Journal of Food Engineering 101 (2010) 8–15

Contents lists available at ScienceDirect

Journal of Food Engineering

journal homepage: www.elsevier.com/locate/jfoodeng

Applying CFD for designing a new fruit cabinet dryer

Y. Amanlou *, A. Zomorodian
Ag. Eng. Dept., Shiraz University, Iran

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

Article history: Cabinet dryers are the most popular equipment for fruit drying. One of the drawback of this dryer can be
Received 20 February 2010 non-uniformity in the desired moisture content of end product. A new cabinet dryer with a side mounted
Received in revised form 31 May 2010 plenum chamber was designed, constructed and evaluated. To obtain a uniform distribution of drying air
Accepted 1 June 2010
flow and temperature taking into account the overall operating conditions, seven different geometries of
Available online 8 June 2010
cabinet dryer were envisaged theoretically (computational fluid dynamics (CFD) by using fluent soft-
ware) and experimentally. Experiments were conducted on the most appropriate sketch with acceptable
Keywords:
uniform air flow and temperature distribution. The experimental results showed that the new cabinet
Green fig
Cabinet drying
dryer illustrated an even distribution of air velocity and temperature throughout the dryer. Comparing
Numerical and experimental investigation the experimental and predicted (extracted for the CFD analysis) data revealed a very good correlation
coefficient of 99.9% and 86.5% for drying air temperature and air velocity in the drying chamber,
respectively.
Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction This valuable tool is capable of analyzing the flow pattern of the air
conditioning system in short span of time, which was previously
Cabinet dryers are the most favorite equipment used in farms impossible from experimental and theoretical methods (Anderson,
for fruit drying. These dryers are simple in structure, low in cost 1995; Yongson et al., 2007).
installation and can be employed in almost any environmental Computational fluid dynamics (CFD) have been widely used for
conditions. Non-uniformity in the moisture content of the end predicting the air velocity and temperature in drying chambers
product is an inherent drawback in applying the cabinet dryer; (Norton and Sun, 2006). With the development of low cost, power-
hence producers are not usually interested in utilizing this drying ful computers, and commercial software packages such as Fluent,
system (Adams and Thompson, 1985; Mathioulakis et al., 1998; Star-CD and CFX in last decade, CFD has been increasingly de-
Mirade, 2003). ployed in the food industry (Scott and Richardson, 1997; Verboven
In conventional cabinet dryers, hot air is usually introduced un- et al., 2000a,b; Xia and Sun, 2002).
der first tray (bottom tray) and passes through the other trays nor- Accuracy of prediction can be strongly improved by including
mally. Therefore the fruits located on the bottom trays would some pertinent physical properties of fruits such as: air flow resis-
receive the highest energy and could be over dried, while the upper tance, kernel and bulk densities and porosity in the study (Fluent
trays may not receive enough energy to be dried due to increase in 6.2 user’s guide, 2005).
drying air relative humidity and decrease in air velocity and tem- The main objectives of this research were:
perature. Dehydration rate shows a very strong relationship with
drying air temperature and velocity (Mulet et al., 1987; Karathanos 1. to investigate the effect of different geometries of cabinet dry-
and Belessiotis, 1997). ing chamber on dryer performance,
Uniform air flow distribution inside dryers is of paramount 2. to compare the various designs by CFD and to select the best
importance because it determines both the efficiency and the design regarding uniform air flow and temperature distribution
homogeneity of the products being dried (Mirade, 2003). Control- throughout the dryer,
ling all of these parameters experimentally is very tedious and dif- 3. to fabricate and evaluate the best design and comparing the
ficult. Although computational fluid dynamics technique (CFD) experimental data with the CFD extracted data.
cannot replace physical experiments completely but it can signifi-
cantly reduce the amount of time needed for experimental works.
2. Materials and methods

* Corresponding author. Tel.: +98 2414244660; fax: +98 7112286104. In conventional cabinet drying system the hot drying air is
E-mail address: [email protected] (Y. Amanlou). completely introduced at the bottom of the plenum chamber.

