Journal of Food Engineering 74 (2006) 247–252

www.elsevier.com/locate/jfoodeng

Performance of a solar dryer with limited sun tracking capability

Gikuru Mwithiga *, Stephen Njoroge Kigo
Biomechanical and Environmental Engineering Department, Jomo Kenyatta University of Agriculture and Technology, P.O. Box 62000, Nairobi, Kenya

Received 3 December 2004; accepted 7 March 2005

Available online 26 April 2005

Abstract

A small solar dryer with limited sun tracking capabilities was designed and tested. The dryer had a mild steel absorber plate and a
polyvinyl chloride (pvc) transparent cover and could be adjusted to track the sun in increments of 15
. The performance was tested
by adjusting the angle the dryer made with the horizontal either once, three, five or nine times a day when either loaded with coffee
beans or under no load conditions. The temperature distribution in the plenum and also the drying rate of parchment coffee were
determined. The temperature inside the plenum chamber could reach a maximum of 70.4
C and the dryer could lower the moisture
content of coffee beans from 54.8% to below 13% (w.b.) in 2 days as opposed to the 5–7 days required in sun drying. Tracking the
sun though allowing a faster rate of drying did not offer a significant advantage in terms of length of drying duration.
 2005 Elsevier Ltd. All rights reserved.

Keywords: Solar dryer; Tracking; Coffee; Temperature distribution; Drying

1. Introduction placed in an enclosure, with a transparent cover on it.

The principle of solar crop drying dates back to the
advent of civilization. The methods then used were sim-
ple and often crude but were reasonably effective. The
two oil crises of the early and late 1970s were also an
important impetus that highlighted the need to have
solar dryers to save on energy (Karekezi & Ranja,
1997). Use of solar energy today is much more desirable
because most of the other alternative sources of energy
have adverse effects on the environment and are in most
cases more expensive. Consumers of solar dried prod-
ucts such as fruits and vegetables are also willing to
pay more for solar dried products (Agribusiness Devel-
opment Centre, 2001).
Solar dryers can broadly be categorised into direct,
indirect and specialised solar dryers (Kreider & Kreith,
1981). Direct solar dryers have the material to be dried
*
Corresponding author. Tel.: +25 406752029; fax: +25 415152164.
E-mail address: gikurum@yahoo.com (G. Mwithiga).
Heat is generated by absorption of solar radiation on
the product itself as well as on the internal surfaces of
the drying chamber. In indirect solar dryers, solar radi-
ation is not directly incident on the material to be dried.
Air is heated in a solar collector and then ducted to the
drying chamber to dry the product (Rabl, 1985). Specia-
lised dryers are normally designed with a specific prod-
uct in mind and may include hybrid systems where
other forms of energy are also used (Whitfield, 2000;
Mumba, 1995; Foster & Mackenzie, 1980). Although
indirect dryers are less compact when compared to
direct solar dryers, they are generally more efficient.
Hybrid solar systems allow for faster rate of drying by
using other sources of heat energy to supplement solar
heat.
The energy collected by a flat plate solar collector
depends on the solar isolation at the site as well as the
positioning of the solar collector with respect to the
sun. Maximum energy can be collected only if the sun-
rays strike collector perpendicularly at all time through-
out the day. This of course can be very expensive since

