ARTICLE IN PRESS

Energy 33 (2008) 1277– 1287

Contents lists available at ScienceDirect

Energy
journal homepage: www.elsevier.com/locate/energy

Heat transfer and friction factor correlations for a duct

having dimple-shape artificial roughness for solar air heaters
R.P. Saini , Jitendra Verma
Alternate Hydro Energy Centre, Indian Institute of Technology Roorkee, Uttarakhand 247667, India

a r t i c l e in fo abstract

Article history: The heat transfer coefficient between the absorber plate and air can be considerably increased by using
Received 26 June 2007 artificial roughness on the underside of the absorber plate of a solar air heater duct. Under the present
work, an experimental study has been carried out to investigate the effect of roughness and operating
Keywords: parameters on heat transfer and friction factor in a roughened duct provided with dimple-shape
Artificial roughness roughness geometry. The investigation has covered the range of Reynolds number (Re) from 2000 to
Solar air heater 12,000, relative roughness height (e/D) from 0.018 to 0.037 and relative pitch (p/e) from 8 to 12. Based
Nusselt number on the experimental data, values of Nusselt number (Nu) and friction factor (fr) have been determined
Friction factor
for different values of roughness and operating parameters. In order to determine the enhancement in
heat transfer and increment in friction factor values of Nusselt number and friction factor have been
compared with those of smooth duct under similar flow conditions. Correlations for Nusselt number
and friction factor have been developed for solar air heater duct provided such artificial roughness
geometry.
& 2008 Elsevier Ltd. All rights reserved.

1. Introduction elements on the absorber plate. A review on roughness

geometries used in solar air heater ducts was presented by
Several investigators have investigated forced convection heat Varun et al. [23].
transfer in smooth and roughened ducts, and lots of useful The application of artificial roughness in the form of dimples
information is available in the literature. The use of artificial on the heat transfer surface may be attractive roughness geometry
roughness on a surface is an effective technique to enhance heat in solar air heater ducts as it does not require complicated
transfer to fluid flowing in ducts. Artificial roughness can be manufacturing process, particularly if the dimple shape is a
provided by fixing wires, ribs, wire mesh or expanded metal mesh spherical indentation. Because of this characteristic, dimples do
and by providing roughness in dimple-shape geometry. Several not add extra weight to the absorbing plate, especially if surface
investigations have been carried out to study the effect of artificial material is removed to form the dimples. On thinner materials, an
roughness on heat transfer and friction factor for two opposite indentation-producing device can be used to form the dimples,
roughened surfaces by Han [1,2], Han et al. [3–5], Lau et al. [6–8], which then gives protrusions on the opposite side. This then leads
Taslim et al. [9,10], Liou and Hwang [11], Han and Park [12] and to another advantage because substantial heat transfer augmen-
correlations for heat transfer coefficient and friction factor were tations are also provided by the array of protrusions. A number of
developed by these investigators. Prasad and Mullick [13], Gupta investigations have been carried out on the heat transfer
[14], Saini and Saini [15], Karwa [16], Jaurker et al. [17], Bhagoria characteristics of channels or pipes with dimple roughness on
et al. [18], Chaube et al. [19], Verma and Prasad [20] and Gupta et the surface. Schukin et al. [24] studied the effects of channel
al. [21] have carried out investigations on rib-roughened absorber geometry (converging and diverging channels) on heat transfer for
plates of solar air heaters that form a system with only one downstream flow over a single hemispherical cavity. Chyu et al.
roughened wall and three smooth walls. Correlations for heat [25] reported overall heat transfer rate as 2.5 times more in case
transfer coefficient and friction factor have been developed for of duct having dimple-shape roughness geometry than that of a
such systems. Mittal et al. [22] compared the effective efficiency smooth surface. Lin et al. [26] presented corresponding flow and
of the solar air heater having different types of roughness heat transfer predictions in order to explain the observed heat
transfer behavior. Moon et al. [27] investigated the effect of
channel height on the heat transfer and friction in a dimpled
 Corresponding author. Tel.: +911332 285841; fax: 911332 273517. passage. Heat transfer enhancement of the order of 2.0 and
E-mail address: [email protected] (R.P. Saini). pressure drop penalties of the order of 1.6–2.0 times over smooth

