Journal of Food Engineering 338 (2023) 111250

Contents lists available at ScienceDirect

Journal of Food Engineering

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

Effects of thermal boundary conditions on the performance of spray dryers

Sara Moradi Maryamnegari a, *, Ali Ashrafizadeh a, Egbert Baake b, Mattia Guglielmi b
a
DOS Computational Lab, Faculty of Mechanical Engineering, K. N. Toosi University of Technology, Iran
b
Institute of Electrotechnology, Leibniz University of Hannover, Germany

A R T I C L E I N F O A B S T R A C T

Keywords: In this paper, three kinds of wall conditions are numerically simulated to investigate deposition, powder recovery
Spray drying and energy aspects of spray dryers. A co-current dryer with a pressure nozzle is chosen as a base dryer and two
Energy performance dryers with fully insulated, and cooled walls are compared with the base one. Governing equations of a transient
Deposition
flow field are solved through the Eulerian approach, while Lagrangian particle tracking predicts particles’ mo­
Skim milk
Sticky point curve
tion. The sticky point curve is employed as a criterion of bouncing or stickiness to model the deposition pattern of
CFD simulation skim milk particles on the surfaces. Results show that the spray dryer with cooled walls has better deposition
characteristics, while the dryer with insulated surfaces works with higher drying efficiency. Overall, it is
observed that the wall conditions can be changed to improve the drying efficiency and wall deposition. However,
changing the thermal boundary conditions does not seem to be effective in improving the powder recovery.

(continued )
Nomenclature
A Surface (m2 ) Sc Schmidt number ( − )
aw Water activity ( − ) Sm , SF , Sh Source terms
b Constant in GAB model ( − ) t Time (s)
CD Drag coefficient ( − ) T Temperature (K)
Cp Specific heat capacity (J/Kg.K) uj Velocity components (m /s)
C Concentration (mol/m3 ) V Volume (m3 )
D Diffusivity (m2 /s) w Mass fraction
D Deposition (%), Thermophoretic coefficient ( − ) Y dissipation
DE Drying efficiency (%) X Moisture content (kg/kg)
d Diameter (m) xj Cartesian direction
E energy (J) Greek symbols
F Force (N) α Volume fraction ( − )
G Generation of kinetic energy (kg/m.s3 ) μ Dynamic Viscosity (Kg/m.s)
g Gravity acceleration (m/s2 ) ρ Density (kg /m3 )
h Heat transfer coefficient (W/m2 K) ω Turbulent dissipation (1 /s)
hm Mass transfer coefficient (m/s) η Thermal loss factor ( − )
H Enthalpy (J/kg) ε Relative error ( − )
η Thermal loss factor (%) δ Absolute error (%)
k Turbulent kinetic energy Subscripts
k Constant in GAB model ( − ),ratio of specific heat of solute to water ( − ) cell Cell
m Mass (kg) d Droplet
Nu Nusselt number ( − ) D Drag
P Pressure eff Effective
Pr Prandtle number ( − ) ev evaporation
PR Powder recovery (%) f Fluid
Re Reynolds number ( − ) F Related to momentum
Sh Sherwood number ( − ) g Glass transition temperature (K)
h Related to energy
(continued on next column)
(continued on next page)

* Corresponding author.
E-mail address: [email protected] (S. Moradi Maryamnegari).

https://doi.org/10.1016/j.jfoodeng.2022.111250
Received 6 May 2022; Received in revised form 8 August 2022; Accepted 13 August 2022
Available online 20 August 2022
0260-8774/© 2022 Elsevier Ltd. All rights reserved.
S. Moradi Maryamnegari et al. Journal of Food Engineering 338 (2023) 111250

