Renewable Energy 145 (2020) 1494e1506

Contents lists available at ScienceDirect

Renewable Energy
journal homepage: www.elsevier.com/locate/renene

Modeling and comparative analysis of solar drying behavior of

potatoes
Ahmed Djebli a, b, Salah Hanini a, Ouassila Badaoui a, b, *, Brahim Haddad b,
Amina Benhamou c
a
Biomaterials and Transport Phenomena Laboratory (LBMPT), Universit
eYahia Fares Medea, street Aïnd'Heb, 26000, Algeria
b
Unit

e de D
eveloppement des Equipements Solaires, UDES/Centre de D

eveloppement des Energies Renouvelables, CDER, Bou Ismail, 42415, W. Tipaza,
Algeria
c
Materials Research Unit, Process and Environment (URMPE), University of M'Hamed Bougara, Boumerdes, 35000, Algeria

a r t i c l e i n f o a b s t r a c t

Article history: Potatoes are a very high demand product in the consumer market and are widely available everywhere in
Received 10 January 2019 Algeria. With the extensive levels of solar energy across the country, solar drying becomes a natural
Received in revised form avenue for the production of solar dried potatoes as a new prospective commercial product. This paper is
3 June 2019
a study of the drying process of potatoes using an indirect and a mixed solar dryer with forced con-
Accepted 15 July 2019
vection and a comparison of their respective performances. Experimental results indicated that the
Available online 16 July 2019
mixed solar dryer had the slowest drying rate compared to the indirect solar dryer with drying times of
4 h 45min and 3 h 40min respectively, even though the temperature at the tray level in the mixed solar
Keywords:
Drying models
dryer was higher than that of the indirect solar dryer. Fourier series and Laplace transforms were used to
Indirect solar dryer solve the diffusivity equation. Seven theoretical models were investigated and their relative accuracy
Mixed solar dryer validated against experimental data. Two new dimensionless models were found to best fit the drying
Potatoes curve of potato slices using the two dryers. The proposed models were found to be more accurate for
charactering the drying process of Sultana grapes, chili peppers, bananas, mangoes and cassava, ac-
cording to models reported in the literature under different operating conditions using various solar
dryer designs.
© 2019 Elsevier Ltd. All rights reserved.

1. Introduction process can reduce product storage costs. Moreover the drying
process preserves the flavor and nutritive benefits of the product
Potatoes, an herbaceous plant, are a product of great importance [6].
economically and ranked as the fourth food crop worldwide after Drying is a technique that can be applied by several ways, the
maize, wheat, and rice. In Algeria, potatoes are a strategic product most common is the open sun drying. However, this technique has
whose production increased remarkably over the last three de- some drawbacks, namely; the product to be dried is exposed to dust
cades. In 2014, the production reached more than 4 million tones and insects as it is spread on the floor or on a mat and directly
covering an area of 156,176 ha. exposed to the sun and wind [7]. In addition, the dried product may
Drying is one of the oldest techniques that permits food pres- be subject to contamination and of poor quality [8]. A different
ervation, such as grains [1], fruits [2], vegetables [3] and seafood [4]. approach that could alleviate the drawbacks of open sun drying is
It prolongs the lifespan of the product by reducing its moisture that drying can be carried out using fuel and electricity, but these
content to stop bacterial activity, eliminates microorganisms that drying methods may not be cost efficient and detrimental to the
cause rot or fermentation, stops mold proliferation and decreases environment. On the other hand, the use of tailor designed solar
product weight to facilitate transport and storage [5]. The drying drying equipment can be a good alternative and the most viable
solution from the economic and environmental perspectives [9]. In
principle, solar drying consists of converting solar energy to ther-
* Corresponding author. Biomaterials and Transport Phenomena Laboratory
mal energy which is then used to remove moisture from the
(LBMPT), Universite
Yahia Fares Medea, street Aïnd'Heb, 26000, Algeria. product directly [10] or indirectly [11,12]. The most important factor
E-mail address: [email protected] (O. Badaoui).

https://doi.org/10.1016/j.renene.2019.07.083
0960-1481/© 2019 Elsevier Ltd. All rights reserved.
 A. Djebli et al. / Renewable Energy 145 (2020) 1494e1506 1495

Nomenctalture r2 correlation coefficient

RMSE root Mean Squared Error
A, A1, A2, k, k1, k2, n, m, p models parameters t drying time (h)
Deff effective diffusivity coefficient (m2/s) T temperature ( C)
Dmeff average effective diffusivity coefficient (m2/s) Tmax maximum temperature ( C)
DM dry Matter V drying rate
e thickness of the potato slices WM wet Matter
ISD indirect solar dryer M moisture content (kg water/kg dry matter)
MAE mean Absolute Error M0 initial moisture content (kg water/kg dry matter)
mb mass of the Beaker (kg) Meq equilibrium moisture content (kg water/kg dry
md mass of the dried product in the solar dryer (kg) matter)
mf mass final of the sample (kg) MR moisture ratio (kg water/(kg dry matter.h))
mfb mass final of the sample with the Beaker (kg) Mexp,i experimental moisture ratio
mi mass initial of the sample (kg) Mpre,i predicted moisture ratio
MSD mixed solar dryer c2 reduced chi-square
N number of observations

