DESALINATION

ELSEVIER Desalination 168 (2004) 101-109

www.elsevier.com/locate/desal

Drying of agricultural crops by solar energy

Saber Chemkhi a*, Fethi Zagrouba a, Ahmed Bellagi b
"INRST, LEPT, BP 95, 2050 Hammam Lif, Tunisia
Tel. +216 (97) 332997; Fax +216 (671) 430934; emaiI : saberchemkhi@yahoo.fr
bENIM, Avenue Ibn Eljazzar, 5000 Monastir, Tunisia

Received 13 February 2004; accepted 20 February 2004

Abstract
The main application for solar energy in southern Mediterranean countries in agriculture is the drying of agricultural
crops. The optimisation of dryers necessitates complete knowledge of the whole drying process, thus leading to energy
savings and avoiding environmental pollution by using renewable sources of energy. We present a study concerning
thermal behaviour of a solar air heater as a source of energy for drying agricultural products. In order to simulate the
functioning of our solar collector, we chose a simple model based on the evaluation of different heat transfer
coefficients in the collector and within the external environment, which allows determining the overall collector heat
loss coefficient. Our purpose was to calculate the fluid outlet temperature, the output energy used and thermal
efficiency as a function of ambient temperature, incident solar radiation, wind speed and air mass flow rate. The present
suggested model's results demonstrated that the air heater has paid offsince we can attain the temperature on the order
of 80°C at the fluid outlet and 60% in thermal performance.

Keywords: Solar energy; Climatic factors; Flat-plate collector; Heat loss coefficient; Thermal efficiency; Drying
agricultural crops

1. Introduction a solar plan collector, a drying box and a

In this study we present an analysis o f a
drying convective pilot using solar energy. This
solar dryer is essentially composed o f three parts:
*Corresponding author.
chimney. We are interested in analysing the dif-
ferent components o f this dryer and its energy
balance. The solar dryer is designed to meet cer-
tain demands: essentially agricultural use such as
for harvesting, a modest dried production and
simple maintenance.

Presented at the EuroMed 2004 conference on Desalination Strategies in South Mediterranean Countries: Cooperation
between Mediterranean Countries of Europe and the Southern Rim of the Mediterranean. Sponsored by the European
Desalination Society and Office National de l'Eau Potable, Marrakech, Morocco, 30 May-2 June, 2004.
0011-9164/04/$- See front matter © 2004 Elsevier B.V. All rights reserved
doi; 10.1016/j.desal.2004.06.174
102 S. Chemkhi et al. / Desalination 168 (2004) 101-109
In this paper we present a detailed, but simple,
study concerning the thermal behaviour of a solar
air heater as a source of energy for drying agri-
cultural products. This analysis is based on
experimental measurements carried out at the
Solar Applications Laboratory in the National
Institute of Scientific and Technology Research
(Tunisia).
2. Choice and conception of the dryer
In our conception of the solar collector, its
length is chosen so that we get a maximal output
temperature of air. If we study the distribution of
the temperature on its length, we get an evolution
that converges toward a constant maximal value.
The particular shape of the absorber is not
arbitrary: we tried to maximise the surface of
exchange in order to increase the thermal
exchanges by convection. In the case of a fiat
absorber, we would have a surface of 2.5 m2; for
our case the surface reached 8.6 m 2. We can also
benefit from the shape of the collector which
seems to be an important mass flow since the
other shapes (too groove and wavy) benefit the
half of the maximal mass flow seen only that the
section of air passage will be reduced. The
insulator is a layer of glass wool characterised by
its weak conductivity (Z= 0.045 Wm-LK-~). The
inner part is covered with a silvered surface that
has an important role in thermal loss reduction
towards the rear while reflecting the part of the
radiance transmitted by the absorber. The cavity
of stagnant air between cover and absorber serves
as a trap of the radiance transmitted through the
cover. It is the role of a greenhouse that stores the
heat and stocks it until saturation of the thermal
inertia of the stagnant air to nourish the absorber
and that also nourishes the fluid in heat [1-3].
3. Heat balance equations
3.1. Hypotheses
• The regime is permanent.
• The lateral losses to the level of the collector
are negligible.
• The soil temperature is equal to the ambient
temperature: Tso~l= To.
• We disregard the conduction through the
cover and through the absorber.
• The temperature on a horizontal slice of the
absorber is constant.
3.2. The overall collector heat loss transfer
coefficient
We attempted to evaluate the energy balance
of the collector. At first, we defined the overall
collector heat loss coefficient (Ut) that leads us
first to calculate the thermal losses and then to
evaluate the well stocked useful energy [2,3].
We define the following fluxes with regard to
an element of surface of the collector:
• exchanged flux between the transparent cover
and the ambient middle:
,o,dS (hcv_o+ (r:o) (1)
• exchanged flux between the transparent cover
and the absorber:
dS (hc._.+ (2)
• exchanged flux between the absorber and the
intern face of the insulator:
(P3=dS(hCp_isl+ hrp_isl)(Tp-Tisl) (3)
• exchanged flux through the insulator by
conduction:
q~4 = dShed(Tisl-Tu2) (4)
• exchanged flux between the insulator and the
ambient middle:
q~5 = dS(hcu2 + hru2_~)(T,;2-T~) (5)
 S. Chemkhi et al. / Desalination 168 (2004) 101-109 103