(continued ) able to redesign the system and reduce the exergy destruction. Several
k Related to turbulent kinetic energy other studies have also been carried out in the category of heat recovery
l solute from the exhaust air claiming reduction of energy consumption in spray
m Related to mass dryers (Moejes et al., 2018; Ai et al., 2016). Optimization of process
p Particle
P Pressure gradient
parameters is the essence of second category of methods for improving
t Turbulent the energy efficiency of dryers. Cheng et al. (2018) examined factors
tot total relating to the feedstock and heating system to analyze thermal effi­
T Thermophoretic ciency of spray dryers. Their calculations showed that increasing the
w water
solid concentration and temperature of the liquid feed improve the
λ Thermal diffusivity
ω Related to dissipation thermal efficiency. Further studies was carried out to determine the
effects of process parameters on energy efficiency in systems with
air-to-air heat exchangers and recirculation of exhaust air (GolmanW.J.,
1. Introduction 2014). The investigations show that increasing slurry feed rate, and
decreasing slurry feed concentration, inlet air temperature and flow
The performance of spray dryers can be examined by evaluation of rate, improve the energy efficiency of a spray dryer with recirculation of
powder recovery, deposition, and energy consumption. These three exhaust air. Moreover, they developed a simulation code via Visual Basic
significant factors have been individually investigated in many studies of Applications (VBA) within Microsoft Excel (Golman and Yermu­
to find a solution for enhancing the dryers’ efficiency. A survey of khambetova, 2019). The third category is a new method based on the
literature on this subject is written as follows. monodisperse atomization for forming droplets with less energy con­
From a deposition perspective, the accumulation of particles on the sumption (Jaskulski et al., 2020). This kind of atomization is based on
spray dryer chamber is an unfavorable phenomenon that leads to safety ink-jet technology, which can atomize high viscous fluids. A liquid
hazards (e.g. fire outbreak), and waste of powders (Keshani et al., feedstock with a maximum viscosity of 350 mPa.s can be sprayed within
2015a). Many researchers tackle this issue by optimizing the input pa­ conventional nozzles i.e., pressure and rotary atomizers; however,
rameters such as inlet air temperature, air and feed mass flow rates, and state-of-the-art monodisperse nozzles can spray fluids with viscosity up
additives (Adhikari et al., 2009a; Chegini and Ghobadian, 2007). For to 500 mPa.s. It means that feedstocks with high solid concentrations
example, results of some studies show that adding protein to an initial can be atomized; therefore, less energy is consumed for evaporating of
feed reduces the stickiness of particles and simultaneously increases the water (Atuonwu and Stapley, 2017). A great deal of energy loss from
quality of sugar-rich materials (Adhikari et al., 2009b). The second chamber walls has been reported in publications (Jaskulski et al., 2017).
possible remedy for the deposition problem is to change the chamber Therefore, in the fourth category of methods, it is suggested that dryers
configuration. Using a parabolic shape of the bottom section of the dryer with thermally-insulated chambers may have higher energy efficiency
leads to the flow field stabilization in the conical section, and conse­ (Birchal et al., 2006). Mezhericher (2008) compares fully-insulated and
quently reduction of the deposition (Keshani et al., 2015b). As another non-insulated chambers using 3D simulations of industrial-scale dryers.
attempt to prevent excessive deposition, a new design of dryer with a It is found that thermal wall boundary condition affects the temperature
vapor-permeable membrane was introduced in (Soltani et al., 2015). distribution and vapor mass fraction, while it has no influence on the
The second factor for evaluating the dryers’ efficiency is powder velocity flow field near walls. Roustapour et al. (2013) simulated a 2-D
recovery. Due to the importance of this factor, several experimental and co-current spray dryer with a cooling air jacket to examine the flow
numerical studies have been carried out to increase collected powder pattern and temperature field within the dryer.
(Adhikari et al., 2009c; Truong et al., 2005). For instance, the impact of Given the above literature review, it seems that a thorough investi­
five carrier agents (maltodextrin, gum Arabic, starch sodium octenyl gation on the effects of chamber’s boundary conditions on the perfor­
succinate, whey protein concentrate, and egg albumin) on the powder mance of spray dryers has not been carried out or published yet.
yield of persimmon pulp powders indicate that all agents improve the Therefore, the aim of this study is to investigate the performance of
powder recovery (Du et al., 2014). Langrish et al. (2020) proposed spray dryers with fully-insulated and cooled walls from deposition,
geometrical modifications of traditional dryers with a box-shape powder recovery, and energy efficiency point of views. The sticky point
chamber. They found that powder recovery of skim milk with 30% curve for skim milk is used for modelling the deposition on chamber
concentration increased by 13.5% in the new design. walls. Numerical simulation results in this study provide some useful
From an energy point of view, spray drying is an intensive energy information which can be used by designers and users of spray dryers.
consuming technology, in which vast amount of water evaporates from
liquid feedstock to produce dry particles. Therefore, several studies have 2. Mathematical model
been carried out to improve the thermal efficiency of spray dryers
(Petersen et al., 2017; Khuenpet et al., 2016). The first category includes The Eulerian-Lagrangian (EL) multiphase approach is used for solv­
the attempts which aim to improve the thermal efficiency of spray dryers ing the flow field in the dryer. In this method, hot air and particles/
by recovering energy from the exhaust gases. For example, Golman and droplets are considered as a continuous phase and dispersed phase,
Julklang (2014) analyzed energy efficiency and energy saving of two respectively. The interactions between the gas and particles/droplets are
industrial dryers; one of them was equipped with an air-to-air heat coupled via the Particle_Source_in Cell (PSI_CELL) algorithm (Crowe,
exchanger and the other one had a system for recirculating exhaust 1983). In this algorithm, source terms are considered in all equations in
gases. Overall, the second approach showed a better performance than the continuous phase. By tracking particles, source terms are updated in
the first one, i.e. 34% increase in energy efficiency compared to 16% every time step.
increase. In this line of thought the sustainability assessment, energy and
exergy analysis of a spray dryer with a heat exchanger unit were carried 2.1. Continuous phase equations
out for five different dead state temperatures ranging from 0 ◦ C to 20 ◦ C.
It was found that the efficiency of the heat exchanger is higher than the The mass, momentum, and energy conservation equations for
spray dryer, so that combination of the heat exchange and spray dryer incompressible and transient flow are introduced as follows (Ludwig
increases the total efficiency of the system (Caglayan and Caliskan, et al., 2022):
2017). In another study, Caglayan et al. (2018) stated that evaluation of
∂( ) ∂ ( )
thermo-ecologic parameters of a spray dryer was beneficial in the energy ρ + ρ u =0 (1)
∂t f ∂xj f j
and environment field and by calculating these parameters, they were

2
S. Moradi Maryamnegari et al. Journal of Food Engineering 338 (2023) 111250

√̅̅̅̅̅̅̅̅
∂(
ρu +
) ∂ ( )
ρ uu = −
∂P ∂τij
+ + ρf g i + S F (2) Here, ζ, and úi új are a normally distributed random number, and
∂t f i ∂xj f i j ∂xi ∂xj local root mean square, respectively.
[ ( )] For computing the characteristic lifetime (τe), the Lagrangian inte­
∂( ) ∂ ( ) ∂ ∂T gral time (TL) is used as shown in the following expression:
ρ cp T + ρ cp uj T = λeff + Sh (3)
∂t f ∂xj f ∂xj ∂xj
τe = − TL log(r) (11)
In Eq. (1), x, ρ, and u are position, density, and velocity, respectively.
In the momentum equation, P, τij, g, and SF, are pressure, stress tensor, In Eq. (11), r is a random number between 0 and 1. For k-ω turbulent
gravity, and source term, respectively. In the energy equation, T, Cp, λeff, model, TL = 0.3/ω.
and Sh are temperature, specific heat transfer, effective thermal diffu­
sivity, and source term, respectively. 2.3.1. Particle diameter distribution
The distribution of droplet size is obtained based on the Rosin-
2.2. Turbulence model Rammler Diameter Distribution method. This method is based on an
exponential equation, which predicts the droplet diameter and mass
The turbulence model for simulating the flow field in the chamber fraction of droplets:
should be carefully selected since inappropriate models result in phys­
Yd = e− (d/d)n
(12)
ically unacceptable solutions. In this study, k-ω SST is used for the
prediction of flow turbulence effects since investigations show that this
, where d is the mean diameter, n is spread parameter, and Yd is the mass
model was slightly closer to the benchmark data (Jubaer et al., 2019).
fraction of droplets (Ludwig et al., 2022).
The governing equations in this model are as follows:
(( ) )
∂( ) ∂ ( ) ∂ μ ∂k
ρf k + ρf uj k = μ+ t +G̃ k − Yk + Sk (4) 2.4. Drying kinetic model
∂t ∂xj ∂xj σk ∂xj
(( ) ) The Characteristic Drying Curve (CDC) model is adopted to calculate
∂( ) ∂ ( ) ∂ μt ∂ω
ρω + ρ uj ω = μ+ + Gω − Yω + Dω + Sω (5) the moisture evaporation from particles. Based on this model, a
∂t f ∂xj f ∂xj σω ∂xj
nonlinear coefficient corrects the drying rate. This correction coefficient
is a function of the particle moisture content and is defined as follows
, where k and ω are turbulent kinetic energy and turbulent dissipation,
(Hussain et al., 2022):
respectively. Also, μ and μt are dynamic viscosity and turbulent viscos­
ity, respectively. G
̃ k and Gω denote the generation of turbulence kinetic dmd ( )
= − hm Af Ci,s − Ci,f (13)
energy due to the mean velocity, and the generation of dissipation, dt
respectively. Yk, and Yω are dissipation of k and ω due to turbulence. In ⎧
addition, Sk, and Sω are source terms for turbulent kinetic energy and ⎨1
⎪ if X ≥ Xcr
( )n
dissipation, respectively. f = X − X eq (14)
⎪
⎩ if X < Xcr
Xcr − Xeq
2.3. Discrete phase motion
, where hm, A, and f are the mass transfer coefficient, droplet surface, and
Particle trajectories are calculated by Newton’s second law as correction factor, respectively. Ci,s, and Ci,f are mass concentrations at
expressed in the following equation (Ludwig et al., 2022): the droplet surface, and the bulk air, respectively. X is the moisture
( ) content and subscripts cr and eq refer to the moisture contents at the
dup ( ) ρp − ρ gi critical and equilibrium states, respectively. Tran and et all. (Tran et al.,
= FD u − up + + FT + Fp (6)
dt ρp 2017) developed a model for the second drying stage, and “n” is
calculated by a linear function based on the velocity of the surrounding
, where FD, FT, and FP are drag, thermophoretic, and pressure gradient air.
forces, respectively. Forces in Eq. (6) are mathematically modeled as The equilibrium moisture content is retrieved from the Guggenheim-
follows (Zolfagharnasab et al., 2021): Anderson-de Boer (GAB) model (Jaskulski et al., 2017):
πdp2 ⃒ ⃒( ) Xm bkaw
FD = CD ρf ⃒u − up ⃒ u − up (7) Xeq = (15)
8 1 − kaw [1 + b − kaw ]

