Application of Euler Langrange Approach To Predict, Ricardo 2015 PDF

Subscriber access provided by CMU Libraries - http://library.cmich.
edu
Article
Application of Euler-Lagrange Approach to Predict
the Droplet Solidification in a Prilling Tower
Guilherme Antonio Novelletto Ricardo, Dirceu Noriler, Waldir Martignoni, and Henry F. Meier
Ind. Eng. Chem. Res., Just Accepted Manuscript DOI: 10.1021/acs.iecr.5b01430 Publication Date (Web): 14 Sep 2015
Downloaded from http://pubs.acs.org on September 15, 2015
Just Accepted
Just Accepted manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides Just Accepted as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. Just Accepted manuscripts
appear in full in PDF format accompanied by an HTML abstract. Just Accepted manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI). Just Accepted is an optional service offered
to authors. Therefore, the Just Accepted Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the Just
Accepted Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these Just Accepted manuscripts.
Industrial & Engineering Chemistry Research is published by the American Chemical

Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.
Page 1 of 52 Industrial & Engineering Chemistry Research
1
2
3
4
5
6
7
8
Application of Euler-Lagrange Approach to Predict
9
10
11
12 the Droplet Solidification in a Prilling Tower
13
14
15
16
17 Guilherme A. N. Ricardo, Dirceu Noriler, Waldir P. Martignoni, Henry F. Meier,*
18
19
20
21
22
23
Department of Chemical Engineering, Regional University of Blumenau, So Paulo St. 3250,
24
25
26
89030000, Blumenau, Brazil
27
28

29 PETROBRAS/SIX/PQ, Xisto Road 143, 83900000, So Mateus do Sul, Brazil
30
31
32
33
34
35
36
37 KEYWORDS: prilling process, particle solidification, Euler-Lagrange approach
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52 *
53
Corresponding Author. Tel.: +55 47 32216060; fax: +55 47 32216001.
54
55
56 E-mail address: meier@furb.br (H.F. Meier)
57
58
59 1
60
ACS Paragon Plus Environment
Industrial & Engineering Chemistry Research Page 2 of 52
1
2
3
ABSTRACT
4
5
6
7
8
9
10
11 This paper deals with the phenomenological characteristics of the multiphase flow in a three-
12
13
14 dimensional urea prilling tower under conditions close to the industrial scale process,
15
16 considering the simultaneous heat and momentum transfer and the droplet solidification, using
17
18 the computational fluid dynamics technique. The Euler-Lagrange approach is used to represent
19
20
21 the multiphase flow. Solidification of the droplets was considered by means of an unsolidified
22
23 core model, which predicts the thickness of the solid phase formed from the surface to the center
24
25
of the droplets, during the evolution of their trajectories in the continuous phase. The global
26
27
28 operational performance of the process was evaluated based on the solidification and breaking
29
30 indexes defined from local variables obtained from numerical simulations. Solidification index
31
32
33
values in the range of 0.83 and 1.00 were obtained from simulations and the internal resistance to
34
35 heat transfer was found to strongly affect the solidification phenomenon. A non-symmetrical
36
37 flow was used to evaluate the breaking index of droplets due to impact with the walls. The global
38
39
40 thermal properties associated with the droplets, the surface and core temperature and the overall
41
42 heat-transfer coefficient had values of 34.37 C, 115.79 C and 19.29 W/m2K, respectively, for
43
44 droplets with a diameter of 1.2 mm. The air temperature at the top of the tower was 74.26C for
45
46
47 droplets of this diameter.
48
49
50
51
52
53
54
55
56
57
58
59 2
60
1
2
3
1 INTRODUCTION
4
5
6 The prilling process is widely applied to produce urea and ammonium nitrate in granular
7
8 particles. This simple process consists of spraying a substance in the liquid state at the top of a
9
10
11
very high tower. The spraying process produces small spherical droplets which, due to the
12
13 gravitational force, fall counter-currently with a stream of cooling air (at room temperature) fed
14
15 from the bottom of the tower. Through heat exchange with the cooling air, the urea droplets
16
17
18 change from the liquid to the solid state, forming particles with approximately spherical shape.
19
20 In practice, many prilling towers have operational problems caused by the incomplete
21
22 solidification of the droplets. The low efficiency of the solidification process also results in poor
23
24
25 quality grains being produced and, consequently, losses in relation to productivity and profit. The
26
27 operational problems are mainly associated with two factors: the characteristics of the air flow
28
29
inside the tower and the incomplete solidification of the droplets. The first factor is strongly
30
31
32 affected by the geometric characteristics of the air outlet at the top of the tower, which
33
34 contributes to an asymmetric air flow. This asymmetry significantly affects the trajectory of the
35
36
37 droplets, which may be directed toward the walls of the tower. After hitting the walls, the second
38
39 factor, incomplete solidification, takes place. This factor contributes to the breakup of droplets,
40
41 which leads to the formation of a kind of "cake" which grows over time. Also, the incompletely
42
43
44 solidified droplets may break by hitting against each other. A general description of the urea
45
46 production processes can be found in Bertn et al.1, Bertn et al.2, Cotabarren et al.3 and Heinrich
47
48 et al.4
49
50
51 A publication by Bakhtin et al.5 is one of the oldest papers related to the application of
52
53 mathematical models and numerical methods to evaluate the thermal characteristics and phase
54
55
56
changes which occur in prilling towers. The authors neglect the interactions between phases and
57
58
59 3
60
1
2
3
assume the velocity of the continuous phase to be constant. The mathematical model includes
4
5
6 three sequential thermal periods to which the urea droplets are subjected: cooling of urea droplets
7
8 in the liquid state; solidification of the droplets at a constant temperature; and cooling of the solid
9
10
11
particles. It was concluded that the optimum tower height is strongly dependent on the droplet
12
13 diameter and that there is an optimum tower height which guarantees the complete solidification
14
15 of the droplets.
16
17
18 Yuan et al.6 performed a numerical study aimed at reducing the height of prilling towers, by
19
20 coupling a fluidized bed to the base of the tower. A major problem is the formation of "cake" on
21
22 the walls of the tower due to the breaking of the droplets that hit the walls. To avoid this
23
24
25 problem, the complete solidification of the droplets and also the cooling of the solidified
26
27 particles need to be ensured. Thus, high towers must be used. The height of the tower required is
28
29
dependent on the diameter of the droplets. The larger the diameter of the droplets, the greater
30
31
32 their terminal velocity and internal resistance to heat transfer will be. Consequently, the
33
34 residence time of the particles inside the tower and the heat transfer between the phases are
35
36
37 reduced. The above-cited authors proposed a one-dimensional model to represent the thermal
38
39 behavior of the two phases and used a shrinkage unsolidified core model to represent the
40
41 increase in the solid phase thickness of the droplets during solidification. They considered a
42
43
44 constant terminal velocity for the droplets and three thermal periods, as considered by Bakhtin et
45
46 al.1 It was concluded by the authors that about 90% of the total tower height is needed to ensure
47
48 complete solidification of the droplets and the use of a fluidized bed may reduce the tower height
49
50
51 by up to 50%.
52
53 Mehrez et al.7,8 performed numerical simulations of a prilling tower considering the
54
55
56
simultaneous transfer of mass, energy and momentum between the two phases. They analyzed
57
58
59 4
60
1
2
3
the temperature and humidity profiles along the radial position of the urea droplets and also the
4
5
6 average temperature and humidity profiles along the axial position of the tower. The weight,
7
8 buoyancy and drag forces were considered as the forces that govern the velocity of the droplets.
9
10
11
The same classical three thermal periods for the thermal behavior of the droplets were
12
13 considered. The boundary conditions of symmetry and convective heat and mass transfer were
14
15 applied at the center and the outer surface of the droplets, respectively. The model solution was
16
17
18 obtained by the discretization of the radial domain of the droplets and the axial domain of the
19
20 tower, together with finite approximation schemes. One of the main results obtained by the
21
22 authors was the finding that, at the bottom of the tower, the temperature in the center of the
23
24
25 droplets was higher than the melting temperature of urea, i.e., there is a portion of the volume of
26
27 the droplets that does not solidify completely. Also, the formation of a "cake" can occur due to
28
29
the breaking of the incompletely solidified droplets.
30
31
32 Although numerous studies involving multiphase flows with the simultaneous transfer of
33
34 energy and momentum have been conducted, there is a notable lack of scientific literature
35
36
37 regarding the application of CFD techniques to the industrial use of the prilling process, which
38
39 was the primary objective of this study. Thus, in this research, numerical simulations were
40
41 performed to obtain data and information on the phenomena involved in the urea prilling process
42
43
44 based on Euler-Lagrange (EL) approach focusing on the conservation of momentum and energy
45
46 and the solidification of urea droplets.
47
48
49
50
51 2 MATHEMATICAL FORMULATION
52
53 The urea prilling tower is modeled according to the following main assumptions:
54
55
56
57
58
59 5
60
1
2
3
4
Cooling air represents the continuous phase and it is modeled on a Euler
5
6 framework9;
7
8 Urea droplets represent the discrete phase and are modeled on a Lagrange
9
10
11 framework10;
12
13 The continuous phase under turbulent conditions is modeled by RANS equations;
14
15
16 There are no chemical reactions in either of the phases;
17
18 Phases interact only by exchanging heat and momentum. Due to the low
19
20 volumetric ratio between the droplets and the gas, the interactions between the phases are
21
22
23 modeled by a two-way coupling11,12;
24
25 The solidification phenomenon occurs at constant temperature;
26
27
28 The droplets are spherical;
29
30 Only heat transfer by convection and conduction (during solidification) are
31
32 considered, because these are the most important mechanisms in this process;
33
34
35 An unsolidified core model6 is used to predict the change in the liquid radius
36
37 inside the droplet during the solidification. This model considers that the solidification
38
39
40 progresses from the outside surface to the inside of the droplet until reaching the center,
41
42 as shown in Figure 1. Conversely, the melting occurs from the center of the droplet until
43
44 reaching the outside surface. In order to avoid physical inconsistency, one percent of the
45
46
47 total volume of the droplet its considered to remain liquid (section 2.2);
48
49 The Biot number is lower than 0.1 for the liquid core of the droplets and,
50
51
52
consequently, there is no temperature gradient in the liquid core, but there is through the
53
54 thickness of the solid.
55
56
57
58
59 6
60
1
2
3
The thermal behavior of the droplets during the prilling of urea can be divided into three
4
5
6 distinct periods5-8: (i) the cooling of the droplets by the surrounding air from their initial
7
8 temperature until the temperature at which solidification starts is reached,; (ii) solidification of
9
10
11
the droplet surface at constant temperature, progressing to the center of the droplets until
12
13 complete solidification, in which the resistance to heat transfer due to the formation of the solid
14
15 phase cannot be neglected; (iii) cooling of the solid particles. Hereinafter, the subscripts g, p, l, s
16
17
18 and eff represent the continuous phase, the dispersed-phase, liquid state, solid state and effective,
19
20 respectively.
21
22
23 Figure 1. Representation of the unsolidified core model.
24
25
26
2.1 EULERIAN MODEL
27
28
29 In the Eulerian approach, the physical phenomena are dependent on the temporal and spatial
30
31 domain and the observer analyzes the changes in the properties at a fixed point in space. The
32
33
34
instantaneous and local conservative equations are derived from a method that considers a fixed
35
36 volume in space, in which scalar or vector quantities are transported through this control volume
37
38 using a fixed coordinate system9. These conservative equations are well known and further
39
40
41 details can be found in Drew13 and Enwald et al.14 The conservative averaged equations used for
42
43 the numerical experiments in this study are given below:
44
45
46 Conservation of mass,
47
48

