Accelerat ing t he world's research.

Thermo-hydraulic performance of
solar air heaters having integral
chamfered rib roughness on absorber
plates
RAJENDRA KARWA
Energy

Cite this paper Downloaded from Academia.edu 

Get the citation in MLA, APA, or Chicago styles

Related papers Download a PDF Pack of t he best relat ed papers 

Art ificially roughened solar air heat er: Experiment al invest igat ions
Dalip Bera

Performance evaluat ion of solar air heat ers having v-down discret e rib roughness on t he absorber pla…
RAJENDRA KARWA

T hermo-hydraulic performance of a solar air heat er wit h n-subcollect ors in series and parallel configu…
Rajendra Karwa
 Energy 26 (2001) 161–176
www.elsevier.com/locate/energy

Thermo-hydraulic performance of solar air heaters having

integral chamfered rib roughness on absorber plates
a,*
Rajendra Karwa , S.C. Solanki b, J.S. Saini b

a
Department of Mechanical Engineering, Faculty of Engineering, Jai Narain Vyas University, Jodhpur 342 011,
India
b
Department of Mechanical and Industrial Engineering, University of Roorkee, Roorkee 247 667, India
Received 28 April 2000

Abstract

This paper presents results of an experimental investigation of the performance of solar air heaters with
chamfered repeated rib-roughness on the airflow side of the absorber plates. The roughened elements have
a relative roughness pitch of 4.58 and 7.09 while the rib chamfer angle is fixed at 15°. For the airflow
duct depths of 21.8, 21.5 and 16 mm, the relative roughness heights for the three roughened plates used
are 0.0197, 0.0256 and 0.0441, respectively. The airflow rate per unit area of absorber plate has been
varied between 0.024 to 0.102 kgs21 m22 (flow Reynolds number ranges from 3750 to 16 350). The study
shows substantial enhancement in thermal efficiency (10 to 40%) over solar air heaters with smooth
absorber plates due to the enhancement in the Nusselt number (50% to 120%). The thermal efficiency
enhancement is also accompanied by a considerable enhancement in the pumping power requirement due
to the increase in the friction factor (80% to 290%). At low flow rates, corresponding to applications
requiring air at a high temperature, the solar air heater with roughness elements having a high relative
roughness height, yields a better performance. However, at high flow rates the increase in the pumping
power is greater than the relative gain in the energy collection for a greater relative roughness height and,
hence, the net gain is higher for smaller roughness heights. At still higher flow rates, the smooth duct air
heater has better effective efficiency. A mathematical model for thermal performance prediction of solar
air heaters with absorber plate having integral chamfered rib-roughness has been presented. The experi-
mental and predicted values of thermal efficiency lie within ±7% with a standard deviation of ±5.8%. 
2001 Elsevier Science Ltd. All rights reserved.

* Corresponding author. Fax: +91-291-433733.

E-mail address: karwaFr@yahoo.com (R. Karwa).

0360-5442/01/$ - see front matter  2001 Elsevier Science Ltd. All rights reserved.
PII: S 0 3 6 0 - 5 4 4 2 ( 0 0 ) 0 0 0 6 2 - 1
162 R. Karwa et al. / Energy 26 (2001) 161–176

Nomenclature
A smooth absorber plate area = WL, m2
Cp specific heat of air, J/kg K
Dh hydraulic diameter of duct =4WH/ {2(W+H)}, m
e rib height, mm
e+ roughness Reynolds number
e/Dh relative roughness height
f friction factor
FR heat removal factor
g heat transfer function
G mass flow per unit area of plate = M/(WL), kg/m2 s
h heat transfer coefficient from absorber plate to air, W/m2 K
H air flow duct depth, m
I solar radiation on the collector plane, W/m2
k thermal conductivity of air, W/m K
L length of test section, m
m mass velocity of air = M/(WH), kg/m2 s
M mass flow rate, kg/s
N number of glass covers
Nu Nusselt number = hDh/k
p rib pitch, mm
P pumping power, W
p/e relative roughness pitch
Pr Prandtl number
Qu useful heat collection, W
QL heat loss from the collector, W
R roughness function
Re Reynolds number
St Stanton number
Ta ambient temperature, °C
Ti inlet air temperature, °C
Tm mean air temperature, (To+Ti)/2, °C
To outlet air temperature, °C
Tp mean plate temperature, °C
UL overall loss coefficient, W/m2 K
v velocity of air in the duct, m/s
w width of rib, mm
W width of the duct
W/H duct aspect ratio
 R. Karwa et al. / Energy 26 (2001) 161–176 163

