tube size. Trans. Am. Soc. Mech.
et al. Engrs, Vol. 72, pp. 881-888, 1950
12. BERGELIN, O.P. Heat transfer and pressure drop during viscous flow across unbaffled
COLBURN, A.P. tube banks. Univ. Delaware, Engng Exp. Stn, Bull.2, 1950.
HULL, H.L.
13. SNYDER, N.W. Heat transfer in air from a single tube in a staggered-tube bank.
A.I.Ch.E. Symp. Ser.5, Vol.49, pp.11-20, 1950.
14. BERGELIN, O.P. Heat transfer and fluid friction during flow across banks of tubes - IV: a
ESDU Copyright material. For current status contact ESDU.
BROWN, G.A. study of the transition zone between viscous and turbulent flow. Trans.
DOBERSTEIN, S.C. Am. Soc. Mech. Engrs, Vol. 74, pp. 953-960, 1952.
15. FRITSCHE, A.F. Gestaltung und Berechnung von Ölkühlern. Verlag Leeman, Zürich,
1953.
16. KAYS, W.M. Heat transfer and friction characteristics for gas flow normal to tube
LONDON A.L. banks – use of a transient test technique. Trans. Am. Soc. Mech. Engrs,
LO, R.K. Vol. 76, pp. 387-396, 1954.
17. DWYER, O.E. Crossflow of water through a tube bank at Reynolds numbers up to a
SHEEHAN, T.V. million. Ind. Engng Chem., Vol. 48, pp. 1836–1846, 1956.
et al.
18. BERGELIN, O.P. Heat transfer and pressure drop during viscous and turbulent flow
LEIGHTON, M.D. across baffled and unbaffled tube banks. Univ. Delaware, Engng Exp.
et al. Stn, Bull. 4, 1957.
19. JONES, C.E. Convection heat transfer and pressure drop of air flowing across in-line
MONROE, E.S tube banks. Part I – apparatus, procedures, and special effects. Trans.
Am. Soc. Mech. Engrs, Vol. 80, pp. 18-24, 1958.
20. GRAM, A.J. Convection heat transfer and pressure drop of air flowing across in-line
MACKEY, C.O. tube banks. Part II – correlation of data for ten-row-deep tube banks.
MONROE, E.S. Trans. Am. Soc. Mech. Engrs, Vol. 80, pp. 25-35, 1958.
21. FAIRCHILD, H.N. Convection heat transfer and pressure drop of air flowing across in-line
WELCH, C.P. tube banks at close back spacings. Am. Soc. Mech. Engrs, Paper
61-WA-250, 1961.
22. WELCH, C.P. Individual heat transfer in a crossflow in-line tube bank. J. Heat Transf.,
FAIRCHILD, H.N. Vol. 86, pp. 143-148, 1964.
23. AUSTIN, A.A. Convective heat transfer in flow normal to banks of tubes. Ind. Engng
BECKMANN, R.B. Chem., Vol. 4, pp. 379-387, 1965.
et al.
24. HAMMEKE, K. Wärmeübergangs-und Druckverlustmessungen an Querangeströmten
HEINECKE, E. Glattrohrbündeln, insbesondere bei hohen Reynoldszahlen. Int. J. Heat
SCHOLZ, F. Mass Transf., Vol. 10, pp. 427-446, 1967.
10
ESDU 73031
25. SCHOLZ, F. Einfluss der Rohrreihenzahl auf den Druckverlust und Wärmeübergang
von Rohrbündeln bei hohen Reynoldszahlen. Chem. Ing. Techn, Heft
40, pp. 988-995, 1968.
26. ZHUKAUSKAS, A. Heat transfer in banks of tubes with crossflow of fluid. (In Russian.)
et al. Mintis, Vilnius, USSR, 1968.
27. BOISSIER, A. Les pertes de change et le transfert thermique, côté gaz, dans les
de CHATILLON, M. échangeurs de chaleur à tubes lisses, à circulations orthogonales.
et al. Electricité de France Bull. Dirn des Etudes et Rech., Série A, No. 2/3,
pp. 1-242, 1971.
28. SMITH, E.M. Thermal performance of cross-inclined tube bundles measured by a
COOMBS, A. transient method. J. Mech. Engng Sci., Vol. 14, No. 32, pp. 205-220,
1972.
ESDU Copyright material. For current status contact ESDU.
29. ZHUKAUSKAS, A. Heat transfer from tubes in crossflow. Advances in heat transfer, Vol.8,
pp.93-160, 1972.
30. LIS, J. Unpublished data. Central Electricity Research Laboratories,
Leatherhead, England, 1973.
11
ESDU 73031
8. TABLES
TABLE 8.1
Values of the Parameters to be used in Equation (3.2)
Curve no. in
Range of Re b Xt Xl Xt ⁄ Xl a m
Figure 2
*
Curve 1 in Figure 2 is also applicable to cases when X t /X l = 1.0 for values of X t ≥ 1.05 .