0260-8774/$ - see front matter  2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jfoodeng.2005.03.018
248 G. Mwithiga, S.N. Kigo / Journal of Food Engineering 74 (2006) 247–252
it requires a continuous tracking system. An alternative
to continuous tracking is to periodically adjust the
position of the collector so that sunlight rays are as
close as possible to the perpendicular position. Where
manual tracking is necessary due to the high cost of
tracking devices, as is the case in many developing
countries, one has to balance between maximising net
energy absorbed and reducing labour costs (Chua &
Chou, 2003).
Many researchers have used the exponential drying
model in describing the drying behaviour of agricultural
materials (Afzal & Abe, 2000; Doymaz, 2004a, 2004b;
Inci & Pehlivan, 2004; Simal, Femenia, Garau, & Ros-
sello, 2005; Velić, Planinić, Tomas, & Bili, 2004). This
model is presented in form of Eq. (1).
MR ¼ expð
ktÞ
MR ¼ ½M ðtÞ
M e =½M ð0Þ
M e 
where MR = moisture ratio, decimal; M(t) = instanta-
neous moisture content, % d.b.; M(0) = initial mois-
ture content, % d.b.; Me = equilibrium moisture content,
% d.b.; k = drying rate constant, per hour; t = time, h.
For estimating the thin layer drying of a material this
model is easy to use and fairly accurate. However, the
temperature of the drying medium is one of the factors
that affects the drying rate constant of this model. This
fact introduces a problem in its use in prediction of dry-
ing rate under solar drying conditions since the temper-
ature of the drying medium under solar drying
conditions is rarely constant. Nevertheless a few
researchers have successfully used it among other mod-
els under solar drying conditions with reasonable suc-
cess (Basunia & Abe, 2001; Bennamoun & Belhamri,
2003; Inci & Pehlivan, 2000).
The objective of this study was to study the effect of
four sun tracking alternatives on the temperature in
the drying chamber of a box-type solar dryer and to
determine the drying rates of parchment coffee.
Fig. 1. The solar dryer used in drying experiments; (a) Pictorial view, (b) cross-section view.
2. Materials and methods
2.1. Equipment
The dryer used in this study is shown in Fig. 1. It con-
sisted of a gauge 20 mild steel flat absorber plate formed
into a topless box 1.0 m long · 0.8 m wide and 30 cm
high. The inner surface was painted black for maximum
absorption of solar radiation while the outer surface was
painted white. The underside of the collector plate was
insulated with a 15 mm thick block board while the side-
walls were un-insulated. The space enclosed by the walls
of the box (hereafter referred to as the drying chamber)
was divided into four vertical compartments of equal
size along the length of the box using wire mesh. The
drying rack was also made of wire mesh and was 5 cm
ð1Þ above the base absorber plate. The top of the drying
chamber was covered with a transparent polyvinyl chlo-
ð2Þ ride sheet of thickness of 20 lm. The transparent sheet
was fixed onto a wooden frame that allowed it to snugly
fit into the top of the drying chamber and yet be easily
removable during loading and sampling. Cold air
entered the drying chamber through 3 mm holes drilled
at the base of the wall of the box and escaped from the
enclosure through a 5 mm wide slot cut at the top of the
wall opposite the entry point as can be seen in Fig. 1(b).
Air flow was entirely by natural convection.
The entire drying unit was then bolted onto a shaft
which was in-turn mounted on to a stand as shown in
Fig. 1(a) so as to face east–west direction. A selector disc
on the stand allowed the tilt angle that the drying unit
made with the horizontal to be easily adjusted in incre-
ments of at least 15
. This way the collector plate could
be intermittently adjusted in-order to track the sun dur-
ing the day.
2.2. Experimental design
Solar dryer testing was carried out in the months of
October and November 2003 at the Jomo Kenyatta Uni-
 G. Mwithiga, S.N. Kigo / Journal of Food Engineering 74 (2006) 247–252
versity of Agriculture and Technology (Kenya) located
approximately 1
south of the equator. The maximum
and minimum dry bulb temperatures recorded at the
weather station during the testing period were 28–
30
C and 12.5–15
C respectively. The average relative
humidity was 61.5%.
There were four dryer settings for tracking the sun
hereafter referred to as T1, T3, T5 and T9. Data collec-
tion was made with the dryer empty in which case only
the temperature of the drying chamber was recorded
and also when the dryer was loaded with coffee in which
case the moisture content was monitored. Therefore 8