Greek symbols

dp pressure drop, Pa
f rib chamfer angle, degree
h thermal efficiency
he efficiency of electric motor
hfan efficiency of blower or fan
hth efficiency of thermal-electric conversion process
heff effective thermal efficiency
m dynamic viscosity, Pa s
r density of air, kg/m3
ta transmittance-absorptance product
xt electric transmission loss coefficient

Subscript

s smooth duct

1. Introduction

Flat plate solar air heaters (solar collectors) have been employed to deliver heated air at low
to moderate temperatures for space heating, crop drying and industrial applications. The air to be
heated is passed through a rectangular cross-section duct below a metal absorber plate with the
sun-facing side blackened to facilitate absorption of solar radiation incident on the absorber plate.
Transparent covers are placed over the absorber plate to reduce the thermal losses from the heated
absorber plate. The thermal efficiency of a solar air heater is significantly low because of the low
value of the convective heat transfer coefficient between the absorber plate and the air, leading
to high absorber plate temperature and greater heat losses to the surroundings. It has been found
that the main thermal resistance to the heat transfer is due to the formation of a laminar sub-
layer on the absorber plate heat-transferring surface. Efforts for improving the heat transfer rate
have been directed towards artificially destroying or disturbing this sub-layer. An artificial rough-
ness on the heat transfer surface in the form of projections mainly creates turbulence near the
wall or breaks the laminar sub-layer and thus enhances the heat transfer coefficient with a mini-
mum pressure loss penalty. Artificial roughness in the form of wires or ribs of various geometrical
shapes have been employed for the enhancement of the heat transfer coefficient in heat exchangers
[1–3] and on the under side of the absorber plate to enhance the heat transfer coefficient between
the plate and the air flowing below the plate in solar air heaters [4–6]. From the experimental
study of the effect of artificial roughness, various investigators have shown that the geometry of
the roughness (roughness shape, pitch, height, etc.) has a marked influence on the heat transfer
and friction characteristics of the surface. Williams et al. [3] have reported strong effect of cham-
fering of the heads of the rectangular section ribs provided on the outer surface of the inner tube
164 R. Karwa et al. / Energy 26 (2001) 161–176

for the flow through a circular annulus. They observed that the Stanton number for a 13° cham-
fered surface was higher than that of 0° and 22° chamfered surfaces by about 8 and 11%, respect-
ively. Karwa et al. [7] carried out experimental investigation of heat transfer and friction for flow
of air in rectangular ducts with repeated chamfered rib-roughness on one broad wall. The rough-
ened wall was uniformly heated with the remaining three walls insulated. These boundary con-
ditions correspond closely to those found in solar air heaters. They observed that the Stanton
number and friction factor take their maximum values at the chamfer angle of 15°. Furthermore,
they [3,7] have shown that the closely-spaced chamfered ribs (relative roughness pitch #7) are
more effective in enhancement.
As implied from the discussion of the effect of rib-roughness, the heat transfer coefficient
enhancement is also accompanied with an enhancement in the friction factor. Thus, an appropriate
way to evaluate performance, in the case of solar air heaters with a roughened absorber plate, is
to take both heat collection rate and pumping power requirement into account, i.e. to carry out
a thermo-hydraulic performance evaluation.
The object of the present investigation was to study, both theoretically and experimentally, the
effect of chamfered integral rib-roughness on the absorber plate transverse to the flow of air as
shown in Fig. 1(a) on the thermo-hydraulic performance of solar air heaters. The theoretical
investigation has been carried out using the heat transfer coefficient and friction factor correlations
of Karwa et al. [7]. A direct experimental comparison of the performance of the roughened solar
air heater has been made with the performance of a solar air heater with a smooth absorber plate.
This has been accomplished by using two identical parallel ducts, one with the roughened absorber
plate and the other with a smooth one. The airflow rate per unit area of absorber plate was varied
between 0.024 to 0.102 kg s21 m22 (flow Reynolds number ranging from 3750 to 16350). Fig.
1(b) shows the rib geometry investigated.