FT = − DT,p
1
∇T (8) , where b and k are the temperature-dependent constants, aw is the water
mp T activity, and k is the ratio of vapor density of bulk air to the vapor
density at saturation state. Furthermore, Xm is the monolayer moisture
πdp3 content equal to 0.06156 for skim milk. The Ranz-Marshall correction is
FP = ρ f up (9)
6 used to determine the heat transfer coefficient:
, where CD is the drag coefficient, which depends on the flow regime Nu = 2 + 0.6Re1/2 Pr1/3 (16)
(Arjmandi et al., 2022). Moreover, DT,p is the thermophoretic
coefficient. Sh = 2 + 0.6Re1/2 Sc1/3 (17)
The size of the particles is in the order of microns, so the effect of the
turbulence dispersion of the continuous phase should be considered on In Eq. (16) and Eq. (17), Nu, Re, Pr, Sh, and Sc are Nusselt, Reynolds,
the particles. In this study, the Discrete Random Walk (DRW) model Prandtl, Sherwood, and Schmidt number, respectively.
based on the stochastic model determines the fluctuating continuous By calculating the mass transfer from the particles, the heat transfer
equation can be solved using the energy balance.
phase velocity (u ) by using the Gaussian probability distribution as
′

shown in Eq. (10) (Gao et al., 2012): dTd ( ) dmd

md Cp,d = hA Tf − Td + ΔHv (18)
√̅̅̅̅̅̅̅̅ dt dt
úi = ξ úi új (10)
, where t, h, and ΔHv are time, heat transfer coefficient, and latent heat of

3
S. Moradi Maryamnegari et al. Journal of Food Engineering 338 (2023) 111250

vaporization, respectively.

2.5. Dryer’s performance measures

The percentage of Deposition (D), Powder Recovery (PR), Drying

Efficiency (DE), and thermal loss factor (η) are quantities that are
commonly used to describe the performance of spray dryers. These
factors are introduced in Eqs. (19–22) (Khuenpet et al., 2016; Woo,
2017):
mwall
D = × 100 (19)
msprayed

, where mwall and msprayed refer to the mass of particles on walls and total
mass of sprayed solids, respectively.
moutlet
PR = × 100 (20)
msprayed

, where moutlet is the mass of produced powder.

Eev
DE = × 100 (21)
Ein
In Eq. (21), Eev and Ein are consumed energy for evaporation of
moisture content and supplied energy for drying, respectively.
Eloss
η= (22)
Etot
Here, Eloss and Etot present the waste energy and total supplied en­
ergy, respectively.
The absolute (δ) and relative errors (ε) in the error analysis sections
are defined as follows:
δ = |xi − x| (23)
⃒x − x⃒
⃒ ⃒
ε=⃒ i ⃒ × 100 (24)
x

, where xi, and x are measured and actual values, respectively.

Fig. 1. Configuration of dryers.

3. Model implementation

The numerical setups, and assumptions for the present work are shows the configuration of dryers in detail.
discussed in the following section.
3.3. Grid generation
3.1. Simulation assumptions
The computational domain is discretized with hexahedral grids in
the ICEM-CFD software. This structured mesh generation technique
In the present study, several assumptions are considered, which are
creates a high-quality mesh and makes the control of the element size
reliable in this simulation and reduce the complexities of the flow field
much easier. As summarized in Table 1, two important criteria for
within the chamber.
hexahedral grid structure i.e., quality and skew are checked. The quality
metric is normalized and vary between 0 (worst quality) and 1 (best
• The interaction between particles within the chamber is neglected.
quality) (ANSYS ICEM CFD, 2021). The grid is refined in areas with high
• The particles deposit on the clean wall.
heat and mass transfer rates i.e., in the core center as well as near the
walls. The mesh structure for the cooled-wall dryer with focus on
3.2. Geometry boundary layers is illustrated in Fig. 2.