0360-5442/$ - see front matter & 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.energy.2008.02.017
 ARTICLE IN PRESS

1278 R.P. Saini, J. Verma / Energy 33 (2008) 1277–1287

Nomenclature Nu Nusselt number for the roughened duct

Nus Nusselt number for the smooth duct
A area of the flow (m2) p roughness pitch (m)
AT throat area of the orifice (m2) Pr Prandtl number
AC area of the absorber plate (m2) DP pressure drop across the orifice meter (N/m2)
Cd coefficient of discharge for the orifice meter Rav average radial distance of the duct
CP specific heat of air (kJ/kg/K) Re Reynolds number
D equivalent diameter of the duct (m) (¼ 2WH/(W+H)) W width of the duct (m)
e height of the roughness element (m) b diameter ratio, D2/D1
DF pressure drop across the test section (N/m2) y tilt angle of the manometer
fr friction factor for roughened absorber plates r density of fluid (kg/m3)
fs friction factor for the smooth absorber plate n kinematic viscosity of fluid (m2/s)
H height of the duct (m) ti inlet temperature of air (1C)
h̄ average heat transfer coefficient (W/m2/K) to outlet temperature of air (1C)
k thermal conductivity of air (W/m/K) t̄ a average temperature of air (1C)
L length of the collector (m) t̄ p average temperature of the absorbing plate (1C)
m mass flow rate (kg/s)

surfaces were reported. Mahmood et al. [28] carried out a detailed provided in order to minimize the heat losses from the topside of
flow and heat transfer study on a dimpled plate and identified the heater assembly. A 4-mm-thick galvanized iron (GI) sheet
specific vortex structures responsible for augmenting heat having roughened surface on its under side was provided as heat
transfer from the downstream rims of each dimple. Heat transfer absorber plate. This plate formed the top wall of the duct. Fig. 2
enhancement ranging from 1.8 to 2.4 times over smooth plates shows the cross-section of the duct. The top of the entry and exit
was observed. Investigations made by various investigators sections of the duct was covered with 8-mm-thick wooden panels.
[25–28] have been for heat transfer and friction characteristics In order to ensure that no heat losses occur due to conduction
in ducts for the flow of high Reynolds number. The range of effect, the duct was insulated properly by providing glass wool at
Reynolds number investigated by them is not suitable for the the outer surface of the duct covering entry length, test length and
airflow in solar air heater ducts. Further most of the investigations exit length.
made so far in rectangular ducts were having more than 1 The mass flow rate of air was measured by means of a
roughened surface walls provided with such artificial roughness calibrated orifice meter connected with an inclined manometer.
geometry. Control valves were provided in the flow lines to control the flow.
The present investigation was, therefore, carried out with the An orifice meter plate was designed for the flow measurement in
objective of extensive experimentation on dimple-shape geome- the pipe having inner diameter of 80 mm. Calibrated copper–con-
try as artificial roughness provided on to the underside of one stantan thermocouples were used to measure the temperature of
broad wall of the duct of a solar air heater, in order to collect data air and absorbing plate temperature at different locations. A
on heat transfer and fluid flow characteristics. The experimental digital micro-Volt-millimeter was used to indicate the output of
data are presented in the form of Nusselt number and friction the thermocouples. The pressure drop across the test section was
factor plots as function of Reynolds number for different rough- measured with the help of a micro-manometer.
ness parameters. Using experimental data collected during
extensive experimental study, correlations for Nusselt number 2.2. Instrumentations
and friction factor have also been developed.
2.2.1. Temperature measurement
Calibrated copper–constantan thermocouples (24 SWG) with
2. Experimental program digital micro-Volt-millimeter, indicating output in degree centi-
grade with an accuracy of 0.1 1C were used to measure air
2.1. Experimental setup temperature and the temperatures of the absorber plate at
different locations. The thermocouples were calibrated with a
A schematic diagram of the experimental setup including the standard thermometer. Fig. 3 shows the calibration curve for the
test section is shown in Fig. 1. The flow system consisted of an thermocouples used. Fifteen thermocouples have been pasted on
entry section, a test section, an exit section, a flow meter and an the top surface of the absorber plate to record the plate
air blower. The duct having the dimensions of inner cross-section temperature. The locations of thermocouples are shown in
as 2400 mm  300 mm  25 mm was made of wooden panels. The Fig. 4. The temperature of air inside the duct was recorded by
test section having a length of 1000 mm was provided. The length 10 thermocouples inserted inside the duct at different locations as
of entry and exit section was provided as 900 mm and 500 mm, shown in Fig. 4. Thermocouple output was measured with the
respectively. These lengths were taken as per ASHRAE Standard help of a digital micro-Volt-millimeter through a selector switch.
93–97 [29]. An electric heater having a size of 1500 mm  290 mm
was fabricated by combining series and parallel loops of heating 2.2.2. Airflow measurement
wire on a 5-mm-thick asbestos sheet. A mica sheet of 1 mm thick The mass flow rate of air was measured by means of a
was placed on the electric heater wire, in order to get uniform calibrated orifice meter connected with an inclined manometer.
radiation between the electric heater and absorber plate. The The control valves were provided in the line to control the flow.
heat flux may be varied from 0 to 1000 W/m2 with the help of a The description of the orifice-meter is shown in Fig. 5. The orifice
variac connected across it. A 50 mm thick layer of glass wool as plate was calibrated with a Pitot tube. The calibration curve of the
an insulating material and a 12 mm thick wooden panel was same is given in Fig. 6.
 ARTICLE IN PRESS