2. Thermo-hydraulic performance evaluation criterion

Cortes and Piacentini [8] argued that the electric power required for pumping air through the
solar air heater duct is produced elsewhere from thermoelectric sources and then transmitted. A

Fig. 1. (a) Solar air heater with roughened absorber plate; (b) Roughness geometry.
 R. Karwa et al. / Energy 26 (2001) 161–176 165

considerable portion of the energy is lost in conversions and transmission, etc. Hence, they used
effective thermal efficiency heff for collector performance evaluation, based on the net energy
output of collector as (Qu2P/C), defined as
heff5(Qu2P/C)/(IA) (1)
where C is a conversion factor used for calculating equivalent thermal energy for obtaining the
pumping power. It is a product of the efficiencies of fan, electric motor, transmission, and thermo-
electric conversion, i.e.
C5hfanhe(12xt)hth (2)

As the operating cost of a collector depends on the pumping power spent, the effective
efficiency defined by Eq. (1) and based on the net energy gain is a suitable criterion for perform-
ance evaluation of the solar air heaters. Gupta [9] and Saini [10] also used the criterion defined
by Eq. (1) to investigate the thermo-hydraulic performance of roughened solar air heaters. This
criterion has been used in the present work for the thermo-hydraulic performance investigation
of the roughened solar air heaters.

3. Mathematical model

The energy balance of a solar air heater gives the distribution of incident solar energy, I, into
energy gain, Qu, and various losses. It is given by
Qu5A{I(ta)2UL(Tp2Ta)}5AFR{I(ta)2UL(Ti2Ta)} (3)
where A is the area of the absorber plate and the product (ta) is the transmittance-absorptance
product of the cover-absorber plate combination. FR is termed as heat removal factor [11].
The energy gain of the solar air heater equals the heat transferred to the air flowing through
the duct of the collector and is given by
Qu5MCp(To2Ti)5GACp(To2Ti) (4)

The convective heat transfer equation gives

Qu5hA(Tp2Tm) (5)

The term Tp is the mean absorber plate temperature and Tm is the mean air temperature. UL in
Eq. (3) is known as overall loss coefficient of a solar collector and is given by
UL5QL/{A(Tp2Ta)} (6)

The overall heat loss, QL, in Eq. (6), is the sum of losses from the top, Qt, the back, Qb, and
the edge Qe. The top, back, and the edge losses are calculated from modified equation [11] of
166 R. Karwa et al. / Energy 26 (2001) 161–176

Klein [12]. The wind heat transfer coefficient correlation of McAdams [13] has been used in
these equations.
The friction factor and heat transfer coefficient correlations for the integral chamfered rib-
roughness are given [7] in terms of roughness function, R{=√(2/f)+2.5 ln(2e/Dh) +3.75}, and heat
transfer function, g[={f/(2St)21}√(2/f)+R], respectively. These correlations are given as a function
of the roughness Reynolds number, e+{=√(f/2) Re (e/Dh)} and roughness parameters. e/Dh is the
ratio of the rib height, e, and the hydraulic diameter, Dh [=4WH/{2(W+H)}], of the duct. These
correlations are:
For 7#e+,20
R51.66e−0.0078f(W/H)−0.4(p/e)2.695exp[20.762{ln(p/e)}2](e+)−0.075 (7a)
when W/H.7.75 use W/H=7.75
g5103.77e−0.006f(W/H)0.5(p/e)−2.56exp[0.7343{ln(p/e)}2](e+)−0.31 (7b)
when W/H.10 use W/H=10
For 20#e+#60
R51.325e−0.0078f(W/H)−0.4(p/e)2.695exp[20.762{ln(p/e)}2] (7c)
when W/H.7.75 use W/H=7.75
g532.26e−0.006f(W/H)0.5(p/e)−2.56exp[0.7343{ln(p/e)}2](e+)0.08 (7d)
when W/H.10 use W/H=10
Knowing the Stanton number from the above correlations, the heat transfer coefficient, h, is
found from
h5(St Re Pr)k/Dh (8)

The equations for determination of pressure loss and pumping power are
dp5{(4fL)/(2rDh)}m2 (9)

P5(M/r)dp (10)

The thermal efficiency, h, of a solar air heater is the ratio of the useful heat gain, Qu, and the
incident solar radiation, I, on the plane of the solar air heater, i.e.
h5Qu/(IA) (11)

