Energy Vol 20, No 10, pp 1041-1047, 1995

Pergamon

0360-5442(95)00053-4

Copyrtght 1995 Elsevter Science Ltd

Printed m Great Britain All rights reserved
0360-5442/95 S9 50 + 0 00

THE EFFECT OF COLLECTOR ASPECT RATIO ON THE

COLLECTOR EFFICIENCY OF FLAT-PLATE SOLAR AIR
HEATERS
HO-MING Y E H t and TONG-TSHIEN LIN
Department of Chemical Englneenng, Tamkang Umverslty, Tamsul, Talwan, Repubhc of Chma

(Recewed 13 December 1994)

Abstract--The effect of collector aspect ratio of the collector efficiency of fiat-plate solar air
heaters has been mvesngated theoretically and experimentally. With constant collector area, the
collector efficiency increases when the collector aspect ratio increases. The theoretical predictions
agree reasonably well with experimental results

INTRODUCTION

The main applications of a solar air heater are space heating and drying. The solar air heater occupies
an important place among solar heating systems because of mimmal use of materials. Furthermore,
direct use of air as the working substance reduces the number of components required in the system.
The primary disadvantage of solar air heaters is the need for handling relatively large volumes of air
with low thermal capacity as working fluid.
Our design of a solar air heater has an extended heat-transfer area, j an arrangement for producing
free convection, z,3 creation of air turbulence past the heating surface, 4 and inclusion of important forced
convection. It is well known that the collector configuration will influence the fluid velocity as well as
the strength of forced convection. A simple procedure for changing the fluid velocity and also the
strength of forced convection involves adjusting the aspect ratio of a rectangular flat-plate collector
with constant flow rate. It is our purpose in the present studies to investigate theoretically and experimentally the effect of aspect ratio of a rectangular flate-plate collector on the collector efficiency for
constant collector area and constant flow rate.
THEORY

Collector efficiencies

The conveniem representation of efficiency for sheet and tube solar energy collector
= Q J A I o = FR[ap~'g - UL(Tr., - T a ) l l o ] ,

is 5-7

(l)

in which the heat-removal factor is defined as

FR = ( i n C p / A c U c ) { 1 - exp [ - (ULF'AJFnCp)]} .

(2)

Equation (1) was derived from energy balances under the following assumptions: the temperatures of
the absorbing plate and bulk fluid are mainly functions of the flow direction and the radiant energies
absorbed by the glass cover and fluid are negligible.
Although there are many different designs of fiat-plate collectors available, it is fortunately not necessary to develop a new analysis for each situation. 6-s The generalized relations, Eqs. (1) and (2), which
are developed for the tube and sheet case, apply to most collector designs. It is only necessary to derive
the appropriate form of the collector efficiency factor F', and Eqs. (1) and (2) may then be used to
predict the thermal performance.
tTo whom all correspondence should be addressed.
1041

1042

Ho-MingYeh and Tong-TshienLin

Accordingly, the collector-efficiency factor for the flat-plate solar energy collectors shown in Fig. 1
with h = hi = h2 is9
F' = { 1 + ULI{h + [(l/h) + ( llhr)] -I } }-~,

(3)

in which the collector overall loss coefficient UL is the sum of the top (Ut) and bottom and edge (Ub)
loss coefficients, i.e.
UL = Ut + Ub.

(4)

Heat-transfer coefficients

The resistance to energy loss through the bottom and edges of the collector is mainly the resistance
to heat flow through the insulation by conduction, i.e.

Ub=ks/~.

(5)

An empirical equation for Ut was developed by Klein l following the basic procedure of Hottel and
Woertz) I For the horizontal collector shown in Fig. 1,
(To.m/520
Ut =

1 t -1
] o 43o _ lOO/Tp.m+ ~---~w

(rp.m/Ta)

(6)

1 + (1 + 0.089hw - 0.1166hw%)(1 + 0.07866)J

o'(Tp,m+ Ta)(7~p,m+ T~a)
+ (%+0.00591hw)- 1+[2+(1+0.089hw-0.1166hwep)( 1+0.07866)-1+0.133%]/%-1 "
The heat-transfer coefficients between the air and two duct walls were assumed to be equal in Eq. (3).
In the study of solar air heaters and collector-storage walls, it is necessary to know the forced convection
heat-transfer coefficient between two flat plates. For air, the following correlation 12 may be derived
from Kay's data for fully developed turbulent flow with one side heated and the other side insulated:
Nu = h D d k = 0.0158 Re 8

(7)

The characteristic length is the equivalent diameter of the duct

(8)

De = 4 HB/2 (B + H) .;

the Reynolds number for the rectangular duct is then defined by

!,

v,
I

/
Glasscover
Absorbingplate

:
Airflow
C--h Ik ' h l

Bottonplate
~" Insulauon

z
Us

Fig. 1. Flat-platesolar air heater.