0.
49
t
50 (1)
51
52
53
54 Conservation of momentum,
55
56
57
58
59 7
60

p ,
3
t
4 (2)
5
6
7
8 Conservation of thermal energy,
9
10

C T
C T T S .
11

12
(3)
13
14
15 Note that the last terms of Eqs. (2) and (3) represent the source terms of momentum and energy
16
17 that couple the dispersed and continuous phases. These terms, for both phases, are equal but
18
19
20 opposite in sign. Thus:
,
21
22
(4)
23
24
S S .

25
26 (5)
27
28
29 Since no mass transfer is considered in this phenomenon, the mass source in Eq. (1) is
30
31
32 neglected. The reader should refer to section 2.3 for details on the modeling of these terms.
33
34
35
36
37 2.2 LAGRANGIAN MODEL
38
39 In this approach, the droplets or particles are tracked by calculating the trajectories of each
40
41
42
individual droplet or particle through solving the equations of motion10. In this kind of
43
44 simulation, the discrete phase and the continuous phase are coupled, and computational portions
45
46 of particles or droplets are traced through the field of the continuous phase. During the
47
48
49 simulation, in each time step or iteration, a new velocity for the discrete phase is calculated using
50
51 the equations of motion15. In the Lagrangian approach, the dispersed-phase can be represented by
52
53 a number of particles or droplets, N# , using Lagrangian coordinates. In this case, the Lagrangian
54
55
56
57
58
59 8
60
1
2
3
representation of the dispersed-phase at time t is given by the following set of properties, which
4
5
6 is characterized by an eight-dimensional random vector:
7
$% t,
t, T t, R (&) t, i 1, , N- t..
8 & & & &
9
(6)
Here, N# represent the number of physical particles or droplets. However, in Lagrangian

10
11
12
simulations the number of physical particles is represented by N/ (computational number of
13
14
particles or droplets). The value for N/ is usually lower than the real number of physical
15
16
17
18
particles. This numerical strategy reduces significantly the computational cost of EL simulations.
19
20
21 The position of the particle or droplet element at time t is generally expressed by a simple
22
23 position evolution model, described by the following first-order ordinary differential equation
24
25
26
(ODE)10:
d%
&
27

.
&
28
dt
29 (7)
30
31
32 The simplest model to describe the velocity of a particle or droplet i at time t is the following
33
34 acceleration model:
35
d

36 &
1 .
&
dt
37 (8)
38
39
40
41 The density of urea is approximately 1000 times higher than the density of air. Consequently,
42
43 the main forces acting on the droplets are the drag and gravitational forces16. Thus, the dispersed-
44
45 phase acceleration, considering spherical droplets, can be expressed as follows11,12,17:
46
1 t F3 4

5 .
& & &
47
48 (9)
49
50
51 The first term on the right side of Eq. (9) is the drag force per unit particle mass acting on the
52
53 particle. The second term is the gravitational force. The characteristic particle response
54
55
56
57
58
59 9
60
1
2
frequency, F3 , is dependent on the drag coefficient and the Reynolds number for the particle,
3 &
4
5
6 and it can be expressed by12:
7
1 C3 Re
8 & &
F3 & .
&
24
9

10 (10)
11
The particle response time, , is:

12
13 &
14
:
15
4d 5
16 &
,
&
18
17 (11)
18
19
Where Re is the Reynolds number of the particle and C3 is the drag coefficient calculated
&
20
21
22
23
24
through the correlation of Morsi and Alexander17.
25
26 The model for the temperature evolution over time of the dispersed-phase is generally
27
28 expressed as the rate of heat exchange between the dispersed-phase and the continuous phase
29
30
31 applying the following first-order ODE6:
dT
&
32
H .
&
33
dt
34 (12)
35
36
37 Equation (12) is applied to the three thermal periods described in section 2.1. The term on the
38
39 right side is modeled considering convection as the main mechanism of heat exchange between
40
41
42 the two phases and thus:
43
H F> 4T T 5.
44 & & &
(13)
45
46
47 Similarly to the momentum conservation of the particles, the thermal particle response
48
frequency, F> , is dependent on the Nusselt number of the particle and thermal particle response
49 &
50
51
52 time, and can be expressed by:
53
1
F> 6Nu .
& &
54
>
&
55
(14)
56
57
58
59 10
60
1
2
The thermal particle response time, > , is given by:

3 &
4
5
:
A CpA 4d 5
6 &
> ,
7 &
k
(14)
8
9
10
11 Thus, the relation between the Nusselt number and the Reynolds and Prandtl numbers for the
12
13 particle can be expressed by the classical constitutive equation of Ranz and Marshall18.
14
15
16 Due to the solidification phenomenon, the three thermal periods should be modeled
17
18 independently using Eqs. (12) and (17)6. In the first thermal period a loss of the sensible heat of
19
20 the droplets in the liquid state to the air occurs through the mechanism of convective heat
21
22
23 transfer. Through a combination of expressions, the energy equation for the droplet i can be
24
25 rewritten as:
26
6h 4T T 5
& &
dT
&
27
.
28
dt A CpA d
&
29 (16)
30
31
The last equation is valid for the temperature range of TD > T > T .
32 & &
33
34
35 The second thermal period consists of the solidification of the droplets, for which the essential
36
37
38 problem to be considered is the heat transfer between the gas phase and the unsolidified core of
39
40 the droplets. As the solidification evolves with time, the change in the radius of the liquid phase
41
42 inside the droplet is related to the rate of radius change due to solidification as follows:
43
dR (&)
&
44
.
&
45
dt
46
(17)
47
The term in the right side of Eq. (17) is modeled based on the fact that the latent heat of
&
48
49
50
51
52
melting of the droplet is removed and transferred to the air through the mechanism of convective
53
54 heat transfer. As a consequence, it is possible to relate the rate of solidification of the droplets to
55
56
57
58
59 11
60
1
2
3
the mechanisms of heat transfer between the air and the unsolidified core of the droplets through
4
the overall heat-transfer coefficient. Thus, can be expressed as follows:

&
5
6
7
F- T T .
8 & &
9 (18)
10
The particle solidification response frequency, F- , is related to the particle solidification
11 &
12
13
14 response time and to a function which is dependent on the solidification rate. As a result,
15
V#
&
1
16
F- : ..
17 &
- d 2&
&
18 (19)
19
20
where the term - can be written as follow:
21 &
22
H m#
&
23
- .
&
24
U & A
&
25 (15)
26
27
28
29
Accordingly, the change in the radius of the liquid phase for a droplet i due to solidification
30
31 can be described by the combination and mathematical arrangement of Eqs. (17)-(20),
32
dR (&)
&
U & d: T T
33
:.
dt
34
H # d 2&
(21)
35
36
Equation (21) is valid only during the solidification phenomenon, when T T .

37
38
39
40
41
Due to the formation of a solid shell around the droplets during the solidification phenomenon,
42
43 both convective and conductive resistances are considered. For spherical droplets, the overall
44
45 heat-transfer coefficient is expressed by6:
46
1
U & .
47
1 d &
48

49
h& d 2& k #
(22)
50
51
52
53 In Eq. (22), it can be noted that at the end of the solidification the thickness of the solid shell is
54
55 equal to the outer radius of the droplets. This causes a physical inconsistency in which the
56
57
58 overall heat-transfer coefficient tends to zero. The unsolidified core model considers that a
59 12
60
1
2
3
portion of the droplet remains in the liquid state. In this study, it is assumed that 1% of the total
4
5
6 volume of droplets remains in the liquid phase and thus the conductive resistance to heat transfer
7
8 does not tend to infinity. On applying this assumption, the temperature of the liquid core
9
10
11
decreases from the melting temperature to the final temperature but the core is assumed to
12
13 remain in the liquid state. Therefore, the minimum radius of the liquid phase that can be achieved
14
15 is defined as a function of the maximum thickness of the solid phase, as follows:
16
3V( N&O
17
R & N&O P
R
.
18
4
19 (23)
20
21
The minimum liquid volume, V( N&O , is determined by:
22
23
d T
24
V( N&O 0.01 S U.

25
6
26 (24)
27
28
29 In the third thermal period, the sensible heat from the completely solidified droplet is
30
31 transferred to the air through the conductive and convective mechanisms of heat transfer. Thus,
32
33
34 Eqs. (12) and (13) are written as follows:
35
6U & 4T T 5
&
dT
36 &
.
dt
37
# C# d
&
(25)
38
39
Equation (25) is valid when the droplets are completely solidified, in others words, for T < T .
40
41
42
43
44 In this period, the droplets are completely solidified and the thickness of the solid phase reaches
45
46 the maximum value while the liquid radius reaches the minimum value. Thus, the overall heat-
47
48 transfer coefficient is estimated by6:
49
1
U & .
50
d NWX
&
51
1

52
h
(26)
&
4d 2NWX 5k #
&
53
54
55
56
57
58
59 13
60
1
2
3
The next section describes the mathematical coupling between the continuous and the discrete
4
5
6 phases.
7
8
9
10
11 2.3 TWO-WAY COUPLING
12
13 Two-way coupling can be considered when the particle concentration is large enough to cause
14
15 global disturbances in the turbulent nature of the flow and the turbulence of the flow affects the
16
17
18 trajectory of the particles19. In this case, from the mathematical viewpoint, modeling the source
19
20 terms becomes more complex, as does the solution of the mathematical model. There are cases
21
22 where the particle loading is too high, when it may also be necessary to account for the
23
24
25 interactions between particles. These cases are more complex and represent four-way coupling20.
26
27 In this context, the reader is referred to a paper by Souza et al.21, where two-way coupling
28
29
30
between the phases is considered.
31
32
33 2.3.1 Interphase Transfer of Momentum
34
35 The source term for the momentum in Eq. (2) performs coupling between the Lagrangian
36
37
dispersed-phase and the Eulerian continuous phase. The total force exerted by the fluid on a body
38
39
40 immersed in that fluid can be expressed by22:
O] O]
41
1 1
42
Y Z[\W Y m F3 4

t5 with i 1, , n ,
& & &
V V
43 (27)
44
45 &^_ &^_
46
47
48 where np is the total number of particles. Other terms can be included in Eq. (27) to represent the
49
50 contributions of, for instance, lift, virtual mass and Basset history, but these are neglected in this
51
52
53 study.
54
55
56
57
58
59 14
60
1
2
3
2.3.2 Interphase heat transfer
4
5
6 The energy source term in Eq. (3) is modeled considering the convective heat transfer between
7
8 phases, which characterizes the thermal coupling between the Eulerian and Lagrangian phases.
9
10
11
The convective mechanism is a function of the overall heat-transfer coefficient. Finally, the
12
13 energy source term can be expressed as follows22:
14
O] O]
1 1
15
S Y Q/cO/& Y U & A 4T t T 5 with i 1, , n .
& &
V V
16
17
(28)
18 &^_ &^_
19
20
21 Due to the low temperature of the dispersed-phase, the heat transfer due to radiation is also
22
23 neglected.
24
25
26
27
28
29 2.4 TURBULENCE MODEL
30
31 The airflow inside a prilling tower is turbulent and the nature of the turbulence is associated
32
33
34 with high values for the Reynolds number. In practice, the turbulence, which is characterized by
35
36 instantaneous fluctuations in the scalar properties of the gas phase, such as velocity, pressure and
37
38 temperature, significantly increases the rates of mass, momentum and energy transfer and,
39
40
41 consequently, the droplet solidification rate. The turbulence in the continuous phase was
42
43 modeled using the standard k- model, initially proposed by Launder and Spalding23. This model
44
45
46
is widely used in practical engineering problems due to its low computational cost. The k-
47
48 model uses the Boussinesq approximation, which introduces the concept of turbulent viscosity.
49
50 The reader is referred to Speziale24 for a review of turbulence models. The k- model introduces
51
52
53 two conservation equations for turbulence, one for its kinetic energy and the other for its
54
55 dissipation rate, which are, respectively, expressed by:
56
57
58
59 15
60
1
2