Niles et al. [14] used the following equation for the calculation of the outlet air temperature
when the ambient air passes through a solar air heater (i.e. Ti=Ta), which has been derived by
equating the right-hand sides of Eqs. (3) and (4),
To5Ta1I(ta)e/UL (12)
 R. Karwa et al. / Energy 26 (2001) 161–176 167

where e=12exp[2(UL/(GCp){1+UL/h}−1]=(FRUL/GCp)and they calculated the plate temperature

using the following equation derived from Eq. (3)
Tp5Ti1I(ta)/UL{12GeCp/UL} (13)

Eqs. (12) and (13) have been used for the cross-check of the values of Tp and To calculated
from Eqs. (3)–(11).
The transmittance–absorptance product for the absorber-cover combination has been estimated
at the experimental conditions to be 0.75. Thermal conductivity of the material used for the back
insulation has been taken as 0.037 Wm21 K 21. Long wave emissivity values of the glass cover
and plate with flat black paint have been taken as 0.88 and 0.9, respectively [12,15,16].
Eqs. (3)–(13) constitute a non-linear model for the roughened solar air heater that has been
used for computation of the useful heat gain, thermal efficiency, pressure loss, and pumping
power. The model has been solved by following an iterative process [9,11,17,18] using a per-
sonal computer.

4. Experimental program

The experimental test facility has been designed and fabricated according to the guidelines
given in ASHRAE standard 93-77 [19], for testing of solar collectors using an open loop system.
A schematic of the experimental setup is shown in Fig. 2. The test facility consists of two identical

Fig. 2. Schematic diagram of experimental setup.

168 R. Karwa et al. / Energy 26 (2001) 161–176

rectangular section wooden ducts 2640 mm long set on the suction side of a centrifugal blower.
The duct cross-section was varied by changing the duct depth, H, from 16.5 to 21.8 mm (while
the width of the duct was 150 mm). The lower and the two side wooden walls of the entire length
of the duct have a smooth surface. One of the ducts carries a smooth absorber plate while the
other carries the roughened absorber plate. Unheated entrance duct length in each duct is 400
mm (L/Dh<10 to 14, depending upon the duct depth). A short entrance length was chosen because
for a roughened duct the thermally fully developed flow establishes in a short length of 2 to 3
hydraulic diameters [20,21]. A 200 mm long exit section is provided to reduce any downstream
effect on the test section. For the turbulent flow regime, ASHRAE standard 93-77 recommends
minimum entry and exit lengths of 5√(WH) and 2.5√(WH), respectively, i.e. 280 mm and 140
mm, respectively, for the maximum duct depth employed in the present case.
The test section length for each duct is 1.64 m resulting in length to hydraulic diameter ratios
of 43 to 57 depending upon the duct depth. Glass plates of 3.25 mm thickness have been used
as covers over the absorber plates. The topsides of the entry and exit lengths of the ducts are
covered with smooth faced 10 mm thick plywood. The outside of the entire setup, from test
section inlet to the orifice plate, is insulated with 50 mm thick foamed polystyrene. The test setup
was installed on the rooftop.
The scheme and the instruments used for the measurement of absorber plate temperature, static
pressure drop and the mass flow rate of air through the each duct are the same as used in previous
work of the authors [7]. Three thermocouples arranged span-wise, as shown in Fig. 3(a), measure
the exit air temperature after the mixing section. Fig. 3(b) shows the location of 18 thermocouples
affixed along 10 axial stations to measure the temperature of the plate.
A calibrated precision pyranometer was used for the measurement of the intensity of the solar
radiation. The pyranometer was installed near the solar air heater with the horizontal guard plate
of the pyranometer at the level of the glass cover of the solar air heater.
The measurement of wind velocity has been carried out using a wind speed and direction
measuring instrument consisting of a micro-logger, wind speed and direction sensor, and a display
unit. The instrument was installed in the vicinity of the collector at a height corresponding to the
collector height.
Table 1 gives the details of the roughness and duct parameters. Each quoted rib height in Table
1 is the mean of three measurements taken along rib-length at each of 10 equally spaced axial
stations along the length of the plate. Each quoted rib pitch is the mean value determined from
a count of complete pitches over an accurately measured distance of about 300 mm. The variation

Fig. 3. (a) Location of thermocouples for exit air temperature measurement; (b) Location of thermocouples on the
absorber plate.
 R. Karwa et al. / Energy 26 (2001) 161–176 169