Collectoraspect ratio effect on the collectorefficiency

Re

Devp/l~

[4 HBI(2B + 2H)][in/(pBH)]p
=

1043

(9)

2th

-/x(B + H)"
The radiation coefficient between the two air-duct surfaces may be estimated by assuming a mean
radiant temperature equal to the mean fluid temperature, i.e.
hr ~ 4trT~f.m/[(l/%)+ l / e R ) - 1].

(10)

The convective heat-transfer coefficient hw for air flowing over the outside surface of the glass cover
depends primarily on the wind velocity V. McAdams t3 obtained the experimental result
hw = 5.7 + 3.8 V,

(11)

where the units of hw and V are W/m z C and m/sec, respectively.

Mean temperatures
The temperature distribution of air along the flow direction of the rectangular duct can be obtained
from energy balances a s TM

Ty(Z) - Ta - IorgCtp/UL
We,, - Ta - I o T g O l p / U L

[ AcF' UL(Z/L)]
=

exp -

~aCp

j .

(12)

The mean fluid temperature is then found by integrating Eq. (12) from z = 0 to z = L, viz.,

Tf.m=(1/L) f 2 T f ( z ) d z .

(13)

Performing this integration and substituting r I and FR from Eqs. (1) and (2), respectively, the mean
fluid temperature is obtained as

Tf.m = Tf.~+ ( ~llo/ ULFR )[1 -- ( FR/ F' ) ] .

(14)

This is also the proper temperature for evaluation of the fluid properties.
The mean plate temperature will always be greater than the mean fluid temperature due to the heattransfer resistance between the absorbing surface and the fluid. The mean plate temperature may be
used to calculate the collector efficiency, i.e.

= TgCtp-- UL(Tp:n - Ta)/lo.

(15)

If we equate the relations for ~ given in Eqs. (15) and (1) and solve for the mean plate temperature,
we find
Tp, m =

Tf., + (*llolULFR)(1 -- FR).

(16)

Calculation method for collector efficiencies

The procedure for calculation of theoretical values of a9 will now be described. First, with known
collector geometries (L, B, H) and system properties (rs, a, Cp, p, /x, k, ks, Is, %, eR, eg), as well as
the given operating conditions (Io, T~, V, m, Te.,), a temporary value of rl is estimated from Eqs. ( 1 ) (11) once Tp.m and Tf.m are assumed. The values of Tf.m and Tp,m are then checked by using Eqs. (14)
and (16), respectively, and new values of Tf.m and Tp.m may be obtained. If the calculated values of
Tf, m and Tp.m are different from the assumed values, continued calculations by iteration is needed until

1044

Ho-Mmg Yeh and Tong-Tshien Lin

the last assumed values meet the finally calculated values, and thus the corresponding value of 7 / i s
also finally obtained.
EXPERIMENTAL STUDIES

Apparatus and procedure

Four solar air heaters with the same collector surface area Ac o f 0.42135 m 2 but with different aspect
ratios (L/B) were built. The L/B ratios are 6:1 (1.59 m x 0.265 m), 3:2 (0.795 m x 0.53 m), 2:3
(0.53 m x 0.795 m), and 1:6 (0.265 m x 1.59 m). In order to maintain steady climatic conditions, experiments were carried out with an artificial simulator, as shown in Fig. 2. One set of heat sources consists
of 84 electrical energy supplies (110V, 125W). During operations, the strengths o f heat sources were
adjusted by using a set of o n - o f f switches. The insolations Io were measured and recorded with an Epply
Laboratory pyranometer. Wind was provided by a fan and the wind velocity V was set at 1.0 m/sec (as
indicated by a Chicago Cenco anemometer). The temperatures were measured at both the outside glass
cover and the absorbing surface at several points with a No. 709 probe from the Yellow Springs Instrument Co., while the ambient temperature Ta was controlled by an air conditioner and was measured at
a position 15 cm above the outside glass cover with a No. 705 probe. In addition, eight mercury thermometers were employed to measure the air temperatures within the heater, as well as at the inlet
and outlet.
Except for the glass cover, all parts of the solar air heater were well insulated thermally to make
the energy loss as small as possible. For each heater, an air box with a preheater was used for air
distribution and temperature control at the inlet. Air to be heated was supplied steadily by a blower
(Redmond Co. Model BL Model 552) and the flow rate was controlled by a transformer, while the
flow rate was measured with a No. 3132 Taylor anemometer from Fisher Scientific Co.