0260-8774/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jfoodeng.2010.06.001
 Y. Amanlou, A. Zomorodian / Journal of Food Engineering 101 (2010) 8–15 9

Nomenclature

C, D prescribed matrices rk turbulent Prandtl numbers for k

C0, C1 empirical coefficients re turbulent Prandtl numbers for e
C ij prescribed matrices E total energy
Dij mass diffusion coefficient ti velocity vector
q density of fluid t mag velocity magnitude
k turbulent kinetic energy sij eff deviatoric stress tensor
e rate of dissipation p pressure
l dynamic viscosity Prt Prandtl number
lt turbulent viscosity T temperature
Gk generation of turbulent kinetic energy due to the mean cp specific heat capacity at constant pressure
velocity gradients u velocity magnitude in x direction
Gb generation of turbulent kinetic energy due to buoyancy t time
YM contribution of the fluctuating dilatation in compress- Sk ; Se ; Sh user-defined source terms
ible turbulence to the overall dissipation rate Si source term for ith momentum equation
C 1e , C 2e , C 3e constants used in turbulent model

The drying potential of the hot air is gradually diminished as it inside the dryer (Fluent user guide, 2005).There are many turbu-
passes through the upper trays. lent models embedded in commercial codes of CFD and it is left
In order to surmount the problem of non-uniformity in air flow to the user to assert which one is appropriate for the case under
distribution in conventional cabinet dryers, a new design of cabinet study (Norton and Sun, 2006). Among turbulent models, the stan-
dryer was introduced. In this design a side plenum chamber with a dard k  e model still remains an industrial standard and its suc-
trapezoid cross-section was added to conventional structure of a cessful applications are reported in recent literature (Foster et al.,
cabinet dryer. Part of fresh drying air was introduced between 2005; Margaris and Ghiaus, 2006). The standard k  e model is a
the upper trays to keep the air drying potential high and uniformly semi-empirical model based on model transport equations for
distributed within the trays, Fig. 1. It is believed that this change the turbulent kinetic energy (k) and its dissipation rate (e). The
would result in more even air flow distribution and more uniform transport equation for k is derived from the exact equation, while
moisture removal from all the trays simultaneously in the cabinet the transport equation for e is obtained using physical reasoning
dryer. It is supposed that the geometry of side mounted plenum and bears little resemblance to its mathematically exact counter-
chamber could have some crucial effect on hot air distribution in- part (Yongson et al., 2007). The turbulent kinetic energy, k, and
side the drying chamber. Thus evaluating the effect of important its rate of dissipation, e, are obtained from the following transport
geometrical changes on air flow distribution patterns seemed to equations:
be very interesting in the dryer.
  
@ @ @ lt @k
ðqkÞ þ ðqkui Þ ¼ lþ þ Gk þ Gb  qe  Y M þ Sk
2.1. Design of a new cabinet drying chamber @t @xi @xj rk @xj
ð1Þ
In order to find the most appropriate geometrical shape of cab-
inet dryer for achieving more uniform distribution of drying air @ @
flow inside the dryer, after investigating many sketches with dif- ðqeÞ þ ðqeui Þ
@t @xi
ferent shapes and dimensions, seven designs were selected.   
@ l @e e e2
It is obvious that the geometry of trapezoid plenum chamber ¼ lþ t þ C 1e ðGk þ C 3e Gb Þ  C 2e q þ Se ð2Þ
(air inlet size, angle of side plenum chamber with respect to hori-
@xj re @xj k k
zon, divergence angle, a, and width of side plenum chamber)
would affect the air flow pattern in the dryer chamber. In many Convective heat and mass transfer modeling in the k  e models is
conventional dryers installing hood-type ceiling and air deflector given by the following equation (Yongson et al., 2007):
are common, thus the effects of these two parameters were also   
studied (Pabis et al., 1998). Some other important dimensions of @ @ @ cp lt @T
ðqEÞ þ ½ui ðqE þ pÞ ¼ kþ þ ui ðsij Þeff þ Sh
the dryer cabinet could be: air exit size and its location, distance @t @xi @xi Prt @xi
of fruit trays from dryer ceiling. Information about different ð3Þ
sketches is summarized in Fig. 1 and Table 1. Depth of drying
chambers normal to the paper is constant and equal to 30 cm in
2.3. Tray simulation
all cases, Fig. 3.
Fruit trays are porous media for air flow. Porous media are mod-
2.2. Basic governing equations for designing the cabinet dryer
eled by the addition of a momentum source term to the standard
fluid flow equations. The source term is composed of two parts:
There are three categories of basic equations, which are derived
a viscous loss term (the first term on the right-hand side of Eq.
from three basic laws of mass and energy conservation. The mass,
(4)), and an inertial loss term (the second term on the right-hand
momentum and energy conservation result in the continuity equa-
side of Eq. (4))
tion, Navier–Stokes equation and energy equation, respectively
!
(Norton and Sun, 2006). Reynolds number was calculated for the X
3 X
3
1
seven plans at the dryer entrance. In all cases Reynolds number Si ¼  Dij ltj þ C ij qtmag ti ð4Þ
j¼1 j¼1
2
was higher than 2000, thus the air flow showed a turbulent pattern
10 Y. Amanlou, A. Zomorodian / Journal of Food Engineering 101 (2010) 8–15

