Proceedings of ICES2006

ASME Internal Combustion Engine Division 2006 Spring Technical Conference

May 8-10, 2006, Aachen, Germany

Proceedings of ICES2006
ASME Internal Combustion Engine Division 2006 Spring Technical Conference
ICES2006-1333
May 8-10, 2006, Aachen, Germany

ICES2006-1333
Experimental Study of Thermal Performance and Pressure Drop in Compact Heat
Exchanger Installed in Automotive
M. H. Saidi A. A. Mozafari
School of Mechanical Engineering School of Mechanical Engineering
Sharif University of Technology, Tehran, Iran Sharif University of Technology, Tehran, Iran
(11365-9567) (11365-9567)
[email protected] [email protected]

A. R. Esmaeili Sany J. Neyestani

School of Mechanical Engineering School of Mechanical Engineering
Sharif University of Technology, Tehran, Iran Sharif University of Technology, Tehran, Iran
(11365-9567) (11365-9567)
[email protected] [email protected]

ABSTRACT is also fundamental to have good knowledge about the

In this Study, radiator performance for passenger car has efficiency of the cooling system at an early stage of truck
been studied experimentally in wide range of operating development in order to save time and costs.
conditions. Experimental prediction of Nusselt number and Considerable success has been reported about improving heat
heat transfer coefficient for coolant in radiator tubes are also transfer effectiveness, declutching or modulating the cooling
performed with ε − NTU method. The total effectiveness fan [1], improving the efficiency of the ram path [2,3] and the
coefficient of radiator and heat transfer coefficient in air side is management of air inside the engine component [3,4]. Ram air
calculated via try and error method considering experimental effect on the airside cooling system performance has been
data. The Colburn factor and pressure drop are also estimated reported by Schaub and Charles [5]. There are several related
for this heat exchanger. Examples of application demonstrate investigations about heat transfer (Colburn factor) and friction
the practical usefulness of this method to provide empirical correlation for compact louvered fin-and-tube heat exchangers
data which can be used during the design stage. in [6,7].

INTRODUCTION NOMENCLATURE
The new emission regulations (Euro4 and EPA02) will require
higher performance and efficiency of a truck cooling system. A Total heat transfer area on Cmin side (m2)
Ac Minimum free flow area (cross-sectional are) (m2)
The truck cooling system could be divided into two parts, heat
C Cold fluid (air)
exchangers, (radiator, charge air cooler, oil cooler, etc) and air Specific heat capacity (kJ/kg.K)
Cp
flow management components (fan, fan clutch and drive,
shroud).The engine cooling radiators are basic components on G Mass velocity (kg/s)
which the correct engine operation and its performance depend. g Gravitational constant (m/s2)
h Hot fluid (coolant), Heat transfer coefficient(W/m2.K).
Predicting the radiator behaviour is an important issue that
i Inlet
helps the radiator designers to project them correctly once they Entrance loss coefficient
Kc
know the real boundary condition. Due to limited space at the
front of the engine the size of the heat exchangers is restricted Ke Exit loss coefficient
and cannot be essentially increased. Therefore, to meet these k Thermal conductivity of wall (W/m.K)
higher cooling demands it is very important to work with m& Mass flow rate (kg/s)
optimization of other components of the cooling system such as o Outlet
fan and fan shroud. Nevertheless the optimization of the truck Q& Overall heat transfer (kW)
cooling system involves not only in optimization of each single T Temperature ( o C )
component but even in analysis of interaction between theme It

1 copyright@ ICES2006 by ASME

1 Copyright © 2006 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 02/18/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use
 t Wall thickness (m) FLOW AND TEMPERATURE MALDISTRIBUTION
U Overall heat transfer coefficient (W/m2.K) In calculating heat exchanger performance it is normally
w Wall assumed that flow is uniform in both rate and temperature
δ Fin thickness distribution. In practice neither is likely to be true. While the
σ Ratio of free flow area to frontal area effects might be relatively small on the coolant side on the
ε Total effectiveness coefficient
airside the effects can be very large. The causes of this
maldistribution include [8]:
TEST EQUIPMENTS - Heat load and flow resistance from air-conditioning and
In the experimental tests, the automobile is connected to Eddy
auxiliary heat exchangers placed in front of the radiator.
current dynamometer and all actual moving conditions are
- Non-uniform inlet conditions caused by the radiator grille.
imposed on it. As shown in Fig. 1 two axial fans with
- Inherent mismatch between the annular flow of the fan and
associated channels employed in order to model the moving car
the rectangular heat exchanger. (The fan shroud will never be
on road. The air flow rate is controlled by moving gate installed
deep enough to allow full transition).
at the inlet of channel. All tests are performed based on SAE
- The proximity of objects placed downstream of the fan,
J2082 JUN92 standard.
cause asymmetric fan airflow.
In this research, the Pitot tube and wane anemometer are used
- Hot air recirculation caused by low pressure between the
to measure air velocity. K type thermocouples and
grille and the heat exchangers.
thermometers connected to wane anemometer were used for
Fig. 2 and 3 show histograms of radiator airflow
temperature measurement. Coolant flow rate was obtained by
maldistribution obtained from CFD analysis of a ford F350
rotameters. The multimeter was employed to estimate the
Vehicle, with ducting and seals removed.
electro power consumption. The Diagnostic system was also
employed to determining the motor parameters that controlled
by ECU.