k d v k S U k h P ,
3
t g
4 (29)
5
6

7
d v S U h C_ P C: .
8
t k k
9 (30)
10
11
12
13
14
15
16
17
2.5 TURBULENT DISPERSION OF PARTICLES
18
19 Turbulent effects associated with the continuous phase can be accounted for in the dispersed-
20
21 phase, if the instantaneous velocity of the continuous phase at each point of the dispersed-phase
22
23
24 trajectory is known25. This velocity is a function of the fluctuations that occur in the gas phase
25
26 velocity. In this study, a stochastic model is used to predict the instantaneous velocity, in which
27
28 the fluctuations are predicted by a discrete random walk (DRW) model26. In this procedure, the
29
30
31 particle interacts with eddies over the shortest lifetime and crossing time, and when this value is
32
33 reached a new instantaneous velocity is calculated. Thus, the turbulent dispersion of the particles
34
35 (TDP) is predicted by integrating the trajectory equation for each of the particles using the
36
37
38 instantaneous velocity of the continuous phase.
39
40
41
42
43 2.6 DEFINITION OF PERFORMANCE PARAMETERS
44
45
46 In order to quantify the number of completely solidified droplets and the number of broken
47
48 particles, two industrial performance parameters are defined as follows.
49
50
51 2.6.1 Solidification Index
52
53
54 The aim of determining the solidification index is to quantify the percentage of solidification
55
56 for each droplet. This index varies from 0, where no solidification occurs, to 1, where the droplet
57
58
59 16
60
1
2
3
is completely solidified. This parameter is defined as the ratio between the solid volume formed
4
5
6 during the solidification and the total solid volume that could be formed. As discussed
7
8 previously, one percent of the total volume of the droplet remains in the liquid phase. Thus, the
9
10
11
solidification index for the droplet i is defined as:
V#c(
&
12
I#c( ,
&
13
VcW(_#c(&[
14 & (31)
15
16
17
18 or
19
T
& T
r 4r & 5
&
20
I#c( .
&
21
& T
r 1
22 (32)
23
0,01
24
25
26
27 Since the solidification index can vary from one droplet to another, it is convenient to define
28
29 the averaged solidification index,
30
I#c(
&
31
I#c( i 1, , N/ .
N/
32 (33)
33
34
35
36
2.6.2 Breaking Index
37
38
39 This parameter is defined to compute the percentage of droplets that break after hitting the
40
41 walls of the prilling tower. In this case, the term "break" means that the CFD program will cease
42
43
44
to integrate the trajectory of the droplets and abort them from the discrete domain. Besides the
45
46 breaking of liquid droplets, partially solidified droplets may also break after impact.
47
48 To account for this phenomenon, it was defined as a breaking criterion that liquid droplets will
49
50
51 break after hitting the walls while partially solidified droplets may or may not break after impact.
52
53 This definition considers that droplets with a solid phase thickness of < 30% of the droplet radius
54
55 will break, while the other droplets will be reflected. During the reflection, the normal restitution
56
57
58
59 17
60
1
2
3
coefficient is 1 and the tangential restitution coefficient is 0.3. This criterion is implemented in
4
5
6 the code developed. Thus, the breaking index is defined as:
N/qpc\[
7
Ip\Wg&O .
8
N/
9 (34)
10
11
12
13
14
15
16 3 NUMERICAL PROCEDURE
17
18
19
ANSYS FLUENT22 was used as the CFD code in this study and, by default, it is not possible
20
21 to carry out Lagrangian simulations considering the solidification of liquid droplets. However,
22
23 the software allows the user to write specific functions and insert them into the software, as user-
24
25
26 defined functions (UDF)27.
27
28 In its general form, the code developed integrates the thickness of the solid phase formed in the
29
30 droplets along the droplet trajectory and informs the software which of the three thermal periods
31
32
33 relates to the urea droplets. To provide this information, a solidification law was implemented,
34
35 and a criterion for switching between the standard laws of the software and the solidification law
36
37
38
was also created. In addition, a source term to take into account the latent heat released during
39
40 the solidification of the urea droplets to the gas phase was implemented along with a break
41
42 criterion (section 2.6.2).
43
44
45 Before the numerical experimentation evolved to industrial cases, simpler cases were
46
47 considered to test the physical and numerical consistency of the code implemented in the
48
49 software, as discussed by Ricardo et al.28 The hydrodynamic characteristics of the airflow, the
50
51
52 solidification phenomenon and the trajectory of the dispersed phase were analyzed considering
53
54 the effect of the TDP model as well as the solidification and breaking indexes of the particles
55
56 after hitting the walls of the tower. However, in a first analysis, the effects of the TDP model and
57
58
59 18
60
1
2
3
the breaking of particles were neglected in order to allow a comparison of the results. For each
4
5
6 analysis three droplet diameters were considered. Table 1 shows the numerical scenarios of this
7
8 research.
9
10
11 Table 1. Numerical scenarios*.
12
13
14
15 All simulations were performed in the same manner, by solving the equations of motion and
16
17 continuity for the continuous phase in the steady state. Initially, a first-order upwind scheme for
18
19 the equations of motion was used and convergence was then obtained using a second-order
20
21
22 upwind scheme. For the pressure-velocity coupling the SIMPLEC algorithm was used. After a
23
24 stable numerical solution for the continuous phase had been obtained, injections of urea were
25
26
included in the calculation. In the Eulerian-Lagrangian procedure, the equations of motion,
27
28
29 continuity and energy were solved in the steady state using a pseudo-transient formulation. The
30
31 pressure-velocity coupling scheme was used for the solution.
32
33
34
3.1 GEOMETRY AND NUMERICAL MESH
35
36
37 Two geometries and numerical meshes were generated for the numerical analysis. In order to
38
39 represent a real industrial case the bottom of the prilling tower has two air inlets, as shown in
40
41
42
Figure 2. The geometry used to analyze the industrial case considering the effect of the TDP
43
44 model and the particle breaking is represented in Figure 2a. The geometry used to compare the
45
46 results neglecting the TDP model is shown in Figure 2b. The prilling tower consists of nine spray
47
48
49 nozzles to ensure the required urea production rate. The light blue areas at the top represent the
50
51 air outlets. These geometries differ only in terms of the air outlets. Both numerical meshes have
52
53 hexahedral and tetrahedral elements. The hybrid mesh considerably reduces the time required to
54
55
56
57
58
59 19
60
1
2
3
perform the simulations. The numerical properties related to the two meshes are shown in Table
4
5
6 2.
7
8
9 Figure 2. Characteristics of the prilling tower used: (a) for the analysis considering the
10
11 breaking of particles and the TDP model; and (b) for the comparison neglecting the TDP model.
12
13
14 All dimensions are expressed in meters.
15
16
17
18
19
Table 2. Details of the numerical meshes.
20
21
22
23
24 3.2 BOUNDARY CONDITIONS
25
26
The boundary conditions used for the inlet and outlet were the mass flow inlet and pressure
27
28
29 outlet, respectively. The temperature of the air at the inlet was the same as the backflow
30
31 temperature at the outlet, that is, 301.15 K. The turbulence intensity for all inlets and outlets was
32
33
34
5%. The hydraulic diameters for the bottom inlet, the side inlet and the outlets were 8.415 m,
35
36 2.11211 m and 1.8 m, respectively (see Figure 2 for further details regarding the geometry). A
37
38 temperature of 413.15 K and a velocity of 0.02 m/s were used for all urea injections. The
39
40
41 droplets injections were made in nine circular surfaces with 0.3 m in diameter, one of them
42
43 located on the center of the tower, and others located in a circle with radius of 3 m at 0.5 m
44
45 below the top of the tower. The walls of the tower were considered to be stationary, free-slip and
46
47
48 without roughness. The physical and thermal properties of the urea used in all simulation are
49
50 contemplated in Table 3, whereas the physical and thermal properties of the air are summarized
51
52
in Table 4.
53
54
55 Table 3. Properties of urea.
56
57
58
59 20
60
1
2
3
Table 4. Properties of air*.
4
5
6
7
8
9
10
11
12 3.3 NUMERICAL ANALYSIS OF THE NUMBER OF COMPUTATIONAL PARTICLES
13
14 Based on the mass flow of urea and the droplet diameter of 1.2 mm, the real number (N# ) of
15
16
droplets is around 14 million per second. It is practically impossible to perform numerical
17
18
19 simulations considering this number of particles due to the computational effort required. Thus,
20
