Applied Thermal Engineering 96 (2016) 240–249

Contents lists available at ScienceDirect

Applied Thermal Engineering

j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / a p t h e r m e n g

Research Paper

Tower characteristics correlation and parameter retrieval in wet-

cooling tower with expanded wire mesh packing
Rohit K. Singla, Kuljeet Singh, Ranjan Das *
School of Mechanical, Materials and Energy Engineering, Indian Institute of Technology Ropar, Punjab 140001, India

H I G H L I G H T S

• Experimental study of forced draft cooling tower with expanded wire mesh packing.
• Correlation for Merkel number is proposed.
• Feasible combinations of controlling parameters are estimated.
• Differential evolution-based optimization method is applied.

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

Article history: In this work, an experimental investigation is performed on a counter flow forced draft cooling tower
Received 31 May 2015 with expanded wire meshed fill as packing material. The effect of controlling parameters such as the air
Accepted 15 November 2015 and water flow rates on the performance parameters such as the range, water evaporation rate, heat
Available online 3 December 2015
rejection/gain rates, Merkel number, effectiveness and global heat and mass transfer coefficient has been
analyzed. A correlation is then subsequently developed for the Merkel number as a function of water
Keywords:
and air flow rates. Next, an optimization problem is solved to simultaneously estimate the controlling
Cooling tower
parameters in order to satisfy a given Merkel number using differential evolution. It is found from the
Merkel number
Correlation present study that many feasible combinations of controlling parameters satisfy a given value of the Merkel
Parameter retrieval number, which may be useful to the operator for regulating desired conditions.
Differential evolution © 2015 Elsevier Ltd. All rights reserved.

1. Introduction to know the actual performance. Many studies dealing with the
mathematical modeling, experimental investigation with differ-
Cooling tower is pertinent to many industrial applications which ent packing such as splash, honeycomb, ceramic tile, corrugated,
is based upon the concept of evaporative and convective cooling and film are reported in literature [6–10]. For example, an exper-
where the ambient air is brought into contact with hot water [1]. iment was conducted by Facão and Oliveira [11] to propose mass
Such systems are known as wet type cooling towers. However, in transfer coefficient correlation for indirect contact cooling towers.
a dry cooling tower there is no direct contact between the two fluids, Gao et al. [12] proposed correlations for range and efficiency against
so the heat is transferred only through convection. Based upon air parameters such as inlet temperature and flow rate of water. Mondal
direction, cooling towers can be classified as natural and forced draft et al. [13] developed correlations for performance parameters such
types. In natural draft, an upward movement of air is created through as temperature drop, effectiveness and tower characteristic ratio with
the tower due to the density differences, whereas forced draft cooling water-air mass flow rate. Additionally, several studies [14–16] are
tower uses a power driven fan to draw the air. Due to popularity available in literature for development of correlations of temper-
of wet type cooling tower, its performance evaluation and thermal ature drop, effectiveness and Merkel number against water-air flow
behavior study attracted the research community. Merkel [2], Poppe rate.
[3] and e−NTU [4] are some popular methods used to evaluate the It is seen that the cooling tower performance may be approxi-
performance of wet cooling towers and it is found that Merkel and mated by parameters such as Merkel number, Me, effectiveness, e,
e−NTU methods produce identical results [5]. Mathematical mod- global heat and mass transfer coefficient, Ka and evaporation rate,
eling of systems has many advantages such as saving money, M ev , which are again influenced by parameters such as flow rates
manpower and time, but experimental investigation is also necessary and temperatures of fluids. Many studies on cooling towers are based
upon the evaluation of performance characteristics using the knowl-
edge of input parameters such as water and air flow rates, tower
* Corresponding author. Tel.: +91 1881242245; fax: +91 1881223395. height, and inlet water temperature. These approaches fall under
E-mail address: ranjandas81@gmail.com (R. Das). the purview of forward problems. However, a different situation can

http://dx.doi.org/10.1016/j.applthermaleng.2015.11.063
1359-4311/© 2015 Elsevier Ltd. All rights reserved.
 R.K. Singla et al./Applied Thermal Engineering 96 (2016) 240–249 241

(a) (b)
Fig. 1. (a) Experimental setup, (b) Schematic diagram: Sensor locations: (i) L1, water tank; (ii) L2, DBT of inlet air; (iii) L3, WBT of inlet air; (iv) L4, cold water; (v) L5, hot
water; (vi) L6, DBT of outlet air; (vii) L7, WBT of outlet air; (viii) L8, ambient temperature.