In an established and permanent regime, the The balance of energy is obtained by writing that
conservation of the fluxes gives: the energy absorbed by the collector during the
time dt is distributed in the following way:
%=qo 2 and t03=~04=~05 (6)
Qabs dt = Qudt + Qpdt + de e (10)
The linearity of the thermal losses of the
collector requires the hypothesis of the uniformity where d e c is the growth of the energy stored in
of the temperature of the collector. the collector.
The thermal efficiency of the collector is
Qp =dSUt(Tp-Ta) (7) defined as

Ut is the overall collector heat loss coefficient:

q = +%% G =qo-UtT,. (11)
1 1 where 1]0 is the optic efficiency and T, is the
Ut- +
1 + 1 1 + 1 + 1 (8) reduced temperature [2,3 ].
hv-a hp-v hp-isl hcd his2-a
3.4. Node method
Hereunder in the balance of energy exposition we
suppose Tp and TI as uniform for an element of We can approach this problem respecting the
surface dS. hypotheses while applying the electric analogy
method. This approach establishes Fig. 1, called
the node method [4-7], relative to a vertical
3.3. Useful energy recovered by the collector
section of the collector.
G is the incidental solar flux in m 3of the cover This method is based on the electric analogy
surface. The fraction absorbed by the collector is of the collector. A node placed in its centre repre-
G~pO~v. The well stocked useful energy to the sents every element of the collector. For every
fluid (air) is: node, the energy balance is written:

ou = mcp(rs-re)- Gds-o. (9) dT

CP,-~ : Pi + d S E ~is+ Fij +Ro.)(Tj-Ti) (12)
i

Cover
hrP" v
I Absorber
hop4 hrp-is
Air flow =-
~:~ hcis.f

Fiiiiiii!iiiiiiii!iiiiiiiii!iiiiiii]liiiiiiiiiiiiiiiiiiiiiiiii!iii!it
Insulator
Fig. 1. The heat transferprocessin a vertical
hcis-a hcd section of the solar collector.
104 S. Chemkhi et aL / Desalination 168 (2004) 101-109

where K U, C o and R o. are, respectively, the 3.5. Numerical resolution

coefficient of coupling conductive, convective
We have a system of five equations with five
and radiance between the i and t h e j nodes. P,. is
unknowns: T~, Tp, T:, T,;l and T~sz.It is put under
the heat source term. As we work in a permanent
the matrix shape A *X=B. We have a linear sys-
regime, the first term of our equation will be
tem to solve, but its constants depend themselves
negligible.
on the variables; we then use an iterative method.
For our collector we have five nodes: the
To have precise results, the collector is
cover, absorber, fluid, the internal face of the
divided into N fine horizontal slices in order to
insulator and its external face. We make the heat
solve the system on every slice while considering
balance of every node only while considering an
that the exit of the ith slice is the entrance for the
element of the surface dS of the collector:
following slice. A program of resolution on
• For the cover:
MATLAB has been developed for this problem.