21 numerical simulations were carried out considering a number of computational particles (N/ ) that
22
23
24 is much smaller than the real number of particles. The sum of these computational particles
25
26 represent the total mass flow of the real particles, but only the trajectories of the computational
27
28 particles are computed in the entire domain. Four initial cases were performed considering
29
30
31 different numbers of particles to verify if the number of computational particles could
32
33 significantly influence the numerical results. The geometry shown in Figure 2b was simulated
34
35 taking into account the TDP model and the breaking criterion, and the results are shown in Table
36
37
38 5.
39
40 It can be observed that on increasing the number of computational particles there is no
41
42
43
significant changes in the properties, for instance, the overall and convective heat transfer
44
45 coefficients, the core and surface temperature of the droplets, and the solidification index. On the
46
47 other hand, the breaking index, from 30,000 to 45,000 particles, shows a variation that almost
48
49
50 disappears by increasing the number of computational particles. This could be attributed to the
51
52 use of the TDP model, since the DRW model generates random numbers to satisfy the equations.
53
54 Since the variation in the breaking index is very small, the simulations in this research were
55
56
57 carried out considering 45,000 computational particles.
58
59 21
60
1
2
3
Table 5. Influence of the number of computational particles on process variables at the bottom
4
5
6 of the tower.
7
8
9
10
11
12 4 NUMERICAL RESULTS
13
14
15 This section discusses the main results of this study. Firstly, a brief discussion on using the
16
17 TDP model and the effect of the geometry of the tower on the droplet trajectories is presented.
18
19 The following variables and properties are then discussed:
20
21
22
23 Particle-related variables and properties: solidification and breaking indexes;
24
25 overall heat-transfer coefficient; core and surface temperatures; evolution of the solid
26
27 phase thickness and surface temperature along the particle trajectories; and the average
28
29
30 temperature profile at the radial position of the particle;
31
32 Air-related variables: temperature and velocity fields.
33
34
35
36
37
38 4.1 EFFECT OF THE TURBULENT DISPERSION OF PARTICLES
39
40 The aim of this initial analysis was to demonstrate the influence of the TDP model on the
41
42
43
trajectories of the urea droplets, as shown in Figure 3. It should be noted that although 45,000
44
45 droplets were simulated, only 100 are demonstrated, colored according to the evolution of the
46
47 solid phase thickness.
48
49
50 Figure 3a shows trajectories of the droplets without TDP model where they follow straight
51
52 from top to the bottom of the tower without spreading of the spray. On the other hand, Figure 3b
53
54 which is applied TDP model, illustrates a more realistic scenario in which the droplets follow
55
56
57 different trajectories depending on their contact with the air phase. Although this model is
58
59 22
60
1
2
3
generally used for very small particles, the results show a significant change in the droplet
4
5
6 trajectories due to the fluctuations in the air velocity. Global variables are analyzed in the
7
8 following, since there is no reason to demonstrate the results obtained disregarding the TDP
9
10
11
model.
12
13
14 Figure 2. Trajectories of the droplets: (a) disregarding the TDP model, (b) applying the TDP
15
16 model.
17
18
19
20
21
22
23 4.2 EFFECT OF THE GEOMETRY ON THE DROPLET TRAJECTORY
24
25 Figure 4 compares a symmetric airflow caused by symmetric air outlets (a) and an asymmetric
26
27 airflow caused by asymmetric air outlets. The droplet trajectories are colored according to the
28
29
30 solid phase thickness. Figure 3 clearly shows the effect of the air outlets on the trajectories of the
31
32 urea droplets, even those with a relatively large particle diameter. Thus, the tests to analyze the
33
34 urea prilling process were carried out using asymmetric air outlets and considering the TDP
35
36
37 model.
38
39
40 Figure 3. Trajectories of the droplets: (a) symmetric flow, (b) non-symmetric flow.
41
42
43
44
45
46
4.3 ANALYSIS OF THE GEOMETRY AND THE BREAKING OF THE PARTICLES
47
48
49 The objective of this numerical analysis was to cause non-uniformity in the airflow through
50
51 changes in the air outlets at the top of the tower, as observed in an industrial setting. In Figure 5
52
53
54
it can be noted that the droplet trajectories are directly influenced by the fluid dynamics
55
56 characteristics of the air stream. Many droplets hit the walls of the tower in Figure 5a and an
57
58
59 23
60
1
2
3
increase in the droplet diameter reduces these impacts, as can be seen in Figure 5c. In Figure 5a,
4
5
6 most droplets are completely solidified in the first half of the tower (left side) while the
7
8 remainder are solidified only at the bottom. Figure 5b shows that 2/3 of the tower was required
9
10
11
for part of the droplets to solidify (left side) and basically the whole tower was needed for the
12
13 rest of droplets. Finally, in Figure 5c, only a few droplets are completely solidified and for this
14
15 the whole tower was required.
16
17
18
19
Figure 4. Droplet trajectories colored according to core temperature: (a) dp = 1.2 mm; (b) dp =
20
21 1.465 mm; and (c) dp = 1.73 mm.
22
23
24 The droplet trajectories, colored according to the surface temperature, for droplets with
25
26
diameters of 1.2, 1.465 and 1.73 mm are shown in Figure 6. It can be observed in Figures 6a and
27
28
29 6b that the droplets collected at the bottom of the tower show only slight differences in their
30
31 surface temperatures. On the other hand, for droplets with a diameter of 1.73 mm the surface
32
33
34
temperature is 22% higher. Thus, increasing the diameter of the urea droplets increases their
35
36 surface temperature at the base of the tower. Even at low surface temperatures, droplets for
37
38 which the core temperature is equal to the melting temperature can be collected, due to the high
39
40
41 conductive resistance to heat transfer, as discussed below.
42
43
44 Figure 5. Droplet trajectories colored according to surface temperature: (a) dp = 1.2 mm; (b) dp =
45
46 1.465 mm; (c) dp = 1.73 mm.
47
48
49
Figure 7 shows the air temperature field. The air temperature, from the middle to the top of the
50
51
52 tower, decreases with an increase of the droplet diameter. This is because on increasing the
53
54 droplet diameter, the solidification index decreases, lowering the release of the latent heat of
55
56
57
solidification and, consequently, the air temperature decreases. From the middle to the bottom of
58
59 24
60
1
2
3
the tower, the opposite occurs. In Figure 7a, the low temperature in the lower region of tower
4
5
6 results from the complete solidification of the droplets, which reduces the overall coefficient due
7
8 to the increase in the internal resistances to the heat transfer. In Figure 7b and 7c, the droplets
9
10
11
continue to solidify in the lower region and lose heat, which increases the air temperature. These
12
13 aspects can be easily observed in Figure 8.
14
15
16 Figure 6. Thermal field for the air: (a) dp = 1.2 mm (b) dp = 1.465 mm (c) dp = 1.73 mm.
17
18
19 Figure 7. Average air temperature: (a) dp = 1.2 mm; (b) dp = 1.465 mm; and (c) dp = 1.73 mm.
20
21
22
23 In relation to the velocity fields in Figure 9, it was observed from the stream lines (not shown
24
25 herein) that there are recirculating zones near the reducing-conical region at the bottom of the
26
27 tower. These zones arise due to the geometrical characteristics of the tower. The low-velocity
28
29
30 regions result from the concentration of droplets, which reduces the velocity of the air by
31
32 friction. The velocity near the walls of the tower is close to zero, as expected. The regions of
33
34 higher velocity appear due to the side inlet and outlets for the air (see Figure 2). The asymmetric
35
36
37 air outlets change the direction of the airflow, directing it toward the top-right side of the tower,
38
39 since there is no outlet on the top-left side. Thus, the asymmetry of the airflow leads the droplets
40
41
42
or particles toward the walls of the tower.
43
44
45 Table 6 indicates that on increasing the droplet diameter and adjusting the air mass flow rate
46
47 and the geometric characteristics of the tower, the solidification index decreases significantly.
48
49
The solidification index for droplets with a diameter of 1.2 mm reached a maximum of 1, but
50
51
52 some droplets with 1.465 and 1.73 mm did not completely solidify. As previously discussed,
53
54 many droplets hit the walls of the tower. As shown in Table 6, the breaking index for droplets
55
56
57