that has an impact on yield is the daily solar radiation intensity [13]. drying red peppers and sultana grapes was investigated by
The solar drying technique is widely used in the food industry ELkhadraoui at al [24]. Using this system, the drying time of red
[14]. Many examples of solar dryers can be cited, Şevik [15] utilized peppers and grapes was shortened by 7 and 17 h respectively. The
a solar dryer with a double pass solar air collector, heat pump and criterion of payback period was used as an economic evaluation,
photovoltaic unit to dry carrots. Under sunny weather conditions where the payback period was found to be short, that is 1.6 years. A
the system can be easily operated without the need for a heat v-corrugated double pass plate solar air heater with a blower is
pump. But during night time and cloudy weather the dryer can be used to force the heated air to the drying chamber. Blanco-Cano
operated using a heat pump. Also, Reyes et al. [16] dehydrated et al. [25] analyzed experimentally the kinetics of the thin-layer
mushrooms in a hybrid solar dryer. Their results showed that slice drying process of Granny Smith apples. The Drying tests were
thickness and air recycling levels had statistically significant effects performed under different drying temperatures and a novel model
on the product final moisture content. Air recycling percentages was proposed to predict the evolution of the moisture ratio during
help minimize energy supply for heating the drying air. Also, To- the drying process.
matoes were dried by Ringeisen et al. [17] in Davis, California under In Algeria, solar drying use is not widespread, it is not yet used in
variable climatic conditions throughout the summer and fall using the industrial and in the agricultural sectors although the country
a solar dryer with an energy concentrator. Note that the drying time possesses one of the highest potential of solar thermal energy in the
decreased by 21% on average when using a concave concentrator. world. As highlighted by Haddad et al. [26], solar energy can play a
Thermal storage with desiccant makes solar drying more effective key role to ensure social welfare. Algeria receives on average
in terms of drying time and energy consumption as indicated by 2500 KWhm
2 per year of solar energy ranking the country among
Dina et al. [18] experimental results when drying cocoa beans in a the first countries and the one with the highest solar potential in the
solar dryer. As for the case of potatoes, Tesfamichael and Assefa [19] Mediterranean basin with an estimated yearly potential of 168 TWh
reported an analysis of the drying curves and the time requirement [27]. Experimental studies can lead to the integration of solar drying
for drying potato slices in a natural convection solar dryer. The to the Algerian economy and thus bring new social benefits.
results showed that the drying time can be reduced by 30% This paper presents an experimental and theoretical analysis of
compared to drying under direct sunlight. Likewise, Barubara et al. solar drying of potatoes using new mixed and indirect solar dryers
[20] employed a finned forced convection, flat plate solar collected under Algerian coastal climate conditions. The results were used to
to dry potatoes in Indonesia. Different slice thickness (0.01, 0.015 determine the drying characteristics curves and the rate of drying.
and 0.02 m) were tested, their average initial water content was Furthermore, an analysis of the effects of temperature in the two
76.48%.The experimental results showed that the slice thickness of solar dryers and their performances were compared. Experimental
1 cm is best for the drying process. data were fitted using seven models: Five from the literature and
Drying agricultural product using indirect and mixed solar dryer two proposed models from this study to describe the drying ki-
processes has been the subject of many researchers. Forson et al. netics of potatoes. The Two proposed empirical models are
[21] developed a model in parallel with experimental work using a dimensionless and in accordance with the boundary conditions.
mixed-mode natural convection solar crop-dryer. The model These two models were compared to other models in the literature
permitted to predict the performance of the dryer. A new prototype using experimental data from this work and from the literature to
of an indirect active hybrid solareelectrical dryer for agricultural assess their technical viability for different products, operating
products using a high range of air mass flow rates between 0.04 and conditions and solar drying equipments. The aim is to encourage
0.08 kg/m2s was examined by Boughali et al. [22]. Their results the use of solar dryers in the Algerian industrial and agricultural
showed that low airflow rates are preferable for the drying process. sectors by offering full analysis of solar drying technology to apply
El-Sebaii and Shalaby [23] experimentally investigated the thermal to various agricultural products.
performance of an indirect-mode, forced convection solar dryer.
The drying experiments were performed using thyme and mint. It 2. Materials and methods
was found that the drying times of the two products depend on the
mass of the product to be dried and the temperature of the drying This study was carried out using two solar dryers, indirect and
air. Their drying kinetics follow the constant and failing rate phase. mixed. The indirect and mixed dryers were designed and fabricated
The performance of a mixed mode solar greenhouse dryer for at UDES facility as part of a reasearch project using local materiels.
1496 A. Djebli et al. / Renewable Energy 145 (2020) 1494e1506

Door Outlet

Tray
Fan
Duct
PV

Glass

Intlet

Fig. 1. The indirect solar dryer with forced convection.

2.1. Indirect solar dryer polycarbonate walls. Furthermore, additional heating is provided
by a solar collector connected to two fans.
The indirect solar dryer is composed of two main parts; a solar Four fans permit air circulation inside the greenhouse. The hot
collector and a drying chamber as shown in Fig. 1. The dimensional air passes through carts and trays and dries the product. The
characteristics of the equipment are shown in Table 1. Air gets released moisture is evacuated using a fan on the top of the right
heated up in the solar collector then forced into the drying chamber wall as illustrated in the figure below.
through 0.15 m diameter ducts using two electric fans powered by a
photovoltaic panel. The hot air dries the product by discharging the 2.3. Data collection
released product moisture to the outside via the exhaust outlet.
The duration of the drying process depends on the type of
product to be dried. The operation is deemed complete when no
2.2. Mixed solar dryer
more weight loss is detected. A RADWAG Balance is used to weigh
the product, accuracy ±0.0005 kg is used to monitor the drying
Two main parts compose the mixed solar dryer, a greenhouse
process by recording water weight loss in the product as a function
and a solar collector (Figs. 2 and 3). The geometric configuration of time until complete dehydration; namely, the balance indicating
and dimensions of the dryer are summarized in Table 1. The
constant weight.
greenhouse receives solar radiation through the transparent Global radiation received on the collector plane is measured
during the day of the experiment. The hourly or daily tilted radia-
Table 1 tion is measured using Kipp&Zonen pyranometer. The air velocity
Solar dryer's characteristics. was measured using Thermo-anemometer C.A 1224.
Temperature is measured with K-type thermocouples with a
Solar dryer Solar dryers components Number Specification
type diameter of d ¼ 0.001 m and an accuracy of 0.5  C. All the ther-
mocouples were connected to a data logger (Keysight 34972A LXI
ISD Dry chamber volume 1 0.46 m3
Tray surface 1 0.57 m2 Data acquisition/Switch unit) which sends the data to the data
Duct 2 Diameter: acquisition computer.
0.08 m
Length: 0.8 m
2.4. Product preparation
Collector surface 1 1.73 m2
Tmax ¼ 69  C
Fan (flow rate) 2 0.01 m3/s Table 2 gives the product details before drying. The sliced po-
MSD Greenhouse volume 1 30.6m3 tatoes were set on wire trays in the indirect solar dryer and the
Carts dimensions 9 Length:1.1 m
Width: 0.65 m
High: 1.5 m
Trays surface 45 Surface:
0.65 m2
Collector surface 1 1.82 m2
Tmax ¼ 77  C
Equipped with
fins
Fans (flow rate) 2 0.0225 m3s-1
(used for collector)
circulation and moisture extractor Fans 5 0.19 m3s-1
(flow rate)
Fig. 2. Mixed solar dryer from inside and outside.
 A. Djebli et al. / Renewable Energy 145 (2020) 1494e1506 1497

Carts

Outlet

Fans

Polycarponate

Inlet of the
greenhouse

Duct

Air collector

Tray

Fig. 3. Schematic diagram of the new mixed solar dryer.

Table 2 where mf is calculated using the equation below:

Product data.

Location of the drying operation Bou-ismail, Algeria mf ¼ mfb
mb (2)

Weight purchased 2 kg Two types of curves allow describing the evolution of the
Weight of sliced potato used in each tray 0.825 kg product during the drying:
The potato variety Spunta
Price of 1 kg 0.38 Dollar
- The drying rate V

mixed solar dryer respectively. They were spread on the drying tray V ¼ dM/dt (3)
in a way that allows for air passage. In order to monitor the drying
process, the products were carefully weighed during the drying Where the moisture content M is defined as:
operation.
WM
DM
M¼ (4)
2.5. Drying kinetics DM