In this study, a modified version of the Mobile Minor spray dryer is

3.4. Boundary conditions
chosen for the simulation. Simulated dryers have a cylinder-on-cone
design with diameter of 0.8 m and height of ∼ 2 m. Regarding the
The sticky point curve of skim milk is employed to model the particle
geometrical specifications, more details can be found in the paper of
Kota and Langrish (2007). The configurations of the base dryer and
insulated-wall dryer are exactly the same, while the cooled-wall dryer is Table 1
Mesh resolution and quality measures.
covered with double walls around the cylindrical as well as conical
sections. In the cooled-wall dryer, cooling air enters from the bottom of Case number Number of Quality Skew
elements (0.95–1)
the dryer (conical section) and then exits from the top of the dryer. A
pressure nozzle is located at the top of dryers to atomize the skim milk Base design/insulated-wall 409Κ >0.61 87%
chosen as the liquid feedstock. Hot air enters from top of dryers with dryer
Cooled-wall dryer 530K 88.2%
vane angle of 0◦ i.e., there is no swirl for agent gas in the entrance. Fig. 1
>0.61

4
S. Moradi Maryamnegari et al. Journal of Food Engineering 338 (2023) 111250

Table 2
Initial and boundary conditions.
Continuous phase

Inlet mass flow rate (kg/s) 0.025

Inlet mass flow rate for cooled wall 0.05
(kg/s)
Inlet temperature (◦ C) 100, 140, 180
Inlet temperature of cooled wall (◦ C) 25
Free stream temperature (◦ C) 27
Inlet pressure (Pa) 0
Turbulent intensity (%) 0.2
Turbulent length scale (m) 0.02
Fig. 2. Generated mesh in a dryer with double walls. Heat flux for insulated wall (W/m2) 0
Wall thickness (mm) 2
DPM wall boundary condition The Sticky Point Curve through a UDF (
deposition on walls. This curve has been obtained through an experi­ Appendix A)
ment for specified materials and is associated with the glass transition Wall material Stainless steel
temperature. The curve is described by the Gordon- Taylor equation Convection heat transfer coefficient 3.5
(W/m2•K)
given below:
Heat flux for insulated wall (W/m2) 0
wl ⋅Tgl + K⋅ww ⋅Tgw
Tg = (21)
wl + K⋅ww
Table 3
, where wl, and ww are the mass fractions of the solute and water. Tgl, and Injection properties.
Tgw are the glass temperatures of the two mentioned components.
Discrete phase
Furthermore, K is the ratio of specific heat of solute to water. The sticky
point temperature is about 20 ◦ C above the glass transition temperature Injection velocity (m/s) 20
Mass flow rate (kg/s) 0.0008
(Hennigs et al., 2001). Based on this criterion, and the level of temper­
Mean diameter (μm) 32
ature and moisture content, particles may rebound or deposit on the Spread parameter 1.33
walls. Particles below the curve retain their momentum and rebound Number of streams 10
from walls, otherwise, they stick to the walls. The areas of sticky and Solid concentration (%) 8.8
non-sticky particles are shown in Fig. 3. Boundary conditions and in­ Spray angle 76◦
Injection type cone
jection properties for continuous and discrete phase are summarized in Material Multicomponent
Table 2 and Table 3, respectively.

3.5. Material
Table 4
Thermo-physical properties.
In this study, the skim milk with five components i.e., protein, fat,
carbohydrate, ash, and water is considered as the material to be dried. properties values

The thermo-physical properties of air and water in liquid and vapor Air (Afshar et al., 2019)
states are given in Table 4. Furthermore, the temperature-dependent Density (kg/m3) − 10− 10Τ4+2 × 10− 7Τ3− 10− 4Τ2+0.0206Τ
Specific heat capacity − 4 × 10− 8Τ4+6 × 10− 5Τ3− 0.036Τ2+9.94Τ
properties of solids and the formulation of the mixture properties are
(J/Kg•K)
presented in Appendix B. Thermal conductivity 10− 12Τ4− 10− 9Τ3+4 × 10− 7Τ2+3 × 10− 5Τ
(W/m•K)
Viscosity (m2/s) 4 × 10− 16Τ4− 4 × 10− 13Τ3+9 × 10− 11Τ2+6 × 10− 8Τ
Water in liquid state (Afshar et al., 2019)
Density (kg/m3) − 0.0035Τ2+1.81Τ+765.33
Specific heat capacity 1.85 × 10− 9Τ4− 2.48 × 10− 6Τ3+1.25 ×
(J/Kg•K) 10− 3Τ2− 0.2817Τ+28.07
Water in vapor state (Afshar et al., 2019)
Density (kg/m3) 5 × 10− 11Τ4− 10 ×
10− 7Τ4+10− 4Τ3− 0.045Τ2+9.73Τ− 837.59
Specific heat capacity − 6 × 10− 7Τ4+10− 3Τ3− 0.598Τ2+151.79Τ− 11968
(J/Kg•K)
Thermal conductivity 0.0017Τ2− 2.301Τ+797.35
(W/m•K)
Viscosity (m2/s) 8 × 10− 4Τ2+1.81Τ+765.33

3.6. Solver setup

CFD simulations are carried out using the commercial software

ANSYS FLUENT. Steady-transient strategy is utilized for simulations in
two sections. At first, computations are carried out in a steady-state
mode without injections until a semi-convergence is achieved, then
the model is switched to the transient mode with injections i.e., results of
the steady state flow are used as initial conditions for the transient flow.
This strategy has been recommended for this model (Woo, 2017) since
the computational algorithm converges more easily and the run time
decreases considerably. The coupled procedure is chosen for coupling
pressure and velocity in the transient mode; however, in the steady-state
Fig. 3. The sticky point curve of skim milk (Hennigs et al., 2001).

5
S. Moradi Maryamnegari et al. Journal of Food Engineering 338 (2023) 111250

step SIMPLEC algorithm is used for speeding up the convergence. Mo­

mentum, energy, turbulence, and species equations are discretized using
second order schemes, while PRESTO Scheme is applied for the pressure
gradient terms. The time step for continuous phase is considered 0.05 s.
In this simulation, the preconditioning is highly recommended to ach­
ieve a better convergence. Multi-grid acceleration technique is also used
to improve the convergence rate. For instance, Bi-Conjugate Gradient
stabilized method (BCGSTAB) is used to enhance the linear solver
robustness of the species equation. In addition, using the Incomplete
Lower Upper smoother (ILU) with AMG coupled solver reduces the
processing time, and facilitates the convergence. Therefore, for solving
the linear system of equations, ILU is beneficial and employed in this
study.
Few points regarding the dispersed phase setups need to be
mentioned in this part of the paper. A 4th order (trapezoidal) scheme is
used for the discretization of the equations for tracking particles. Par­
ticles are tracked in a transient mode with time step of 0.001 s.