with a diameter of 1.2 mm was 0.001355. This value represents a loss of approximately 2.032
58
59 25
60
1
2
3
ton/day. The breaking index of droplets with a diameter of 1.465 mm is 61 times lower than that
4
5
6 of droplets with a diameter of 1.2 mm. The loss of production associated with the diameter is 33
7
8 kg/day. The larger the diameter of the droplets the greater the gravitational force acting on the
9
10
11
droplets will be and thus fewer droplets will hit the walls of the tower. As a result, no droplets
12
13 with a diameter of 1.73 mm broke up after hitting the walls. Higher solidification indexes means
14
15 that a greater amount of latent heat of solidification is released. Consequently, the air
16
17
18 temperature at the top of the tower is higher for a droplet diameter of 1.2 mm compared with
19
20 1.465 and 1.73 mm. Therefore, increasing the droplet diameter reduces the average air
21
22 temperature at the top of the tower. This behavior can be explained by considering the overall
23
24
25 heat-transfer coefficient during the first step of the solidification. In this step, the convective
26
27 resistance has a stronger influence on the conductive resistance, which increases with an increase
28
29
in the droplet diameter and reduces the convective coefficient. Consequently, the overall heat
30
31
32 transfer coefficient is lower. Since the overall heat transfer coefficient decreases during the
33
34 solidification phenomenon, the surface temperature of the droplets increases. For droplet
35
36
37 diameters of 1.2 and 1.465 mm the core temperature is below 132.7C, which means the droplets
38
39 are completely solidified. The overall heat-transfer coefficients for droplet diameters of 1.2 and
40
41 1.465 mm are very similar and for 1.73 mm this coefficient is higher, since the solidification
42
43
44 index is lower due to the greater internal resistance to heat transfer.
45
46
47 Table 6. Solidification/breaking indexes and average thermal variables.
48
49
50 Figure 8. Numerical results for air: velocity fields for (a) dp = 1.2 mm, (b) dp = 1.465 mm and (c)
51
52
dp = 1.73 mm.
53
54
55
56
57
58
59 26
60
1
2
3
Finally, Figure 10 shows the average radial temperature profile within the droplets in six
4
5
6 planes along the height of the tower. These results confirm the strong influence of the conductive
7
8 resistance during solidification. At a height of 55 m, the initial stage of solidification, the
9
10
11
conductive resistance is small; thus, the temperature difference between the core and the outer
12
13 surface of the droplets is not high and the temperature profile is linear. In the course of the
14
15 solidification, the conductive resistance increases significantly; thus, at the bottom of the tower,
16
17
18 the temperature difference between the core and the outer surface is high and the temperature
19
20 profile has non-linear characteristics. It should be noted that although the unsolidified core is
21
22 representative, it represents only 1% of the total droplet volume.
23
24
25 Figure 11 shows the evolution of the solidification index with the air mass flow. Droplets with
26
27 a diameter of 1.2 mm can be completely solidified with 100 kg/s of ambient air. It was observed
28
29
that 0.06% of the total number of droplets with this diameter was dragged by the air stream to the
30
31
32 top of the tower, considering an air mass flow rate of 350 kg/s. This behavior was not observed
33
34 in other simulations. For droplets with a diameter of 1.465 mm, increasing the air mass flow rate
35
36
37 from 150 to 200 kg/s did not significantly increase the averaged solidification index, and this
38
39 operational change is therefore not desirable. Moreover, to achieve the same level of
40
41 solidification for droplet diameters of 1.2 and 1.465 mm, the air mass flow rate should be at least
42
43
44 300 kg/s and droplets with a diameter of 1.73 mm are collected at the base of the tower with a
45
46 liquid core, considering an air mass flow rate of 150 kg/s.
47
48
49 Figure 9. Tridimensional illustration of radial temperature profile inside the solid phase formed
50
51
52 during solidification for droplet diameter of 1.2 mm.
53
54 Figure 10. Dependence of the solidification index on the mass flow of air for droplets with
55
56 diameters of 1.2, 1.465 and 1.73 mm.
57
58
59 27
60
1
2
3
5 CONCLUSIONS
4
5
6 This research was conducted in order to evaluate the urea prilling process considering
7
8 operational and geometric conditions close to those used at industrial plants. The results show
9
10
11
that on increasing the diameter of the droplets the air temperature in the upper region of the
12
13 tower decreases while in the lower region it increases. This behavior is influenced by the overall
14
15 heat-transfer coefficient and thus by the solidification phenomenon itself. The solidification
16
17
18 index is directly influenced by the diameter of the droplets and an increase in the droplet
19
20 diameter results in a reduction in the solidification index. This occurs because, for the same
21
22 operating conditions and geometry, the residence time is shorter since with an increase in the
23
24
25 droplet diameter there is greater gravitational force acting on the droplets. An increase in the
26
27 droplet diameter also reduces the average temperature of the air at the top of the tower. In
28
29
addition, it lowers the probability of the droplets hitting the walls of the tower, which may
30
31
32 contribute to their breaking and the formation of cake. One way to increase the solidification
33
34 index is to increase the height of the tower in order to extend the residence time of the droplets
35
36
37 within the tower; however, very high towers may become unfeasible. Another possibility is to
38
39 increase the air mass flow rate, but caution is needed to ensure that the droplets are not entrained
40
41 by the air stream, as described in section 4.3. Higher solidification indexes contribute to an
42
43
44 increase in the air temperature, since the release of the latent heat of solidification is also greater.
45
46 Although the complete solidification of a few droplets in some cases does not occur, all
47
48 solidification indexes were sufficiently high, reaching maximum values in many cases. Air
49
50
51 recirculation zones were observed through stream lines in regions of low velocity near the
52
53 conical walls at the bottom of the tower. The TDP model produces results which are more
54
55
56
consistent with the physics associated with the problem. It was verified that the fluid dynamics
57
58
59 28
60
1
2
3
characteristics of the air contribute to the breakup of the droplets when the flow in the tower is
4
5
6 non-uniform, which is the case in the industrial process. It was found that on increasing the
7
8 droplet diameter the breaking index of the droplets is reduced. This is explained by the fact that
9
10
11
larger droplets have greater gravitational force, in other words, the influence of the airflow is
12
13 diminished.
14
15
16
17
18
19
ACKNOWLEDGEMENTS
20
21 The authors are grateful to CAPES, CNPq (Process Number 310504-2012-0) and
22
23
24
PETROBRAS for the financial support of this research project.
25
26
27
28
29 ABBREVIATIONS
30
31
32 Symbols
a_ , a : , a T
33
34 dispersed-phase Reynolds number constants, [-]
35
A
&
36
37 dispersed-phase area, m 2
38
1
39 &
40 dispersed-phase acceleration, m/s2
C_ , C: , C
41
42 k- model constants, [-]
43
C3
&
44
45 drag coefficient, [-]
46
47
48
Cp specific heat, J/kg K
d
&
49
50 dispersed-phase diameter, m
51
F3
&
52
53 characteristic dispersed-phase response frequency, 1/s
54
F-
55 &
56 solidification dispersed-phase response frequency per temperature unit,
57
58
59 29
60
1
2
3
1/sK
4
F>
&
5
6 thermal dispersed-phase response frequency, 1/s
7
8
9
g gravitational acceleration, m/s2
h
&
10
11 convective heat-transfer coefficient, W/m2 K
12
H
&
13
14 dispersed-phase rate of heat exchange, K/s
15
H
16
17
latent heat of solidification, J/kg
I#c(
18
19 solidification index, [-]
Ip\Wg&O
20
21
breaking index, [-]
22
k
23
24 turbulent kinetic energy, m2/s2
25
26
27
k# dispersed-phase thermal conductivity, W/m K
m
28
29 dispersed-phase mass, kg
30
31
N- number of physical droplets/particles
32
Ns
33
34 number of computational droplets/particles
35
36 Nu Nusselt number, [-]
37
38
39 p total pressure, Pa
P
40
41 turbulent production due to viscous and gravitational force, J/s m3
42
43
44
Pr Prandtl number, [-]
Re
&
45
46 dispersed-phase Reynolds number, [-]
47
rt
&
48
49 dispersed-phase volumetric rate of solidification, kg/s m3
50
51
R & N&O minimum internal liquid radius, m
52
S
53
54 energy source term, W/m3
55
56
momentum source term, N/m3
57
58
59 30
60
1
2
3
t time, s
4
5
6 Tg continuous-phase temperature, K
7
T
8 &
dispersed-phase temperature, K
9
U &
10
11 global heat-transfer coefficient, W/m2 K
12
v
13
14
velocity vector, m/s
15
16 V volume, m3
17