R.P. Saini, J. Verma / Energy 33 (2008) 1277–1287 1279

Fig. 1. Schematic diagram of experimental setup.

Fig. 2. Cross-section of the duct.

2.2.3. Pressure drop measurement in duct movement was noted which yield the pressure difference across
Pressure drop across the test-section of the duct was measured the two pressure tapping. The butyl-alcohol having specific
by means of a micro-manometer having accuracy 0.01 mm. The gravity 0.8 has been used in the manometer to increase the
micro-manometer consisted of a moveable reservoir, a fixed accuracy further. The locations of the air taps were as shown in
reservoir and an inclined transparent tube connected to these Fig. 4.
reservoirs through flexible tubes. The movable reservoir was
mounted on a sliding arrangement using a micrometer having a
least count of 0.01 mm. The movement of reservoir was recorded. 2.3. Roughness geometry and range of parameters
The two reservoirs were connected with the air taps of the duct
through flexible tubes. The meniscus was maintained at a fixed Fig. 7 shows the roughness geometry of dimple-shape, which
prescribed mark by moving the reservoir up or down and the has been investigated under the present investigation. The range
 ARTICLE IN PRESS

1280 R.P. Saini, J. Verma / Energy 33 (2008) 1277–1287

Fig. 3. Calibration curve for thermocouples.

Fig. 4. Location of thermocouple on absorber plate and thermocouple and air taps in air duct.

of parameters covered under the present experimental study is as (ii) Pressure drop across the test section.
given below; (iii) Pressure difference across the orifice meter.

(i) Reynolds number 2000pRep12,000.

(ii) Relative roughness height 0.0189pe/Dp0.038.
(iii) Relative roughness pitch 8pp/ep12. 4. Data reduction

Steady-state values of the plate and air temperature in the duct

3. Experimental procedure at various locations were used to determine the values of useful
parameters. Mass flow rate, m, velocity of air, v, heat supplied to
The test runs to collect relevant heat transfer and flow friction the air, q and heat transfer coefficient, h, were calculated by using
data were conducted under steady-state conditions. For different following expressions:
airflow rates, the system was allowed to attain a steady state  
before the data were recorded. The following parameters were 2r DP sin y 1=2
m ¼ C d AT (1)
measured. 1  b4

(i) Temperature of the absorber plate and air at inlet and outlet m
v¼ (2)
of the test section. rWH
 ARTICLE IN PRESS

R.P. Saini, J. Verma / Energy 33 (2008) 1277–1287 1281

Fig. 5. Description of the orifice-meter.

Fig. 6. Calibration curve of the orifice-meter.

q ¼ mC p ðt o  t i Þ (3) in the direction normal to the flow were found to be negligible.