Fig. 1. Various designs of cabinet dryer envisaged for study.

Table 1
Geometrical parameters and dimensions of the sketches for the new cabinet dryer.

Geometrical parameters a b c d e f g
Air inlet size, depth of 0.3 m (m) 0.05 0.15 0.15 0.15 0.15 0.15 0.15
Air outlet size, depth of 0.3 m (m) 0.05 0.15 0.15 0.15 0.15 0.15 0.15
Location of outlet Ra Cb R C C C C
Divergence angle, a (degree) 0.25 0.337 0.337 0.337 0.337 0.337 0.337
Distance of fruit tray from dryer ceiling (m) 0.15 0.15 0.15 0.15 0.15 0.25 0.10
Side plenum chamber width (m) 0.643 0.45 0.45 0.45 0.45 0.45 0.45
Hood-type ceiling No No No No Yes Yes No
Existence of deflector No No No Yes No No No
a
Left corner.
b
Center of ceiling.

where Si is the source term for the ith (x, y, or z) momentum equation to the pressure gradient in the porous cell, creating a pressure drop that
and D and C are prescribed matrices. This momentum sink contributes is proportional to the fluid velocity (or velocity squared) in the cell.
 Y. Amanlou, A. Zomorodian / Journal of Food Engineering 101 (2010) 8–15 11

Fluent 6.0.12 allows the source term to be modeled as a power  Inlet: three air velocities (1, 2 and 3 m s1) and air temperatures
law of the velocity magnitude: (20, 30 and 40 °C) were selected. Direction of air flow was nor-
mal to air inlet.
Si ¼ C 0 jtjC 1 ð5Þ  Outlet: assuming gauge pressure = 0 at the outlet, Fluent
where C0 and C1 are user-defined empirical coefficients which are extrapolated the required information from the interior of dry-
related to bed pressure drop of the porous media (Fluent 6.0.12 ing chamber.
user’s guide, 2005). In the present study for defining C0 and C1, a test  Porous media: empirical parameters of pressure drop equation
rig was designed and fabricated in which the pressure drop of dry- and fruit bed porosity were defined (Amanlou and Zomorodian,
ing fruit bed was measured (Amanlou and Zomorodian, 2010). The 2010).
test rig is made of a smooth PVC cylinder with 40 cm inside diam-  Wall: heat transfer coefficients of the chamber walls and envi-
eter. Air velocity and temperature can be adjusted and monitored in ronmental conditions were defined. Thermal conductivity, spe-
the test rig. The pressure drops across a thin layer (2.5 cm depth) of cific heat and density
 of the chamber
 wall were assumed to be:
 