Range of Re b Xt Xl a m
12
ESDU 73031
TABLE 8.2
Ranges of Experimental Values Used in the Correlations and Uncertainty of the Data
Staggered In-line
Re b
tubes tubes
13
ESDU 73031
9. CALCULATION PROCEDURE
In general the inlet temperature, T1 , of each fluid will be given but the outlet temperature, T2 , may or may
not be specified. If the outlet temperatures are not known then it is necessary to assume tentative values in
order to obtain corresponding first estimates of Tb, m and Tw, m . Once a first estimate of the heat transfer
has been obtained a second approximation to the mean temperatures can be made and the calculations
repeated, if necessary.
The following procedure applies where the use of mean temperatures is satisfactory. When very large
changes in temperature are anticipated the tube bank should be divided into groups containing a sufficiently
small number of tubes (no more than about 20) so that the use of mean temperatures does not lead to
significant errors.
(iii) Evaluate Prb and Pr w corresponding to Tb, m and Tw, m and evaluate Prb ⁄ Pr w (for gases
Prb ≈ Prw ).
(iv) Evaluate Gmax = ρb Vmax at minimum flow area between tubes (see Figures 2 and 3).
(vi) Determine Nur using Figures 1, 2 and 4 for in-line banks and from Figures 3 and 4 for staggered
banks. Alternatively, use Equation (3.2) with values of a and m from Table 8.1.
(viii) Given N determine F2 from Figures 6a or 6b, or use Equation (3.4). For separate groups of tubes
between rows n 1 and n 2 the factor F2 is given by
n 2
1
∑
F 2 = ------------------------------ F 4
n2 – n1 + 1 n
1
(xi) Evaluate h = Nu × λ b ⁄ D .
(xii) Evaluate Q· = h∆T∆πDL (number of tubes in bank) where ∆T is the appropriate mean temperature
difference between the tube wall and crossflow fluid (see Section 3.1).
·
If outlet temperatures were not specified then these can now be re-estimated using the value of Q (step(xii))
in a heat balance for the crossflow or tube flow, e.g.
·
Q = Wc pb ( T b2 – T b1 ) .
14
ESDU 73031
New estimates of Tb, m and Tw, m can then be obtained and the calculations repeated until a convergence
of assumed and calculated values is obtained.
·
If the outlet conditions are specified (i.e. Q is predetermined) then the required surface area, and hence
number of tubes, can be obtained knowing the average heat transfer coefficient for the whole bank (or for
separate groups of tubes).
ESDU Copyright material. For current status contact ESDU.
15
ESDU 73031
ESDU Copyright material. For current status contact ESDU.
16
ESDU Copyright material. For current status contact ESDU.
104
8
7
6
5
4
Sl
3 St
2
W
103
8
7
6
5
1
4 2
3
Row number 1 2 3 3
2
Reb = ρb Vmax D / µb = Gmax D / µb where Gmax = W / Amin
(Nur)Pr
b=1
and Amin is the flow cross-sectional area between
102
8 tubes calculated along the full lines
7
6
17
2
Curve Xt
2 Xt Xl
number Xl
1
101 1.2 to 4 >1.15 >0.6
3 1
ESDU 73031
8 >1.05 1.0
7
6 2 >1.2 0.4 to 0.5
5
4 0.35 to
3 1.05 to 1.1
0.55
3
a. Curve numbers 1 to 3
10
10
101 2 3 4 5 6 7 8 102 2 3 4 5 6 7 8 103 2 3 4 5 6 7 8 104 2 3 4 5 6 7 8 105 2 3 4 5 6 7 8 106
Reb
FIGURE 2 NUSSELT NUMBER FOR THE REFERENCE CONDITIONS AND Pr b = 1.0 . IN-LINE TUBE BANKS
ESDU Copyright material. For current status contact ESDU.
104
8
7
6
5 Curve Xt Xl
4 number
3 4 4 to 6 1.4 to 1.7
2 5 2 to 3 1.1 to 1.15
6 4 to 6 1.1 to 1.2
103
8
7
6
5
4
3
2
(Nur)Pr
b=1 For values of Nur at Prb … 1 use Figure 4.
102
8
7
18
6
5
4
3
4
101
8
5
7
6 6
5
ESDU 73031
4
3
b. Curve numbers 4 to 6
10
10
101 2 3 4 5 6 7 8 102 2 3 4 5 6 7 8 103 2 3 4 5 6 7 8 104 2 3 4 5 6 7 8 105 2 3 4 5 6 7 8 106
Re
FIGURE 2a
ESDU Copyright material. For current status contact ESDU.