runs (4 dryer settings · 2 loading conditions) were made.
The four sets of dryer adjustments were as follows:
T1: The dryer was set horizontal at 8.00 am and
remained in that position throughout which
amounted to a single tracking position per day.
T3: The dryer was set at an angle of 45
to the hori-
zontal and facing east at 8.00 am, adjusted to the hor-
izontal at 10.00 am and then set at 45
to the
horizontal facing west at 2.00 pm. This was equiva-
lent to three tracking positions per day.
T5: The dryer was set at an angle of 60
to the hori-
zontal facing east at 8.00 am and then adjusted to an
angle of 30
to the horizontal at 10 am followed by an
adjustment to the horizontal position at 11 am. In the
afternoon the dryer was adjusted to 30
and 60
to
the horizontal facing west at 1.00 pm and 3.00 pm
respectively. Therefore, the total number of tracking
positions per day was five.
T9: The dryer was set at an angle of 60
to the hori-
zontal facing east at 8.00 am and thereafter the angle
was reduced by 15
every 45 min to reach the hori-
zontal position at 11.00 am. The dryer remained at
the horizontal position until 1.00 pm when the angle
it made with the horizontal was systematically
increased by 15
facing west in intervals of 45 min
in order to reach an inclination of 60
at 4.00 pm.
This resulted in a total of nine tracking positions
per day.
2.3. No load condition
Dryer performance was determined under no loading
conditions by measuring the temperature distribution in
the dryer at designated positions. In each of the four
compartments of the drying chamber, three thermocou-
ples were positioned in equal spacing across the com-
partment and at a height of 5 cm above the absorber
plate and used to monitor temperatures. The output sig-
nal for each of the 12 thermocouples was sent to a data
logger (Thermodac, model no. 5001A, Eto Denki com-
pany, Japan) and recorded in intervals of 1 h from
8.00 am in the morning to 4.00 pm in the afternoon.
249
2.4. Coffee drying
Parchment coffee was acquired from a local factory
soon after initial stages of pulping and fermentation.
These are normal processing steps in the wet method of
coffee processing and are normally followed by drying.
The coffee was then put into 1 kg plastic packages and
stored in a refrigerator at 5 ± 2
C. Before a drying
experiment, the coffee was removed from storage and
allowed to reach room temperature overnight. Four sam-
ples of approximately 100 g were then placed in wire-
mesh trays and inserted into the dryer at 8.00 am in the
morning while a fifth sample was reserved for the deter-
mination of the initial moisture content. The coffee in the
wire-mesh trays was spread evenly in thin layers so that
resistance to the passage of air through the dryer was
minimal. Also the area occupied by the trays was less
than 10% of the total area and thus allowed enough solar
energy to reach the absorber plate. The mass of the
sample was determined every 1 h in the first day of drying
until 4.00 pm. The mass was measured using a digital
electric balance, measuring to an accuracy of 0.001 g.
In the second and third day of drying the mass of the
sample was determined at 8.00 am, 12.00 and 4.00 pm.
The ISO routine method as described by Illy and Viani
(1995) was used in moisture content determination.
3. Results and discussion
3.1. Temperature distribution
Fig. 2 shows the change in the average temperature
developed in each of the drying chambers with time
when the dryer was set at horizontal position (T1). It
can be seen that the temperature developed in the middle
chambers was higher than that developed in the other
chambers and this was the case throughout the day.
The lower temperature in both chamber 1 and 4 could
be due the larger un-insulated surface area of side-walls
leading to higher conduction heat losses. The tempera-
ture also increased with change in time to reach a max-
imum in each chamber at around noon. The lower
temperatures near the edge of the absorber plate can
be attributed to the higher transverse heat losses. The
maximum temperature of 70.4
C recorded in chamber
2 is within the range of 37–72
C reported by Whitfield
(2000).
Fig. 3 shows the change in the ambient air tempera-
ture with time. It also shows the change in the drying
chamber temperature (under no load conditions) with
time for sun tracking conditions of T1, T3, T5 and T9.
It can be seen that the ambient temperature remained
relatively constant throughout the day only rising by
2
C. In each case the chamber temperature increased
as the time of the day elapsed to reach a maximum at
250 G. Mwithiga, S.N. Kigo / Journal of Food Engineering 74 (2006) 247–252