+p:,)ds o(r:ro)
d&-hrv_o(rv-r )dS, (13) 4. Experimentation
4.1. Experimental device
+ (hcp_ + hrp_f p,,)(Tp-Tv)dSp = O
The Laboratory of Solar Applications of the
• For the absorber: INRST was equipped with a testing apparatus
permitting the characterisation of solar collectors.
This testing apparatus processes several relative
(hc._. + )
measuring instruments for climatic data. The
(r - dS. - hc._:(r.- r:) (14) bench is oriented towards the south to have the
benefit of the maximum of the incident solar
dSw-hrp-isl G-isl (Tp- Tis I )d~ w = 0 radiation; the collector is fixed to a slant of 36 °,
knowing that it can vary from 20 ° to 70 ° with
regard to the horizontal. Measurementss are
• For the fluid:
raised by an acquisition chain permitting the
storage of these values. The collector is tested
+ hc.,_:(V,.,-r:)dS,.
under two flow regimes: the free and forced flow
(15)
: (r:r.) regime. In the case of forced flow, a fan is used to
circulate air with an adjustable speed. We take the
climatic data (ambient temperature, solar radia-
• For the internal face o f the insulator:
tion and wind speed), the distribution of the
temperature on the collector (under the cover, on
(r: r,. )ds - hC,.l_¢(r,., - r:)dS,,
the internal face of the insulator, on its external
=0 (16) face and along the absorber) and the two input
and output air temperatures.
• and for the external face of the insulator:
4.2. Test of the collector in free out-flow
h~ddSi~ (Tm - T~,2) + (hris2-sot +h%2-a )
(17) We made several measurements for different
dSis o days in order to be able to get an idea on the
 S. Chemkhi et al. / Desalination 168 (2004) 101-109 105

{Ta=24.3 °C; V,~=O.419m/s; m=7.3410"~kg/s} {Ta=20°C; V~d=O.41m/s; m=7 10.3 kg/s}

--,-Te(*C)
100 ~ -..-'rs(*c) 1000 1001 _. 11000

_ oo ; ' - ' - ........ OOooo ,-

v

40- 400 ~"

20- 200
O/ ,,, t , i I I I I t i t i tO
10 11 12 13 14 15
I I I I I I L 0
Olo 11 '12 'la q.' is '16 Time (hours)
Time (hours)
Fig. 2. Solar radiation effect on the outlet temperature for Fig. 3. Solar radiation effect on the outlet temperature for
a sunny day. a cloudy day.

{Ta=29.8 °C; V,,,,a=1.032m/s; m=7.2 I/s}

100. .-~-Te(~C) 1000
occurs with a certain representative delay due to
-+-Ts(OC) he own inertia o f the collector. While comparing
80 8OO the Figs. 2 and 4, we notice the influence of the
G 60 60o wind factor on performance of the collector. If the
F- wind speed increases, the thermal losses toward
40. 400 (.9
the outside increase. By observation of Fig. 4, we
20" 200 notice a fluctuation o f T, that depends on the
speed of wind. To put in evidence the effect of
0- : : : : : : : : : : : : : : : : : : : : : : - 0
10.5 11.5 12.5 13.5 14.5 15,5 the ambient temperature, we chose two days, a
Time (hours) cool and a hot day (Fig. 2 and Fig. 4). An eleva-
tion of the ambient temperature does not improve
Fig. 4. Solar radiation effect on the outlet temperature for
a hot and windy day. the efficiency of the collector: indeed for a cool
day Ta = 24.34°C, the overheating o f air reaches
47.3°C; for another hot day Ta = 29.8°C, we got
behaviour of our collector, its outlet temperature an overheating of 38.9°C.
and its efficiency for different climatic con-
ditions. For the first regime o f out-flow, we
4.3. Test in forced out-flow
studied different climatic agent effects on the
efficiency of our collector. In the same way, we worked in forced out-
For two typical days and by a simple com- flow to put in evidence the effect of the mass
parison between Figs. 2 and 3, we notice that flow of the circulating air in the heater. For a
solar flux G has an essential role and a direct climatic factor effect such as in free out-flow, we
influence on the outlet air temperature. We tried note the existence of a direct relationship between
to value the answer o f the collector in terms of the outlet temperature of the collector and the
temperature according to the incidental solar climatic factors, especially solar radiation. Fig. 5
radiation for two different climatic conditions: T, illustrates this dependence for a typical day. For
follows the evolution o f G; there is a certain con- the climatic conditions nearly constant, from
cordance between the evolutions of G and Ts. The Fig. 6 we can deduce a relationship between Ts
answer of the collector to the solar flux variations and the mass flow in the collector. While
106 S. Chemkhi et al. / Desalination 168 (2004) 101-109

Ta=22.1 *C; V~d=O.482m/s; m=46.3610~ Kg/s} {Ta=24.2 °C; G=950W/m2; V~,~=O.322m/s}

100.
100 ~ lO00 -*-Tg(°C)
80, "*-Ts(*C)
80 "~-Ts(*C) 800 -c-Thermal

60
•-~_~,(Wm-3
6oo
~" 'oo
40 400
~ 40
0
20 . . . . . - " ' - ~ ~ ~ .... ~ 200 ~_ 20
f i g

10.5 11 11.5 12 12.5 13 13.5 14 14.5 oo io 2o go ,io so

Time (hours) Air mass flow (10 .3 kg/s)

Fig. 5. Solar radiation effect on the outlet temperature in Fig. 6. Air mass flow effect on the collector outlet
forced out-flow regime. temperature and its efficiency.