green figs (5.74% db initial moisture content) for random filling 0:16 mWK , 1255 kgJ K and 720 mkg3 , respectively (Incropera and
method were measured. Porosity of figs bed was measured to be De Witt, 1990). Ambient temperature and relative humidity
50.61%. The quantity of C0 and C1 were also calculated to be 0.029 were 15 ± 1 °C, and 30 ± 2%, respectively, in all cases.
and 0.6849, respectively (Amanlou and Zomorodian, 2010).
2.5. Experimental apparatus
2.4. Simulation details
The simulation results of the seven different designs of the new
To solve the governing equations, initial and boundary condi- cabinet drying chamber were compared with each other, Fig. 1, and
tion must be defined around the boundary of system (domain). the best design with the most uniform air temperature and air flow
Since the equations are highly non-linear, they are not solvable distribution for the drying chamber was chosen. In order to verify
by explicit, closed-form analytical methods. The numerical finite the CFD simulation results, the best design was fabricated and
volume method as used in Fluent 6.2 (Fluent6: user’s guide, evaluated under different operating conditions. Figs. 2 and 3 show
2005) has been used for solving the equations on a PC P IV the experimental layout of drying chamber and testing apparatus,
3.2 GHz with 1.5 GB random access memory. respectively. A constant speed centrifugal fan (1400 rpm, 50 Hz,
All the geometrical configurations were displayed in Fig. 1. They made in Italy) was used as an air flow source. A damper was in-
were used to build up a numerical model based on an unstructured serted at the fan outlet to alter the air flow rate. The air flow was
two-dimensional mesh by tetrahedral cell. Table 2 shows the num- measured using a hot wire anemometer (Lutron, ±0.1 m s1, made
ber of tetrahedral cells for each sketch. Double precision solver and in Taiwan) located far enough from the fan outlet in a PVC pipe
segregated method were chosen. In this investigation the pattern (15 cm i.d.) connected to plenum chamber. Inlet air temperature
of air flow during drying process was important and since there to the drying bed was precisely controlled by a thermostatic sensor
was no variable condition in the study, the case was simulated as hanging just before the air was introduced into the dryer chamber
steady state condition. (Atbin, ±0.1 °C, made in Iran). An electrical heating unit (6 kW) was
Setting up of flow-simulating computation involves specific attached to the fan inlet to regulate the air temperature introduced
boundary conditions, in particular at surfaces bounding of the do- into the dryer.
main. In this study various boundary conditions were defined as A long diffuser with rectangular end cross-section was used at
followings: the drying chamber air inlet. Six flow straighteners were also in-
stalled in the attached plenum chamber (three at the floor and
three at the ceiling). These two accessories would help to have a
Table 2 more uniform and straight air flow in the plenum chamber. There
The number of tetrahedral cells for seven sketches.
was no inclination or geometrical changes perpendicular to the air
Sketches Number of cells Number of nodes flow direction in the plenum chamber. These schemes enable us to
a 7002 3834 simplify the simulation procedures and applying two-dimensional
b 6676 3660 analysis for this study.
c 6676 3660 Green fig (Ficus carica L.) was chosen as a drying fruit and was
d 6664 3653
randomly and uniformly spread on perforated stainless steel trays
e 6302 3466
f 6814 3727
(2.5 cm in thickness). Three trays were tightly embedded in the
g 5528 3023 dryer chamber. This horticultural product is grown mostly in Iran,
Turkey and Afghanistan and has been one of the important non-oil

Fig. 2. Schematic diagram of the apparatus used in this research. (A) Electrical heating unit, (B) centrifugal fan, (C) damper, (D) hot wire anemometer, (E) thermostat, (F)
diffuser, (G) thermostatic sensor, (H) side plenum chamber, (I) main plenum chamber, (J) fruit trays, and (K) air exist.
12 Y. Amanlou, A. Zomorodian / Journal of Food Engineering 101 (2010) 8–15

agricultural export commodities in the last three decades in Iran. It were made of pressed wood with the highest dimensional preci-
is widely used in confectionery, snack foods and pastry industries sion and all possible leaks were prevented. Experiments were car-
(Doymaz, 2005). ried out at three levels of inlet air temperature (20, 30, and 40 °C)
The best sketch was fabricated in the Department of Agricul- and three levels of inlet air velocity (1, 2, and 3 m s1). Each exper-
tural Engineering, Shiraz University, Shiraz, Iran. Chamber walls iment was replicated three times.

Fig. 3. Experimental layout of fabricated new drying chamber and testing apparatus.

Fig. 4. Drying air temperature distribution profiles for the seven designs (CFD simulation results), inlet air temperature = 313 K.
 Y. Amanlou, A. Zomorodian / Journal of Food Engineering 101 (2010) 8–15 13