ber of tracking adjustments also developed the highest

temperature at all times.
The differences between drying chamber and ambient
air temperature were also computed using Eq. (3). These
temperature differences (DT) are a measure of the drying
potential of the air in the drying chamber since they af-
80
fects its ability to pick up moisture and to flow in and
70
out of the dryer. The temperature difference was ob-
60
served to increase sharply in all cases from a moderate

4 pm
50 20.8–27.0
C at 8.00 am to 31.0–40.9
C at mid-day.
Temperature, oC

3 pm
40 Thereafter the temperature difference continued to rise

2 pm
30 albeit slowly and reached a peak of 42.2
C at 3 pm be-

1 pm
20 fore dropping marginally. The DTs observed in this
10 1 2 noon study are higher than the average value of 20.4
C re-
1 1 am

0 ported by Schavelis (2000) while drying wood in an

10 am

automated solar lumber dryer.

Chamber 4

9 am

Time DT ¼ T c
T a ð3Þ
Chamber 3

8 am
Chamber 2

where, Tc = drying chamber temperature,
C and,

Chamber 1

Ta = ambient temperature,
C.

Chamber The temperature difference (drying potential) can also
be presented as a function of the number of tracking
positions of the dryer as seen in Fig. 4. It can be seen
Fig. 2. Change in drying chamber air temperature with time as that the drying chamber and ambient air temperature
measured at the centre of each chamber. difference (DT) is highest in the afternoon and that it in-
creases with the number of tracking positions. The
80
change in (DT) has a linear relationship with the number
of tracking positions with coefficient of determination
(R2) ranging from 0.9757 to 0.9982. This linear relation-
70 ship is presented in form of Eq. (4). However, it is
expected that DT will eventually tend to a maximum
60 as the number of tracking adjustment tend to infinity.
DT ¼ A þ BðnÞ ð4Þ
Temperature oC

50

40 45

40
Temperature difference

30

35
20
Ambient T9
T5 T3 30
T1
10
25

0
20
on
am

am

am

am

pm

pm

pm

pm
no
8

10

11

4
12

8.00 am 9.00 am 10.00 am

15 11.00 am 12.00 noon 1.00 pm
Time 2.00 pm 3.00 pm 4.00 pm

Fig. 3. Ambient and drying chamber air temperature as a function of 10

time for different sun tracking conditions. 1 3 5 7 9
No of tracking positions
12 noon and thereafter decreased steadily in the after- Fig. 4. Variation of temperature difference with tracking conditions
noon. The dryer setting T9 which had the highest num- and time.
 G. Mwithiga, S.N. Kigo / Journal of Food Engineering 74 (2006) 247–252 251

where A and B are regression constants and, n = number 1

of tracking positions. 0.9
0.8
3.2. Coffee drying 0.7

Moisture ratio
0.6
The moisture content of coffee beans as a function of 0.5
time is presented in Fig. 5. It can be observed that the
0.4
initial moisture content was approximately 121% d.b.
0.3 T9
(54.8% w.b.) and that in the first 8 h of drying the mois- T5
0.2
ture content fell to between 36.7% and 70.2% d.b. with T3
0.1 T1
dryer setting T9 having the largest reduction in moisture
content. The moisture content of the grain continued to 0
0 2 4 6 8 10
fall even during the night but at a reduced rate. From
Time, hours
Fig. 5 it can be seen that the rate of drying again picked
up on the second day of drying and reduced the mois- Fig. 6. The moisture ratio of coffee beans as a function time and
ture content to levels ranging from 14.6% d.b. (12.7% tracking condition for the first day of drying.
w.b.) for T9 to 22.3% d.b. for T1 by 4.00 pm of the sec-
ond day of drying. The decrease in moisture content on
Table 1
the second day was lower than the decrease observed on The drying rate constants and coefficient of determination for thin
the first day. This is because coffee dries within the fall- layer solar drying of coffee
ing rate period of drying and the rate of water removal Tracking conditions Drying rate constant, Coefficient of
decreases with decrease in moisture content during this per hour determination
period. Thus the acceptable equilibrium moisture con- T9 0.1651 0.9957
tent (emc) for the storage of coffee beans which is 13% T5 0.1326 0.9912
w.b. (Lucia & Assennato, 1994) was reached before T3 0.1033 0.9920
the end of the second day in the case of T9. However T1 0.0711 0.9901
the emc for coffee storage was also achieved in the sec-
ond night of drying for the dryer settings T5 and T3.
Also, the moisture content for tracking condition T1 moisture content for storage. Therefore, the normal
at 8.00 am on the third day of drying was 13.7% w.b., sun-drying period of 5–7 days was reduced to about
which was only 0.7% higher than the recommended 2 days.
It was not possible to fit the drying curves for the en-
tire drying period using the exponential drying model as
described in Eq. (1). However, for first 8 h of drying a
140
reasonably good agreement was found as can be seen
in Fig. 6. In fitting these curves an equilibrium moisture
120 content of 9.5% d.b., which was the lowest moisture
T1
content achieved after 3 days of drying, was assumed.
Moisture content, % d.b.

