Академический Документы
Профессиональный Документы
Культура Документы
An International Journal
To cite this article: T. H. Nguyen, J.-L. Lanoisellé, T. Allaf & K. Allaf (2016): Experimental and
Fundamental Critical Analysis of Diffusion Model of Airflow Drying, Drying Technology, DOI:
10.1080/07373937.2016.1155052
Article views: 8
Download by: [Library Services City University London] Date: 19 March 2016, At: 10:51
Experimental and fundamental critical analysis of diffusion model of airflow
drying
Abstract
the kinetic effect of airflow velocity. Some authors confirmed that it does not trigger any
modification of drying; while some articles tried to establish empirical models of the
effective diffusivity Deff versus the airflow velocity; what is fundamentally erroneous. By
analyzing internal and external transfer phenomena, this research aimed at recognizing
that once air velocity is higher than a Critical Airflow Velocity CAV, the internal
transfers become the limiting phenomenon. CAV depends on the effective diffusivity and
the product size. It was calculated in the cases of two studied raw materials (apple and
carrot), differently textured by Instant controlled pressure drop DIC. Values of CAV
greatly depend on diffusivity of water within the matrix. At temperature T = 40 °C, they
were 1 m/s for untreated carrot and 2.1 m/s for DIC textured carrot, whose Deff values
were 1.31 and about 3 10-10 m2 s-1, respectively. Also, at temperature T = 40 °C, they
were 2.1 m/s for untreated apple and 3 m/s for DIC textured apple, whose Deff were 1.4
1
KEYWORDS: Drying kinetic model, heat transfer, mass transfer, agro-material, Critical
1. INTRODUCTION
external, heat and mass transfers. It is frequently assumed that, based on the drying rate,
Downloaded by [Library Services City University London] at 10:51 19 March 2016
drying operation can be divided into three periods [1]: 1) an initial unsteady rate,
Nevertheless, Rizvi [2] postulated that the constant-rate period, which only would exist in
some rare food systems, should be correlated with internal liquid water movement likely
controlled by weight forces. Though, within structured foods, water transfer is achieved
through complex internal capillary and diffusion liquid movement. Reaching the
interaction surface, water can be evaporated with a rate depending on airflow parameters.
When this last is adequately intensified, surface is not immediately replenished by liquid
water from the interior of the material [3]. Such drying cases are likely to perform without
exhibiting any constant-rate period. In fact, the experimental study on several products
noted that the drying kinetics was only the falling-rate period; such as: wood [4]; grape[5,
6]
; cassava, ginger and okra [7]; tomato [8, 9]; tapioca root [10]; sugar beet root [11]; avocado
[12]
; fish[13, 14]; leaves [15, 16]; herbs [17]; grains and oilseeds [18]; date palm [19]; apple [20, 21];
2
In the analysis of the drying rate, the impact of airflow velocity should take place in
terms of value but also the type and even, for perpendicular flows, the distance between
nozzles and products. However, in the present work, we limited our approach to a first
1. In some studies, no discernible effect of the airflow type and velocity on the
drying rate was observed in the range 0.5-2.5 m/s for potatoes ; 1-2 m/s for pistachios ;
Downloaded by [Library Services City University London] at 10:51 19 March 2016
0.5-1 m/s cross flow for garlics [27]; 1.5-2.6 m/s for vegetables [28], etc.
2. However, other studies revealed that drying rate increases with increasing air
velocity. This was obtained in the range of 0.2-1.5 m/s for apricots [29]; 0.5-1 m/s cross
flow for carrot [30]; 0.25-1 m/s for grape [31]; 0.3-0.7 m/s for banana in a counter-current
dryer [32], 1.4-2.7 m/s for chestnut[33], co-current flow at 0.6-1.8 m/s for apple [34] and at
3. Finally, some works [36-42] stipulated that the effect of this air velocity is efficient
only below a limited value (). Although scarce, these observations are the most coherent
induces a higher drying rate. Nevertheless, once the value of this external resistance is
lower than internal resistance, internal mass transfer becomes the limiting phenomenon.
Higher airflow velocities do not have effect on drying kinetics. From the Table 1, it
would be possible to estimate this critical airflow velocity CAV between 1 and 3 m/s
(depending on the product nature, structure and shape/size, as well as temperature and
air-humidity).