Because of the wide diversity of foodstuffs, analysis of the drying

kinetics is the best way to describe the behavior of a food product
- The moisture ratio (MR)
during the drying process. In general, the experimental results of a
kinetics analysis are presented as curves that express the evolution M
Meq M
of the rate of the drying process as a function of time. These curves MR ¼ z (5)
M0
Meq M0
are generally obtained for different experimental conditions, such
as temperature, hygrometry, and air velocity. Meq is very small and negligible compared to M and M0.
In the case of potatoes, their moisture content was calculated on Where MR is the moisture ratio; Meq is the equilibrium content;
the basis of their dry weight. To measure dry matter, 20 g samples M0 is the initial moisture content.
of potato slices were taken from the products that were dried
separately in the two solar dryers. Then they were put in beakers 2.6. Theoretical models of drying
and dried in the oven at 105  C for 24 h. After that, they were placed
in a desiccator. The beakers were then weighed and the dry mass Several models were used to depict the kinetics of the drying
(DM) was calculated using equation (2): process and to predict the moisture ratio MR as a function of drying
time. Experimental results of potato drying were used to validate
- The dry matter DM is defined as: the theoretical models with the aim of identifying the one that best
md  mf fits the data. To that end, five models were selected from the
DM ¼ (1) literature for their good agreement with the experimental data. The
mi
Midilli model was found to be adequate for predicting the drying
behavior of tomatoes per Boughali et al. [22] and thyme per
1498 A. Djebli et al. / Renewable Energy 145 (2020) 1494e1506

Lahnine et al. [28]. As for the case of the Two Term model, Yaldiz PN  2
et al. [29] and Lahsasni et al. [30] have concluded that the model is MRexp;i
MRpre;i
the one that fits best the experimental data of the drying process of
c2 ¼ i¼1
(12)
N
z
Sultana grapes and Prickly pears. The Page model fitted remarkably
well with the drying curves of thymus [23], raw mango slices [31],
and chili peppers [32]. The logarithmic model satisfactorily - Mean Absolute Error MAE:
described the drying characteristics of pumpkin slices [33], long
green peppers [34] and apricots [35]. Koua et al. [2] found that 1 XN  
MRexp;i
MRpre;i 
MAE ¼ (13)
banana, mango and cassava drying curves are well fitted with the N i¼1
Henderson and Pabis model.
Unlike the literature models, the new proposed dimensionless
models are universal, independent of units of measurement, and - Root Mean Squared Error RMSE:
verifying the boundary conditions per equations (7) and(8) as vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u
follows: u1 X N  2
RMSE ¼ t MRexp;i
MRpre;i (14)
N i¼1


M∞
M t∞
t
¼f (6) Where MRexp,i and MRpre,i are experimental and predicted
M∞
M0 t∞
t0
moisture ratios; Z is the number of constants in model; N is the
number of observation.
M∞ ¼ lim M (7)
t/t∞
2.7. Determination of effective moisture diffusivity using fourier
series
M0 ¼ lim M (8)
t/t0
Diffusivity is an important transport parameter for the calcula-
After simplifications: tion of water transfer inside the product. This coefficient largely
determines the activity of surface water.
MR ¼ f(t) (9) In general, the diffusion coefficient is estimated from drying
kinetics, by solving Fick diffusion equation, eq (15). It has been
M∞
M
Where: MR ¼ M∞
M0 widely used to describe the drying process for most organic
products [42].
t
t¼ (10) v2 M 1 vM
t∞
¼0 (15)
vx2 Deff vt
Table 3 is a recapitulation of the five mathematical models
selected from the literature and the proposed new dimensionless Fourier series were used to solve Crank's equation, equation
models. (15). Assuming that the initial moisture distribution is uniform, the
So, to select the mathematical model that fits best the experi- external resistances are negligible and the process is isothermal,
mental data of potato drying, seven models were evaluated using the initial and boundary conditions are as follows:
statistical parameters [41].
vM e
The experimental data were fit to the models using Origin Mðt ¼ 0; xÞ ¼ M0 ; ðt; x ¼ 0Þ ¼ 0 and Mðt; x ¼ H Þ ¼ Ms
software. The constants of the models, correlation coefficients and
vx 2
the reduced chi-square were computed by the software. The (16)
models are empirical and their constants have no physical meaning. In addition, for long periods of drying, all terms in the series are
Model constants depend on the distribution of measurement negligible compared to the first therefore the solution for equation
points MR as a function of time. The constraints are the initial (15) can be simplified as follows:
values input into the software to fit the model with the experi- !
mental data. 8 p2 Deff
Time values, t, were used in the five models from literature to MR ¼ exp
t (17)
p 2 e2
predict moisture ratio using Excel software. As for the two new
models, t values were calculated using equation (10), which then Thus, taking the natural logarithm in both sides of equation (8)
were used to predict moisture ratio. Equations (13) and(14) were gives the following relationship.
then used to calculate mean absolute error and root mean squared
error. 8 p2 Deff
lnðMRÞ ¼ ln
t (18)
The model that works best was selected according to the p 2 e2
following criteria:
A plot of ln(Xr) as a function of (t) yields a straight line with a
2 slope b:
- Correlation coefficient r :
p2 Deff be2
PN  2 b¼ thusDeff ¼ ; (19)
MRpre;i
MRpre;i e2 p2
r 2 ¼ P i¼1 2 (11)
N
i¼1 MRexp;i
MRexp;i The higher the value of (Deff) the more effective is the drying
process.
Where (Deff) is the effective diffusivity, (e) is the thickness of the
- The reduced chi-square c2: potato slices, (t) the drying time.
 A. Djebli et al. / Renewable Energy 145 (2020) 1494e1506 1499

Table 3
Mathematical models that describe solar drying kinetics.

Model name Model expression References

New model1 1 þ Atn expð
k1 tn Þ Present work

MR ðtÞ ¼
½1 þ Atm expð
k2 tm p
New model2 1 þ Atn lnð1 þ k1 tn Þ Present work
MR ðtÞ ¼
½1 þ Atm lnð1 þ k2 tm p
Page MR(t) ¼ exp(-ktn) [36]
Logarithmic MR (t) ¼ A1exp(-kt)þA2 [37]
Midilli-Kucuk MR (t) ¼ A1exp(-ktn)þA2t [38]
Henderson and Pabis MR (t) ¼ A1exp(-kt) [39]
Two term MR (t) ¼ A1exp(-k1 t)þA2exp(-k2t) [40]

2.8. Determination of effective moisture diffusivity using laplace 800

transforms

It consists of finding the analytical solution for eq (15) by using

Solar radiation (w/m2)

600
Laplace transforms [43]. Expression of the solution is given by:
sffiffiffiffiffiffiffiffiffiffi8 0 19
> >
Mt Deff t < 1 X∞
n B nl C
=
400
¼2 pffiffiffi þ 2 ð
1Þ ierfc@q ffiffiffiffiffiffiffiffiffiffiA (20)
M∞ l >
2
: p n¼1 D t > ; eff

Xt
Where: MR ¼ 1
X∞
200

sffiffiffiffiffiffiffiffiffiffi8 0 19
> >
Mt Deff t < 1 X∞
n B nl C
=
MR¼1
¼1
2 p ffiffiffi þ 2 ð
1Þ ierfc@q ffiffiffiffiffiffiffiffiffiffiA 0
M∞ l2 > : p n¼1 Deff t > ; 0 4 8 12 16 20 24
(21) Time 00:00 to 24:00h (30/07/2017)

Fig. 4. Example of a recording of solar radiation as a function of time of the day.