18 &
19
dispersed-phase velocity, m/s
%
&
20
21 dispersed-phase element position, m
22
23
24 Greek Letters

25
26 density, kg/m3
27

28
29 effective stress tensor, N/m2

30
31 effective thermal conductivity, W/mK
32

33
34
continuous-phase absolute viscosity, Pas

&
35
36 dispersed-phase rate of radius change due to solidification, m/s
37
&
38
39 dispersed-phase solid thickness, m

40
41 continuous phase turbulent viscosity, Pas
42

43
44 turbulent kinetic energy dissipation, m2/s2
45
46
47
k , g k- model constants, [-]

&
48
49 dispersed-phase response time, s
50
-
&
51
52 solidification dispersed-phase response time, sK
53
>
54 &
55 Thermal dispersed-phase response time, s
56
57 Superscript
58
59 31
60
1
2
3
(i) referred to droplet/particles i, [-]
4
5
6 Subscript
7
8 0 initial
9
10
11
f fusion
12
13 g continuous phase
14
15 l liquid state
16
17
18 max maximum
19
20 min minimum
21
22 p dispersed-phase
23
24
25 s solid state
26
27
28
29
30 REFERENCES
31
32
33
34
(1) Bertn, D. E.; Mazza, G. D.; Pia, J.; Bucal, V. Modeling of an industrial fluidized-bed
35 granulator for urea production. Ind. Eng. Chem. Res. 2007, 46, 7667-7676.
36
37
38 (2) Bertn, D. E.; Pia, J.; Bucal, V. Dynamics of an industrial fluidized-bed granulator for urea
39
40 production. Ind. Eng. Chem. Res. 2010, 49, 317-326.
41
42 (3) Cotabarren, I. M.; Bertn, D. E.; Romagnoli, J.; Bucal, V.; Pia, J. Dynamic simulation and
43
44 optimization of a urea granulation circuit. Ind. Eng. Chem. Res. 2010, 49, 6630-6640.
45
46
47 (4) Heinrich, S.; Henneberg, M.; Peglow, M.; Drechsler, J.; Mrl, L. Fluidized bed spray
48
49
granulation: analysis of heat and mass transfers and dynamic particle populations. Braz. J. Chem.
50 Eng. 2005, 22(2), 181-194.
51
52
53 (5) Bakhtin, L. A.; Vagin, A. A.; Esipovich, L. Ya.; Labutin, A. N. Heat-exchange
54
55 calculations in prilling tower. Chem. Pet. Eng. 1978, 14(11), 994-999.
56
57
58
59 32
60
1
2
3
(6) Yuan, W., Chuanping, B., Yuxin, Z. An innovated tower-fluidized bed prilling process.
4
5 Chin. J. Chem. Eng. 2007, 15(3), 424-428.
6
7
8 (7) Mehrez, A.; Ali, A. H. H.; Zahra, W. K.; Ookawara, S.; Suzuki, M. Study on heat and
9
10 mass transfer during urea prilling Process. Int. J. Chem. Eng. Appl. 2012, 3(5), 347-353.
11
12 (8) Mehrez, A.; Ookawara, S.; Ali, A. H. H.; Suzuki, M. A numerical study on cooling-
13
14 solidification of urea particles in prilling tower. J. Chem. Eng. Japan. 2014, 47(1), 1-8.
15
16
17 (9) Bird, R. B.; Stewart, W. E.; Lightfoot, E. N. Transport Phenomena, 2nd ed; Wiley: New York,
18
2001 (revised 2007).
19
20
21 (10) Maxey, M. R.; Riley, J. J. Equation of motion for small rigid sphere in a nonuniform flow.
22
23 Phys. Fluids. 1983, 26(4), 883-889.
24
25
26
(11) Tsuji, Y.; Tanaka, T.; Ishida, T. Lagrangian numerical simulation of plug flow of
27 cohesionless particles in a horizontal pipe. Powder Technol. 1992, 71, 239-250.
28
29
30 (12) Sommerfeld, M. Validation of a stochastic Lagrangian modelling approach for inter-
31
32 particle collisions in homogeneous isotropic turbulence. Int. J. Multiphase Flow 2001, 27,
33
1829-1858.
34
35
36 (13) Drew, D. A. Mathematical modeling of two-phase flow. Annu. Rev. Fluid Mech. 1983, 15,
37
38 261-291.
39
40
(14) Enwald, H.; Peirano, E.; Almstedt, A. E. Eulerian two-phase flow theory applied to
41
42 fluidization. Int. J. Multiphase Flow 1996, 22, 21-66.
43
44
45 (15) El-Behery, S. M.; El-Askary, W. A.; Hamed, M. H.; Ibrahim, K. A. Numerical and
46
47 experimental studies of heat transfer in particle-laden gas flows through a vertical riser. Int.
48
J. Heat Fluid Flow 2012, 33, 118-130.
49
50
51 (16) Rybalko, M.; Loth, E.; Lankford, D. A Lagrangian particle random walk model for
52
53 hybrid RANS/LES turbulent flows. Powder Technol. 2012, 221, 105-113.
54
55
(17) Morsi, S. A.; Alexander, A. J. An investigation of particle trajectories in two-phase flow
56
57 systems. J. Fluid Mech. 1972, 55 (2), 193-208.
58
59 33
60
1
2
3
(17) Squires, K. D.; Eaton, J. K. Particle response and turbulence modification in isotropic
4
5 turbulence, Phys. Fluids. 1990, 2(7), 1191-1203.
6
7
8 (18) Ranz, W. E.; Marshall Jr., W. R. Evaporation from drops. I. Chem. Eng. Prog. 1952, 48(3),
9
10 141-146.
11
12 (19) Zhu, H. P.; Zhou, Z. Y.; Yang, R. Y.;Yu, A. B. Discrete particle simulation of
13
14 particulate systems: A review of major applications and findings. Chem. Eng. Sci. 2008, 63,
15
16 5728-5770.
17
18
(20) Berlemont, A.; Chang, Z.; Gouesbet, G. Particle lagrangian tracking with
19
20 hydrodynamic interactions and collisions. Flow Turbul. Combust. 1998, 60, 1-18.
21
22
23 (21) Souza, F. J.; Silva, A. L.; Utzig, J. Four-way coupled simulations of the gasparticle flow
24
25 in a diffuser. Powder Technol. 2014, 253, 496-508.
26
27 (22) Ansys, Ansys Fluent 13.0 Theory Guide; ANSYS, Inc.: Canonsburg, PA, 2010; v.13.0.5
28
29
30 (23) Launder, B. E.; Spalding, D. B. Lectures in Mathematical Models of Turbulence; Academic
31
32 Press: London, 1972.
33
34 (24) Speziale, C. G. Analytical methods for the development of Reynolds stress closures in
35
36 turbulence. Annu. Rev. Fluid Mech. 1991, 23, 107-57.
37
38
39 (25) Boulet, P.; Moissette, S.; Andreux, R.; Oesterl, B. Test of an Eulerian-Lagrangian
40
simulation of wall heat transfer in a gas-solid pipe flow. Int. J. Heat Fluid Flow 2000, 21,
41
42 381-387.
43
44
45 (26) Gosman, A. D.; Ioannides, E. Aspects of computer simulation of liquid-fueled combustors.
46
47 J. Energy 1983, 7(6), 482-490.
48
49 (27) Ansys, Ansys Fluent 13.0: UDF Manual; ANSYS, Inc.: Canonsburg, PA. 2010;
50
51 v.13.0.5
52
53
54 (28) Ricardo, G. A. N.; Noriler, D.; Martignoni, W. P.; Meier, H. F. Eulerian-Lagrangian
55
analysis of multiphase flow in urea prilling process with phase changing. Chem. Eng. Trans.
56
57 2013, 32, 2173-2178.
58
59 34
60
1
2
3
(29) Saleh, H. N.; Ahmed, S. M.; Al-mosuli, D.; Barghi, S. Basic design methodology for
4
5 prilling tower. Canadian J. Chem. Eng. 2015, 93, 1403-1409.
6
7
8 (30) Kozyro, A. A., Dalidovich, S. V., Krasulin, A. P. Heat capacity, enthalpy of fusion, and
9
10 thermodynamics properties of urea, Zhur. Prikl. Khim. (Leningrad), 59, 1456-1459, 1986.
11
12 (31) Ruzicka, V.; Domalski, E. S. Estimation of the Heat Capacities of Organic Liquids as a
13
14 Function of Temperature Using Group Additivity. II. Compounds of Carbon, Hydrogen,
15
16 Halogens, Nitrogen, Oxygen, and Sulfur. J. Phys. Chem. Ref. Data 1993, 22, 619-657.
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59 35
60
1
2
3
Table 1. Numerical scenarios*.
4
5
6
Mass Flow Mass Flow Droplet
7
8 Rate of Urea Rate of Air Diameter
9 (ton/h) (kg/s) (mm)
10 1.200
11
12 100.0 1.465
13 1.730
14
15 1.200
16 150.0 1.465
17
18 1.730
19 1.200
20
21 200.0 1.465
22 1.730
23 62.5
24 1.20
25 250.0 1.465
26
27 1.730
28 1.20
29
30 300.0 1.465
31 1.730
32
33 1.200
34 350.0 1.465
35
36 1.730
*
37 For all numerical analysis, the air and the initial urea
38 temperature were 301.15 and 413.15 K, respectively;

39 Based on industrial process8,29, the droplet diameters were
40 chosen as: dp = 1.2 mm (minimum); dp = 1.465 mm (central
41 midpoint); dp = 1.73 mm (maximum).