Table 1
Dimensions of roughened elements and test duct

Plate Rib height Rib pitch Chamfer Duct width Duct depth Hydraulic p/e e/Dh Duct
no. e (mm) P (mm) angle f W (mm) H (mm) diameter aspect ratio
(degree) Dh (mm) W/H

1 0.74 5.25 15 150 21.5 37.60 7.09 0.0197 6.98

2 1.68 7.69 15 150 21.8 38.06 4.58 0.0441 6.88
3 0.74 5.25 15 150 16.0 28.90 7.09 0.0256 9.38

in the height along any one rib and between the ribs was negligibly small, the maximum difference
being 0.02 mm on a mean rib height of 0.74 and 1.68 mm.
The integral rib-roughened absorber plates of aluminium were prepared by machining. The
plates were made in two or three pieces and transverse grooves were machined on the plain side
(opposite of the ribbed side) of the plates to reduce the effect of axial conduction of heat.
The minimum roughness height is chosen such that the laminar sub-layer would be of the same
order as roughness height at the lowest flow Reynolds number. The maximum rib height is 1.68
mm so that the fin and flow passage blockage effects are negligible. The rib width to height ratio
varies in a narrow range (1.0 to 1.5).
All readings were recorded within one hour of the solar noon on clear bright days. The data
points have been equally divided about the solar noon so that the transient effects do not influence
the test results. The variation in insolation for 30 min interval has been found to be 5 to 20
Wm22. The wind velocity during the experimentation ranged from 0.22 to 1.4 m s21.

5. Data reduction

The heat transfer coefficient, h, for the heated test section was calculated from Eq. (5), where
the useful heat gain, Qu, in Eq. (5) is the heat transfer rate to the air determined using Eq. (4)
from the measured air mass flow rate, M, and air temperatures, Ti and To, at the test section inlet
and outlet, respectively. In Eq. (5), A is the heat transfer area, assumed to be the corresponding
smooth plate area for the roughened plate, and Tm is the average temperature of the air calculated
as the mean of the measured inlet and outlet temperatures.
The solar insolation gives a condition of constant heat-flux boundary condition and the plate
temperature varies in the direction of the airflow. In all the calculations, a mean plate temperature
designated as Tp has been used, which is calculated as integrated mean value of the plate tempera-
ture variation obtained experimentally.
The heat transfer coefficient has been used to determine the Nusselt number:
Nu5hDh/k (14)

The collector thermal efficiency, h, was calculated from the calculated useful heat gain, Qu,
and the measured solar insolation, I, using the equation:
170 R. Karwa et al. / Energy 26 (2001) 161–176

h5Qu/(IA) (15)

The effective thermal efficiency, heff, has been determined using Eq. (1).
The Reynolds number was determined from the value of the mass flow rate, M, using the equ-
ation:
Re5mDh/m (16)
where m (=M/WH) is the mass velocity of air through the collector duct.
The friction factor was determined from the measured values of pressure drop, dp, across the
test section length, Lf, of 1.2 m (between the two points at 400 mm and 1600 mm from the inlet
of the test section) and the mass flow rate using the equation:
f52(dp)Dh/{4Lfrv2} (17)

The pumping power required to propel the air through the collector duct has been determined
using Eq. (10). The pressure loss dp in Eq. (10) is the pressure loss for the collector length L.
The effective efficiency has been calculated from Eq. (1) with the conversion factor C as 0.2
based on the assumption that the blower-motor combination efficiency is 0.6 and the net efficiency
of thermo-electric conversion process (referred to the consumer point) is 0.33.
The thermophysical properties of air employed in the calculations were picked up from avail-
able tables [22] corresponding to the average fluid temperature Tm. The effect of humidity has
been neglected since the relative humidity values during the experimentation were found to be
low and variation was small, ranging between 20 to 35%.
From the analysis of the uncertainties in the measurements by various instruments [23], the
uncertainties in the calculated values of various parameters are given below.