4. Results and discussion

In this section, the mesh-independency, model validation, and nu­

merical results are discussed.
Fig. 4. Comparison of the computed deposition flux with experimental and
simulation data (Kota and Langrish, 2007).
4.1. Mesh analysis

A grid-independency study is performed so that the outlet air tem­ escape boundary condition is selected. Simulations have been carried
perature was continuously monitored until it did not significantly out for several inlet drying air temperatures to compare with the
change with further grid refinement. The obtained results are reported measured data as shown in Fig. 5.
in Table 5. The outlet temperature of the grid number 3, and 4 are
approximately the same, so the mesh number 3 is chosen since it is ac­ 4.2.1. Error analysis
curate enough and computationally cheaper. The absolute (δ) and relative (ε) errors of the validated models are
reported in Table 6. The maximum deviation of deposition flux happens
at the height of 0.55 m from the bottom of the chamber. The deposition
4.2. Validation
phenomenon is highly complex, and it is very difficult to accurately
model it. Several factors e.g., shear stress, velocity, temperature, and so
In this study, two significant aspects of spray dryers are examined
on affect the particles to stick to or rebound from walls. Therefore,
using CFD simulations. At first, particle deposition results are compared
relatively high error is expected with the existing deposition models.
with experimental data, and simulation results of Kota and Langrish
The case with inlet air temperature of 140 ◦ C gives the maximum per­
(2007).
centage errors of 6.7% and 8.3% for the outlet temperature and relative
Comparison of wall deposition in a conical section of a pilot scale
humidity, respectively. The initial absolute humidity of inlet air has not
dryer is illustrated in Fig. 4. The simulation of Kota and the present study
been mentioned in the reference, so it is assumed zero, which may
have the same trends; however, results of the present study are closer to
contribute to the deviation of numerical results from the experimental
the experimental data. The discrepancy between the simulation results
may be due to the employed approximations and discretization
methods. For example, the one-way coupling with escape boundary
condition is used in the simulation of Kota, while the two-way coupling
between particles and gas is employed in the present study. In addition,
the sticky point curve is used as a boundary condition in this work.
To examine the validity of energy-related aspects of this work,
comparison has been made with the experimental data of (Ozmen and
Langrish, 2003) who calculated the temperature and relative humidity
of exhaust air in a pilot-scale dryer. The duration of their experimental
work was 30 min. However, simulating for this time period (30 min) in
the transient mode requires about one month of CPU time. So, the
steady-state strategy is employed for validating the energy aspect. As
particle build up is a time-dependent process, the deposition pattern is
not considered in this part of validation and it is assumed that particles
leave the computational domain upon their collision to the walls i.e., the

Table 5
Grid-independency study.
Grid number Number of cells Outlet temperature (◦ C)

1 29k 153.10
2 124k 150.43
3 390k 148.77
Fig. 5. Comparison of outlet temperature and relative humidity of this work
4 914k 148.38
with experimental data (Ozmen and Langrish, 2003).

6
S. Moradi Maryamnegari et al. Journal of Food Engineering 338 (2023) 111250

Table 6
Error analysis of the simulated results.
Deposition flux (Kg/m2) Outlet temperature (◦ C) Relative humidity (%)

0.15m 0.35m 0.55m 100 C

◦
140 C◦
180 C
◦
100 C
◦
140 ◦ C 180 ◦ C
− 4 − 5 − 4
δ 1.05e 7e 1.1e 1.94 3.98 3.89 2.1 1.96 1.2
ε (%) 21.9 17.8 51.3 3.8 6.7 5.8 5.4 8.3 7.1

data. However, the overall trends in both numerical and experimental transition temperature, they become rubbery and then deposit on the
results are the same. wall upon the collision. So, the base dryer shows better performance
from the deposition aspect at moderate temperature i.e., 140 ◦ C.
In the insulated-wall dryer, wall deposition increases with increasing
4.3. Investigations regarding dryers’ performance the inlet air temperature. Insulating the walls prevents heat loss, and
therefore makes the dryer chamber warmer compared to the dryers
In this section, three dryer’s performance parameters are investi­ without the insulation. The hot air within the chamber evaporates
gated for the base, insulated wall, and cooled wall dryers. At first, the moisture content and put dry particles in the sticky region of the sticky
performance of each dryer is examined from the deposition point of view curve. So, the insulated-wall dryer experiences much more wall depo­
followed by considerations regarding the potentials of dryers to produce sition in comparison with the base, and cooled-wall dryer for the same
powder. Finally, the dryers are compered from the thermal performance
point of view, i.e., the ability of dryers to remove the moisture from the
droplets.

4.3.1. Deposition performance

The contour plots of wall depositions in all cases are illustrated in
Fig. 6. In the base dryer, particle deposition is minimum at the moderate
temperature i.e., 140 ◦ C. In low and high temperatures, the wall depo­
sition build up is evident in the conical section. It seems that a significant
amount of moisture remains in the base dryer at low inlet air tempera­
ture since wet particles are more likely to stick to the walls. On the other
hand, when the temperature of dry particles is greater than the glass

Fig. 7. Comparison of particle deposition percentages of different dryers.

Fig. 6. Contour plots of wall deposition. Fig. 8. Comparison of powder recovery of different dryers.

7
S. Moradi Maryamnegari et al. Journal of Food Engineering 338 (2023) 111250

Fig. 9. Comparison of drying efficiencies of different dryers.

Table 7
Comparison of thermal loss factors of different dryers. Fig. 10. Snapshot of particles’ temperatures in all cases.

Temperature ( C)◦
Base design Insulated-wall dryer Cooled-wall dryer

100 11.44 0.00 18.73

regarding the percentage of particles that reach the outlet is shown in
140 13.81 0.00 22.04
180 16.36 0.00 24.93 Fig. 8. The overall percentage of solid recovery is low in the dryers (less
than 0.2%), therefore, the importance of designing a dryer with high
efficiency should be emphasized here. As shown in Fig. 8, the highest
powder recovery belongs to the base design at 100 ◦ C, and 140 ◦ C, while
inlet air temperature. the minimum powder recovery is obtained in the insulated dryer at all
The lowest amount of wall deposition is observed in the cooled-wall inlet air temperatures. These results are consistent with the deposition
dryer where cooling air flows through the secondary wall. The cooled data since the highest deposition in the insulated-wall dryer results in
wall reduces droplet temperatures near the surface, so that they satisfy the reduction of the yield. Furthermore, the recoveries of the cooled-
the non-sticky criterion and rebound form the wall upon collision. Un­ wall and base dryer are close at low and high temperatures, however,
like the base dryer as well as the insulated-wall dryer, the best perfor­ the solid recovery for the cooled-wall dryer is less than half of that of the
mance from the deposition point of view is obtained in the high base dryer for the intermediate temperature.
temperatures in the cooled-wall dryer, so that the deposition build up is
completely removed at 180 ◦ C. 4.3.3. Thermal performance
Further investigation regarding the deposition issue is provided by The energy aspects of the three cases i.e., drying efficiency, thermal
comparing the percentage of deposited particles versus inlet air tem­ loss factor, and the amount of remaining moisture content in different
perature as shown in Fig. 7. The percentage of deposition is negligible locations of dryers are evaluated in this section. The drying efficiency is
for the base dryer at 140 ◦ C as compared to 100 ◦ C, and 180 ◦ C. The shown for different gas inlet temperatures in Fig. 9.
deposition for the cooled-wall dryer is insignificant for low, moderate, The maximum performance is observed at 100 ◦ C for all cases. The
and high temperatures, while the insulated-wall dryer experiences insulated-wall dryer shows higher drying efficiency, about 43%, at
deposition up to 25% for the inlet temperature of 180 ◦ C. The analysis of 100 ◦ C, while the performance of the cooled-wall dryer is slightly higher
data is consistent with the contour plots of deposition in Fig. 6. than the insulated-wall dryer at 140 ◦ C. In other words, insulating the
walls is beneficial for low inlet air temperatures. The drying efficiencies
4.3.2. Powder recovery performance of all cases are minimum and approximately equal at high inlet air
The second parameter is the powder recovery. The simulation results temperature (180 ◦ C).