where ierfc is the integral of erfc function:

1 
ierfcðzÞ ¼ pffiffiffi exp
z2
ðzÞerfcðzÞ (22) reached at midday, and then decreased for the rest of the day. The
p scatter of the data is due to cloudiness which also the solar radia-
For n ¼ 1: tion to reach its peak value.

sffiffiffiffiffiffiffiffiffiffi sffiffiffiffiffiffiffiffiffiffi 0 1
!
Deff t Deff t l2 B l C 3.2. Temperature measurement
MR ¼ 1
2 þ4 exp

4erfc@qffiffiffiffiffiffiffiffiffiffiA
pl2 pl2 Deff t Deff t
For the indirect solar dryer, four points were chosen for tem-
(23) perature measurements. The first thermocouple was placed in the
duct, outlet of the air solar heat collector, the second on the tray
The complementary error function is defined as: holding the potato slices, the third at 1 m above the tray, and the
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 ffi fourth is for ambient temperature outside the dryer. For the MSD,
erfcðzÞ ¼ 1
erf ðzÞ ¼ 1
1
exp
z2 (24) three thermocouples were used; one, at the entrance of the
greenhouse to measure the temperature of the hot air coming from
which yields to the following equation: the air solar heat collector, the second on the potato tray and the
sffiffiffiffiffiffiffiffiffiffi sffiffiffiffiffiffiffiffiffiffi third is above the tray. The recorded temperatures are reported in
!
Deff t Deff t l2 Figs. 5e7.
MR ¼
3
2 þ4 exp
Figs. 5 and 6 indicate that temperature in the ISD is slightly
pl 2 pl2 Deff t
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi lower than that in the MSD. Fig. 6 illustrates that temperature in the
u ! drying chamber is lower than that in the duct at the outlet of the air
u 4l2
þ 4t1
exp
(25) collector which was around 8  C. The temperature of the air
pDeff t entering the drying chamber in the case of the ISD dryer is higher
than that of the air entering the greenhouse in the MSD due to the
Where: l ¼ e/2.
lengths of the ducts used for each dryer. Heat loss increases with
the length of the ducts. Duct length is 2 m for the MSD compared to
3. Results and interpretations 0.8 m for the indirect solar dryer. As shown in Fig. 5 by the blue line,
the temperature of the air coming out of the collector and going
3.1. Solar radiation into the greenhouse increases gradually as the sun moves from a
low elevation in the morning to its highest point at noon at which
Global horizontal radiation was measured during 24 h. An time the temperature reaches its maximum level. Similar trends
example of solar radiation variation as a function of time is shown apply to the red and green curves which correspond to the tem-
in Fig. 4. In this particular case, the maximum value, 747 W/m2, was peratures at the tray level and above the trays. At noon time, the
1500 A. Djebli et al. / Renewable Energy 145 (2020) 1494e1506

Fig. 5. The temperature evolution as a function of time for different measurement points, MSD.

70

65

60 T(°C) at the tray level

(potatoes slices)
T(°C)

55 T(°C) above the tray

50
T (°C) at the duct
45

40
0 1 2 3 4
Time 09:00 to 12:40h (30/07/2017)

Fig. 6. The temperature evolution as a function of time for different measurement points, ISD.

60

T(°C) of the potatoes

50 slices in the mixed solar
dryer
T(°C)

40 T(°C) of the potatoes

slices in the indirect solar
dryer
30
T(°C) of the ambient air

20
0 1 2 3 4
Time 09:00 to 12:40h (30/07/2017)

Fig. 7. The temperature evolution as a function of time for different measurement points.

temperature measurement above the trays shows higher values drying chamber. The temperature gain is 26.1 and 27.3 for MSD and
than inside the trays because of the higher elevation of the ther- ISD respectively. Despite the fact that the temperature in the ISD at
mocouple location which is at least 1 m above the trays. the tray level was a little bit lower than that in the MSD, the drying
The air collector used in the mixed solar dryer is equipped with rate was slowest for the MSD. This was due to the insufficient air
fins and gives an average temperature of 50  C at the entrance of circulation inside the greenhouse, where the air velocity average is
the greenhouse using the two fans, which is satisfactory for the 0.15 m/s at tray level and 0.45 m/s in the drying chamber as sum-
needs of this work and enhances the thermal efficiency of the dryer. marized in Table 4 and Fig. 7. At the start of the drying process the
The collector was used to ensure low temperature differences be- air velocity acts positively in the drying kinetics. However, for
tween bottom and top trays on the carts, and enhance the thermal products whose drying kinetics is controlled by the internal
efficiency. migration of water, the influence of air velocity becomes insignifi-
The ambient temperature on average was 26.5  C. Average cant. Despite the long drying time of MSD, 23% higher than ISD, it
temperature was 53.8  C in the greenhouse and 52.6  C in the has a superior loading capacity weight 32 times higher than ISD. As
 A. Djebli et al. / Renewable Energy 145 (2020) 1494e1506 1501

Table 4
Comparison between the performance of MSD and ISD.

Indirect solar dryer Mixed solar dryer

Tmin/Tmax ( C) 45.5/55 45.5/57.4

Average Temperature ( C) 52.6 53.8
Hot air velocity at tray level (m/s) 0.45 0.15
Drying time 3 h 40 min 4 h 45 min
Color of dried potatoes slices No change in color The color of about 35% of the total number of slices turn Little bit grayish
Shape of dried potatoes slices shrinked More shrinked
Loading capacity weight 4 kg 130 kg

Moisture ratio Drying rate a drying time difference of more than 1 h. One phase was observed
1 2.5 which was labeled as phase number 2, corresponding to the de-
MR V(dM/dt) creases of both the drying rate and the product moisture content to
0.8 2 the equilibrium limit.
Bonazzi and Bimbenet [44] have argued that the theoretical
0.6 1.5 drying curves, obtained for non hygroscopic and non-deformable
products such as cellulose, wood, clay or plaster, are character-
0.4 1
ized by the presence of three phases. However, the progress of the
drying process is not the same and the three phases are not always
0.2 0.5
observable as it has been shown by experimental studies. A clas-
0 0
sification of products according to their behavior during drying is
0 1 2 3 4 5 presented by Van Brakel [45].
Time (h) The absence of the product warm-up and the constant drying
phases, most likely due to the structure of agricultural products and
Fig. 8. Evolution of moisture and the drying rate of the potato slices as a function of
the fact that cell walls disrupt rapid migration of moisture to the
time, MSD.
outer surface as noted by Boughali et al. [46]. In general it is due
also to the fact that agro alimentary products are not saturated with
Moisture ratio Drying rate water, ie: they do not contain free water [47]. Transfer of moisture
1 2.5 from the inside of these products to the surface occurs via several
MR V(dM/dt) complicated mechanisms such as liquid diffusion, vapor diffusion
0.8 2 or capillarity [48]. the drying kinetics as it is shown by experi-
mental studies with several products, such as maize, potatoes,
0.6 1.5 plums, apples [47], potato slices [49,50], carrots [51], avocado [52],
dates [53] and apples [54].
0.4 1 Table 5 gives a comparison of drying potatoes between the
present studies with other investigations. The two dryers used to
0.2 0.5 dry potatoes in the present study have a better performance than
the mixed-mode solar dryer used by Tripathy and Kumar [55], in
0 0 New Delhi, India when considering drying time. Chouicha et al. [56]
0 1 2 3 4
used the hybrid indirect solar dryer in their study and concluded
Time (h)
that it has the best drying time and temperature but note that the
Fig. 9. Evolution of moisture and the drying rate of potato slices as a function of time, study was conducted in the Sahara desert with dry environment
ISD. while the present study was conducted in Bou-Ismail with humid
coastal climate.