3
2. MODELING
Only once airflow velocity value is higher than CAV, drying is diffusion-controlled
process. Then, the effective moisture diffusivity inside the matrix can be deduced from
can not perform without proving that external transfer resistance is lower than internal
Downloaded by [Library Services City University London] at 10:51 19 March 2016
diffusion resistance. This inevitably should imply to analyze drying kinetics and achieve
Several surveys on drying of fruits and vegetables under suitable drying conditions
proposed diffusivity values ranged from 10-9 to 10-11 m2/s at temperature 50 °C [43]. They
varied for apricots from 5.59 to 6.51 10-9 m2/s over temperature range of 50–80 °C
dependently on airflow rate range of 0.2–1.5 m/s [29], for red peppers from 3.723 to
air velocity of 0.5 m/s [44], for figs from 1.33 to 3.36 10-9 m2/s at 55–85 °C for airflow
velocity ranged in 0.5–3 m/s [39], and for kiwi from 3.00 to 17.21 10-10 m2/s at 30 and
90 °C, respectively for air velocity of 2.3 m/s [45]. These results proved that, normally the
Simal et al. [46] using the drying experiment performed at 90 °C with 0.020 m broccoli
stems; Tong and Lund [47] on white bread, plain sheet muffin, and baking powder biscuit
at temperatures between 20 and 100 °C; Garau et al. [48] on the orange skin within the
interval of temperatures and moisture contents comprised between 30 and 90 °C, and
0.25 and 3.66 kg water/kg db, respectively, it was presented also depending on water
4
content. In all these studies, the authors should have assumed the process as an internal
presented the effective diffusivity depending on airflow velocity. Such as Akpinar and
Dincer [49] on eggplant slices with 5 mm thickness and 35 mm diameter during drying at
the temperatures of 55, 65, and 75 °C and the velocities of 1.0 and 1.5 m/s, respectively;
Benhamou et al. [50] on marjoram leaves at three temperatures (40, 50 and 60 °C) and two
Downloaded by [Library Services City University London] at 10:51 19 March 2016
air flow (0.0227 and 0.0556 m3/s); Guan et al. [51] on tilapia fillets at the drying
temperature (35, 45 and 55 °C), hot air velocity (1.5, 2.5 and 3.5 m/s) and thickness (3, 5
and 7 mm). Also, Jin Park et al. [52] on mint leaves at the temperatures of 30, 40, and
50 °C and the velocities of 0.5 and 1.0 m/s; Kaya et al. [20] on apple using air
temperatures at 35, 45 and 55 °C, velocities at 0.2, 0.4 and 0.6 m/s and relative humidity
values at 40%, 55% and 70%; Nicoleti et al. [53] on pineapple at the air velocity varied
from 1.5 to 2.5 m/s and the air temperature from 40 to 70 °C; Motevali et al. [54] on jujube
at temperature levels of 50, 60 and 70 ºC and hot air velocities of 0.5, 1 and 1.5 m/s;
Toğrul and Pehlivan [29] on apricot at a variety of flow rates (0.2, 0.5, 1.0 and 1.5 m/s)
and temperatures (50, 60, 70 and 80 °C). These authors presented an effective diffusivity
increasing with surrounding airflow velocity. Even more, some other authors; such as
Abbaszadeh et al. [55] on russian olive at air velocity of 0.5, 1 and 1.5 m/s and three air
temperature levels of 50, 60 and 70 °C; Aghbashlo et al. [56] on beriberi at temperature
from 50 to 70 °C with drying air velocities of 0.5–2 m/s concluded that the effective
5
Additionally, Babalis and Belessiotis [39], Clemente et al. [42], Clemente et al. [57] Blasco
[58]
presented their respective works on figs in the temperature range of 55–85 °C and the
airflow ranged in 0.5–3 m/s, pork under forced convection conditions (25 °C and air
velocity of 0.6 ± 0.1, 2.0 ± 0.1 and 2.8 ± 0.1 m/s), potato cubes of 1 cm side at 60 ºC and
0.5, 1, 1.5, 2, 3, 4.5, 7, 8 and 10 m/s, and curcuma longa rhizomes at 100 ºC and air rates
of 0.2, 0.5, 0.7, 1.2, 2.1, 2.6, 3 and 4 m/s. They concluded that airflow velocity also
Downloaded by [Library Services City University London] at 10:51 19 March 2016
affected effective diffusivity. Undeniably, their work should not allow identifying the
diffusivity because it was performed at airflow velocity lower than CAV (critical airflow
velocity), which was identified to be 2 m/s according to the first three authors and from 1
to 2 m/s according to the last one. However, we do not agree with these authors that they
calculated the effective diffusivity at all values of air velocity. Indeed, drying process
cannot be used to determine the effective diffusivity without proving that the operation is
Thus, increasing velocity cannot accelerate the drying process. Then, it is worth noting
that:
When the airflow velocity is lower than CAV, the drying kinetics does not allow
When the airflow velocity is higher than CAV, the drying kinetics is "controlled"
by the internal transfer and effective diffusion is the limited phenomenon. Then, it is
CAV depends on the effective diffusivity, sample frame and size, and airflow humidity.
However, how can we know when internal diffusion is the limited phenomenon? Little
6
research attempted to answer this question. Marinos-Kouris and Maroulis [59] define a
dimensionless parameter similar to Biot number. Their model takes into account the
controlling mechanisms of heat and mass transfers. The transport properties (water and
thermal diffusivities, boundary heat and mass transfer coefficients) are included in the
ρW
ρD W (1)
t
a1 a2
D a 0 exp exp (2)
W T
ρD W k a we a w (3)
In this Eq. 3, one has to keep in mind that during the drying process the water activity to
be considered is that of the surface, which normally is different than global water activity
of the material[60].
c3
k c0 Wairc1 Tairc2 v air (4)
ρh s
kc T (5)
t
b1 b2
kc b0 exp exp (6)
W T
kc T h T Tair Hs k a we a w (7)
7
d1 d 2 d 3
k d 0 Wair Tair v air (8)
Limited conditions such as initial water content, temperature, and thickness of material,
as well as airflow humidity, temperature, and velocity are also considered in the final
solution. Then the constants of the relative empirical equations are considered as model
parameters. In order to simplify the drying model, the heat and mass transfer–controlled
ρair k eq kcL
Bi m (9)
ρβ0 Deff 0
Wair,i
Wβe dWβi
Wβi
k eq (10)
Wβ 0 Wβ0 Wβe
Where: Bim is the modified mass transfer Biot number proposed by Cdrdova-Quiroz et al.
[62]
resistance of convection mass transfer at the interface is slower than that of diffusion
mass transfer in the interior of food, and therefore the phenomenon is diffusion
controlled. Under these conditions the external properties or variables, such as air
moisture or air velocity have no effect over drying kinetics. If Bim << Keq·(X 0 - X e) the
mass transfer is external convection controlled, and therefore under these conditions air
8
Ruiz-López et al. [63] and Barati and Esfahani [64] assumed that mass transfer mechanisms
2
Wβ Wβ
Deff (11)
t z2
Wβ
Deff ρβ |z d kρair Wj Wair (12)
z
Wβ
Downloaded by [Library Services City University London] at 10:51 19 March 2016
|z 0 0 (13)
z
Wβ z,0 W0 (14)
The mass transfer coefficient k at the food interface is computed versus Re and Sc using
Dair
k 2 0.522 Re0.5 Sc0.33 (15)
L
0.622p w,Ts a w
Wj (16)
Ptotal p w,Ts a w
Here, aw is the water activity of the material (and not only at the surface, as we stipulated
later )). The diffusivity is determined as the second Fick's law coefficient. By using
the equations from to one can calculate the air limit velocity.