Table 8
Percentage of particles’ moisture content in different locations of dryers.
100 ◦ C 140 ◦ C 180 ◦ C

Base design Insulated wall Cooled wall Base design Insulated wall Cooled wall Base design Insulated wall Cooled wall
(%) (%) (%) (%) (%) (%) (%) (%) (%)

Deposit 1.7 1.84 1.46 1.79 0.00 1.85 0.00 0.00 –

In fluid 4.4 0.54 4.31 0.09 0.05 0.07 0.01 0.01 0.02
Outlet 3.73 0.69 2.98 0.00 0.00 0.00 0.00 0.00 0.00

8
S. Moradi Maryamnegari et al. Journal of Food Engineering 338 (2023) 111250

Table 7 compares the thermal loss factors for the three dryers in this
study. In the insulated-wall dryer, the thermal loss factor is zero since
walls are insulated and there is no heat transfer between the dryer and
the environment. The heat loss increases in the base design and cooled-
wall dryer by increasing the inlet supplied energy. Due to the force
convection, the thermal loss factor in the cooled-wall dryer is higher
than the base design. However, most portion of heat loss exit from the
outlet of the double walls instead of the wall parts of the dryer. One
advantage of this scheme is that the cooling air from the outlet can be
recovered and integrated with the inlet air to prevent from heat loss.
Table 8 shows the percentage of remained water in the particles, which
may be carried by the fluid, deposited on the walls, or even reach the
outlet. This table provides useful information regarding the potential of
drying process in the three cases. It is evident that the remaining
moisture in particles with high primary supplied energy is insignificant
compared to particles in dryers with low inlet temperature. Moreover,
insulated walls prevent heat loss, so that the hot gas within the chamber
removes much more moisture from the particles. This phenomenon is
more evident at 100 ◦ C, where the moisture content of outlet particles in
the insulated-wall dryer is about one fifth of the base dryer. It should be
noted that no deposition is reported in the cooled-wall dryer at 180 ◦ C,
so no data is available in this part of the table.
Fig. 11. Particles distribution in the injection area.
The inlet air temperature should be correctly set since high air
temperature leads to thermal degradation especially in particles with

Fig. 12. Histogram chart for particles size distribution.

9
S. Moradi Maryamnegari et al. Journal of Food Engineering 338 (2023) 111250

Fig. 13. Particle residence times. (a): T = 100 ◦ C, (b): T = 140 ◦ C, (c): T = 180 ◦ C.

low or zero moisture content. On the other hand, the heat loss from walls walls is a remedy for this undesirable effect. Therefore, fine skim
reduces the temperature of agent gas within the chamber. In the dryer milk powders will be produced in the cooled-wall dryer.
with fully insulated walls, particles are more likely to encounter thermal ➢ The dryer with the fully insulated walls experiences the lowest solid
degradation since the waste energy is minimum in this case. As given in recovery, while the base dryer still has the maximum potential to
Fig. 10, the particle temperatures in the insulated-wall dryer at 180 ◦ C is produce powders at low and moderate inlet gas temperature.
higher than other cases. ➢ Wall insulation facilitates moisture removal; however, particles are
The distribution of particle size is between 10 μm and 54 μm in the more likely to experience thermal degradation. Therefore, this kind
injection area as illustrated in Fig. 11. The histogram chart of particles of dryer is beneficial when inlet air temperature is low since more
size distribution for particles and droplets in the chamber is given in high-quality particles will be produced and is also useful from the
Fig. 12. The range of particles’ diameters is wider at low temperature, energy saving perspective.
especially in the base dryer, the particles in size of about 42 μm–46 μm ➢ Changing the condition of walls does not improve all performance
are evidence, while all particles’ diameters are less than about 23 μm at factors of a particular dryer. Choosing a suitable dryer requires
140 ◦ C, 180 ◦ C. In addition, the peak of charts for the insulated-wall specification of priorities specified by the user.
dryer consists of fine particles, while the most percentage of particles ➢ Further studies on comparing different materials with constant and
are located at the end of particles’ range for the cooled-wall and base variable properties are recommended for further generalization and
dryer at 180 ◦ C. Therefore, the percentage of fine particles in the dryer validation of the results.
with insulated wall is higher than other cases.
The residence time of particle is an important parameter that should Credit author statement
be considered in drying process since particles with high residence time
suffer from thermal degradation. In this study, this parameter is calcu­ Sara Moradi Maryamnegari: Conceptualization, Methodology,
lated in terms of the percentage of particles. As shown in Fig. 13, the Software, Validation, Formal analysis, Writing- Original Draft. Ali
residence time of particles is drawn with a polar plot in which the angle Ashrafizadeh: Conceptualization, Writing- Review & Editing, Supervi­
value is the residence time, and the radius value is the percentage of sion. Egbert Baake: Writing- Review & Editing, Supervision. Mattia
particles in a specified period. Overall, the residence time of particles is Guglielmi: Writing- Review & Editing, Supervision.
close to each other in all cases. The residence time for the insulated wall
dryer is slightly higher than other cases at 140 ◦ C, and 180 ◦ C (Fig. 13b
and 13c). Declaration of competing interest

5. Conclusions The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
Conclusions regarding the effects of wall boundary condition on the the work reported in this paper.
dryers’ performance measures are as follows:
Data availability
➢ Insulating the chamber wall takes a turn for the worse by increasing
the particle build up on the walls. By contrast, cooling the external The data and code are available in the Appendix section.