Reynolds number = ±2.12%

Friction factor = ±4.56%
Nusselt number = ±4.65%
Thermal efficiency = ±5.35%
Pumping power = ±3.95%

6. Results and discussion

An experimental investigation of the performance of solar air heaters with chamfered repeated
rib-roughness on the airflow side of the absorber plates has been carried out. Three roughened
plates have been tested. These roughened plates have a relative roughness pitch of 4.58 and 7.09
while the rib chamfer angle is fixed at 15°. For the airflow duct depths of 21.5, 16 and 21.8 mm,
the relative roughness heights for the three roughened plates are 0.0197, 0.0256 and 0.0441,
respectively. The airflow rate per unit area of absorber plate has been varied between 0.024 to
0.102 kgs21 m22 (flow Reynolds number ranges from 3750 to 16350). A direct experimental
 R. Karwa et al. / Energy 26 (2001) 161–176 171

comparison of the performance of the roughened solar air heater has been made with the perform-
ance of the solar air heater with a smooth absorber plate. This has been accomplished by using two
identical parallel ducts, one with the roughened absorber plate and the other with the smooth one.
The enhancement in the thermal efficiency, as a result of the use of artificial roughness on the
absorber plate, is shown by plots of efficiency ratio, h/hs, for the three roughened plates in Fig.
4 where hs is the thermal efficiency of the smooth duct solar air heater under identical conditions.
The thermal efficiency of the roughened air heaters is about 10 to 40% higher as compared to
solar air heater with smooth absorber plate. The enhancement depends on the flow rate and the
relative roughness height. It can be seen that the improvement over the smooth duct solar air
heater is greater at lower flow Reynolds number range of about 3500 to 10 000 (the corresponding
mass flow rate per unit area of plate is 0.025 to 0.05 kg m22 s21). The air heater with the greatest
relative roughness height (e/Dh=0.0441) shows the greatest enhancement in the thermal efficiency.
It has been reported [7] that the maximum value of the Stanton number occurs at a lower Reynolds
number for the rib-roughness of a larger relative roughness height. Actually the maxima of the
Stanton numbers for plates with different flow and roughness geometries has been reported [7]
to occur at a nearly fixed value of the roughness Reynolds number. Furthermore, it is reported
that the value of the Stanton number remains nearly constant after attaining its maximum value.
Thus, at low flow rates the Stanton number for larger relative roughness heights is near their
maximum, it is not so for the lowest relative roughness height as can be seen in Fig. 4. At high
flow rates, the Stanton number for the lowest relative roughness height also attains its maximum.
This explains (i) the large difference in the efficiency ratio values for the three plates with different

Fig. 4. Efficiency ratio versus Reynolds number.

172 R. Karwa et al. / Energy 26 (2001) 161–176

relative roughness heights at low flow rates and the decreasing difference in efficiency ratio values
with the increase in the flow rate and (ii) the difference in the trends of the efficiency ratio curves
for the three plates.
The improvement in the thermal efficiency of the solar air heater with chamfered rib-roughness
on the absorber plate is attributed to the increased heat transfer coefficient due to the increased
turbulence near the plate surface. The enhancement in the Nusselt number due to the artificial
roughness is found to be about 50 to 120% as shown in Fig. 5(a). Again the greatest enhancement
in the Nusselt number has been observed for the air heater with the largest relative roughness
height. The enhancement in the heat transfer rate has been found to result in a lower temperature
of the roughened absorber plate and in the increased outlet air temperature as compared to the
smooth plate. Thus, the heat loss from the collector to the surroundings, which is proportional to
the temperature excess of the absorber plate over the ambient, is lower for the roughened collector.
The artificial roughness on the absorber plate that results in the desirable increase in the thermal
efficiency also results in an undesirable increase in the pressure drop due to the increased friction.
The friction factors for the roughened solar air heaters are about 1.8 to 2.3 times of that for the
smooth duct solar air heater for the relative roughness height of 0.0197 and about 3.1 to 3.9 times
for the relative roughness height of 0.0441 as can be seen in Fig. 5(b).
The thermal efficiency, h, and the effective thermal efficiency, heff, data of the solar air heater
with plate no. 3 (p/e=7.09, e/Dh=0.0256, f=15°, with duct aspect ratio of 9.38) have been plotted
as a function of the Reynolds number in Fig. 6. The plots show that though the thermal efficiency
increases monotonously with the Reynolds number and becomes asymptotic at higher Reynolds
number, the effective efficiency attains a certain maximum value and thereafter decreases with
further increase in the Reynolds number. Both the useful heat gain, Qu, and pumping power, P,
increase with increase in the Reynolds number, but the pumping power is a much stronger function
of the flow rate (varies as cube of the flow velocity) than the convective heat transfer coefficient.