2.2. Fundamental
Respecting the phenomenological drying model defined by Allaf and Allaf [65], drying
curve should systematically be divided into three stages: 1) a first surface interaction
9
stage; 2) an internal liquid water movement stage; and 3) an internal vapor water
movement stage.
This last is the famous paradoxical stage where both temperature and vapor pressure
gradients are directed from the product surface towards the core. Drying is then done by
progressive low water activity front. Normally, no shrinkage accompanies this third
Downloaded by [Library Services City University London] at 10:51 19 March 2016
stage.
As described by Allaf and Allaf [65], the first surface interaction stage systematically takes
place whatever the material structure is. The only way to get this stage as a constant-rate
period is to get a drying operation only controlled by the external interaction between
airflow and the product surface as the limiting process. This means that material should
avoid shrinkage and maintain constant both interaction surface and its water activity; it is
In the most conventional drying cases, the intensification of airflow conditions are easily
carried out by increasing temperature, velocity and dryness of hot air. This results in a
rapid and efficient surface drying, which habitually would decrease water activity and
imply shrinkage. Therefore, consistently, the first stage of surface interaction period is
The next stage, in which the internal liquid water movement is controlling (as a limiting
process) the drying operation, is systematically a falling rate phase. Thus, the internal
10
liquid water transfer is usually considered as the limiting process, thus controlling the
main part of the entire drying operation. Various phenomena occur to ensure water
transfer from inside the product to the surface, such as gravity force, capillary, molecular
diffusion, etc.
At this respect, Allaf et al. [66] proposed the operation of drying as divided into four
Downloaded by [Library Services City University London] at 10:51 19 March 2016
The heat transfer from the external environment to the exchange surface,
The transfer of the liquid substance and / or steam, from the products to its
exchange surface,
To understand more easily these processes, authors usually classify them into two
categories: 1) external transfer processes between the outside and the surface of the
product, and 2) process for which internal transfers between the exchange surface and the
product core.
In process 1: External heat transfer (convection) and vapor transport occur at the
exchange surface through the action of the air stream. The drying rate can be constant,
only when a) it takes into account the impact of shrinkage, b) there is no decrease in
water activity on the surface, and c) the evaporation rate at the surface is not limited by
11
the whole internal mass transfer. This last means that the limiting processes should then
This period only concerns the interaction between the airflow and the exchange surface.
It strictly depends on the gradient of vapor pressure between the exchange surface and the
airflow.
Downloaded by [Library Services City University London] at 10:51 19 March 2016
Q hA Tair Ts (17)
However, Allaf [67] defined more precisely this vapor transport achieved at the surface A
by:
A p w,Ts a ws p w,air
W kρ v (18)
md Ptotal
Recently, Perré et al. [68] underlining the direct impact of the external mass and heat
transfers on the calculation of internal diffusion mass transfer coefficient from drying
unsteady regime. It is based on imposing on one face a relative humidity and measuring
the humidity at the back face of the sample. Since such an operation is done
independently on any external air exchange with the surface, this new method is easy and
Usually, several authors tried to determine the internal diffusion mass transfer coefficient
from experimental data of drying kinetics. One can find numerous correlations in the
12
literature. Skelland [69] compiled some of them, which are based on the internal and
external heat and mass transfer analogy related to a particle shape. The most adequate
correlation for experimental conditions of the present work was that of Kays[70]:
The characteristic dimension used for the calculation of Sh and Re numbers is defined as
the total surface area of the body divided by the perimeter of the maximum projected area
Downloaded by [Library Services City University London] at 10:51 19 March 2016
perpendicular to the airflow. The mass transfer coefficient k depends on the average
1 2
0.56
v
3 D 3 0.56
k 0.43 v (20)
L0.44
In airflow drying, the exchange surface temperature quickly increases to a high level and
a gradual spread of heat within the solid occurs. As previously mentioned, the presence of
the holes. The heat transfer phenomenon thus generated has a temperature gradient as the
Only when it can be assumed that external heat and mass transports do not limit the
whole operation, internal transfers become the limiting process [71]. Usually, increasing
13
temperature, velocity and airflow dryness can easily intensify the external transports. In
these conditions, as water transfer within the product becomes the principal restricting
factor for the drying kinetics, the model of Mounir and Allaf [72] can be adopted, with a
ρw ρw
vw vm Deff (21)
ρm ρm
Downloaded by [Library Services City University London] at 10:51 19 March 2016
Heat and water can be transferred between the matrix and its environment in opposite
by microwaves. In the airflow drying process, the heat flux depends on the temperature
and velocity of the surrounding medium (air). Thus, the higher the air temperature, the
higher the gradient of vapor pressure in the surrounding medium. However, the
crusting. Another important airflow parameter is the humidity content, which must be
kept as low as possible, this depends on biological material and technical resources. By
using appropriate desiccators or by lowering the temperature before heating (e.g., with a
heating pump system), this important airflow parameter can be also controlled. However,
an airflow that is too dry may induce a specific degradation of the product surface both in
Nevertheless, airflow velocity is the main parameter that has to properly control to safely
obtain an appropriate external intensification. The higher the rate v, the higher the
Reynolds’s number Re, the heat h and mass convection coefficients k. Conversely,
external intensification is no longer relevant when the internal Fick’s diffusion mass
14
transfer becomes the limiting process. Thus, the impact of airflow rate stops once it
reaches a critical level vl corresponding to the critical value kl of the mass convection
coefficient [65].
To avoid the external convection transfer to become the limiting process, the following
dρ w
Deff Ptotal
dz z zsurface
k lim (24)
ρ v p w,Ts a w,s p w,air
From this last equation, it is possible to identify the Critical Airflow Velocity CAV:
1.78
dρ w
0.56
1 Ptotal Deff
3
L0.44 dz z zsurface
CAV 2.33 2 2/3
(25)
0.56
3
D vapor air p w,Ts a w,s p w,air
v
Once airflow velocity is higher than CAV, drying becomes diffusion-controlled process.