104
8
7 Curve Xt Xl
6
5 number
4 7 1.1 to 1.2 1.0
3 8 1.2 to 1.4 1.0
2
9 1.4 to 1.75 1.0
10 1.75 to 3 1.0
11 3 to 6 1.0
103
8
7
6
5
4
3
2
(Nur)Pr
b=1 For values of Nur at Prb … 1 use Figure 4.
102
8
7
19
6
5
4
3
7
101 8
8 9
7
6 10
11
ESDU 73031
5
4
3
c. Curve numbers 7 to 11
10
10
101 2 3 4 5 6 7 8 102 2 3 4 5 6 7 8 103 2 3 4 5 6 7 8 104 2 3 4 5 6 7 8 105 2 3 4 5 6 7 8 106
Reb
FIGURE 2b
ESDU Copyright material. For current status contact ESDU.
104
8
7
6 Sl
5
Reb = ρb Vmax D / µb = Gmax D / µb where Gmax = W / Amin
4
3 St and Amin is the minimum flow cross-sectional area between
2
tubes calculated along the full lines or the dashed lines
W
whichever is the smaller
103
8
7
6
5 For Nur at Prb … 1 use Figure 4.
4
Row number 1 2 3
3
2
(Nur)Pr
b=1
102
8
7
20
6
5
4
3
2
Applicable for all values of
Xt , Xl between 0.6 and 4
101
8
7
6
5
ESDU 73031
4
3
10
10
101 2 3 4 5 6 7 8 102 2 3 4 5 6 7 8 103 2 3 4 5 6 7 8 104 2 3 4 5 6 7 8 105 2 3 4 5 6 7 8 106
Reb
FIGURE 3 NUSSELT NUMBER FOR THE REFERENCE CONDITIONS AND Pr b = 1 . STAGGERED TUBE BANKS
ESDU Copyright material. For current status contact ESDU.
101
9
8
7
6
4
Nur
(Nur)Pr = (Prb)0.34
3 b=1
Nur
(Nur)Pr
b=1
10
10
0.9
0.8
21
0.7
0.6
0.5
0.4
0.3
ESDU 73031
0.2
10-1
10-1 2 3 4 5 6 7 8 10
10 2 3 4 5 6 7 8 101 2 3 4 5 6 7 8 102 2 3 4 5 6 7 8 103
Prb
FIGURE 4 EFFECT OF Pr b ON VALUES OF ( Nur )Pr = 1 FOR BOTH IN-LINE AND STAGGERED TUBE BANKS
b
ESDU 73031
1.4
1.2
ESDU Copyright material. For current status contact ESDU.
F1
1.0
0.8
0.6
0 1 2 3 4
Prb / Prw
22
ESDU 73031
F2 = NuN / NuN=10
1.1
N
ESDU Copyright material. For current status contact ESDU.
1.1
1.0
F2 Reb > 102
0.9
b. Staggered tube banks Tentative data
0.8
2 4 6 8 10 12 14 16 18 20
N
0.9
0.8
F3 102 < Re < 106
0.7
0.6
0.5
90 80 70 60 50 40 30 20 10 0
θ°
23
ESDU 73031
1.1
1.0
Reb > 2x103, (Xl − 1) / (Xt − 1) > 0.2
0.9
F4
2 3
{10 < Re <10
ESDU Copyright material. For current status contact ESDU.
b
0.5 <, (X − 1) / (X − 1) < 2
l t
0.8
0.7
0.6
2 4 6 8 10 12 14 16 18 20
1.1
0.9
F4
0.8
0.7
b. Staggered tubes
0.6
2 4 6 8 10 12 14 16 18 20
24
ESDU 73031
APPENDIX A
The information given in this Item is based on correlations of experimental data from the sources noted in
Table 8.2. Table 8.2 also gives the ranges of experimental data used to derive the correlations. Outside these
ranges the uncertainties are not known and the data should be used with caution.
The following notes give additional information on the derivation of some of the data.
A1.1 Values of Nu r
The values of the parameters in Table 8.1 to be used in Equation (3.2) for the calculation of Nu r are
dependent on the type of flow (i.e. laminar or turbulent) in the boundary layer on the surface of a tube. For
ESDU Copyright material. For current status contact ESDU.
2
tube banks this is determined mainly by Reynolds number, as indicated in Table 8.1. For Reb < 10 the
5
boundary layer flow is predominantly laminar and for Reb > 2 × 10 the boundary layer is predominantly
turbulent. Most of the available data are for Reynolds numbers between these values.