2.6. Data acquisition procedures analysis were compared with the experimental data. Three statis-
tical criteria namely coefficient of determination (R2), root mean
For measuring the air velocity and temperature during each square error (RMSE) and mean relative percentage deviation mod-
test, at different locations of the drying chamber, 18 points on ulus (P%) were used as indices of goodness of CFD simulation out-
the side wall of the drying chamber were pierced (Figs. 2 and 3). put (Snedecor and Cochran, 1989).
All holes except the one through which hot wire probe was in-
serted for air velocity measurement, were filled tightly by conic
plastic washers. Inlet air temperature and velocity were kept con- 3. Results and discussion
stant in each set of experiment. For reading air temperature and
velocity at each point inside the drying cabinet, hot wire probe 3.1. A suitable design for the new cabinet dryer
was inserted through the hole and adjusted at three different equi-
distant locations along the depth of the drying chamber. The aver- To decide about the suitability of the design of the new cabinet
age of three readings with three replications was selected as the dryer, data analysis of seven proposed plans were carried out using
experimental data for a single point. The corresponding values of CFD. The most feasible design can be assessed by comparing the
air temperature and velocity of 18 holes extracted from the CFD uniformity of air flow distribution in the drying chamber as done

Fig. 5. Air velocity distribution profiles for the seven designs (CFD simulation results) at inlet air velocity = 3 m s1.
14 Y. Amanlou, A. Zomorodian / Journal of Food Engineering 101 (2010) 8–15

by many other researchers (Mulet et al., 1987; Karathanos and ber to the air exit. This canalization can be attributed not only to
Belessiotis, 1997; Mathioulakis et al., 1998). The temperature dis- the small size of air inlet and exit in comparison with chamber
tribution profiles for the seven designs are shown in Fig. 4. dimensions but also due to small divergence angle of a. Therefore,
Hot air distribution is poor in the design (a) because air inlet design (a) is not recommended for dryer chamber. In design (b),
and exit are small in comparison with the tray width. Hood-type increasing the divergence angle a improves the air velocity distri-
ceilings in the designs (e and f), introducing a deflector into the bution and transferring the air exit location to the center causes a
air inlet in design (d) and lowering the distance of fruit trays from more uniform air velocity distribution at the tray zones. Shifting
dryer ceiling in design (g) could not improve the air temperature the air exit to the right of design (c), would weaken air velocity dis-
distribution in the dryer chamber. Except design (a), other designs tribution at the dryer chamber corners. Introducing an air deflector
preserved maximum temperature close to inlet temperature at the in the sketch (d) created an air flow resistance and caused a de-
trays’ zone. Both of (b) and (c) designs show reasonable uniform air crease in air velocity at the dryer right and bottom corner at the gi-
temperature distribution in the drying chamber, Fig. 4. The average ven inlet air velocity. Hood type air exit (designs e and f) could not
air temperature of different designs at 18 points was calculated. improve the velocity distribution in comparison with design (b),
Comparing the results showed that the drying air temperatures (Fig. 5), therefore design b is worth because its structure is simple
in (b) and (c) designs were higher than the air temperatures of and can manufacture more economically then e and f designs. The
the other designs at the tray zones.Changing the geometrical shape designs (e and f) were discarded. Shorter ceiling height and reduc-
of drying chamber displayed some more influence on air velocity ing the inlet-exit distance resulted in disbanding uniform air flow
distribution than on air temperature distribution (Fig. 5). In design distribution in the design (g). Minimum air velocities at the tray
(a), the air flow canalizes through the bottom of the drying cham- zones have been happened in designs (a and d). Therefore the de-
sign (b) seems to be the best taking into account all considerations
such as the simplicity in structure, uniformity in air temperature
and velocity distribution.

3.2. Comparison between the experimental and the CFD analysis data

The experimental results obtained from 18 points of drying

chamber for both air temperature and velocity were compared
with the corresponding results extracted from the CFD analysis
for the best proposed plan (b). The results are reasonably accept-
able and illustrated in Figs. 6 and 7.
The results of statistical analysis show good correlations be-
tween the predicted CFD and the experimental data. This can be
proved by observing the highest R2 and the lowest RMSE and P%
values for different air temperatures and velocities, Figs. 6 and 7.
Indeed knowing the physical properties of drying fruit (pressure
drop and porosity), the k  e model was a very powerful tool for
investigating the air velocity distribution patterns inside the new
drying chamber. Moreover adding a porous media term to the
momentum equation did help a true simulation of drying chamber.
Analysis of experimental data in direction of depth of dryer cham-
ber did not show a significant variation, so two-dimensional CFD
model was propitious for this study.
Fig. 6. Comparison of the results of experimental and CFD predicted drying air
temperature.
4. Conclusions

 Computational fluid dynamic (CFD) was a very powerful tool for

the cabinet dryer system analysis.
 The side mounted plenum chamber improved conventional
cabinet dryer’s performance.
 The standard k  e model using empirical data (fruit trays pres-
sure drop and porosity) showed very accurate results.
 Design (b) was shown to be the most suitable design for obtain-
ing uniform air temperature and velocity distributions in the
drying chamber.
 Sharp angle, a, do not recommend for the side mounted plenum
chamber.
 Comparing the experimental and predicted data (extracted
from the CFD analysis) revealed good correlation coefficients
of 99.9% and 86.5% for drying air temperature and air velocity
in the drying chamber, respectively.