However, another aspect has to be considered; it is the initial interaction between hot air
and the superficial zone of the product. Allaf et al. [66] introduced the concept of “starting
Washing/Diffusion CWD model. The effective diffusivity was then determined based on
the second Fick's law. However, the phenomenological model of diffusion cannot
15
normally fit all the experimental data of drying because of the inevitable presence of an
initial interaction between airflow and the surface of the product during the first stage of
drying [65]. The only experimental data which are able to insert in such measurements of
second Fick’s law should be limited to time t > t1 to remove all washing-stage data, and
the part of drying where internal vapor transfer can be assumed to be negligible compared
to liquid water diffusion. Indeed, internal vapor transfer has a specific behavior with the
Downloaded by [Library Services City University London] at 10:51 19 March 2016
gradient of vapor pressure as driving force. Since vapor pressure is strictly function of
can not be revealed through normal simple diffusion equation. Without excluding this
paradoxical part of drying kinetics, Perré [60] introduced a dimensionless number, the
Drying Intensity Number to chose the right level of modeling, depending on the drying
configuration.
of two types of products (carrot and apple). Our objective was to compare the impacts of
Traditional Hot Air Drying (THAD), and Swell- Drying (SD), which combines Instant
Controlled Pressure Drop (DIC) texturing treatment with conventional hot air drying.
3.1. Materials
16
Three types of materials were studied. They were carrot and apple bought from a local
internal diffusivity is very high. Carrot and apple were washed, peeled and cut into 0.8
cm thickness slices and 1 cm cube, respectively, similar to SCP. Initial water content was
634 ± 62% db (dry basis) for carrot, 656 ± 55% db for apple, and 7200 ± 43% db for
SCP.
Downloaded by [Library Services City University London] at 10:51 19 March 2016
In order to obtain various diffusivity values, carrot and apple were treated at different
DIC conditions. Based on previous studies [73-75] and preliminary trials, the samples of
carrot and apple were dried at 45 °C till a water content of 25 ± 1% db for carrot,
"Instant controlled pressure drop") followed by THAD final drying. THAD was carried
out on 5 samples of these DIC-textured and unprocessed apple and carrot samples (Table
Table 1. DIC operating parameters for the 5 experiment points of apple and carrot.
DIC reactor was defined and presented in various papers [65, 76, 77]. It includes three main
17
DIC treatment includes two main steps: establishment of high steam pressure (between
0.1 and 0.5 MPa), and an instant releasing of pressure towards a vacuum of about 5 kPa.
During the first step, the product is exposed to high temperature for a treatment time of
10 to 20 s. The second stage is carried out with a pressure drop rate higher than
5 105 Pa/s. In the present study, DIC absolute steam pressure in the treatment chamber P
and the high temperature treatment time t were adjusted, while the other DIC operating
Downloaded by [Library Services City University London] at 10:51 19 March 2016
parameters were kept constant. Five samples of differently DIC textured and one
Material was dried in the static dryer (Figure 2). The compressed air passes through
pneumatic valves (1,2) and then goes to the heater (3). Hot airflow is uniformly
distributed (4) perpendicularly to the samples in the tray (5). The GP-Pro Ex 2.5 Software
(Pro-face, Digital Electronics Corporation, Osaka, Japan) was used for programming and
monitoring the dryer. Drying was automatically performed using pre-selected parameters
of airflow temperature and velocity. Mass evolution was evaluated manually during the
operation.
For different situations, specific Design of Experiments (DoE) can be adopted in order to
get the highest efficiency with the lowest number of trials. To study and analyze the
drying operating parameters, central composite experiment method was adopted. In this
research, two operating parameters were studied: the airflow temperature T and velocity
18
v. The air humidity was kept constant at a vapor pressure of 265 Pa. The experiments
After preliminary experiments, we achieved the drying processes following specific DoE
with air velocity ranged from 3 to 15 m/s at various temperature values ranged from 20 to
60 ºC for apple, carrot, and Specific Controlled-Product SCP. Other complement study
was carried out at a constant temperature of 40 ºC, with airflow velocity at different
values ranged from 1 to 15 m/s for different samples of carrot and apple, and for SCP.
Windows® (Statgraphics centurion XV, StatPoint Technologies, Inc., USA), for each
xi (P ≤ 0.10). Pareto chart is introduced to more conveniently reveal this aspect and used
to more clearly present the significance level of independent variables. This can
prediction empirical model of the response parameter Y versus the independent variables
xi:
n n n
Y βo βi x i βii x 2
i βij x i x j ε 0 (26)
i 1 i 1 i 1
19
Where: β o , β i , β ii and β ij are the regression coefficients, and ε 0 the random error. Trends
and response surface, as well as Response Surface Methodology can also be used for
When the airflow velocity is higher than the CAV, drying should be diffusion-controlled
Downloaded by [Library Services City University London] at 10:51 19 March 2016
model was defined through the three parameters of the final equilibrium water content
dry basis W∞, the effective diffusivity Deff, and the “starting accessibility Ws”. This last
is assumed to be carried out through a short time air/surface interaction coupled to a long
The effective diffusivity Deff should be determined based on the second Fick's law, which
uses the relationship with the reduced water content MR as a dependent variable, can be
The samples were assumed to be infinite slab (thickness of 7 mm versus 1500 mm2 as
surface for apple and thickness of 5 mm versus 1000 mm2 as surface for carrot). We
assumed that (1) the system was isotropic and uniform in terms of initial water content,
temperature, and structure, (2) there was no shrinkage of the product [78]. This gives the
20
2
W W 8 1 2n 1 π2 Deff t t1
exp (27)
W1 W π2 n 0 2n 1
2
4d 2
The diffusion model concerns are only for t > t1, the extrapolated value computed at t = 0
of water content Wo has a distinct value from the real initial water content Wi. The
Downloaded by [Library Services City University London] at 10:51 19 March 2016
It represents that the water quickly removed from the surface independently from the
diffusion process. The starting accessibility δWs is defined to reveal the specific effect of
According Allaf [67] the convective mass transfer coefficient k , can be calculated by
equation or:
1681
10.09938
Ts 43.037
p w, Ts 10 (30)
21
δWs
k Ptotal (31)
A
ρv p δt
m d w ,Ts
The t can be easily calculated from the kinetics drying when we carry out continuous
4. RESULTS
Downloaded by [Library Services City University London] at 10:51 19 March 2016
The drying kinetics was studied for the experimental plan in Table 2. Drying curves of
raw material (control) and the DIC sampled at T = 40 °C, v = 9 m/s are given in Figure 3.