Fig. 2- Radiator inlet temperature variation, Ford F350 At

72 (km/h) ambient temperature 38 oC [8]

Fig. 1- Test bench

Table 1 is shown the radiator specification that used in this
study.

Table 1- Radiator specification that used in this study

Fin and tube material Al
Width(mm) 382.4
Dimensional Depth(mm) 27
Height(mm) 320
Fin type Corrugated
Fin pitch(mm) 1.25
Fin total area(m2) 4.7
Tube total area(m2) 0.59 Fig. 3- Radiator inlet velocity variation, Ford F350
Heat transfer total area(m2) 5.29 At 72 (km/h). [8]
Radiator weight(kg) 2.8 It is generally known that the velocity of the airflow through
Cap operation pressure(kg/cm2) 0.9 ± 0.15 the radiator is a function of the vehicle speed and the heat
Radiator volume(lit) 1.14 transferred by a radiator is a function of the airflow rate across
the radiator and also temperature distribution [9]. However, the
non-uniformity is another factor to determine in the engine

2 copyright@ ICES2006 by ASME

2 Copyright © 2006 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 02/18/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use
 cooling. Others like Chiou [10] have suggested that radiator ⎡ 1 − e − NTU .C r ⎤
heat transfer effectiveness "Deteriorates due to two- ε = 1 − exp ⎢− ⎥ Cmin = Cwater (10)
dimensional flow non-uniformity on both the air and coolant ⎣ Cr ⎦
sides". The flow of heat from the hot coolant to the cold air, can be
Therefore temperature distributions as well as the non- represented by Newton's Law of cooling as shown in equation
uniformity of the cooling airflow across the radiator were (11):
measured. Q = UA(Th − Tc ) (11)

RADIATOR PERFORMANCE Where the overall heat transfer coefficient, U comprises of

Two common methods exist for expressing the heat transfer three separate resistances as shown in equation (12):
characteristics of a given heat exchanger surface geometry. 1 1 1
= + RW + (12)
These are known as LMTD approach and the ε − NTU UA U o Ao U i Ai
approach. The performance of a heat exchanger can be
determined by examining the heat loss and heat gain that takes If we assume the fin efficiency, η f , then equation (12) can be
place between its working fluids: rewritten to give:
The heat lost by the coolant can be expressed as:
1 1 1 t 1
(
Qh = m& h .C P .h . Thi − Tho ) (1)
= = + +
U h Ah U c Ac Ah hh K η0 Ac hc
(13)

Heat gained by the air can be expressed as: And total surface efficiency, η o
(
Qc = m& c .C P .c . Tco − Tci ) (2)
η0 = 1 −
Af
(1 − η )
f (14)
And the capacity rate, C is defined as [8]: A
C c = m& c .C p.c A= surface area on which U is based.
(3) For a flat tube heat exchanger it can be shown that [8]:
C h = m& h .C p.h
⎛ 2hc ⎞
tanh⎜⎜ .L ⎟
The smaller of the two values is defined as Cmin and the larger ⎝ Kδ ⎟⎠
Cmax and Cr is the capacity ratio: ηf = (15)
2hc
Cmin .L
Cr = (4) Kδ
Cmax
The most commonly used relationship for laminar flow
The number of transfer units, NTU is defied as [8]: (Re<2300) is the correlation proposed by Seider and Tate in
AU 1 1936 [8].
Cmin Cmin ∫
NTU = = UdA (5) 1 0.14
⎛ D ⎞3 ⎛ µ ⎞
Nu = 1.86⎜ Re . Pr . h ⎟ .⎜⎜ b ⎟⎟ (16)
When the inlet temperature and flow rates are specified, the ⎝ L ⎠ ⎝ µw ⎠
maximum heat transfer rate possible, (for an infinitely sized
counter flow heat exchanger), is given by: Where,