The plate temperature was determined at 15 locations on the
mC p ðt o  t i Þ plate as shown in Fig. 4. The average plate temperature (t̄ p ) was
h̄ ¼ (4) determined by taking average temperature of 15 thermocouples
AC ðt̄ p  t̄ a Þ
located at different locations as shown in Fig. 4.
where the temperature t̄ p and t̄ a are average values of the The air temperature variations were found to be linear. Since
absorber plate and fluid (air) temperature, respectively. The the temperature variations of the absorber plate in the direction
average value of plate temperature (t̄ p ) was determined from normal to the flow were found to be negligible in that direction,
the detailed temperature profile of the absorbing plate. It was the air temperature variations in that direction can be assumed to
found that the temperature of the absorber plate varies pre- be negligible. The temperature variation in the direction normal to
dominantly in the flow direction only and is linear. The variations the absorber plate could not be measured because of very small
 ARTICLE IN PRESS

1282 R.P. Saini, J. Verma / Energy 33 (2008) 1277–1287

Pitch
(p = 12 mm, 15 mm, 18 mm)

Pitch dia
(d = 0.8 mm, 1.3 mm, 1.5 mm, 1.7 mm)

Roughness height
(e = 0.8 mm, 1.3 mm, 1.5 mm, 1.7 mm)
Fig. 7. Schematic diagram of dimple-shape geometry.

duct depth and were assumed to be negligible. The air  Nusselt number ¼ 2.78%
temperature (t̄ a ) was determined as an average of the tempera-  Friction factor ¼ 1.03%
tures measured at six central location of the duct cross-section
along the flow direction as shown in Fig. 4. The convective heat
transfer coefficient was then used to obtain the average Nusselt 6. Results and discussions
number from the following expression:
The effect of flow and roughness parameters on heat transfer
h̄D and friction characteristics for flow of air in a rectangular duct
Nu ¼ (5)
k under the present investigation are presented and discussed
The friction factor was determined from the measured values of under this part of the paper.
pressure drop, DP, across the test section length, L of 1000 mm as In order to conduct the validity test of the duct, experimental
data were collected on conventional duct having smooth absorb-
2DPD
f̄ r ¼ (6) ing surface. The average values of Nusselt number and friction
4rLv2
factor for the duct having smooth plate were determined.
The validity of the test setup was verified by conducting Figs. 8 and 9 show the comparison of the experimental values
experiments for smooth duct under similar conditions. The values of Nusselt and friction factor for roughened solar air heater with
of Nusselt number and friction factor determined for smooth duct those predicted by correlations proposed by Modified Dittus–
were compared with the values obtained from correlations of Boelter correlation for Nusselt number and by Modified-Blasius
Dittus–Boetler equation for the Nusselt number and modified correlation for friction factor [30]. It can be seen from these
Blasius equation for the friction factor [30]. figures that the values obtained experimentally are comparable.
Nusse1t number for a smooth rectangular duct is given by Figs. 10 and 11 have been drawn to represent the Nusselt
Dittus–Boetler equation [30] as number derived from the test results as a function of the
  roughness and operating parameters.
Rav
Nus ¼ 0:023 Re0:8 Pr0:4 2 (7) Fig. 10 shows the Nusselt number as a function of Reynolds
D
number for relative pitch (p/e) values of 8, 10 and 12 and for a
Friction factor for a smooth rectangular duct is given by the fixed value of relative roughness height (e/D) of 0.0322. The plot is
Modified Blasius equation as prepared to compare the heat transfer rate for a duct having
dimple-shape geometry of different values of relative pitch (p/e).
f s ¼ 0:085 Re0:25 (8) Similar trend has been obtained for Nusselt number for all the
values of relative roughness height. It can be seen from the figures
5. Error analysis that for given values of roughness parameters, the Nusselt number
monotonously increases with increase of Reynolds number. It is
An error analysis of experimental measurements has been seen from Fig. 10 that Nusselt number has been found maximum
carried out on the basis of the method proposed by Kline and corresponding to relative roughness pitch, (p/e) of 10 which may
McClintock [31]. The maximum possible measurement errors in the be discussed on similar lines as investigated by other investiga-
values of major parameters of present investigation are given below tors. Due to separation of the flow at the dimple, reattachment of
free shear layer occurs for p/e of 10 and maximum heat transfer
 Reynolds number ¼ 1.06% coefficient occurs in the vicinity of the reattachment region. For
 Heat transfer coefficient ¼ 2.59% relative roughness pitch (p/e) of 8 and 12 reattachment may not
 ARTICLE IN PRESS