shown in Table 1, the configurations of MSD and ISD are different in

term of volume, tray number, etc. The volume of the MSD is 60 3.3. Drying curves
times larger than that of the ISD. The collector in MSD has slightly
larger surface and is more efficient because fins where used in it. The resulting values of the parameters of the models with mean
Figs. 8 and 9 illustrate the variation of the moisture ratio of the absolute error, root mean square error, correlation coefficient, and
potato slices in the two dryers versus time (MR ¼ f (t)) and drying reduced chi-square for MSD and ISD are summarized in Table 6 for
rate (V ¼ f (t)). For both dryers, moisture ratio decreases with time comparison. The values of the parameters fall within the ranges of
until it reaches the equilibrium moisture content. Product moisture 0.9451e0.99993 for the correlation coefficient r2, 8.79∙10
6 to
ratio decreased faster when using the ISD than the MSD resulting in 7.42∙10
3 for reduced chi-square c2, 1.92∙10
3 to 7.15∙10
2 for

Table 5
Comparison of the present study with other investigations.

Research Product Type of dryer Drying time Tmin-Tmax Location of the dryer

C

[55] Potatoes Mixed-mode solar dryer 5h 30min 30.4e50.4 New Delhi, India
[56] Potatoes Hybrid indirect solar dryer 2h 45 min 41.6e61.1
ria
Ouargla, Alge
Present work Potatoes indirect solar dryer 3h 40 min 45.5e55
ria
Bou-Ismail, Alge
Present work Potatoes Mixed solar dryer 4h 45 min 45.5e57.4
ria
Bou-Ismail, Alge
1502 A. Djebli et al. / Renewable Energy 145 (2020) 1494e1506

Table 6
Results obtained from models for potatoes.

Model Solar dryer Coefficient r2 c2 MAE (%) RMSE (%)

New Model1 Mixed A ¼ 0.0102, n ¼ 0.5603, k1 ¼
7.2166, m ¼ 0.6000, k2 ¼
2.9703, p ¼ 42.022 0.9999 2 E
5 0.232 0.331
Indirect A ¼ 0.0384, n ¼ 0.6638, k1 ¼
5.9941, m ¼ 0.6288, k2 ¼
3.2482, p ¼ 10.368 0.99993 1 E
5 0.205 0.274
New Model2 Mixed A ¼ 49.2082, n ¼ 0.3782, k1 ¼
0.0197, m ¼ 0.9235, k2 ¼ 0.00001, p ¼ 2965.1 0.99972 4 E
5 0.986 1.140
Indirect A ¼ 4773.52, n ¼ 0.405, k1 ¼
0.00019, m ¼ 1.0157, k2 ¼ 8.61E-8, p ¼ 4434.7 0.99985 2 E
5 0.623 0.735
Page Mixed k ¼ 0.4226, n ¼ 1.3373 0.99408 4.3 E
4 1. 766 2.086
Indirect k ¼ 0.4803, n ¼ 1.4592 0.99629 4.3 E
4 1.726 1.958
Logarithmic Mixed A1 ¼ 1.2931, k ¼ 0.3452, A2 ¼
0.2819 0.99765 1.8 E
4 1.400 2.022
Indirect A1 ¼ 1.4605, k ¼ 0.3371, A2 ¼
0.4431 0.99639 4.5 E
4 1.722 1.931
MidillieKucuk Mixed A1 ¼ 0.9866, k ¼ 0.3914, n ¼ 1.1750, 0.99921 6 E
5 1.009 2.037
A2 ¼
0.0226
Indirect A1 ¼ 0.9882, k ¼ 0.4256, n ¼ 1.2405, 0.99901 1.3 E
4 0.885 1.010
A2 ¼
0.0348
Henderson and Pabis Mixed A1 ¼ 1.0691, k ¼ 0.5826 0.98082 1.38 E
3 4.353 5.976
Indirect A1 ¼ 1.0869, k ¼ 0.6882 0.97291 3.17 E
3 4.807 5.291
Two term Mixed A1 ¼ 11.361, k1 ¼ 0.2167, 0.95837 3.4 E
3 7.155 7.633
A2 ¼
10.352, k2 ¼ 0.1957
Indirect A1 ¼ 1, k1 ¼ 1, A2 ¼ 0.1929, k2 ¼ 0.3156 0.94511 7.42 E
3 6.567 7.532

mean absolute error MAE and 2.52∙10
3 to 7.63∙10
2 root mean coefficients r2 of 0.99993 and 0.99985 respectively. The low values
square error RMSE for the MSD and ISD. of the mean absolute error MAE, mean square error MSE and
The experimental data is in good agreement with the two sug- reduced chi-square c2 were obtained using new model2 for the two
gested models, based on the high values of the correlation coeffi- dryers.
cient r2 which is equal to 0.9999 and 0.99972 in the case of MSD. For Fig. 10 illustrates the evolution of the moisture ratio for the MSD
the case of the ISD, new models 1 and 2 gave correlation and ISD, as measured experimentally in comparison to predictions

Experimental (MSD) Experimental (ISD)

New model1 New model1
New model2 New model2
1 1

0.8 0.8

0.6
0.6
MR 0.4 MR
0.4

0.2
0.2

0
0 0.2 0.4 0.6 0.8 1 0
0 0.2 0.4 0.6 0.8 1

(a) (b)
Experimental (MSD) Experimental (ISD)
Midilli-Kucuk model Midilli-Kucuk model
1 1

0.8 0.8

0.6 0.6
MR
MR
0.4 0.4

0.2 0.2

0 0
0 1 2 3 4 5 0 1 2 3 4
Time (h) Time (h)
(c) (d)
Fig. 10. Moisture ratio versus (t, time), comparing experimental curve with the predicted one based on: (a, b) Suggested Models; (c, d) Midilli-Kucuk Model.
 A. Djebli et al. / Renewable Energy 145 (2020) 1494e1506 1503

Midilli-Kucuk model Midilli-Kucuk model

New model1 New model1
New model2 New model2
1 1

0.8 0.8
MR Theory

MR Theory
0.6 0.6

0.4 0.4

0.2 0.2

0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
MR Experimental MR Experimental
(a) MSD (b) ISD
Fig. 11. Variation of moisture ratio of the two models with the experimental moisture ratio.

by means of the new models 1 and 2 and the Midilli-kucuk model. Midilli-Kucuk model deviated slightly from straight line.
The New models showed a good fit with the experimental data of
the MSD and ISD. 3.4. Validation of the new models with experimental data in the
Calculated values of moisture ratio MR using the theoretical litterature
models are plotted against the experimentally measured MR
values, see Fig. 11, to gauge the accuracy of the models. A good The experimental data were used to test the robustness of the
agreement between experimental and predicted moisture ratio suggested models according to various operating conditions (air
values has been noted for the two suggested models. All data from velocity, temperature), different designs and different kinds of
the two suggested models are closer to the straight line indicating products. Table 7 presents various dryer designs and operating
the best fit. In the case of the MSD, the data generated from the conditions from the present work and from the literature.