It is clear that the DIC texturing stage have a significant influence on the drying kinetics.
As shown in the Figure 3, the drying time of all SD samples was shorter than control. The
necessary time to reach 5% db for SD sampled at 0.5 MPa for 15 s (DIC 1) was about
93 min, while control sample needs 470 min (Table 3). Nguyen et al. [79] obtained a
higher absolute expansion for higher both saturated steam pressure and processing time.
The higher the product porosity, the higher the water effective diffusivity.
We also studied the effect of air velocity on the drying kinetics of apple and carrot at air
different points, we note that CAV was 1 m/s for untreated carrot and 2.1 m/s for carrot
expanded at the following DIC conditions: 0.35 MPa for 15 s, 0.2 MPa for 15 s,
0.15 MPa for 10 s, or 0.1 MPa for 10 s. CAV reached 3 m/s for carrot DIC expanded at
0.5 MPa for 15 s. Similar results were obtained with apple. CAV was 2.1 m/s for
22
untreated apple and 3 m/s for all DIC textured apple samples between 0.1 and 0.5 MPa
for 15 s to 20 s.
Conversely, SCP has a very specific behavior and its diffusivity was not accessible
because its kinetics systematically depended on air velocity (Figure 4). In all our drying
conditions, the extern transfers controlled drying process of SCP, and Critical Airflow
Downloaded by [Library Services City University London] at 10:51 19 March 2016
Velocity CAV should have been higher than the highest experimental value we
used: 15 m/s.
accessibility, as well as mass transfer coefficient k were calculated from drying curves
(W = f(t)) obtained only for airflow velocity higher than CAV, for both carrot and apple.
It was not possible to reach such condition for Specific Controlled-Product SCP.
4.2.1. Apple
The results presented in Table 3 indicate that the effective diffusivity Deff of dried apple
was ranged from 0.67 to 2.76 10-10 m2 s-1, which is in agreement with those reported by
other authors. Such as Kaya et al. [20] reported Deff value from 0.483 to 2.019 10-10 m2 s-1
and Velić et al. [35] obtained Deff value for apple between 1.7 and 4.4 10−9 m2 s-1.
23
After DIC texturing stage in SD, the effective diffusivity was improved by DIC caused
expanding the structure apple. It varied in the range of 10.02 10-10 to 32.79 10-10 m² s-1.
DIC texturing allowed effective diffusivity to be ten to fifteen times greater than
velocity at 9 m/s increased the effective diffusivity by 12.6 times compared to the
Table 3. Results of evaluated drying kinetics parameters of apple: drying time to reach a
( Ws), mass transfer coefficient (k). The correlation coefficient (R2) between the
experimental and predicted data value of the model is always higher than 97.5%
(dependent variable) and v (m s-1) and 1/T (K-1) as independent parameters. General
trends (Figure 5) illustrates that 1/T as parameter had significant linear influence on
ln(Deff) of apple for DIC 1 sample (P = 0.5 MPa, t = 15 s). ln(Deff) was completely
The empirical equation of the logarithm of the effective diffusivity for DIC 1 on
equation as follows:
2981.86
ln(Deff ) 10.5216 (32)
T
24
Thus, the effective diffusivity for DIC 1 is Arrhenius-similar model:
24791
Deff 2.69510 5 exp (33)
RT
Where Deff is expressed in 10-10 m2 s-1 and T in K. The experimental results fit well this
empirical model.
Downloaded by [Library Services City University London] at 10:51 19 March 2016
According to CWD model, starting accessibility should play a key-role in the airflow
drying process. As presented in Table 3, after DIC texturing stage, the starting
accessibility was systematically greater up to 2.3 times more than untreated sample. From
Pareto chart in Figure 6, it is possible to identify the most significant effects on the
starting accessibility for DIC 1 sample as due to the air temperature T. The higher the
temperature, the higher the starting accessibility. Moreover, the air velocity v and T2
effects were very weak (as we can have from ANOVA: P-value = 0.088 for v and P-
The empirical equation was calculated to express the effective diffusivity of DIC 1 versus
R2 =77.49%:
25
4.2.1.3. Mass Transfer Coefficient K
increases, and be higher for higher airflow velocity because of Ws evolution versus v.
Figure 7 indicates from general trends, the mass transfer coefficient was effected
parameters allowed identifying the prediction models for the masse transfer coefficient
with R2 = 78.02%:
Figure 7 shows that the experimental results fit this empirical model.
4.2.2. Carrot
Carrot Deff values presented in Table 4 are ranged from 0.7 to 2 10-10 m2 s-1. This
correlates literature values located Deff between 10-11 and 10-8 m2 s-1. Srikiatden and
Roberts [80] reported Deff values of carrot core from 6.42 to 14.7 10-10 m2 s-1 and carrot
cortex from 6.68 to 11.8 10-10 m2 s-1. Zielinska and Markowski [81] obtained Deff value for
26
After DIC texturing stage in SD, the effective diffusivity was improved by DIC caused
expanding the structure of carrot and apple. For carrot, DIC 1 (P = 0.5 MPa, t = 15 s)
dried at 60 °C and velocity at 9 m/s increased the effective diffusivity by 4.6 times
Table 4. Results of evaluated drying kinetics parameters of carrot: drying time to reach a
Downloaded by [Library Services City University London] at 10:51 19 March 2016
( Ws), mass transfer coefficient (k). The correlation coefficient (R2) between the
experimental and predicted data value of the model is always higher than 95%
Pareto chart of Figure 8 illustrated the impact of drying operating parameters on the
logarithm of the effective diffusivity for DIC 2 carrot sample (P = 0.35 MPa, t = 15 s).