The value of Reb at which transition from one flow regime to another occurs (giving different values of
the parameters a and m ) is affected by several factors, the most influential being the spacing between tubes
since this influences the onset and spread of turbulence. However, in the variation of the reduced
–0.34
experimental data (in the form Nur × Pr b ) with Reb , trends with varying tube spacing are difficult to
define precisely because of inconsistencies (and the fairly considerable scatter of data) between the results
of various sources, even after the reduction of the data to the common reference conditions. For this reason
the mean values given in Table 8.1 have been selected for various groups of tube spacings and, within a
particular group, the values of the parameters can be taken as constant. The groups were chosen so that the
uncertainty of the data, indicated by the scatter of the majority of experimental values about the selected
mean curves, was within about ± 15 per cent of Nu . In practice the correlations of reduced experimental
–0.34
data, in the form Nu × Prb × 1⁄ ( F1 F2 ) , with Reb indicate that about 90 per cent of the data lie within
scatterbands of ± 15 per cent about the selected mean curves, except at low Reynolds numbers when a
greater uncertainty applies (see Table 8.2) due to the presence of free-convection effects.
For inclined tubes, using simple yaw theory, the Reynolds number, based on the component of flow normal
m
to the cylinder axis ( Vmax sin θ ) , is Reb sin θ . Thus, since Nu can be assumed proportional to Re b for a
given Pr b , then for inclined banks
Nu θ m
- = ( sin θ ) .
------------------------ (A1.1)
Nu θ = 90°
The values given in Table 8.1 indicate that a typical value of m is about 0.6 for most tube banks so that
0.6
F 3 = ( sin θ ) . (A1.2)
Values of F3 given by Equation (A1.2) are the basis of Figure 7 and these values agree with the few available
data (Derivations 7 and 8) for a staggered and in-line tube bank ( Xt = 2 , Xl = 2 ) to within about 5 per
cent for 90° ≥ θ ≥ 30° .
25
ESDU 73031
A1.3 Value of the Factors F 4, F 2
Insufficient data are available to assess precisely the accuracy of Figures 8a and 8b giving row to row
variation of heat transfer. However, the available data, for both staggered and in-line tube banks, indicate
an uncertainty of about ± 20 per cent on the first row, decreasing to about ± 10 per cent at the third row and
to about ± 5 per cent for the remaining rows. The large uncertainty for the first few rows is a reflection of
the large scatter of data between various sources. In particular, data for the first row are greatly influenced
by the turbulence level of the flow approaching the tube bank and Re b .
Values of the factor F2 , giving the effect on Nu of a varying number of tube rows in the bank, are derived
from Figures 8a and 8b assuming
Nu N N
1
Nu N = 10
∑
F 2 = ----------------------- = ---- F .
N n = 1 4
(A1.3)
ESDU Copyright material. For current status contact ESDU.
The factor F 4 is derived from data for tube banks with 10 or more rows. Equation (A1.3) assumes that the
average heat transfer of a tube bank with N rows of tubes can be obtained by summing and averaging the
individual row mean heat transfer of the first N rows of a tube bank with more than N rows. Few data are
available from which to check this assumption for small values of N , although the error is likely to be small
providing N ≥ 4 .
26
ESDU 73031
KEEPING UP TO DATE
Whenever Items are revised, subscribers to the service automatically receive the material required to update
the appropriate Volumes. If you are in any doubt as to whether or not your ESDU holding is up to date, please
contact us.
ESDU International plc Tel: 020 7490 5151 (from the UK)
+44 20 7490 5151 (from outside the UK)
Fax: 020 7490 2701 (from the UK)
+44 20 7490 2701 (from outside the UK)
E-Mail: esdu@ihs.com
Website: www.ihsesdu.com
ESDU Copyright material. For current status contact ESDU.
For users in the USA, please address all Customer Service and Support enquiries and suggestions to:
IHS Engineering Products Tel: 1 800 447 3352 (toll free number)
and Fax: 1 303 397 2599
Global Engineering Documents Website: www.ihs.com
global.ihs.com
ESDU 73031
Convective heat transfer during crossflow of fluids over plain tube
banks
ESDU 73031
ESDU 73031 gives a method for estimating the heat transfer from a bank
of plain tubes in fluid crossflow (as, for example, in a shell-and-tube heat
exchanger). The data apply to single-phase gas or liquid flow and were
obtained by an empirical correlation of experimental results from a wide
range of sources. Graphs enable mean heat transfer from the complete
bank to be estimated when the tubes are in-line or staggered, and allow
for variation of fluid properties with temperature and, if appropriate, tube
ESDU Copyright material. For current status contact ESDU.
inclination to the flow. The heat transfer from individual rows of the bank
can also be predicted. From a comparison of the method with
experimental results it is accurate to within 15 per cent.
All rights are reserved. No part of any Data Item may be reprinted, reproduced, or
transmitted in any form or by any means, optical, electronic or mechanical including
photocopying, recording or by any information storage and retrieval system without
permission from ESDU International plc in writing. Save for such permission all
copyright and other intellectual property rights belong to ESDU International plc.