Table 7
Different dryer designs and operating conditions used in the present work and in the literature.

Dryer design Tm ( C) Air velocity (ms
1) Product Product form

Present work Indirect solar dryer 52.6 0.45 Potatoes slices

Mixed solar dryer 53.8 0.15 Potatoes slices
[29] Indirect solar dryer 60 0.5 Sultana grapes spherical
[32] Solar cabinet dryer 38.9 2 Chilli pepper conical
[2] Direct solar dryer with chimney 52 0 plantain Banana slices
Cassava slices
Mango slices

Experimental data Experimental data

New model1 New model1
New model2 New model2
1 1

0.8 0.8
MR
0.6 0.6
MR
0.4 0.4

0.2 0.2

0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1

(a) Chilli pepper (b) Sultana grapes

Fig. 12. Variation of moisture ratio of the two models with the experimental moisture ratio from litterature.
1504 A. Djebli et al. / Renewable Energy 145 (2020) 1494e1506

1 New model1 A comparison between the suggested models and the best
New model2 models found in literature for Sulatana grapes, bananas, mangoes,
Henderson and Pabis cassava and chilli peppers, using the correlation coefficient,
0.9995 reduced chi-square and root mean square error is presented in
Table 8. The new models have high values of r2 and low values of c2
and RMSE compared to those from the so-called Two term model,
0.999 page & Henderson and Pabis models.
The data confirmed that the two new models are universal and
can characterize perfectly the drying curves of different agricultural
0.9985
products under different operating conditions using any type of
designs.
0.998
3.5. Calculation of effective diffusion coefficient

0.9975
3.5.1. Determination Deff using fourier series
plantain banana cassava mango
The diffusion coefficient was calculated using fick's law. The
Fig. 13. Evaluation of correlation coefficients for the three products used in Koua et al. equation of diffusion was resolved using Fourier series.
[2] study.

3.5.1.1. Calculation with linear fit. Fig. 14 shows the variation of ln

The new suggested models were evaluated using experimental (MR) versus time for the two dryers. The slope of the line, b, for the
data selected from the literature for a variety of food product; MSD and ISD are 2.33 E
4 and 3.01 E
4 respectively. The effective
namely, Sulatana grapes, bananas, mangoes, cassava and chilli diffusion coefficient Deff was calculated using eq (14) and presented
peppers. The selected products have, naturally, different structures in Table 9.
and porosity in comparison to potato slices.
As shown in Fig. 12, the proposed models proved to be in good 3.5.1.2. Calculation using slopes from 2 points. Deff obtained from
agreement with the experimental data. the slope of the line joining every two points and the average Deff
Correlation coefficients for the proposed models are compared were calculated and presented in Table 9. The average value of the
to those obtained from the study of Koua et al. [2] for plantain effective diffusion coefficient obtained by 2 points slope method is
bananas, mangoes and cassava as depicted in Fig. 13. slightly higher than that obtained via the linear fit.

Table 8
Results obtained from suggested models for products from literature.

Reference Product Model Coefficient r2 c2 RMSE (%)

[29] Sultana grapes New Model1 A ¼ 37.308, n ¼ 1.8024, k1 ¼ 5.3469, m ¼ 1.7035, k2 ¼ 1.3649, p ¼ 1.3307 0.999999 1.34E-7 0.02
New Model2 A ¼ 0.2868, n ¼ 6.6147, k1 ¼ 10.136, m ¼ 0.5381, k2 ¼ 0.0009, p ¼ 14024 0.99999 3.23E-6 0.11
Two term A1 ¼ 0.1943, k1 ¼ 6.3513, A2 ¼ 0.8057, k2 ¼ 0.0339 0.95837 3.4 E
3 /
[32] Chilli pepper New Model1 A ¼ 0.4512, n ¼ 1.57E7, k1 ¼ 3.05E6, m ¼ 0.9896, k2 ¼
1.2479, p ¼ 7.0576 0.99953 7 E
5 0.73
New Model2 A ¼ 11.102, n ¼ 0.3765, k1 ¼
0.0870, m ¼ 0.6966, k2 ¼ 0.0005, p ¼ 814.65 0.99953 7 E
5 0.74
Page k ¼ 1.1894, n ¼ 1.3373 0.9991 1.2E-4 1.14
[2] plantain Banana New Model1 A ¼ 0.2712, n ¼ 0.7744, k1 ¼ 5.7142, m ¼ 0.5743, k2 ¼
1.6343, p ¼ 2.5100 0.999999 3.77E-7 0.05
New Model2 A ¼ 5.3709, n ¼ 0.2658, k1 ¼
0.0861, m ¼ 0.6130, k2 ¼ 0.0307, p ¼ 10.145 0.999999 3.92E-7 0.05
Henderson and Pabis A1 ¼ 0.1206, k ¼ 0.9856 0.9997 4.20E-5 0.61
Cassava New Model1 A ¼ 0.3686, n ¼ 2.7026, k1 ¼ 6.6703, m ¼ 1.0176, k2 ¼
0.1655, p ¼ 5.5143 0.99996 4.08E-6 0.16
New Model2 A ¼ 1.3772, n ¼ 0.8813, k1 ¼
0.1026, m ¼ 0.4931, k2 ¼ 0.0104, p ¼ 130.17 0.99995 5.23E-6 0.18
Henderson and Pabis A1 ¼ 0.1250, k ¼ 1.0137 0.9999 1.52E-5 0.37
Mango New Model1 A ¼ 0.47778, n ¼ 1.4934, k1 ¼ 6.5357, m ¼ 0.6653, k2 ¼
1.8269, p ¼ 2.4153 0.99916 8 E
5 0.77
New Model2 A ¼ 30131, n ¼ 0.2684, k1 ¼
0.00003, m ¼ 0.9831, k2 ¼ 0.0002, p ¼ 0.8224 0.99917 7 E
5 0.7
Henderson and Pabis A1 ¼ 0.1119, k ¼ 1.0489 0.9979 4.89E-5 2.14

ln(MR) Linear ln(MR) ln(MR) Linear ln(MR)

0 5000 10000 15000 20000 0 3000 6000 9000 12000 15000
0 0
Time (s) Time (s)
-1 -1

-2 -2
ln(MR)

ln(MR)

-3 -3
y = -2.33E-04x + 4.47E-01 y = -3.01E-04x + 4.82E-01
-4 -4

-5 -5

MSD ISD

Fig. 14. Variation of the ln (MR) over time.

 A. Djebli et al. / Renewable Energy 145 (2020) 1494e1506 1505

Table 9
The effective diffusion coefficient using Fourier series.