10
S. Moradi Maryamnegari et al. Journal of Food Engineering 338 (2023) 111250

Appendix A

User-Defined Functions (UDFs) are employed to implement some necessary modifications on drying kinetic, boundary conditions, material
properties, and post-processing. The CDC model is used through a “DEFINE_DPM_HEAT_MASS” macro to predict the drying rate of multi-components.
The sticky point curve of skim milk is used in all DPM wall boundary conditions as a criterion of particles’ stickiness using a “DEFINE_DPM_BC” macro.
Thermo-physical properties of pure water (Afshar et al., 2019) as an evaporating component is also defined using related macros in both continuous
and discrete phases. Moreover, F_UDM and C_UDM are defined through the UDFs for visualizations of deposition contour plots.

11
S. Moradi Maryamnegari et al. Journal of Food Engineering 338 (2023) 111250

. (continued).

Appendix B

The thermal conductivity of skim milk components is calculated as follows (Chen and Peng, 2005):
kp = 0.17881 + 0.0011958T − 2.7178 × 10− 6 T 2 B.1

kf = 0.18071 + 0.0027604T − 1.7749 × 10− 7 T 2 B.2

kc = 0.20141 + 0.0013874T − 4.3312 × 10− 6 T 2 B.3

ka = 0.32962 + 0.0014011T − 2.9069 × 10− 6 T 2 B.4

12
S. Moradi Maryamnegari et al. Journal of Food Engineering 338 (2023) 111250

, where p, f, c, and a subscript refer to protein, fat, carbohydrate, and ash components, respectively. Furthermore, T refers to the particle temperature.
The overall thermal conductivity of skim milk is obtained based on the average of parallel and series models:
Kserial = εp kp + εf kf + εc kc + εa ka B.5

εp εf εc εa
Kparllel = + + + B.6
kp kf kc ka
( )
Kd = 0.5 Kserial + Kparllel B.7

Here, ε is the volume fraction of each component. The components’ density is calculated as bellow:
ρp = 1329.9 + 0.5185T B.8

ρf = 925.59 + 0.41757T B.9

ρc = 1599.1 + 0.31046T B.10

ρa = 2423.8 + 0.28063T B.11

The density of mixture is calculate in Eq. (B.12) (Chen and Peng, 2005):
1
ρd = cp c B12
ρp + ρf + ρcc + ρca
f c a

, Where c presents the mass fraction of components. The specific heat transfer of skim milk is calculated using the Eq. (B.13) (Chong and Chen, 1999):
cp = 1.09 + 0.00352T − 3.54 × 10− 6 T 2 B.13

References Du, J., Ge, Z.Z., Xu, Z., Zou, B., Zhang, Y., Li, C.M., 2014. Comparison of the efficiency of
five different drying carriers on the spray drying of persimmon pulp powders. Dry.
Technol. 32, 1157–1166. https://doi.org/10.1080/07373937.2014.886259.
Adhikari, B., Howes, T., Wood, B.J., Bhandari, B.R., 2009a. The effect of low molecular
Gao, N., Niu, J., He, Q., Zhu, T., Wu, J., 2012. Using RANS turbulence models and
weight surfactants and proteins on surface stickiness of sucrose during powder
Lagrangian approach to predict particle deposition in turbulent channel flows. Build.
formation through spray drying. J. Food Eng. 94, 135–143. https://doi.org/
Environ. 48, 206–214. https://doi.org/10.1016/J.BUILDENV.2011.09.003.
10.1016/J.JFOODENG.2009.01.022.
Golman, B., Julklang, W., 2014. Simulation of exhaust gas heat recovery from a spray
Adhikari, B., Howes, T., Bhandari, B.R., Langrish, T.A.G., 2009b. Effect of addition of
dryer. Appl. Therm. Eng. 73, 899–913. https://doi.org/10.1016/j.
proteins on the production of amorphous sucrose powder through spray drying.
applthermaleng.2014.08.045.
J. Food Eng. 94, 144–153. https://doi.org/10.1016/j.jfoodeng.2009.01.029.
Golman, B., Yermukhambetova, A., 2019. An Excel VBA-based educational module for
Adhikari, B., Howes, T., Wood, B.J., Bhandari, B.R., 2009c. The effect of low molecular
simulation and energy optimization of spray drying process. Comput. Appl. Eng.
weight surfactants and proteins on surface stickiness of sucrose during powder
Educ. 27, 1103–1112. https://doi.org/10.1002/cae.22139.
formation through spray drying. J. Food Eng. 94, 135–143. https://doi.org/
Golman, B., WJ, 2014. Management, analysis of heat recovery from a spray dryer by
10.1016/j.jfoodeng.2009.01.022.
recirculation of exhaust air. Energy Convers. Manag. 88, 641–649. https://doi.org/
Afshar, S., Metzger, L., Patel, H., Selomulya, C., Woo, M.W., 2019. A practical CFD
10.1016/j.enconman.2014.09.012.
modeling approach to estimate outlet boundary conditions of industrial multistage
Hennigs, C., Kockel, T.K., Langrish, T.A.G., 2001. New measurements of the sticky
spray dryers: inert particle flow field investigation. Dry. Technol. 37, 824–838.
behavior of skim milk powder. Dry. Technol. 19, 471–484. https://doi.org/10.1081/
https://doi.org/10.1080/07373937.2018.1464473.
DRT-100103929.
Ai, S., Wang, B., Li, X., Shi, W., 2016. Analysis of a heat recovery system of the spray-
Hussain, F., Jaskulski, M., Piatkowski, M., Tsotsas, E., 2022. CFD simulation of
drying process in a soy protein powder plant. Appl. Therm. Eng. 103, 1022–1030.
agglomeration and coalescence in spray dryer. Chem. Eng. Sci. 247, 117064 https://
https://doi.org/10.1016/j.applthermaleng.2016.04.108.
doi.org/10.1016/j.ces.2021.117064.
ANSYS ICEM CFD, 2021. Help System, ANSYS ICEM CFD Help Manual, Release, vol. 21.
Jaskulski, M., Atuonwu, J.C., Tran, T.T.H., Stapley, A.G.F., Tsotsas, E., 2017. Predictive
ANSYS Inc., U.S.A, p. 1.
CFD modeling of whey protein denaturation in skim milk spray drying powder
Arjmandi, H., Amini, R., Kashfi, M., Abikenari, M.A., Davani, A., 2022. Minimizing the
production. Adv. Powder Technol. 28, 3140–3147. https://doi.org/10.1016/j.
COVID-19 spread in hospitals through optimization of ventilation systems. Phys.
apt.2017.09.026.
Fluids 34. https://doi.org/10.1063/5.0081291.
Jaskulski, M., Tran, T.T.H., Tsotsas, E., 2020. Design study of printer nozzle spray dryer
Atuonwu, J.C., Stapley, A.G.F., 2017. Reducing energy consumption in spray drying by
by computational fluid dynamics modeling. Dry. Technol. 38, 211–223. https://doi.
monodisperse droplet generation: modelling and simulation. Energy Proc. 123,
org/10.1080/07373937.2019.1633541.
235–242. https://doi.org/10.1016/j.egypro.2017.07.251.
Jubaer, H., Afshar, S., Xiao, J., Chen, X.D., Selomulya, C., Woo, M.W., 2019. On the effect
Birchal, V.S., Huang, L., Mujumdar, A.S., Passos, M.L., 2006. Spray dryers: modeling and
of turbulence models on CFD simulations of a counter-current spray drying process.
simulation. Dry. Technol. 24, 359–371. https://doi.org/10.1080/
Chem. Eng. Res. Des. 141, 592–607. https://doi.org/10.1016/j.cherd.2018.11.024.
07373930600564431.
Keshani, S., Daud, W.R.W., Nourouzi, M.M., Namvar, F., Ghasemi, M., 2015a. Spray
Caglayan, H., Caliskan, H., 2017. Sustainability assessment of heat exchanger units for
drying: an overview on wall deposition, process and modeling. J. Food Eng. 146,
spray dryers. Energy 124, 741–751. https://doi.org/10.1016/j.energy.2017.02.097.
152–162. https://doi.org/10.1016/j.jfoodeng.2014.09.004.
Caglayan, H., Şöhret, Y., Caliskan, H., 2018. Thermo-ecologic evaluation of a spray dryer
Keshani, S., Montazeri, M.H., Daud, W.R.W., Nourouzi, M.M., 2015b. CFD modeling of
for ceramic industry. Energy Proc. 144, 164–169. https://doi.org/10.1016/j.
air flow on wall deposition in different spray dryer geometries. Dry. Technol.
egypro.2018.06.022.
https://doi.org/10.1080/07373937.2014.966201.
Chegini, G.R., Ghobadian, B., 2007. Spray dryer parameters for fruit juice drying. World
Khuenpet, K., Charoenjarasrerk, N., Jaijit, S., Arayapoonpong, S., Jittanit, W., 2016.
J. Agric. Sci. 3, 230–236.
Investigation of suitable spray drying conditions for sugarcane juice powder
Chen, X.D., Peng, X., 2005. Modified biot number in the context of air drying of small
production with an energy consumption study. Agri. Nat. Res. 50, 139–145. https://
moist porous objects. Dry. Technol. 23, 83–103. https://doi.org/10.1081/DRT-
doi.org/10.1016/j.anres.2015.08.003.
200047667.
Kota, K., Langrish, T.A.G., 2007. Prediction of wall deposition behaviour in a pilot-scale
Cheng, F., Zhou, X., Liu, Y., 2018. Methods for improvement of the thermal efficiency
spray dryer using computational fluid dynamics (CFD) simulation. Chem. Prod.
during spray drying. E3S-Conferences. 53, 1031. https://doi.org/10.1051/e3sconf/
Process Model. 2 https://doi.org/10.1631/jzus.2007.A0301.
20185301031.
Langrish, T.A.G., Harrington, J., Huang, X., Zhong, C., 2020. Using CFD simulations to
Chong, L., Chen, X., 1999. A mathematical model of the self-heating of spray-dried food
guide the development of a new spray dryer design. Processes 8. https://doi.org/
powders containing fat, protein, sugar and moisture. Chem. Eng. Sci. 54, 4165–4178.
10.3390/PR8080932.
Crowe, C.T., 1983. Droplet-gas interaction in counter-current spray dryers. Dry. Technol.
1, 35–56. https://doi.org/10.1080/07373938308916769.