Fig. 5. (a) Nusselt number ratio versus Reynolds number; (b) Friction factor ratio versus Reynolds number.
 R. Karwa et al. / Energy 26 (2001) 161–176 173

Fig. 6. Variation of thermal efficiency and effective thermal efficiency with Reynolds number.

Thus, with the increase in the Reynolds number beyond a certain value, the increase in the heat
collection rate is much smaller than the increase in the pumping power requirement and the net
energy gain (Qu2P/C) and, hence, the effective efficiency decreases. It is to be noted that as the
power needed to propel air through the collector duct is not significant at the low flow rates
(corresponding to applications requiring air at high temperatures such as space heating), the differ-
ence between h and heff is negligible. The effect of the relative roughness height on the effective
efficiency at high flow rates (corresponding to applications requiring air at low or moderate
temperatures) is interesting. While at low flow rates the pumping power requirement is not sig-
nificant and the net energy collection is quite high for the larger rib height due to the higher heat
transfer coefficient, the same is not true at high flow rates. At high flow rates, the increase in the
pumping power is greater than the relative gain in the energy collection for a larger relative
roughness height. Thus, at the high flow rates the net gain is higher for smaller roughness heights.
At still higher flow rates, it can be seen from Fig. 6, where the thermal efficiency and effective
thermal efficiency data for the smooth duct collector are also plotted, that the smooth duct air
heater has better effective efficiency and, hence, the smooth duct collector must be a choice at
such flow rates. It is to be noted that the pumping power requirement for the roughened solar air
heaters has been found to be of the order of 0.05% of the energy collected at the lowest flow
Reynolds number of about 3500 which increased rapidly with the Reynolds number to 5% at the
highest Reynolds number of about 16 000 in the present study.
In Fig. 7, the experimental values of the thermal efficiency of the three roughened collectors
have been plotted against the predicted values of the thermal efficiency of these collectors using
174 R. Karwa et al. / Energy 26 (2001) 161–176

Fig. 7. Comparison of experimental values of thermal efficiency with predicted values.

the mathematical model presented. The plot of 39 data shows that the experimental and predicted
values lie with in ±7% with a standard deviation of ±5.8%. Thus, the mathematical model compris-
ing the correlations developed by Karwa et al. [7] can be utilized with confidence for prediction
of the performance of solar air heaters with absorber plates having integral chamfered rib-rough-
ness whose relative roughness height ranges from 0.0197 to 0.0441 and relative roughness pitch
from 4.58 to 7.09 for airflow rate per unit area of the absorber plate ranging from 0.025 to 0.102
kg m22 s21.
It may be noted that the improved performance of the solar air heater with roughened absorber
plate has been achieved at an increase of approximately 20% in the capital cost of the collector.
This additional cost of the collector is likely to be reduced in case of mass production of the
roughened absorber plates.

7. Conclusions

The main findings of the thermo-hydraulic performance study of solar air heaters having cham-
fered repeated rib-roughness on the air flow side of the absorber plates are:

1. There is an appreciable increase in the thermal efficiency (10 to 40%) of solar air heaters with
artificially roughened absorber plate due to the enhancement in the Nusselt number of the order
of 50 to 120% over the smooth absorber plate.
 R. Karwa et al. / Energy 26 (2001) 161–176 175

2. The artificial roughness on the absorber plate also causes 1.8 to 3.9 times increase in the
friction factor.
3. The enhancements in the Nusselt number, friction factor and thermal efficiency are found to
be strong functions of the relative roughness height. The greatest enhancement is observed for
the air heater with the highest relative roughness height.
4. The effective efficiency at first increases with the increase in the Reynolds number, attains a
maximum value at a roughness dependent Reynolds number and thereafter it decreases because
of a substantial increase in the pumping power that varies as the cube of the flow velocity.
5. At low flow rates, corresponding to applications requiring air at a high temperature, the solar
air heater with the roughness elements having a large relative roughness height yields a better
performance. However, at high flow rates the increase in the pumping power is greater than
the relative gain in the energy collection for a greater relative roughness height and, hence,
the net gain is higher for smaller roughness heights. At still higher flow rates, the smooth duct
air heater has better effective efficiency.
6. A mathematical model for thermal performance prediction has been presented. The experi-
mental and predicted values of the thermal efficiency lie within ±7% with a standard deviation
of ±5.8%. Thus, the mathematical model given can be utilized with confidence for prediction of
the performance of solar air heaters with integral chamfered rib-roughness on the absorber plate.