Type of dryer The effective diffusion coefficient Deff The average effective diffusion coefficient Dmeff

MSD 5,90E-10 6,07E-10

ISD 7,62E-10 7,82E-10

Table 10 Acknowledgements
The effective diffusion coefficient using Laplace transforms.

Type of dryer The effective diffusion coefficient Deff This work is supported by EPST CDER/UDES e Algeria. We
MSD 3.25E-10
gratefully thank our colleagues Miss A. Chalal and Dr. A. Diaf from
ISD 4.01E-10 EPST CDER/UDES for their support and help.

References
3.5.2. Determination deff using laplace transformation
[1] P. Kumar, S. Shanmugam, A.R. Veerappan, An experimental study on drying of
The equation of diffusion was solved using Laplace transforms.
non-parboiled paddy grains using an oscillating bed solar dryer, Energy Eng.
The experimental data were fit using Origin software with equation 108 (2) (2011) 69e80.
(21) to obtain the effective diffusion coefficient. Table 10 shows the [2] K.B. Koua, W.F. Fassinou, P. Gbaha, S. Toure, Mathematical modelling of the
values of Deff using Laplace transforms. thin layer solar drying of banana, mango and cassava, Energy 34 (2009)
1594e1602.
The effective diffusion coefficient for the ISD was higher than [3] S. Şevik, M. Aktas, H. Dogg ^an, S. Koçak, Mushroom drying with solar assisted
that of the MSD. heat pump system, Energy Convers. Manag. 72 (2013) 171e178.
[4] B.K. Bala, M.R.A. Mondol, Experimental investigation on solar drying of fish
using solar tunnel dryer, Dry. Technol. 19 (2) (2001) 427e436.
[5] V. Belessiotis, E. Delyannis, Solar drying, Sol. Energy 85 (2011) 1665e1691.
[6] D. Jain, P. Tewari, Performance of indirect through pass natural convective
solar crop dryer with phase change thermal energy storage, Renew. Energy 80
4. Conclusion (2015) 244e250.
[7] A. Kumar, G.N. Tiwari, Effect of mass on convective mass transfer coefficient
during open sun and greenhouse drying of onion flakes, J. Food Eng. 79 (2007)
An experimental and theorical study was carried out on solar
1337e1350.
drying of potatoes, using an indirect solar dryer and a new mixed [8] P. Barnwal, G.N. Tiwari, Grape drying by using hybrid photovoltaic-thermal
solar dryer in Algerian coastal climate. To caracterize drying (PV/T) greenhouse dryer: an experimental study, Sol. Energy 82 (2008)
kenetics and moisture, different theoretical models were tested to 1131e1144.
[9] S.M. Shalaby, M.A. Bek, Experimental investigation of a novel indirect solar
gauge their accuracy. dryer implementing PCM as energy storage medium, Energy Convers. Manag.
The drying rate of potato slices in the MSD was lower than in the 83 (2014) 1e8.
[10] M.C. Te
llez, I.P. Figueroa, B.C. Te

llez, E.C.L. Vidan
~ a, A.L. Ortiz, Solar drying of
ISD. The results showed a good color retention of potato slices in
Stevia (Rebaudiana Bertoni) leaves using direct and indirect technologies, Sol.
the ISD, whilst in the MSD, 35% of the total number of slices turned Energy 159 (2018) 898e907.
a Little grayish. The MSD has a better thermal effiency in compar- [11] O.V. Ekechukwu, V. Heinzel, S. Gordeev, CFD analysis of a thermo-hydraulic
ison to the ISD. in terms of drying time the ISD performs better than flow distribution in generic indirect-type natural-circulation solar-energy
dryers, Int. J. Sustain. Energy 33 (1) (2014) 159e178.
the MSD due to advantageous air velocity 0.45 m/s for the ISD [12] A. Sreekumar, P.E. Manikantan, K.P. Vijayakumar, Performance of indirect
compared to 0.15/s for the MSD. solar cabinet dryer, Energy Convers. Manag. 49 (2008) 1388e1395.
The porous texture of potatoes facilitates heat transfer, which [13] A. Fudholi, K. Sopian, M.H. Yazdi, M.H. Ruslan, M. Gabbasa, H.A. Kazem, Per-
formance analysis of solar drying system for red chili, Sol. Energy 99 (2014)
favors evaporation of the water. As a result the rate of drying is high.
47e54.
The effective diffusion coefficient was calculated using tow [14] M. Kumar, S.K. Sansaniwal, P. Khatak, Progress in solar dryers for drying
methods, Fourier series and Laplace transforms.The values of the various commodities, Renew. Sustain. Energy Rev. 55 (2016) 346e360.
[15] S. Şevik, Design, experimental investigation and analysis of a solar drying
effective diffusion coefficient obtained by Fourier series for the two
system, Energy Convers. Manag. 68 (2013) 227e234.
dryers is higer than the values obtained using Laplace transforms. [16] A. Reyes, A. Mahn, F. Va
squez, Mushrooms dehydration in a hybrid-solar
Seven models were used to fit the experimental data and their dryer, using a phase change material, Energy Convers. Manag. 83 (2014)
comparison shows that the suggested dimensionless models are 241e248.
[17] B. Ringeisen, D.M. Barrett, P. Stroeve, Concentrated solar drying of tomatoes,
the most appropriate models for describing the drying behavior of Energy. Sustain. Develop. 19 (2014) 47e55.
potatoes for the two considered dryers. [18] S.F. Dina, H. Ambarita, F.H. Napitupulu, H. Kawai, Study on effectiveness of
The new proposed models were validated with experimental continuous solar dryer integrated with desiccant thermal storage for drying
cocoa beans, Case Studies in Thermal Engineering 5 (2015) 32e40.
data from the literature and compared to models that best describe [19] A. Tesfamichael, A. Assefa, Experimental analysis of potato slices drying
the drying process of Sultana grapes and chilli peppers the results characteristics using solar dryer, J. Appl. Sci. 13 (2013) 939e943.
indicate that they adequately describe the solar drying behavior of [20] F. Batubara, E. Misran, S.F. Dina, Heppy, Solar energy dryer kinetics using flat-
plate finned collector and forced convection for potato drying, AIP Conference
Sultana grapes, chilli peppers, bananas, mangoes and cassava. The Proceedings 1855 (2017) 070002 1e8.
two new proposed models are independent of units of measure- [21] F.K. Forson, M.A.A. Nazha, H. Rajakaruna, Modelling and experimental studies
ment, design, operating conditions, and take into account the on a mixed-mode natural convection solar crop-dryer, Sol. Energy 81 (2007)
346e357.
boundary conditions which is a significant advantage over existing [22] S. Boughali, H. Benmoussa, B. Bouchekima, D. Mennouche, H. Bouguettaia,
literature models. D. Bechki, Crop drying by indirect active hybrid solar e electrical dryer in the
This study was conducted in the north of algeria where air hu- eastern Algerian Septentrional Sahara, Sol. Energy 83 (2009) 2223e2232.
[23] A.A. El-Sebaii, S.M. Shalaby, Experimental investigation of an indirect-mode
midity is high. It is expected that the dryers would perform more
forced convection solar dryer for drying thymus and mint, Energy Convers.
efficiently in arid or semi-arid regions like in the Sahara desert or Manag. 74 (2013) 109e116.
the highlands where the air is dry. Because of the ready availability [24] A. ELkhadraoui, S. Kooli, I. Hamdi, A. Farhat, Experimental investigation and
and low cost of natural gas locally, using it as a backup source of economic evaluation of a new mixed mode solar greenhouse dryer for drying
of red pepper and grape, Renew. Energy 77 (2015) 1e8.
energy for the drying process is an attractive alternative to solar [25] L. Blanco-Cano, A. Soria-Verdugo, L.M. Garcia-Gutierrez, U. Ruiz-Rivas,
energy for extended periods of cloudy weather. Modeling the thin-layer drying process of Granny Smith apples: application in
1506 A. Djebli et al. / Renewable Energy 145 (2020) 1494e1506