Ta=24.2 °C; G=950W/m2; V~nn=0,322m/s} collector and fluid; it increases the gap of temp-
100
erature between them, thus reducing the thermal
efficiency of the collector.
80

-"- 60
0o...,
20 -B.m3=46.05 10-3 kg/s
-o-m4=0
0
0 28 56 B4 112 14o 168 196 224
X (cm)
Fig. 7. Distributionof the temperature along the absorber
for different air mass flows.
increasing the mass flow, the outlet temperature
decreases and by following, the overheating
decreases. Observing Fig. 7 we notice, on the one
hand, that as the air mass flow increases, the
temperature on the collector decreases. On the
other hand, the air benefits more from the energy
stocked in the first half of the collector; that is,
the heat exchange is more important in this part.
The experience of the closed collector (we closed
the two extremities of the collector) gives the
temperature of balance on the collector; the circu-
lation of air pulls the energy out of the collector.
The weak thermal conductivity of air contributes
to the resistance of the heat transfer between
5. Results and discussion
The model was tested on the outside climatic
conditions to which the collector is submitted.
The results allowed the evaluation of perfor-
mances and limits of the model conceived to
study different thermal exchange methods inside
and outside of the collector. The model calculates
the distribution of the temperature, the different
coefficients of thermal exchange, useful output
energy as well as the efficiency. Compared to the
experimental results, there was satisfactory agree-
ment with all of the parameters. The difference
between the experimental results and those of the
model are due to the lack of precise information
on losses through the insulation, the cover and the
lttteral surfaces. Figs. 8-11 show the permanent
variation of the climatic parameters. This varia-
tion has some direct effects on the results. On one
hand, we notice that results given by the model
are especially steady and more meaningful, but
the experimental results present remarkable fluc-
tuations. For different air mass flows, from the
previous model we can calculate the outlet temp-
erature, the overall collector heat loss coefficient,
 S. Chemkhi et al. / Desalination 168 (2004) 101-109 107

trra=29,8 °G; Vwind=1.032m/s;m=7.210"~kgls} {Ta=29.8 °C; Vwind=1.032m/s; m=7.2103 kg/s}

100

S0
1000

800
°1
30 t
/ "~" experimental efficiency "~
tlOOO

t soo
1
~. | ~ simulatedefficiency ~. ~L600
G 6O 600
0
v
i-.- ~iXmP~ur:7~lntal "~
CO
40 -:- G (Win~) _~ 400

20 200
ot . . . . . . . . . . . . ,100
] I I I I 10.5 11.5 12.5 13.5 14.5 15.5
0.5 ' 1'1,5 12.5 1~3.5 14.5 1~5.5 Time (hours)
Time (hours)

Fig. 8. Comparison between measured and estimated Fig. 9. Comparison between measured and estimated
temperature in free outflow regime. efficiency in free outflow regime.

Ta=22,1 °C; Vv~nd=0,482m/s;m=46.3610"~kg/s} {Ta=22.1 °C; Vwind=0.482m/s;m=46.3610"3kg/s

100 1000 t000

80 ~,-~-Ts experimental "~ 800 800

Tssimulated ~ g...
ft.,, 1600
60 600
p-
40 400 4OO
l --,-- efficiency
200 20 1 "=" simulated efficiency 2OO
20
or"
10.5 11,5
. .
12.5
.
13.5
. .
145
. . .
15.5
0
010.5 ' 11.5
' ' 12.5
' ' 1'3.5' 14.5
' ' 1'5.5'
Time (hours)
Time (hours)

Fig. 10. Comparison between measured and estimated Fig. 11. Comparison between measured and simulated
temperature in forced outflow regime. efficiency in forced outflow regime.