Acknowledgements

This work was carried out at the University of Roorkee, Roorkee (India). The authors acknowl-
edge the facilities provided by the university.

References

[1] Webb RL, Eckert ERG, Goldstein RJ. Heat transfer and friction in tubes with repeated rib roughness. Int J Heat
Mass Transfer 1971;14:601–17.
[2] Han JC, Park JS, Lei CK. Heat transfer enhancement in channels with turbulence promoters. J of Engineering
for Gas Turbines and Power 1985;107:628–35.
[3] Williams F, Pirie MAM, Warburton C. Heat transfer from surfaces roughened by ribs. In: Bergles A, Webb RL,
editors. Augmentation of convective heat transfer. New York: ASME, 1970:36–43.
[4] Prasad K, Mullick SC. Heat transfer characteristics of a solar air heater used for drying purposes. Applied Energy
1983;13:83–93.
[5] Prasad BN, Saini JS. Effect of artificial roughness on heat transfer and friction factor in a solar air heater. Solar
Energy 1988;41(6):555–60.
[6] Gupta D, Solanki SC, Saini JS. Enhancement of heat transfer in rectangular solar air heater duct having artificial
roughness on absorber plate. In: Balakrishnan AR, Srinivasa Murthy S, editors. Proceedings of 11th National Heat
and Mass Transfer Conference, December 21–23, 1991, Madras (India). New Delhi: Tata McGraw-Hill Publishing
Co Ltd, 1991:113–6.
[7] Karwa R, Solanki SC, Saini JS. Heat transfer coefficient and friction factor correlations for the transitional flow
regime in rib-roughened rectangular ducts. Int J of Heat and Mass Transfer 1999;42:1597–615.
[8] Cortes A, Piacentini R. Improvement of the efficiency of a bare solar collector by means of turbulent promoters.
Applied Energy 1990;36:253–61.
176 R. Karwa et al. / Energy 26 (2001) 161–176

[9] Gupta D. Investigations on fluid flow and heat transfer in solar air heaters with roughened absorbers. Ph.D. Thesis,
University of Roorkee, Roorkee (India), 1993.
[10] Saini RP. Study of enhancement of energy collection rates of solar air heaters using artificial roughness in the
air ducts. Ph.D. Thesis, University of Roorkee, Roorkee (India), 1996.
[11] Duffie JA, Beckman WA. Solar energy thermal processes. New York: Wiley and Sons, 1980. p. 211.
[12] Klein SA. Calculation of flat-plate collector loss coefficients. Solar Energy 1975;17:79–80.
[13] McAdams WC. Heat transmission. 3rd ed. New York: McGraw Hill Book Co, 1954. p. 249.
[14] Niles PW, Carnegie EJ, Pohl JG, Cherne JM. Design and performance of an air collector for industrial crop
dehydration. Solar Energy 1978;20:19–23.
[15] Lenel UR, Mudd RR. A review of materials for solar heating systems for domestic hot water. Solar Energy
1984;32(1):109–20.
[16] Edlin F. Optical properties of materials used in solar systems. In: Kreider JF, Kreith F, editors. Solar energy
handbook. Chapter 5. New York: McGraw Hill Book Co, 1981.
[17] Woodman TP. The effect of design and operating parameters on the performance of flat plate solar collectors —
calculation method and detailed appraisal. Solar Energy 1977;19:263–70.
[18] Philips WF. A simplified nonlinear model for solar collectors. Solar Energy 1982;29(1):77–82.
[19] ASHRAE Standard 93-77. Methods of testing to determine the thermal performance of solar collectors. New
York, 1977.
[20] Han JC, Park JS. Developing heat transfer in rectangular channels with rib-turbulators. Int J Heat Mass Transfer
1988;31:183–95.
[21] Liou TM, Hwang JJ. Turbulent heat transfer augmentation and friction in periodic fully developed channel flow.
Journal of Heat Transfer 1992;114:56–63.
[22] Bhatti MS, Shah RK. Turbulent and transition flow convective heat transfer. In: Kakac S, Shah RK, Aung W,
editors. Handbook of single-phase convective heat transfer. New York: John Wiley and Sons Inc, 1987. Chapter 4.
[23] Kline SJ, McClintock FA. Describing uncertainties in single sample experiments. Mechanical Engineering
1953;75:3–8.