Using an empirical equation to express the logarithm of the effective diffusivity for
DIC 1 versus the drying operative parameters, the following regression model could be
2206.46
ln Deff 14.7478 (36)
T
18344
Deff 3.93610 7 exp (36)
RT
27
4.2.2.2. Starting Accessibility Ws
Table 4 shows that after optimized DIC texturing stage, the starting accessibility was 1.5
The standardized Pareto (Figure 9) shows that the most significant effects on the starting
accessibility for DIC 1 (P = 0.5 MPa, t = 15 s) sample were due to the air temperature T,
Downloaded by [Library Services City University London] at 10:51 19 March 2016
The empirical equation was calculated to express the starting accessibility of DIC 1
versus the drying operative parameters. It was established with a regression coefficient
R2 = 77.88%:
coefficient k.
parameters allowed identifying the prediction models for the masse transfer coefficient of
28
The difference of behaviors of apple and carrot ( and , respectively) would be connected
with the excessive impact of temperature directly acting on vapor pressure, while the
airflow velocity takes place at factor about v0.56 . Future research work shall be conducted
5. CONCLUSION
Downloaded by [Library Services City University London] at 10:51 19 March 2016
The drying kinetics of carrot, apple and a very high-diffusivity Specific Controlled-
Product SCP were studied. Our objectives were to identify the strict conditions of
adopting hot air drying HAD as diffusion-controlled operation. The concept of Critical
Airflow Velocity CAV has been defined. As long as hot air velocity is lower than CAV
drying kinetics depends on airflow velocity, and drying operation cannot be a diffusion-
controlled process. This essential condition should systematically be justified prior to any
determination of effective diffusivity from drying curve. This was not respected in
In the present work, we defined Critical Airflow Velocity CAV as well as Coupled
Washing/Diffusion CWD as phenomenological common model for hot air drying HAD.
After studying the kinetics versus the hot air flow conditions (temperature, relative
The first is when the air flow velocity is less than the limiting value CAV. In these
conditions, drying operation should have the external heat and mass transfers at the
interaction surface as the limiting processes. Drying kinetics should depend on airflow
29
velocity and the internal water transfer process (“effective diffusion” from the core to the
The second is when the airflow velocity is greater than the Critical Airflow
Velocity CAV. Drying kinetics does not depend on airflow velocity and the internal
transfer (effective diffusion) of liquid water becomes the limiting process. Drying should
then reveal the effective diffusion phenomenon, which becomes visible from drying
Downloaded by [Library Services City University London] at 10:51 19 March 2016
experimental data, and the effective moisture diffusivity Deff can be determined. So the
effective moisture diffusivity Deff should depend on the material structure and the air
The Critical Airflow Velocity CAV is a function of the internal water diffusion within the
matrix. In this study for carrot, CAV was 1 m/s for untreated carrot and 2.1 m/s for carrot
treated at 0.35 MPa for 15 s, 0.2 MPa for 15 s, 0.15 MPa for 10 s and for 0.1 MPa for
10 s. CAV reached 3 m/s for DIC treated carrot at 0.5 MPa for 15 s. Similar results were
obtained with apple. CAV was 2.1 m/s for untreated apple and 3 m/s for all DIC textured
apple samples between 0.1 and 0.5 MPa for 15 s to 20 s. The higher the effective
diffusivity, the higher the CAV. For Specific Controlled-Product SCP, since the effective
diffusivity was very high, Critical Airflow Velocity CAV could not be identified. It
should have been higher than the highest experimental value we used: 15 m/s.
NOMENCLATURE
Ws Starting accessibility
30
W Rate of water content evolution (% db s-1)
aws Water activity at the surface. Normally, it is different than global water activity aw
k Mass transfer convection coefficient between the surface and the surrounding
medium (m s-1)
31
R universal gas constant
vL
Re Reynolds number: Re (-)
kL
Sh Sherwood number (-) Sh (-)
Dvapor air
Downloaded by [Library Services City University London] at 10:51 19 March 2016
T Temperatures (°C)
t Time (s)
Y Predicted response
Greek letters
32
ρm Apparent density of dry material (kg m-3)
Fick’s number
0 Offset term
ε0 Random error
Subscripts
at time t
e at equilibrium
in food
1 at time t1
i initial
p product
s surface
33
Acknowledgements
The authors acknowledge the Government of Vietnam for the financial support given to
this research.
References
1. Mujumdar AS. Handbook of industrial drying. CRC Press: Florida, United States, 2006;
Downloaded by [Library Services City University London] at 10:51 19 March 2016
1308 p.
Properties of Foods; Rao MA, Rizvi SSH, Datta AK, Eds., CRC Press: Florida, United
milk powder treated by DIC (Instant Controlled Pressure Drop) technology. Dairy
5. Fadhel MI; Kooli S; Farhat A; Belghith A. Séchage du raisin en plein air, dans un séchoir
SMSTS’08 2008,193-202
34
7. Ahouannou C; Jannot Y; Lips B; Lallemand A. Caractérisation et modélisation du
séchage de trois produits tropicaux : manioc, gingembre et gombo. Science des aliments
2000, 20(4/5),413-432.
Expérience du séchage solaire dans le Sahara septentrional est algérien. Revue des
9. Queiroz R; Gabas AL; Telis VRN. Drying Kinetics of Tomato by Using Electric
10. Chirife J; Cachero RA. Through-circulation drying of tapioca root. Journal of Food
11. Vaccarezza LM; Lombardi JL; Chirife J. Kinetics of moisture movement during air
drying of sugar beet root. International Journal of Food Science & Technology 1974,
9(3),317-327.
12. Alzamora SlM; Chirife J. Some factors controlling the kinetics of moisture movement
13. Bellagha S; Amami E; Farhat A; Kechaou N. Drying kinetics and characteristic drying
curve of lightly salted sardine (Sardinella aurita). Drying Technology 2002, 20(7),1527-
1538.