R.P. Saini, J. Verma / Energy 33 (2008) 1277–1287 1283

occur hence the values of heat transfer coefficient correspond to ment of free shear layer hence the maximum value of heat
these values of p/e are found to be minimum. Similarly, larger transfer is found to be corresponding to relative roughness height
relative roughness height (e/D) might produced more reattach- (e/D) of 0.0379.
Fig. 11 shows the Nusselt number as a function of Reynolds
number for relative roughness height (e/D) values of 0.0189,
0.0289, 0.0322 and 0.0379 and for a fixed value of relative pitch
(p/e) of 10. It is seen that Nusselt number increases with increase
in relative roughness height, e/D and has been found maximum
corresponding to relative roughness height of 0.0379.
Figs. 12 and 13 show the variation in friction factor with
Reynolds number for different values of roughness parameters. It
is seen from the figures that friction factor decreases mono-
tonously with increase in Reynolds number for different rough-
ness parameters. Fig. 12 shows the friction factor as a function of
Reynolds number for relative pitch (p/e) values of 8, 10 and 12 and
for a fixed value of relative roughness height (e/D) of 0.0322. The
plot is prepared to compare the friction factor for a duct having
dimple-shape geometry of different values of relative pitch (p/e).
Similar trend has been obtained for friction factor for all the
values of relative roughness height. Friction factor has been found
minimum for relative roughness pitch corresponding to 10.
Similarly Fig. 13 shows the variation of friction factor for relative

Fig. 8. Comparison of experimental and predicted values of Nusselt number.

90
0.025 80 p/e = 10
70
0.02 e/D-.0379
Experimental value 60
Nusselt number

e/D-0.0289
Friction factor

50 e/D-0.0189
0.015 e/D-0.0322
40 smooth

0.01 30

Blasius equation 20
0.005 10

0
0 2000 4000 6000 8000 10000 12000
0 2000 4000 6000 8000 10000 12000
Reynolds number
Reynolds number
Fig. 11. Variation of Nusselt number with Reynolds number for different values of
Fig. 9. Comparison of experimental and predicted values of friction factor. p/e and for fixed value of e/D.

60

e/D = 0.0322
50

40
Nusselt number

30 p/e 10
p/e 8
p/e 12
20 smooth

10

0
2000 4000 6000 8000 10000 12000
Reynolds number

Fig. 10. Variation of Nusselt number with Reynolds number and for different values of e/D and for fixed value of p/e.
 ARTICLE IN PRESS

1284 R.P. Saini, J. Verma / Energy 33 (2008) 1277–1287

0.08 100
0.07 e/D = 0.0322

0.06
Friction factor

0.05
p/e 10
0.04 p/e 8
p/e 12
0.03 A0 Re1.27

Nusselt number
smooth
0.02
0.01
0
0 2000 4000 6000 8000 10000 12000
Reynolds number

Fig. 12. Variation of friction factor with Reynolds number and for different values
of e/D and for fixed value of p/e. 10

0.1
p/e = 10 10000
0.09
Reynolds number
0.08
Fig. 14. Plot of Nusselt number versus Reynolds number.
0.07
Friction factor