Experimental results
At the end of each experimental run, the air temperatures in the interior, as well as at the inlet and
outlet of the collector, the ambient temperature, and the mass-flow rate were measured. It was found
that the temperatures o f the absorbing plate and bulk fluid are mainly functions of the flow direction.
The experimental values of the collector efficiency were then calculated from
rl = mCp/AJo)(Tf,

o -

(17)

Tf,,) .

The results are plotted in Figs. 3 and 4.

\ \ ~ \ \ \\ \
\
\ \
N x x \
x
\\\
\\\\
\
N\
\ \ \ N \ \\
\\\\
", k \ \
\
\\\XB\\\

x,

"C',.

\ \

,A

F~g. 2. Schematic dmgram of a solar air heater with amficlal s~mulation. A = a~r heater, B = artificial sun;
C = blower; D = air mixer; E = differential manometer, F = air conditioner, G = temperature indicator; H = fan;
I =atr box; J thermocouples; K = transformer
=

Collector aspect ratm effect on the collector efficiency

1045

0.8
EX~
O
+
O

THEO.

L~

.....

~1
3/2
2/3

0.6

I=" 0.4

0.2

O'-

1o

15

20

25

o.o

x 103 (kglsec)
Fig 3 Effect of collector aspect ratio on collector efficiency, 1o = 830 W / m 2

O.S

EXE
o
40

THEO.

L~
~1
3/2
2/3

.......

0.6

O
.. 0.4

4-

0.2

0.o
5

10

15

20

25

x 10 3 ( k s / s e e )
Fig 4. Effect of collector aspect raUo on the collector efficiency, Io = 1100 W / m 2

Theoretical prediction

The method for theoretical prediction of the collector efficiencies was described in the last paragraph
of the section on theory. The following and Table 1 are the experimental conditions and physical
properties employed in this study: Ac = BL = 0.42132 m2; L = 1.59, 0.795, 0.53, and 0.265 m; B = 0.265,
0.53, 0.795, and 1.59 m; L/B =6/1, 3/2, 2/3, and 1/6; H = 5.5 cm; zg=0.875; a p = 0.95; eR =0.94;
eg = 0.94; Ub = ks~Is ~ 0; Io = 830 and 1100W/m2; Ta = 30C; Tf.~= (35 -+ 0.1) C; V = 1.0 m/sec;
th = 0.0107, 0.0161 and 0.0214 kg/sec.
By substituting the specified values into the appropriate equations, theoretical predictions were
obtained. The results are also plotted in Figs. 3 and 4 for comparison.

Ho-Ming Yeh and Tong-Tshien Lin

1046

Table 1. Physical properties of air at 1 atm.

T
(C)

p
(kg/ms)

Cp
(kJ/kg.K)

p~x l0s
(kg/m.sec)