14. Sarr J; Sall M; Sow ML; Kane MM. Étude expérimentale des courbes de séchage du
35
15. Mohamed LA; Kouhila M; Jamali A; Lahsasni S; Kechaou N; Mahrouz M. Single layer
solar drying behaviour of citrus aurantium leaves under forced convection. Energy
16. Aghfir M; Kouhila M; Jamali A; Ait Mohamed L. Séchage solaire convectif pour la
18. Brooker DB; Bakker-Arkema FW; Hall CW. Drying and storage of grains and oilseeds.
19. Falade KO; Abbo ES. Air-drying and rehydration characteristics of date palm (Phoenix
20. Kaya A; Aydın O; Demirtaş C. Drying kinetics of red delicious apple. Biosystems
21. Doymaz İ. An Experimental Study on Drying of Green Apples. Drying Technology 2009,
27(3),478-485.
22. Arora S; Bharti S; Sehgal VK. Convective Drying Kinetics of Red Chillies. Drying
23. Gupta P; Ahmed J; Shivhare US; Raghavan GSV. Drying characteristics of red chilli
36
24. Pardeshi IL; Arora S; Borker PA. Thin-Layer Drying of Green Peas and Selection of a
26. Tavakolipour H. Drying kinetics of pistachio nuts (Pistacia vera L.). World Applied
27. Madamba PS; Driscoll RH; Buckle KA. The thin-layer drying characteristics of garlic
28. Krokida MK; Karathanos VT; Maroulis ZB; Marinos-Kouris D. Drying kinetics of some
29. Toğrul İT; Pehlivan D. Modelling of drying kinetics of single apricot. Journal of Food
30. Doymaz İ. Convective air drying characteristics of thin layer carrots. Journal of Food
31. Sawhney RL; Pangavhane DR; Sarsavadia PN. Drying studies of single layer thompson
International Solar Food Processing Conference Indore, India, January 14-16, 2009; 19 p.
37
32. Karim MA; Hawlader MNA. Mathematical modelling and experimental investigation of
tropical fruits drying. International Journal of Heat and Mass Transfer 2005, 48(23-
24),4914-4925.
kinetics of chestnut (Castanea sativa Mill.): Influence of the natural shells. Food and
36. Mulet A; Sanjuán N; Bon J; Simal S. Drying model for highly porous hemispherical
37. Mitchell TJ; Potts CS. Through-circulation drying of vegetable materials. III.—carrots.
38. Mulet A; Berna A; Borr M; Pinaga F. Effect of air flow rate on carrot drying. Drying
39. Babalis SJ; Belessiotis VG. Influence of the drying conditions on the drying constants
and moisture diffusivity during the thin-layer drying of figs. Journal of Food Engineering
2004, 65(3),449-458.
38
40. Mathioulakis E; Karathanos VT; Belessiotis VG. Simulation of air movement in a dryer
by computational fluid dynamics: Application for the drying of fruits. Journal of Food
41. Rossello C; Canellas J; Simal S; Berna A. Simple mathematical model to predict the
drying rates of potatoes. Journal of Agricultural and Food Chemistry 1992, 40(12),2374-
2378.
Downloaded by [Library Services City University London] at 10:51 19 March 2016
(Capsicum annuum L var Jaranda). Journal of the Science of Food and Agriculture
2003, 83(7),697-701.
45. Simal S; Femenia A; Garau MC; Rosselló C. Use of exponential, Page's and diffusional
models to simulate the drying kinetics of kiwi fruit. Journal of Food Engineering 2005,
66(3),323-328.
47. Tong CH; Lund DB. Effective moisture diffusivity in porous materials as a function of
39
48. Garau MC; Simal S; Femenia A; Rosselló C. Drying of orange skin: drying kinetics
49. Akpinar EK; Dincer I. Moisture transfer models for slabs drying. International
et cinétique de séchage solaire en convection forcée des feuilles de marjolaine. Revue des
Downloaded by [Library Services City University London] at 10:51 19 March 2016
51. Guan Z; Wang X; Li M; Jiang X. Mathematical modeling on hot air drying of thin layer
fresh tilapia fillets. Polish Journal of Food and Nutrition Sciences 2013, 63(1).
52. Jin Park K; Vohnikova Z; Pedro Reis Brod F. Evaluation of drying parameters and
desorption isotherms of garden mint leaves (Mentha crispa L.). Journal of Food
53. Nicoleti JF; Telis-Romero J; Telis VRN. Air-drying of fresh and osmotically pre-treated
pineapple slices: Fixed air temperature versus fixed slice temperature drying kinetics.
Jujube (Zizyphus jujube Mill). Journal of Agricultural Science and Technology 2012,
14(3),523-532.
velocity and temperature on energy and effective moisture diffusivity for russian olive
40
(Elaeagnusan gastifolial L.) in thin-layer drying. Iranian Journal of Chemistry and
the effective moisture diffusivity, energy of activation and energy consumption during
57. Clemente G; Bon J; Sanjuan N; Mulet A. Drying modelling of defrosted pork meat under
58. Blasco M. Effect of blanching and air flow rate on turmeric drying. Food Science and
of Industrial Drying, Third Edition; Mujumdar A, Eds., CRC Press: Florida, United
60. Perré P. The Proper Use of Mass Diffusion Equations in Drying Modeling: Introducing
61. Pavón-Melendez G; Hernández JA; Salgado MA; Garcia MA. Dimensionless analysis of
the simultaneous heat and mass transfer in food drying. Journal of Food Engineering
2002, 51(4),347-353.
mass transfer equation with interfacial resistance in food drying. Drying Technology
1996, 14(7-8),1815-1826.
41
63. Ruiz-López II; Córdova AV; Rodr guez-Jimenes GC; Garc a-Alvarado MA. Moisture
and temperature evolution during food drying: effect of variable properties. Journal of
64. Barati E; Esfahani JA. A new solution approach for simultaneous heat and mass transfer
65. Allaf T; Allaf K. Instant Controlled Pressure Drop (DIC) in food processing. Springer-
Downloaded by [Library Services City University London] at 10:51 19 March 2016
66. Allaf K; Mounir S; Allaf T. Swell-drying : séchage et texturation par DIC des végétaux
67. Allaf K. Transfer phenomena and industrial applications. Teaching book, Lebanese
Coefficient Based on the Relative Humidity Measured at the Back Face of the Sample
69. Skelland AHP. Diffusional mass transfer. John Wiley & Sons Inc: Hoboken, United
70. Kays WM; Crawford ME; Weigand B. Convective heat and mass transfer. McGraw-Hill
71. Al Haddad M; Mounir S; Sobolik V; Allaf K. Fruits and vegetables drying combining hot
air, DIC technology and microwaves. International Journal of Food Engineering 2008,
4(6).