(
Qmax = C min Thi − TCi ) (6) Nu =
(x k )
(1h ) (17)
If effectiveness, ε is defied as the ratio of actual dissipation to
maximum dissipation we get [8]: Reynolds number, Re is defined as:

ε=
Q
=
Ch Thi − Tho (=
)
Cc Tco − Tci ( ) Re =
ρuDh
(18)
Qmax Cmin Thi − Tci ( )
Cmin Thi − Tci ( ) (7) µ
Prandtl number, Pr is defied as:
Effectiveness can be expressed as a function of NTU , capacity
µ.CP
ratio and heat exchanger configuration: Pr = (19)
ε = ε (NTU , Cr , Flow Arrangment )
K
(8)
And
There are many equations for ε that are obtained by different
4 × Flow Area
authors, but in real condition and for one row tube radiators, we Dh = (20)
can use equations (9) and (10): Wetted Perimeter
ε=
[ (
1 − exp − Cr . 1 − e − NTU )] Cmin = Cair (9)
Modification to the Petukhov model, by Gnielinski using
Cr experimental data has extended the correlation to include the

3 copyright@ ICES2006 by ASME

3 Copyright © 2006 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 02/18/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use
 transitional range (2300<Re<10000) where most automotive - Stanton Number can then be determined using equation
radiators operate [8]: (23) and Colburn factor using equation (24).
⎛ f ⎞
⎜ ⎟.(Re b − 1000 ). Prb AIR-SIDE PRESSURE DROP IN RADIATOR
Nu = ⎝ ⎠
2 Following Kays and London results, air-side pressure drop can
(21)
⎛ f ⎞ ⎛ ⎞
0 .5 2
be calculated from the following relationship [8]:
1 + 12.7⎜ ⎟ .⎜⎜ Pr 3 − 1⎟⎟
⎝2⎠ ⎝ ⎠ G 2 1 ⎡entrance effect + flow acceleration ⎤
∆P =
2 g ρ i ⎢⎣ + core friction - exit effect⎥⎦
.
Where fanning friction factor, f is given by [8]: (29)
f = (1.58 ln Re b − 3.28 )
−2
(22)
⎡ ⎛ ρi ⎞ ⎤
To facilitate the comparison of different heat exchanger ⎢ K c + 1 − δ + 2⎜⎜
2
2
ρo ⎠
(
− 1⎟⎟ + )⎥
geometries, the heat transfer coefficient is normally made into a ∆P =
G 1 ⎢
. ⎝ ⎥
2 g ρi ⎢ A ρi ρi ⎥ (30)
non-dimensional number using Stanton number, St or Colburn ⎢ + f
Ac ρ m
− 1 − ρ − Ke
2
⎥
ρ o ⎦⎥
( )
factor, j and compared at constant Reynolds number, Re [8]. ⎣⎢
convective heat flux hc
St = = (23)
fluid heat capacity rate ρ .u.C p
RESULTS AND DISCUSSION
Colburn factor is defined as [8]: As shown in Fig. (4), heat transfer rate has been shown via
2 coolant Reynolds number at different gears and maximum load.
j = St. Pr 3 (24)
At special coolant Reynolds number, increasing the car velocity
Using the following procedure, Colburn factor (j) can be leads to increase the airflow rate through radiator and with
determined: reference to equation (1) the heat transfer rate increase. In the
- For best accuracy select data for a coolant flow rate where other hand, increasing the water pump speed is the main reason
Re>4000. for increasing the heat transfer rate because it affects coolant
- As the inlet and outlet temperatures and flow rates are flow rate therefore the heat transfer rate.
known on both sides of the heat exchanger calculate The value of Nusselt number versus coolant Reynolds number
effectiveness, ε could be calculated using equation (7). has been shown in Fig. (5).The coolant heat transfer coefficient
- Using the appropriate ε − NTU relationship, NTU , and has been depicted versus Reynolds number in Fig. (6). As is
shown, increasing the Reynolds number leads to increasing the
using equation (5) UA , could be calculated.
coolant heat transfer coefficient that strongly affects total heat
- Using the Gnielinski correlation of equations (21) and
transfer rate of radiator. The value of Nusselt number has been
(22) the coolant side Nusselt number is obtained.
shown in Fig. 7 versus inlet air velocity at different gears. As is
- Then equation (17), convert Nusselt number to heat
shown in Fig. 7 airflow distribution and inlet air velocity to
transfer coefficient, hh . radiator have significant effect on heat transfer rate.
- Substitute for hh , Ah , t and k into equation (25):
20
1 1 1 t 1
= = + + (25)
UA U c Ac Ah hh K η 0 Ac hc 18