10.0
37.8
65.6
93.3

1 246
I. 137
1.043
0.964

1.0048
1.0048
1.0090
1.0090

1.78
1.90
2.03
2 15

DISCUSSION AND CONCLUSIONS

The effect of collector aspect ratio on the collector efficiency of flat-plate solar air heaters has been
investigated both theoretically and experimentally. It is seen in Figs. 3 and 4 that the theoretical predictions agree reasonably well with the experimental results. The theoretical values are generally lower
than the experimental data, except for the highest aspect ratio (6:1). Perhaps the theoretical predictions
are greater than the experimental results for higher aspect ratios because of the assumption we made
that the loss coefficient Ub from the surfaces of edges and the bottom of the solar collector to the
ambient is negligible. This assumption is obviously unreasonable when the collector aspect ratio is
large since the surface area of the edges from which part of the energy is lost becomes large and this
energy loss becomes significant although the surfaces are somewhat insulated.
We also found from Figs. 3 and 4 that the constant collector area Ac, the collector efficiency increases
when the collector aspect ratio increases. This result is obtained because increasing the aspect ratio
decreases the cross-sectional area of the air duct and thus increases the velocity of air flow and also
the convective heat-transfer rate from the surface of the absorbing plate to the flowing air. However,
increasing the aspect ratio also increases the fan power and thereby leads to increased operating cost.
Consequently, a proper increase of the collector aspect ratio should be economically feasible in the
design of a solar air heater.
REFERENCES
1. M. K. Seluck, Solar Air Heaters and Their Applications, pp. 155-182, A. A. M. Sayigh ed., Academic Press,
New York, NY (1977).
2. F. Kreith and J. F. Kreider, Principles of Solar Engineering, Chap. 4, pp. 203-309, McGraw-Hill, New York,
NY (1978).
3. H. M. Yeh and Y. C. Ting, Appl. Energy 22, 145 (1985).
4. H. M. Yeh and Y. C. Ting, Energy-The International Journal 13, 543 (1988).
5. A. Whillier, Solar Energy Collection and Its Utilization for House Heating, ScD Thesis, MIT (1953).
6. H. C. Hottel and A. Whillier, Transactions of the Conference on the Use of Solar Energy 2, Part I, 74,
University of Arizona Press (1958).
7. A. Whillier, Applications of Solar Energy for Heating and Cooling of Buildings, ASHRAE, New York, NY
(1977).
8. R. W. Bhss, Sol. Energy 3, 55 (1959).
9. J. A. Duffle and W. A. Beckman, Solar Engineering of Thermal Processes, 3rd edn., pp. 238, 245, Wiley,
New York, NY (1980).
10. S. A. Klein, Sol. Energy 17, 79 (1979).
11. H. C. Hottel and B. B. Woetz, Trans. Am. Soc. Mech. Engnrs 64, 91 (1942).
12 W. M. Kays, Convective Heat and Mass Transfer, McGraw-Hill, New York, NY (1966).
13. W. H. McAdams, Heat Transmission, 3rd edn., McGraw-Hill, New York, NY (1954).
14. J. A. Duffle and W. A. Beckman, Solar Engineering of Thermal Processes, 3rd edn., p. 223, Wiley, New
York, NY (1980).
NOMENCLATURE
A~ = Surface area of the collector, LB

(m2)

B = Collector width (m)

Cp = Specific heat of air at constant
pressure (kJ kg-I K-I)

De = Equivalent diameter of the conduit

(m)
F' = Efficiency factor of the solar air
heater

Collector aspect ratio effect on the collector efficiency

FR = Heat-removal factor for the solar
air heater
H = Height of the air tunnel in the solar
collector (m)
h, h I, h E = Convecuve heat-transfer coefficient
for fluid flowing over a flat plate
(10 h -I m-2)
hr = Radiant heat-transfer coefficient
between two parallel plates
(10 h -I m -2 K)
hw = Convective heat-transfer coefficient
for mr ftowmg over the outside
surface of glass cover (10 h -~ m -2)
lo = Incident solar radiation (10 m -2 h -~)
k, k, = Thermal conducuwty of air,
insulator (10 h -~ m -2 K)
L = Collector length (m)
l, = Thickness of the insulator (m)
rh = Mass-flow rate of mr (kg h -~)
Qu = Useful gain of energy carded away
by air per umt time (kg h -I)
Re = Reynolds number
Ta = Ambient temperature (K)
Tf(z) = F l m d temperature (K)
Tf,, Tf,o = Tf at the inlet and outlet of the
solar air heater (F)

1047

Tf.m = Average value of Tf(K)

Tp, TR = Temperature of the absorbing plate,
bottom plate (K)
Tp.m = Average value of Tp (K)
Ub = Loss coefficient from the surfaces
of edges and the bottom of the
solar collector to the ambient
(10 h-' m 2 K)
UL = Overall loss coefficient
(ld h-I m e K)
U, = Loss coefficient from the top of the
solar collector to the ambient (10
h -~ m -2 K)
v = Air velocity in the tunnel (m h-~)
V= Wind velocity (m h -~)
z = Axial coordinate along the flow
direction (m)
tp = Absorptivity of the absorbing plate
Collector efficiency
% = Transmittance of the glass cover
Eg, ep, ER = Emissivity of the glass cover,
absorbing plate, bottom plate
p = Air density (kg m -3)
/~ = Air viscosity (kg m -~ h -l)
ty = Stefan-Boltzmann constant,
2.04 x 10-7 10 h - I m -2 K -4