0.06
analysis to fit a straight line through these data points may be
0.05
e/D-.0379 represented as
0.04 e/D-0.0289
e/D-0.0189 Nu ¼ A0 ðReÞ1:27 (11)
0.03 e/D-0.0322
smooth The coefficient A0 will in fact be a function of other influencing
0.02 parameters. Now, taking the parameter roughness pitch (p/e) into
consideration, the value of fðNu=Re1:27 Þ ¼ A0 g corresponding to all
0.01
values of (p/e) is plotted against (p/e), as shown in Fig. 15.
0 Regression analysis to fit a second-order quadratic curve fit may
2000 4000 6000 8000 10000 12000 be represented by the following expression:
Reynolds number    p  p2 
Nu
log ¼ log B0 þ B1 log þ B 2 log (12)
Fig. 13. Variation of friction factor with Reynolds number and for different values Re1:27 e e
of p/e and for fixed value of e/D.
The Eq. (12) can be rearranged as
  p3:15   p2 
roughness height, e/D of 0.089, 0.0289, 0.0322 and 0.079 and a fix Nu
1:27
¼ B0 expð2:12Þ log (13)
value of p/e of 10. It is seen that friction factor corresponding to Re e e
relative roughness height, e/D of 0.0189 has been found maximum
Further, the constant (B0) will be the function of relative rough-
while friction factor for e/D of 0.0289 is found to be minimum.
ness height (e/D).
This may be discussed on the similar lines as discussed above in
" #
case of heat transfer coefficient. ðNu=Re1:27 Þ
B0 ¼ (14)
7. Correlations for Nusselt number and friction factor ðp=eÞ3:15 ½expð2:12Þ ðlogðp=eÞÞ2 

As discussed earlier that the Nusselt number and friction factor Finally, plot for B0 versus roughness height (e/D) has been drawn
are strong functions of flow and roughness parameters, namely as shown in Fig. 16. A second-order quadratic curve fit in the
flow Reynolds number (Re) and the roughness dimensions of following form is obtained:
   p2 
relative pitch (p/e) and relative roughness height (e/D). The Nu p3:15
functional relationships for Nusselt number and friction factor log expð2:12Þ log
Re1:27 e e
can therefore be written as e   e 2
 ¼ log C 0 þ C 1 log þ C 2 log (15)
p e D D
Nu ¼ f n Re; ; (9)
e D
This can be rewritten as
 p e    p2 
f r ¼ f n Re; ;
e D
(10) Nu p3:15
expð2:12Þ log
Re1:27 e e
 e 0:033    e 2 
¼ C0 expð1:30Þ log (16)
D D
7.1. Correlation for Nusselt number
The values of the coefficients are obtained as given below:
All data points of Nusselt number from experimental data were
plotted against Reynolds number as shown in Fig. 14. A regression A0 ¼ 8:2  104 ; B0 ¼ 3:1  104 ; C 0 ¼ 1:98  103
 ARTICLE IN PRESS

R.P. Saini, J. Verma / Energy 33 (2008) 1277–1287 1285

A11(Re)-0.423

friction factor
B0(Re)1.27(p/e)3.15[exp(-2.12)(log(p/e))2]
Nu/Re1.27

10000
Reynolds number
10 Fig. 17. Plot of friction factor versus Reynolds number.
p/e

Fig. 15. Plot of (Nu/Re1.27) versus (p/e).

Nu/(Re)1.27(p/e)3.15exp[(-2.12(log(p/e))2]

C0(Re)1.27(p/e)3.15[exp(-2.12)(log(p/e))2]
(e/D)0.033[exp(-1.30)(log(e/D))2] B11(Re)-0.423(p/e)-0.465[exp(0.054)(log(p/e))2]
fr/Re-0.423

0.01

10
p/e

Fig. 18. Plot of (fr/Re0.423) versus (p/e).

0.01
e/D
7.2. Correlation for friction factor
1.27 3.15
Fig. 16. Plot of Nu/(Re) (p/e) [exp(2.12)(log(p/e))2] versus (p/e).
A similar procedure has been employed to develop correlation
for friction factor as shown in Figs. 17–19. The final form of
These values result in the following correlation for Nusselt correlation for friction factor is obtained as follows:
numbers:
p0:465   p2 
p3:15 f r ¼ 0:642 Re0:423 expð0:054Þ log
4 1:27 e e
Nu ¼ 5:2  10 Re
 e  e 0:0214    e 2 
 p2   e 0:033
 expð0:840Þ log (18)
 expð2:12Þ log D D
e D
   e 2 
 expð1:30Þ log (17) Fig. 20 shows the comparison between the experimental values
D of Nusselt number and those of predicted by the correlation
 ARTICLE IN PRESS