42 Only the solidification index is analyzed considering these
43 mass flows, the remainder of the analysis is performed only
44 considering 150 kg/s of air.
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59 36
60
1
2
3
4
5 Table 2. Details of the numerical meshes.
6
7
8 Element Quality/Volume (m3)
9 Nodes Elements
10 Minimum Maximum Average
11 Comparison
12 neglecting
13 201,000 501,000 0.25/2.6710-7 0.99/7.6710-2 0.89/1.2110-2
the TDP
14
15 model
16 Analysis of
17 TDP model
18 and the
19 402,000 735,000 0.24/2.7710-7 0.99/3.3110-2 0.87/8.2310-3
breaking and
20
21 solidification
22 of particles
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59 37
60
1
2
3
Table 3. Properties of urea.
4
5
6
7 Variable Value
8 *
9 Density of solid urea 1,335.00 [kg/m3]
10
11
Melting point 132.7 [C]
12 Thermal conductivity of
13 0.0454 [W/mK]
solid urea
14
15 Melting heat 224,000.00 [J/kg]
16
Specific heat of solid urea Cp# 640.5 0.83T 1.175 10x: T : 1.435 10xy T T [J/kgK]
17
Cp( 8060.242 18.23T 8.40917 10x{ T : [J/kgK]
18 Specific heat of liquid
19 urea
20 *
Used the same value for the liquid urea.
21
Expression proposed by Kozyro30. The range of validity of the equation is between 240 and 400 K.
22
The group contribution Ruzicka-Domalski31 method was used to obtain an expression for estimating the specific
23 heat of urea in liquid phase.
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59 38
60
1
2
3
4
Table 4. Properties of air*.
5
6
7
8 Variable Value
9
10 Density 1.225 [kg/m3]
11
12
Viscosity 1.789410-5 [kg/ms]
13 Thermal conductivity 0.0242 [W/mK]
14
15 Specific heat 1006.43 [J/kgK]
16 *
Adapted from the solver database.
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59 39
60
1
2
3
4
5
6 Table 5. Influence of the number of computational particles on process variables at the bottom
7
8 of the tower.
9
10
N/
11
12
Averaged on particle number at the bottom of the tower
13 U h Solidification Breaking
Tcore (C) Tsurface (C)
14 (W/m2K) (W/m2K) Index Index
15
16 30,000 19.2909 269.6304 114.7437 34.2970 0.9899 0.00596
17
18
45,000 19.2909 269.6273 114.6679 34.2918 0.9899 0.00677
19 60,000 19.2904 269.5212 114.7776 34.3022 0.9899 0.00685
20
21 70,000 19.2907 269.5804 114.4087 34.2740 0.9899 0.00689
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59 40
60
1
2
3
Table 6. Solidification/breaking indexes and average thermal variables.
4
5
6
7 Solidification Index Thermal Variables (average*)
8 Breaking
Diameter Tair* Tcore Tsurface U
9 Minimum Average Maximum Index
10
(mm) (C) (C) (C) (W/m2K)
11 1.200 1.0000 1.0000 1.0000 0.0013550 74.26 115.79 34.37 19.29
12
1.465 0.9377 0.9966 1.0000 0.0000222 73.04 129.51 35.33 18.11
13
14 1.730 0.8288 0.9296 1.0000 0.0000000 71.44 132.7 43.13 34.82
15 *
The average procedure was made in a cross section at the top of the tower for air temperature, and over all particles
16
exiting of the domain for others thermal variables.
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59 41
60
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25 Figure 1. Representation of the unsolidified core model.
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59 42
60
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Figure 2. Characteristics of the prilling tower used: (a) for the analysis considering the breaking
44
45 of particles and the TDP model; and (b) for the comparison neglecting the TDP model. All
46
47 dimensions are expressed in meters.
48
49
50
51
52
53
54
55
56
57
58
59 43
60
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
.
36
37 Figure 3. Trajectories of the droplets: (a) disregarding the TDP model, (b) applying the TDP
38
39
40
model.
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59 44
60
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37 Figure 4. Trajectories of the droplets: (a) symmetric flow, (b) non-symmetric flow.
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59 45
60
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34 Figure 5. Droplet trajectories colored according to core temperature: (a) dp = 1.2 mm; (b) dp =
35
36
1.465 mm; and (c) dp = 1.73 mm.
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59 46
60
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34 Figure 6. Droplet trajectories colored according to surface temperature: (a) dp = 1.2 mm; (b) dp =
35
36
1.465 mm; (c) dp = 1.73 mm.
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59 47
60
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37 Figure 7. Thermal field for the air: (a) dp = 1.2 mm (b) dp = 1.465 mm (c) dp = 1.73 mm.
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59 48
60
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22 Figure 8. Average air temperature: (a) dp = 1.2 mm; (b) dp = 1.465 mm; and (c) dp = 1.73 mm.
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59 49
60
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39 Figure 9. Numerical results for air: velocity fields for (a) dp = 1.2 mm, (b) dp = 1.465 mm and (c)
40
41 dp = 1.73 mm.
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59 50
60
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41 Figure 10. Tridimensional illustration of radial temperature profile inside the solid phase formed
42
43
during solidification for droplet diameter of 1.2 mm.
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59 51
60
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24 Figure 11. Dependence of the solidification index on the mass flow of air for droplets with
25
26
27 diameters of 1.2, 1.465 and 1.73 mm.
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59 52
60

Application of Euler Langrange Approach To Predict, Ricardo 2015 PDF

Transféré par

Informations du document

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Application of Euler Langrange Approach To Predict, Ricardo 2015 PDF

Transféré par

Droits d'auteur :

Formats disponibles

Subscriber access provided by CMU Libraries - http://library.cmich.

Industrial & Engineering Chemistry Research is published by the American Chemical

Here, N# represent the number of physical particles or droplets. However, in Lagrangian

The particle response time, , is:

The thermal particle response time, > , is given by:

the overall heat-transfer coefficient. Thus, can be expressed as follows:

Equation (21) is valid only during the solidification phenomenon, when T T .

V( N&O 0.01 S U.

Vous aimerez peut-être aussi

Application of Euler Langrange Approach To Predict, Ricardo 2015 PDF

Transféré par

Informations du document

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Application of Euler Langrange Approach To Predict, Ricardo 2015 PDF

Transféré par

Droits d'auteur :

Formats disponibles

Subscriber access provided by CMU Libraries - http://library.cmich.

Industrial & Engineering Chemistry Research is published by the American Chemical

Here, N# represent the number of physical particles or droplets. However, in Lagrangian

The particle response time,  , is:

The thermal particle response time, > , is given by:

the overall heat-transfer coefficient. Thus,  can be expressed as follows:

Equation (21) is valid only during the solidification phenomenon, when T T .

V( N&O 0.01 S U.

Vous aimerez peut-être aussi

The particle response time, , is:

the overall heat-transfer coefficient. Thus, can be expressed as follows:

Equation (21) is valid only during the solidification phenomenon, when T T .

V( N&O 0.01 S U.