13
S. Moradi Maryamnegari et al. Journal of Food Engineering 338 (2023) 111250

Ludwig, W., Ligus, G., Korman, P., Sędłak, A., Zając, D., 2022. CFD modelling of a powder Soltani, S., Gerde, P., Acevedo, F., Rasmuson, A., 2015. Counter-current spray drying
spraying nozzle sed for dry coating. Chem. Eng. Res. Des. 178, 550–566. https://doi. with stream separation: computational modeling of a novel dryer design. Chem. Eng.
org/10.1016/J.CHERD.2022.01.001. Res. Des. 93, 163–173. https://doi.org/10.1016/j.cherd.2014.05.023.
Mezhericher, M., 2008. Drying of Slurries in Spray Dryers Beer-Sheva. Ben-Gurion Tran, T.T.H., Jaskulski, M., Tsotsas, E., 2017. Reduction of a model for single droplet
University of the Negev, pp. 3–4. drying and application to CFD of skim milk spray drying. Dry. Technol. 35,
Moejes, S.N., Visser, Q., Bitter, J.H., van Boxtel, A.J.B., 2018. Closed-loop spray drying 1571–1583. https://doi.org/10.1080/07373937.2016.1263204.
solutions for energy efficient powder production, innovative. Food Sci. Emerging Truong, V., Bhandari, B.R., Howes, T., 2005. Optimization of cocurrent spray drying
Technol. 47, 24–37. https://doi.org/10.1016/j.ifset.2018.01.005. process for sugar-rich foods. Part II—optimization of spray drying process based on
Ozmen, L., Langrish, T.A.G., 2003. A study of the limitations to spray dryer outlet glass transition concept. J. Food Eng. 71, 66–72. https://doi.org/10.1016/J.
performance. Dry. Technol. 21, 895–917. https://doi.org/10.1081/DRT-120021691. JFOODENG.2004.10.018.
Petersen, L.N., Poulsen, N.K., Niemann, H.H., Utzen, C., Jørgensen, J.B., 2017. An Woo, M., 2017. Computational Fluid Dynamics Simulation of Spray Dryers: an Engineer’s
experimentally validated simulation model for a four-stage spray dryer. J. Process Guide. Taylor & Francis Group.
Control 57, 50–65. https://doi.org/10.1016/j.jprocont.2017.05.001. Zolfagharnasab, M.H., Salimi, M., Zolfagharnasab, H., Alimoradi, H., Shams, M.,
Roustapour, O., Tahhavor, A., Afsari, A., Sayyari, A., 2013. Computational fluid dynamic Aghanajafi, C., 2021. A novel numerical investigation of erosion wear over various
simulation of flow pattern and heat transfer in a pilot plant spray dryer with cooling 90-degree elbow duct sections. Powder Technol. 380, 1–17. https://doi.org/
air jacket. J. Agric. Eng. Res. 14, 89–106. 10.1016/J.POWTEC.2020.11.059.

14