References

Adams, R.L., Thompson, J.F., 1985. Improving drying uniformity in concurrent flow
Fig. 7. Comparison of the results of experimental and CFD predicted air velocity. tunnel dehydrators. Transactions of ASAE 28 (3), 890–892.
 Y. Amanlou, A. Zomorodian / Journal of Food Engineering 101 (2010) 8–15 15

Amanlou, Y., Zomorodian, A., 2010. Evaluation of air flow resistance across a green Mulet, A., Berna, A., Borras, M., Pinaga, F., 1987. Effect of air flow rate on carrot
fig bed for selecting an appropriate pressure drop prediction equation. Food and drying. Drying Technology 5 (2), 245–258.
Bioproducts Processing. doi:10.1016/j.fbp.2010.03.011. Norton, T., Sun, D.W., 2006. Computational fluid dynamics (CFD) – an effective and
Anderson, J.D., 1995. Computational Fluid Dynamics. International Edition. efficient design and analysis tool for the food industry: a review. Trends in Food
McGraw-Hill, New York. Science and Technology 17 (2006), 600–620.
Doymaz, I., 2005. Sun drying of figs: an experimental study. Journal of Food Pabis, S., Jayas, D.S., Cenkowski, S., 1998. Resistance to air flow through grains and
Engineering 71, 403–407. oilseeds. In: Grain Drying: Theory and Practice. John Wiley and Sons, Now York,
Fluent 6.2 User’s Guide, 2005. Fluent Inc. pp. 229–246.
Foster, A.M., Madge, M., Evans, J.A., 2005. The use of CFD to improve the Scott, G., Richardson, P., 1997. The application of computational fluid dynamics in
performance of a chilled multi-deck retail display cabinet. International the food industry. Trends in Food Science and Technology 8, 119–124.
Journal of Refrigeration 28, 698–705. Snedecor, G.W., Cochran, W.G., 1989. Statistical Methods. Iowa State University
Incropera, F.P., De Witt, D.P., 1990. Fundamentals of Heat and Mass Transfer. John Press, USA.
Wiley and Sons, New York. Verboven, P., Scheerlinck, N., De Baerdemaeker, J., Nicolaï, B.M., 2000a.
Karathanos, V.T., Belessiotis, V.G., 1997. Sun and artificial air drying kinetics of some Computational fluid dynamics modeling and validation of the temperature
agricultural products. Journal of Food Engineering 31 (l), 35–46. distribution in a forced convection oven. Journal of Food Engineering 43, 61–73.
Margaris, D.P., Ghiaus, A.G., 2006. Dried product quality improvement by air flow Verboven, P., Scheerlinck, N., De Baerdemaeker, J., Nicolaï, B.M., 2000b.
manipulation in tray dryers. Journal of Food Engineering 75, 542–550. Computational fluid dynamics modeling and validation of the isothermal
Mathioulakis, E., Karathanos, V.T., Belessiotis, V.G., 1998. Simulation of air airflow in a forced convection oven. Journal of Food Engineering 43, 41–53.
movement in a dryer by computational fluid dynamics: application for the Xia, B., Sun, D.W., 2002. Applications of computational fluid dynamics (CFD) in the
drying of fruits. Journal of Food Engineering 36 (1998), 183–200. food industry: a review. In: Sun, D.W. (Ed.), CFD Applications in the Agri-Food
Mirade, P.S., 2003. Prediction of the air velocity field in modern meat dryers using Industry (Special Issue) Computers and Electronics in Agriculture, 34(1–3), 5–24.
unsteady computational fluid dynamics (CFD) models. Journal of Food Yongson, O., Badruddin, I.A., Zainal, Z.A., Narayana, P.A.A., 2007. Airflow analysis in
Engineering 60 (2003), 41–48. an air conditioning room. Building and Environment 42 (2007), 1531–1537.