42
72. Mounir S; Allaf K. Study and modeling of dehydration and rehydration kinetics within
73. Albitar N; Mounir S; Besombes C; Allaf K. Improving the drying of onion using the
74. Téllez-Pérez C; Sabah MM; Montejano-Gaitán JG; Sobolik V; Martínez CA; Allaf K.
Downloaded by [Library Services City University London] at 10:51 19 March 2016
42,978-1003.
77. Allaf K; Bouvier J-M; Forget M; Louka N; Parent F. Procédé de traitement de produits
78. Crank J. The mathematics of diffusion, 2 ed. Clarendon Press-Oxford: London, England,
1975; 421 p.
79. Nguyen TH; Lanoisellé JL; Allaf K. Effect of instant controlled pressure drop DIC
80. Srikiatden J; Roberts JS. Measuring moisture diffusivity of potato and carrot (core and
cortex) during convective hot air and isothermal drying. Journal of Food Engineering
2006, 74(1),143-152.
Downloaded by [Library Services City University London] at 10:51 19 March 2016
81. Zielinska M; Markowski M. Air drying characteristics and moisture diffusivity of carrots.
44
Table 1. The limited value of air velocity
(m/s)
[31]
Broccoli Through-bed Hemispherical 0.8-3.5 2.5 Mulet et al.
circulation
[33]
Carrot Cross flow cube of 1cm 0.24-2.17 1.44 Mulet et al.
[36]
Potato Cross flow 1 - cm cube 0.2-6.43 2.05 Rossello et al.
[37]
Potato Through-bed cubes of 1 cm 0.5-10 2 Clemente et al.
flow
45
Table 2.The experiment points with steam pressure P and thermal treatment time t
Apple-DIC 1 0.5 15
Apple-DIC 2 0.35 15
Apple-DIC 3 0.2 15
Downloaded by [Library Services City University London] at 10:51 19 March 2016
Apple-DIC 4 0.15 10
Apple-DIC 5 0.1 10
Carrot-DIC 1 0.5 15
Carrot-DIC 2 0.35 15
Carrot-DIC 3 0.2 15
Carrot-DIC 4 0.15 20
Carrot-DIC 5 0.1 20
46
Table 3. Levels of independent variables of airflow drying (temperature, velocity) in a 2-parameter, 5-level
Coded levels -α -1 0 +1 +α
4
α (axial distance)= 2 k k is the number of orthogonal design variables
47
Table 4. Results of evaluated drying kinetics parameters of apple.
Nº T v DIC1 Control
-10 2 -1 -3 -1 -10 2 -1 -3 -1
min 10 m s % db 10 m s min 10 m s % db 10 m s
0
Downloaded by [Library Services City University London] at 10:51 19 March 2016
4 2
6 2
P 0 5 3 8 4 0 9 5 5
48
td3%: drying time to reach a final water content of 3% db
49
Table 5. Results of evaluated drying kinetics parameters of carrot.
Nº T v DIC1 Control
-10 2 - -3 -1 -10 2 - -3 -1
min 10 m s % db 10 m s min 10 m s % db 10 m s
1 1
4 2
6 2
50
P 0 5 2 3 5 1 5
51
Downloaded by [Library Services City University London] at 10:51 19 March 2016
52
Figure 2. Dryer set up: 1 - Compressed air valve, 2 - Flow valve, 3 - Heater, 4 - Air distributor; 5 -
Drying plateau
Downloaded by [Library Services City University London] at 10:51 19 March 2016
53
Figure 3. Drying curves at air temperature T = 40 °C, and velocity v = 9 m/s of apple raw
54
Figure 4. Drying curves at air temperature T = 40 °C and velocity varying from 1 to 15 m/s of
55
Figure 5. Statistical analysis of the impact of drying parameters; temperature (°C) and velocity
(m/s) on the logarithm of the effective diffusivity of apple at point DIC 1; (Left) General trends,
and (Right) Comparison between experimental results values calculated from the empirical
model.
Downloaded by [Library Services City University London] at 10:51 19 March 2016
56
Figure 6. Statistical analysis of the impact of drying parameters; temperature (°C) and velocity
(m/s) on the starting accessibility of apple at point DIC 1; (Left) Pareto chart (p-value = 0.1), and
(Right) Comparison between experimental results values calculated from the empirical model.
Downloaded by [Library Services City University London] at 10:51 19 March 2016
57
Figure 7. Statistical analysis of the impact of drying parameters; temperature (°C) and velocity
(m/s) on the mass transfer coefficient k of apple at point DIC 1; (Left) General trends, and (Right)
Comparison between experimental results values calculated from the empirical model.
Downloaded by [Library Services City University London] at 10:51 19 March 2016
58
Figure 8. Statistical analysis of the impact of drying parameters; temperature (°C) and velocity
(m/s) on the logarithm of the effective diffusivity of carrot at point DIC 2; Pareto chart (p-
value = 0.1).
Downloaded by [Library Services City University London] at 10:51 19 March 2016
59
Figure 9. Statistical analysis of the impact of drying parameters; temperature (°C) and velocity
(m/s) on the starting accessibility of carrot at point DIC 1; Pareto chart (p-value = 0.1).
Downloaded by [Library Services City University London] at 10:51 19 March 2016
60
Figure 10. Statistical analysis of the impact of drying parameters; temperature (°C) and velocity
(m/s) on the mass transfer coefficient of carrot at point DIC 3; Pareto chart (p-value = 0.1).
Downloaded by [Library Services City University London] at 10:51 19 March 2016
61