the useful stocked energy by the collector and its Table 1

efficiency. Table 1 presents some results at dif- Thermal efficiency as a function of the mass flow
ferent mass flow of air for the following constant variation
climatic conditions (T, = 24.2°C; Vwi,d,= 0.322
Air mass flow (10 3 kg/s)
m/s; G = 950 W/m2). We notice that the overall
collector heat loss coefficient is nearly constant 8.3 18.3 37.77 46.36
on the order of 3.15 Wm-ZK-L It is a charac-
teristic of the collector for constant climatic T,, °C 72.7 69.4 57.1 54
conditions (solar irradiation, wind speed and Ut, Wm-2K-I 3.16 3.15 3.11 3.11
ambient temperature). Q,. w 405.3 832.1 1248.8 1525.8
On the other hand, we noticed that the rt, % 15.2 31.2 46,9 57.3
developed model suffers from many weaknesses,
especially to the level o f the simplifying hypo- permanent, it is not the case because we have a
theses. We have supposed that the regime is source o f energy, "solar radiation", that is not
108 S. Chemkhi et al. /Desalination 168 (2004) 101-109

steady but rather sometimes has sudden varia- hcd Conductive heat transfer coefficient,
tions: (cloudy sky influences the regime). Also W/m2.°K
we did not hold any of the inertia of the collector hr Radiative heat transfer coefficient,
since we only worked after its stake in regime. W/m2.°K
We disregarded the lateral losses with com- m -- Mass flow rate, kg/s
parison to the front and back losses in the Q, Thermal losses, W
collector, which is not quite justifiable since the O,, -- Stocked useful energy, W
lateral surface is meaningful (approximately 10% S Surface, m E
of the total losses). The nonuniformity of the T -- Temperature, °K
temperature of the absorber (there is a gradient of Ut -- Global heat loss coefficient, W/m 2.
temperature between the plate and the extremity oK
of the small wing) is disrupted as a result and V -- Air flow speed, rn/s
somewhat reduces the assessment of the convec-
tive exchange coefficient in the collector. Greek
Therefore, this work is too simple, but complete,
to model and comprehend the thermal behaviour -- Absorption
of the solar collector submitted to real climatic -- Emittance
-- Thermal efficiency, %
conditions.
q~ -- Thermal flux, W
P Density of air, kg/m 3
T, Transmittance of cover
6. Conclusions

The use of solar collectors for the heating of lndices

air is very profitable since some important outlet
temperatures are obtained at the exit (on the order 0 Initial
of 80°C in free outflow and 60°C in forced out-
a -- Ambient
flow). We perceived that the efficiency of the c -- Sky
collector is weak in free outflow (brings in 10% e -- Input
and 15%); to benefit more from the available f Fluid "air"
energy, we can increase the air mass flow which isl Internal face of the insulator
is2 -- External face of the insulator
increases the efficiency in a remarkable way (on
the order of 50-60%). P -- Absorber
This study will be useful in a later stage to S -- Output
examine the thermal behaviour of the whole dryer soil Soil
and to develop solar drying of agricultural crops. V Cover
W -- Wing

7. Symbols
Acknowledgements
@ -- Specific heat of air, J/kg.°K
F Shape factor of the grey body This study was carried out at the Solar Appli-
G Global solar radiance on the plane cation Laboratory at the National Institute of
of the collector, W/m 2 Science Research and Technology, Tunisia. The
hc Convective heat transfer coefficient, authors would like to thank Dr. Chakib Kerkeni
W/m2.°K for his material and scientific help.
 S. Chemkhi et al. / Desalination 168 (2004) 101-109
References
[1] S. Chemkhi, F. Zagrouba, C. Kerkeni, D. Mihoubi
and A. Ballagi, A simulation model for analysing the
thermal behaviour of a solar convective dryer, 15th
International Congress of Chemical and Process
Engineering CHISA, Praha, Czech Republic, 2002.
[2] H.C. Hottel and A. Whillier, Evaluation of flat-plate
solar collector performance. Trans. Conf. on Solar
Energy, Vol. 2, Part 1, Section A, Tucson, Arizona,
1955, p. 74.
[3] J.A. Duffle and W.A. Beckman, Solar Engineering
of Thermal Processes, Wiley Interseience, New
York, 1980, p. 237.
109
[4] R. Clarke, A. Achong and R. Saunders, Theoretical
thermal performance predictions of a solar air heater
with the absorber having natural roughness. World
Renewable Energy Congress VI, 2000.
[5] D. Protoctor, A generalised method for testing all
classes of solar c o l l e c t o r - I: Attainable accuracy.
Solar Energy, 32(3) (1984) 377-386.
[6] D, Protoctor, A generalised method for testing all
classes of solar collector - - II: Evaluation of col-
lector thermal constants. Solar Energy, 32(3) (1984)
387-394.
[7] D. Protoctor, A generalised method for testing all
classes of solar collector-- III: Linearized efficiency
equations. Solar Energy, 32(3) (1984) 395-399.