T9
100 T5
The drying rate constants and the coefficients of deter-
T3
mination for the first day of drying and four tracking
conditions are presented in Table 1. Clearly the rate of
80
drying is dependent on the number of tracking adjust-
ments since the drying rate constant increases with the
60 number of tracking positions. Also, the drying rate con-
stants presented in Table 1 fall within the range of those
40 complied by Krokida, Foundoukidis, and Maroulis
(2004).
20

0
0 5 10 15 20 25 30 35 40 45 50 55 60
Time elapsed, hours
Fig. 5. Change in moisture content of coffee beans with time when
drying coffee in the solar dryer over a three day duration (0 = 8.00 am
of day 1; 22 = 6.00 am of day 2; 46 = 6.00 am of day 3).
4. Conclusion
The study shows that the solar dryer can be used to
successfully dry grains. Drying of parchment coffee
could be reduced to 2–3 days as opposed to sun drying
which takes 5–7 days. Although tracking the sun has
the advantage of higher drying rates the difference in
252 G. Mwithiga, S.N. Kigo / Journal of Food Engineering 74 (2006) 247–252
the time taken to dry the coffee might not justify the
added requirement in terms of labour and construction
cost. Sun tracking during the drying of high value high
moisture products might be justified but this awaits fur-
ther research.
References
Afzal, T. M., & Abe, T. (2000). Simulation of moisture changes in
barley during far infrared drying. Computers and Electronics in
Agriculture, 26(2), 137–145.
Agribusiness Development Centre (2001). ADC commercialisation
bulletin #11: Dried Fruit. Kampala, Uganda.
Basunia, M. A., & Abe, T. (2001). Thin layer solar drying character-
istics of rough rice under natural convection. Journal of Food
Engineering, 47(4), 295–301.
Bennamoun, L., & Belhamri, A. (2003). Design and simulation of a
solar dryer for agricultural products. Journal of Food Engineering,
59, 259–266.
Chua, K. J., & Chou, S. K. (2003). Low cost drying methods for
developing countries. Trends in Food Science and Technology,
14(12), 519–528.
Doymaz, I. (2004a). Effect of dipping treatment on air drying of plums.
Journal of Food Engineering, 64(4), 465–470.
Doymaz, I. (2004b). Pre-treatment effect on sun drying of mulberry
fruit (morus albal). Journal of Food Engineering, 65(2), 205–209.
Foster, H. & Mackenzie, B. A. (1980). Solar heat for grain drying;
Selection, performance, management. Perdue University, Cooper-
ative Extension Service, West Lafayette, IN 47907.
Illy, A., & Viani, R. (1995). Expresso coffee: The chemistry of quality.
San Diego, CA: Academic Press Limited.
Inci, T. T., & Pehlivan, D. (2000). Mathematical modelling of solar
drying of apricots in thin layers. Journal of Food Engineering, 55(3),
209–216.
Inci, T. T., & Pehlivan, D. (2004). Modelling the thin layer drying
kinetics of some fruits under open air sun drying process. Journal of
Food Engineering, 65(3), 413–425.
Karekezi, S., & Ranja, T. (1997). Renewable energy technologies in
Africa. Cumbria, UK: Longhouse Publishing Services.
Kreider, S. F., & Kreith, F. (1981). Solar energy handbook. New York,
USA: McGrawHill Book Co.
Krokida, M. K., Foundoukidis, E., & Maroulis, Z. (2004). Drying
constants: Literature data compilation for foodstuffs. Journal of
Food Engineering, 61(3), 321–330.
Lucia, M. de & Assennato, D. (1994). Agricultural services bulletin no.
93; Post-harvest operations and management of food grains. FAO,
Rome.
Mumba, J. (1995). Development of a photovoltaic powered forced
circulation grain dryer for use in the tropics. Renewable Energy,
6(7), 855–862.
Rabl, A. (1985). Active solar collectors and their applications. New
York, USA: Oxford University Press.
Schavelis, M. (2000). Application of a G.A.B. mathematical model for
predicting E.M.C. in an automated solar lumber dryer. In:
Proceedings of the second workshop on quality drying of hardwood,
Sopron, Hungary.
Simal, S., Femenia, A., Garau, M. C., & Rossello, C. (2005). Use of
exponential, pages and diffusional models to simulate the drying
kinetics of kiwi fruit. Journal of Food Engineering, 66(3), 323–328.
Velić, D., Planinić, M., Tomas, S., & Bili, M. (2004). Influence of
airflow velocity on kinetics of convection apple drying. Journal of
Food Engineering., 64(1), 97–102.
Whitfield, D. E. (2000). Solar dryer systems and the Internet;
Important resources to improve food preparation. International
conference on solar cooking. Kimberly, South Africa.