16
To give,
14
Heat Transfer (kW)

1 1
=k+ (26)
UA η 0 Ac hc 12

10
Where constant k is given by:
8
1 t
k = + (27) 6
Ah hh K First Gear
4
Second Gear
Substituting equation (14) into equation (27), gives equation 2 Third Gear
(28).
0
−1
⎛
⎛ 1 ⎞
)⎞⎟⎟
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 5500 6000 6500
⎜ − k⎟ = hc Ac ⎜⎜ 1 −
Af
(1 − η f (28) Coolant Side Reynolds Number
⎝ UA ⎠ ⎝ A ⎠
Fig. 4- Radiator heat transfer versus coolant Reynolds
- Using the fin efficiency relationships equations of (14) number in different gears
and (15) we can solve for fin efficiency, η f and air side heat
transfer coefficient, hc . As fin efficiency is a function of
heat transfer coefficient, the solution will be iterative.

4 copyright@ ICES2006 by ASME

4 Copyright © 2006 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 02/18/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use
 a and b are coefficients obtained from experimental results
33
(Fig. 8). Equation (31) is a linear behavior in logarithmic plane.
30
The radiator Colburn coefficient for available test case is:
27

24 j = 0.5457 Re −0.277 (32)

Nusselt Number

21

18 1
15

12

Colburn Number
6 First Gear
Second Gear -0.277
3 Third Gear j = 0.5457 Re
0.1
0
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 5500 6000 6500

Coolant Side Reynolds Number

Fig. 5- Nusselt number versus coolant Reynolds number

0.01
5000 100 1000 10000
Water Heat Transfer Coefficient (W/m^2.K)

4500 Coolant Side Reynolds Number

4000
Fig. 8- Colburn factor via Reynolds number for radiator in
3500
actual condition installed in automobile
3000

2500
With reference to equation (29), the air pressure drop is
affected by core friction, flow acceleration, inlet and outlet
2000
effects, inlet air density and inlet air velocity. The distribution
1500
of air pressure drop through radiator has been shown versus
1000
vehicle speed in Fig. 9. Increasing the pressure drop through
500
radiator is reasonable because increasing the air velocity
0
0 1000 2000 3000 4000 5000 6000
through radiator makes the airflow more turbulent between
radiator fins and therefore the influence of core friction is more
Coolant Side Reynolds Number
appeared.
Fig. 6- Coolant heat transfer coefficient versus coolant Air heat transfer coefficient affected by many parameter
Reynolds number because of complex flow behaviour between radiator fins. One
of the important parameters is the air velocity through radiator
33 or car velocity. As shown in Fig. (10), the influence of car
30 velocity on air heat transfer coefficient has been shown. As
27
shown the heat transfer coefficient is related to many
24
parameters and because of that it has a different trend at
different speeds but increasing the air velocity leads to increase
Nusselt Number

21
the heat transfer coefficient at all speeds.
18

15
800
12
750
9 700

6 First Gear 650

Air Pressure Drop (Pa)

Second Gear
3 Third Gear 600

550
0
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 500
Vehicle Speed (km/h) 450

400
Fig.7-Nusselt number versus vehicle speed in different gears
350

300
Radiator Colburn coefficient can be obtained as a function of 250
Reynolds number as follows: 200
20 25 30 35 40 45 50 55 60
j = a. Re b
(31) Vehicle Speed (m/s)

Fig. 9- Air pressure drop versus vehicle speed

5 Copyright © 2006 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 02/18/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use
 The total effectiveness coefficient of radiator has been shown
450
versus car velocity in Fig. 12 at different speeds. As is shown
Air Heat Transfer Coefficient (W/m^2.K)

400
when the car speed increase from a specified level the
350 effectiveness coefficient does not depend on car velocity and
300 converges to 0.7.
250
The radiator based on its dependency to airflow distribution is
200
approaches to maximum capacity when the speed increased. In
similar diagram, the total effectiveness coefficient of radiator
150
can be shown versus vehicle speed at different coolant flow
100
1000 rpm rates (Fig. 13). This diagram is a general state incorporating the
2000 rpm
50
3000 rpm total radiator performance at different flow rate and can be used
0 in future designing.
0 10 20 30 40 50 60 70
Vehicle Speed (km/h)