1286 R.P. Saini, J. Verma / Energy 33 (2008) 1277–1287

0.10

0.09
fr/(Re)-0.423(p/e)-0.465exp[0.054(log(p/e))2]

0.08
+15%
0.07

frPredicted
0.06

-15%
0.05
C11(Re)-0.423(p/e)-0.465[exp(0.054)(log(p/e))2]
(e/D)–0.0214[exp(0.840)(log(e/D))2] 0.04

0.03
0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
0.01 0.10 frExperimental
e/D
Fig. 21. Comparison of predicted values of friction factor with experimental
Fig. 19. Plot of ðf r =Re0:423 Þðp=eÞ0:465 ½expð0:054Þðlogðp=eÞÞ2  versus (e/D). values.

80 heaters having roughened duct provided dimple-shape roughness

geometry as artificial roughness. It can be concluded that heat
transfer can be enhanced considerably as a result of providing
dimple-shape roughness geometry on the absorber plate of a solar
air heater duct. Nusselt number and friction factor are the strong
60 function of the system and operating parameters.
The maximum value of Nusselt number has been found
corresponds to relative roughness height (e/D) of 0.0379 and
+15%
relative pitch (p/e) of 10. While minimum value of friction factor
has been found correspond to relative roughness height (e/D) of
NuPredicted

40 0.0289 and relative pitch (p/e) of 10. It is therefore, roughness

parameters of the geometry can be selected by considering the net
-15%
heat gain and corresponding power required to propel air through
the duct. Further different arrangement mode of the dimple-shape
artificial roughness on the absorbing plate may be investigated in
20 order to get the optimal arrangement mode of such artificial
roughness.
Based on experimental data collected during investigation,
correlations for Nusselt number and friction factor have been
developed for the investigated system. These correlations devel-
oped for such systems have been found to predict the values of
0
0 20 40 60 80 Nusselt number and friction factor reasonably well under the
NuExperimental given range of parameters.

Fig. 20. Comparison of predicted values of Nusselt number with experimental References
values.

[1] Han JC. Heat transfer and friction characteristics in rectangular channels with
developed for Nusselt number. Fig. 21 shows a comparison of rib turbulators. ASME J Heat Transfer 1988;110:321–8.
experimental values and the values of friction factor obtained [2] Han JC. Heat transfer and friction in channels with two opposite rib-
roughened walls. ASME J Heat Transfer 1984;106:774–81.
from the developed correlation. The average absolute percentage [3] Han JC, Glicksman LR, Rohsenow WM. An investigation of heat transfer and
deviations between the experimental and predicted values have friction for rib-roughened surfaces. Int J Heat Mass Transfer 1978;21:1143–56.
been found to be 7.58 and 4.68 for Nusselt number and friction [4] Han JC, Park JS, Lei CK. Heat transfer enhancement in channels with
turbulence promoters. ASME J Eng Gas Turbines Power 1985;107:628–35.
factor, respectively.
[5] Han JC, Zhang YM, Lee CP. Augmented heat transfer in square channels with
parallel, crossed and V-shaped angled ribs. ASME J Heat Transfer 1991;113:
590–6.
8. Conclusions [6] Lau SC, Kukreja RT, McMillin RD. Effects of V-shaped rib arrays on turbulent
heat transfer and friction of fully developed flow in a square channel. Int J
Heat Mass Transfer 1991;34:1605–16.
An extensive experimental study has been carried out to [7] Lau SC, McMillin RD, Han JC. Turbulent heat transfer and friction in a square
predict the heat transfer and flow characteristics of a solar air channel with discrete ribs turbulators. ASME J Turbomach 1991;113:360–6.
 ARTICLE IN PRESS