an indirect solar dryer, Appl. Therm. Eng. 108 (2016) 1086e1094. experimental and theoretical evaluation, Sol. Energy 159 (2018) 722e732.
[26] B. Haddad, A. Liazid, P. Ferreira, A multi-criteria approach to rank renewables [42] O. Badaoui, S. Hanini, A. Djebli, B. Haddad, A. Benhamou, Experimental and
for the Algerian electricity system, Renew. Energy. 107 (2017) 462e472. modelling study of tomato pomace waste drying in a new solar greenhouse:
[27] L. Bennamoun, A. Belhamri, Contribution a  l’e

tude de faisabilite

de se

choirs evaluation of new drying models, Renew. Energy 133 (2019) 144e155.
solaires pour les produits agricoles, Revue des Energies Renouvelables, CER’07 [43] J. Crank, The Mathematics of Diffusion, second ed., Claredon press, Oxford,
Oujda (2007) 201e204. 1975.
[28] L. Lahnine, A. Idlimam, M. Mahrouz, S. Mghazli, N. Hidar, H. Hanine, A. Koutit, [44] C. Bonazzi, J.-J. Bimbenet, Se
chage des produits alimentaires e Principes,
Thermophysical characterization by solar convective drying of thyme Techniques de l'Inge
nieur F3000, 2003, pp. 1e14.
conserved by an innovative thermal-biochemical process, Renew. Energy 94 [45] J. Van Brakel, Mass transfer in convective drying, in: Advances in Drying,
(2016) 72e80. Hemisphere Publishing Corporation, 1980.
[29] O. Yaldiz, C. Ertekin, H.I. Uzun, Mathematical modeling of thin layer solar [46] S. Boughali, B. Bouchekima, N. Nadir, D. Mennouche, H. Bouguettaia, D. Bechki,
drying of sultana grapes, Energy 26 (2001) 457e465. Expe
rience du se

chage solaire dans le Sahara septentrional est alge

rien, Revue
[30] S. Lahsasni, M. Kouhila, M. Mahrouz, A. Idlimam, A. Jamali, Thin layer des Energies Renouvelables SMSTS’08 Alger (2008) 105e110.
convective solar drying and mathematical modeling of prickly pear peel [47] J.J. Bimbenet, J.D. Daudin, E. Wolff, Air drying kinetics of biological particles,
(Opuntia ficus indica), Energy 29 (2) (2004) 211e224. in: R. Toei, A.S. Mujumdar (Eds.), Drying ’85, Springer Berlin Heidelberg, 1985,
[31] R.K. Goyal, A.R.P. Kingsly, M.R. Manikantan, S.M. Ilyas, Thin-layer drying ki- pp. 178e185.
netics of raw mango slices, Biosyst. Eng. 95 (1) (2006) 43e49. [48] Z.W. Cui, Z.W. Shi-Ying Xu, Z.W. Da-Wen Sun, Effect of microwave vacuum
[32] T.Y. Tunde-Akintunde, Mathematical modeling of sun and solar drying of chilli drying on the carotenoids retention of carrot slices and chlorophyll retention
pepper, Renew. Energy 36 (2011) 2139e2145. of Chinese chive leaves, Dry. Technol. 22 (3) (2004) 563e575.
[33] I. Doymaz, The kinetics of forced convective air-drying of pumpkin slices, [49] T.R.A. Magee, C.P.d. Wilkinson, Influence of process variables on the drying of
J. Food Eng. 79 (2007) 243e248. potato slices, International Journal of Food Science & Technology 27 (5) (1992)
[34] E.K. Akpinar, Y. Bicer, Mathematical modelling of thin layer drying process of 541e549.
long green pepper in solar dryer and under open sun, Energy Convers. Manag. [50] Z. Yusheng, K.P. Poulsen, Diffusion in potato drying, J. Food Eng. 7 (4) (1988)
49 (2008) 1367e1375. 249e262.
_ Tog
[35] I.T. rul, D. Pehlivan, Mathematical modelling of solar drying of apricots in [51] A.M. Sereno, G.L. Medeiros, A simplified model for the prediction of drying
thin layers, J. Food Eng. 55 (2002) 209e216. rates for foods, J. Food Eng. 12 (1) (1990) 1e11.
[36] G. Page, M.S. thesis, Factors Influencing the Maximum Rates of Air Drying [52] S.L.M. Alzamora, J. Chirife, Some factors controlling the kinetics of moisture
Shelled Corn in Layers, Purdue University, 1949. movement during avocado dehydration, J. Food Sci. 45 (6) (1980) 1649e1651.
[37] K. Sacilik, A.K. Elicin, The thin layer drying characteristics of organic apple [53] K.O. Falade, E.S. Abbo, Air-drying and rehydration characteristics of date palm
slices, J. Food Eng. 73 (2006) 281e289. (Phoenix dactylifera L.) fruits, J. Food Eng. 79 (2) (2007) 724e730.
[38] A. Midilli, H. Kucuk, Z. Yapar, A new model for single layer drying, Dry. [54] A. Kaya, O. Aydın, C. Demirtaş, Drying kinetics of red delicious apple, Biosyst.
Technol. 20 (7) (2002) 1503e1513. Eng. 96 (4) (2007) 517e524.
[39] S.M. Henderson, S. Pabis, Grain drying theory. I. Temperature effect on drying [55] P.P. Tripathy, S. Kumar, Determination of temperature dependent drying pa-
coefficient, J. Agric. Eng. Res. 6 (3) (1961) 169e174. rameters for potato cylinders and slices during solar drying, Energy Convers.
[40] P.S. Madamba, R.H. Driscoll, K.A. Buckle, The thin layer drying characteristics Manag. 49 (2008) 2941e2948.
of garlic slices, J. Food Eng. 29 (1996) 75e97. [56] S. Chouicha, A. Boubekri, D. Mennouche, M.H. Berrbeuh, Solar drying of sliced
[41] B. Ameri, S. Hanini, A. Benhamou, D. Chibane, Comparative approach to the potatoes, an experimental investigation, Energy Procedia 36 (2013)
performance of direct and indirect solar drying of sludge from sewage plants, 1276e1285.