Fig. 10- Airside heat transfer coefficient versus vehicle 0.9

speed 0.85
0.8
The dependency of heat transfer rate to airflow distribution can
Total Effectiveness Efficiency
0.75

be determined from difference between real conditions in wind 0.7

0.65
tunnel and obtained results from experiment. As shown in Fig. 0.6
(11), the heat transfer results in wind tunnel improved about 50 0.55
percent because of better air distribution in radiator comparing 0.5

with actual conditions. 0.45

0.4
0.35
55 Water Flowrate 4.5 lit/min
0.3 Water Flowrate 20 lit/min
Flowrate (20 Lit/min)- Wind Tunnel Test
50 0.25 Water Flowrate 27 lit/min
Flowrate (40 Lit/min)- Wind Tunnel Test
Flowrate (20 Lit/min)-Actual Test 0.2
45 Flowrate (40 Lit/min)-Actual Test
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70
40
Vehicle Speed (km/h)
Heat Transfer (kW)

35

30 Fig. 13- Total effectiveness coefficient versus vehicle speed

25 for different coolant flow rate
20

15 As shown in Fig. 13 increasing the vehicle speed at special

10 coolant flow rate (constant speed of engine), leads to decrease
5 the total effectiveness coefficient with reference to equation (3)
0 and (7), minimum total heat capacity is related to airside,
10 14 18 22 26 30 34 38 42 46 50 54 58 62 therefore in equation (7) with increasing the vehicle speed, the
Vehicle Speed (m/s)
mass flow rate through radiator increases and in constant
Fig. 11- Heat transfer comparison between wind tunnel coolant flow rate, the total effectiveness coefficient decreases.
results and actual condition in automobile
CONCLUSIONS
1
In this study, radiator performance for passenger car has been
0.9
studied experimentally in wide range of operation conditions.
The total effectiveness coefficient of radiator and heat transfer
Total Effectiveness Efficiency

0.8

0.7
coefficient in air side is calculated via try and error method
0.6
considering experimental data. The Colburn factor and pressure
drop were estimated for this heat exchanger. The dependency
0.5
of heat transfer rate to airflow distribution determined from
0.4
difference between real conditions in wind tunnel and obtained
0.3
results from experiment. Heat transfer results in wind tunnel
0.2 First Gear
Second Gear
improved about 50 percent due to better air distribution in
0.1 Third Gear
radiator, comparing with actual conditions in automobile.
0
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70

Vehicle Speed (km/h)

Fig. 12- Total effectiveness coefficient versus vehicle speed

for different gears

6 Copyright © 2006 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 02/18/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use
 REFERENCES
[1] D.G. Stratton, R.E. Stringer, S.R.G. Taylor, “Engine
Cooling System Design and Development”, The Institution
of Mechanical Engineers, Automobile Division,
Proceedings 65-66, Volume 180, Part 2A, pp. 221-235.
[2] K.D. Emmenthal, W.H. Hucho, “A Rational Approach
to Automotive Radiator System Design”, SAE 740088,
1974.
[3] M.E. Olson, “Aerodynamic Effects of Front End
Design on Automobile Engine Cooling Systems”, SAE
760188, 1976.
[4] N. Hay, S.R.G. Taylor, “The Effects of Vehicle Cooling
System Geometry on Fan Performance”, Conference on
Fan Technology and practice, Institution of mechanical
Engineers, pp.176-179, 18-19 April 1972.
[5] U.W. Schaub, H.N. Chales, “Ram air Effects on the Air
Side Cooling System Performance of a Typical North
American Passenger Car”, SAE 800032, 1980.
[6] C.C. Wang, K.U. chi, Y.J. Chang, Y.P. Chang, “An
Experimental Study of Heat Transfer and Friction
Characteristics of Typical Louver Fin-and-Tube Heat
Exchangers”, Int. J. of Heat and Mass Transfer, Vol. 41,
pp. 817-822, 1998.
[7] C.C. Wang, C.J. Lee, C.T. Chang, S.P. Lin, “Heat
Transfer and Friction Correlation for Compact Louvered
Fin-and-Tube Heat Exchangers”, Int. J. of Heat and Mass
Transfer, Vol. 42, pp. 1945-1956, 1999.
[8] P. Kanefsky, Nelson, A. Valerie, M. Ranger, "A
Systems Engineering Approach to Engine Cooling
Design", SAE, SP-1541, 1999.
[9] Y.L. Lee, Y.T. Yong, “Analysis of Engine Cooling
Airflow including Non-uniformity over a Radiator”, Int. J.
of Vehicle Design, 24, pp. 121-135, 2000.
[10] J.P. Chiou, “The Effect of Flow Non-uniformity on
the Sizing of Engine Radiators”, SAE 800035, 1980.

7 copyright@ ICES2006 by ASME

7 Copyright © 2006 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 02/18/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use