R.P. Saini, J. Verma / Energy 33 (2008) 1277–1287 1287

[8] Lau SC, McMillin RD, Han JC. Heat transfer characteristics of turbulent flow in [20] Verma SK, Prasad BN. Investigation for the optimal thermohydraulic per-
a square channel with angled discrete ribs. ASME J Turbomach 1991;113: formance of artificially roughened solar air heaters. Renew Energy 2000;
367–74. 20(1):19–36.
[9] Taslim ME, Bondi LA, Kercher DM. An experimental investigation of heat [21] Gupta D, Solanki SC, Saini JS. Thermohydraulic performance of solar air
transfer in an orthogonally rotating channel roughened with 45 deg. Criss- heaters with roughened absorber plates. Solar Energy 1997;61(1):33–42.
cross ribs on two opposite walls. ASME J Turbomach 1991;113:346–53. [22] Mittal MK, Varun, Saini RP, Singal SK. Effective efficiency of solar air heaters
[10] Taslim ME, Li T, Kercher DM. Experimental heat transfer and friction in having different types of roughness elements on the absorber plate. Energy
channels roughened with angled, V-shaped, and discrete ribs on two opposite 2007;32(5):739–45.
walls. ASME J Turbomach 1996;118:20–8. [23] Varun, Saini RP, Singal SK. A review on roughness geometry used in solar air
[11] Liou TM, Hwang JJ. Effect of ridge shapes on turbulent heat transfer and heaters. Solar Energy 2007;81(11):1340–50.
friction in a rectangular channel. Int J Heat Mass Transfer 1993;36:931–40. [24] Chukin AV, Koslov AP, Agachev RS. Study and Application of hemispherical
[12] Han JC, Park JS. Developing heat transfer in rectangular channels with rib cavities for surface heat transfer augmentation. In: ASME 40th international
turbulators. Int J Heat Mass Transfer 1988;31:183–95. gas turbine and aero congress, No. 95-GT-59, Houston, 1995.
[13] Prasad K, Mullick SC. Heat transfer characteristics of a solar air heater used for [25] Chyu MK, Yu Y, Ding H, Downs JP, Soechting F. Concavity enhanced
drying purposes. Appl Energy 1983;13:83–93. heat transfer in an internal cooling passage. In: ASME 42nd international
[14] Gupta D. Investigations on fluid flow and heat transfer in solar air heaters gas turbine and aero congress, ASME paper no. 97-GT-437, Orlando, FL,
with roughened absorbers. Ph.D. thesis, University of Roorkee, India, 1993. 1997.
[15] Saini RP, Saini JS. Heat transfer and friction factor correlations for artificially [26] Lin YL, Shih TIP, Chyu MK. Computations of flow and heat transfer in a
roughened ducts with expanded metal mesh as roughness element. Int J Heat channel with rows of hemispherical cavities. In: ASME Turbo Expo, ASME 99-
Mass Transfer 1997;40:973–86. GT-263, Indianapolis, 1999.
[16] Karwa R. Investigation of thermo-hydraulic performance of solar air heaters [27] Moon HK, O’Connell T, Glezer B. Channel height effect on heat transfer and
having artificially roughened absorber plate. Ph.D. thesis, University of friction in a dimpled passage. In: ASME Turbo Expo, ASME 99-GT-163,
Roorkee, India, 1997. Indianapolis, 1999.
[17] Jaurker AR, Saini JS, Gandhi BK. Heat transfer and friction characteristics of [28] Mahmood GI, Hill ML, Nelson DL, Ligrani PM, Moon HK, Glezer B. Local heat
rectangular solar air heater duct using rib-grooved artificial roughness. Solar transfer and flow structure on and above a dimpled surface in a channel.
Energy 2006;80(8):895–907. [29] ASHRAE Standard 93–97, Methods of testing to determine the thermal
[18] Bhagoria JL, Saini JS, Solanki SC. Heat transfer coefficient and friction factor performance of solar collectors, 1997.
correlations for rectangular solar air heater duct having transverse wedge [30] Han JC, Park JS, Lei CK. Heat transfer enhancement in channels with
shaped rib roughness on the absorber plate. Renew Energy 2002;25(3):341–69. turbulence promoters. Trans ASME J Eng Gas Turbines Power 1985;107:
[19] Chaube A, Sahoo PK, Solanki SC. Analysis of heat transfer augmentation and 628–35.
flow characteristics due to rib roughness over absorber plate of a solar air [31] Kline SJ, Mcclintock FA. Describing uncertainties in single sample experi-
heater. Renew Energy 2006;31(3):317–31. ments. Mech Eng 1953;75:3–8.