be considered where the objective is to evaluate either single or mul- Firstly, the calibration of thermocouples, rotameter and ma-
tiple parameters satisfying a given performance objective, which nometer is carried out. In experimental procedure, first the tank
may be Me. This type of analysis falls under the domain of inverse water heater is lighted to raise the water tank temperature to 45 °C,
problems [17] and these are pertinent to the design of engineer- after which the blower and the centrifugal pump are switched ON
ing systems. The parameters which can be easily controlled by the to force the air through tower and to circulate water, respectively.
operator are considered to be the focus of inverse analysis [18]. The warm water is then distributed over the packing, which comes
It is evident that Me is a critical parameter in the design and per- into contact with air. Experiments are performed at three differ-
formance of cooling towers, but a correlation of Me with mass flow ent ambient temperatures (27 ± 1.5°C, 31.5 ± 0.75°C and 36 ± 1°C ). As
rate of air (ma) and water (mw) for expanded wire mesh packing air passes through the tower, it gains heat from water and in-
seems to be unavailable, and its investigation is very limited [19]. creases its humidity. All measurements are taken under steady-
Therefore, the present study involves two objectives, (i) develop- state in which water inlet along with tank temperatures is
ment of correlation for calculating Me as a function of ma, mw and maintained at 40 °C. The air flow rate, ma is measured by an orifice
(ii) solving an inverse problem to simultaneously estimate ma and plate using differential pressure indicated by a manometer.
mw satisfying a given Me. Since the solution of an inverse problem The experiment is repeated for five levels of ma and for each of
requires an optimization algorithm, thus, in this study the differ- them, four levels of mw are considered. Using the experimental data,
ential evolution (DE) technique has been used. The experimental different performance parameters are studied and subsequently a
setup and procedure is discussed in Section 2. correlation for Me is proposed. Furthermore, for attaining a partic-
ular value of Me, feasible combinations of ma and mw are evaluated
2. Experimental setup and procedure using DE-based optimization. Performance characteristics studied
in the present work are described next.
The experimental setup is a counter flow forced draft cooling
tower (Fig. 1). The water is heated in a tank by an electric heater
(2 kW) and a centrifugal pump is used to pump the hot water 3. Performance characteristics
through a rotameter (range: 0–11 LPH). The water is supplied at the
top to droplet water distributor, which distributes water uni- Performance parameters studied in this work are briefly de-
formly. An expanded wire mesh packing is used to increase the scribed below.
interfacial area between water and air. The cross sectional area ( A fi )
and height of packing (L fi ) are 0.09 m2 and 1.2 m, respectively. Air
is supplied at the bottom with the help of a centrifugal blower and 3.1. Range
its flow rate is measured by an orifice using the differential pres-
sure depicted by manometer. The temperature measurement at The difference between water inlet temperature and water outlet
various points is done with K-type thermocouples along with DAQ temperature is known as range of cooling tower [1], i.e.,
system by National Instruments. The experimentation procedure is
described next. R = Tw ,1 − Tw ,2 (1)
242 R.K. Singla et al./Applied Thermal Engineering 96 (2016) 240–249

3.2. Water evaporation rate The procedure of uncertainty analysis is discussed below.

When air and water come in contact, the water evaporates. The 4. Experimental uncertainty analysis
rate of evaporation can be calculated as below [1],
Considering uncertainties of individual instruments, it is im-
M ev = m a (ω 2 − ω1 ) (2)
portant to study their effect on the final result. The uncertainty
involved in the results is estimated using the theory given by Moffat
3.3. Heat rejection rate from water and heat gain by air [24], according to which, if ζ is a function of independent vari-
ables x 1, x 2, x 3, … x n , i.e.,
Considering convective and evaporative heat transfer along with
water loss through evaporation, Kloppers [20] proposed the fol- ζ = f ( x 1, x 2, x 3,… x n ) (10)
lowing relation (improved Merkel equation) to calculate the heat
rejection rate form water, Then uncertainties of independent variables can be accounted
as u1, u 2, u 3, … un . Therefore, considering the uncertainty of each in-
Q w = mw C pw Tw ,1 − (mw − M ev )C pw Tw ,2 (3) dependent variable, the uncertainty in the final result (uζ ) is
estimated in the following manner,
The heat gained by air can also be estimated as below [20,21],
⎧ ∂ζ ⎞ 2 ⎫
(uζ )2 = ∑ ⎪⎨⎛⎜⎝ 2⎪
n

Q a = m a (ha ,2 − ha ,1 ) (4) ⎟ (ui ) ⎬ (11)

⎪ ∂x i ⎠
i =1 ⎩ ⎪⎭

3.4. Merkel number The above equation is expanded as below,

In the cooling tower, two types of heat transfer processes occur ⎧⎪⎛ ∂ζ ⎞ 2 ⎛ ∂ζ ⎞ 2 ⎛ ∂ζ ⎞ 2 ⎛ ∂ζ ⎞ ⎫⎪
2

uζ = ⎨⎜ u1 ⎟ + ⎜ u2 ⎟ + ⎜ u3 ⎟ + … + ⎜ un ⎟ ⎬ (12)
⎩⎪⎝ ∂x 1 ⎠ ⎝ ∂x 2 ⎠ ⎝ ∂x 3 ⎠ ⎝ ∂x n ⎠ ⎭⎪
simultaneously, i.e. sensible and evaporative heat transfer. This was
introduced in Merkel theory [2] which proposed the following equa-
tion, Based on the above, the relative uncertainty is expressed in the
following manner [24],
mw C pw dT = KadV (hsw − ha ) = m adha (5)

Using this, the Merkel number is evaluated as [2], uζ ⎧⎪⎛ 1 ∂ζ ⎞ 2 ⎛ 1 ∂ζ ⎞ 2 ⎛ 1 ∂ζ ⎞ 2 ⎛ 1 ∂ζ ⎞ ⎫⎪

2

= ⎨⎜ u1 ⎟ + ⎜ u2 ⎟ + ⎜ u3 ⎟ + … + ⎜ un ⎟ ⎬
ζ ⎪⎩⎝ ζ ∂x 1 ⎠ ⎝ ζ ∂x 2 ⎠ ⎝ ζ ∂x 3 ⎠ ⎝ ζ ∂x n ⎠ ⎭⎪
w ,2 T
KaV C pw dT (13)
Me = = ∫ (6)
mw Tw ,1 hsw − ha
For the present work, uncertainties involved in different instru-
For Eq. (6), the numerical integration is done using Chebyshev ments used in this experiment for measurements are evaluated as
method as mentioned below [22]: below,

w ,2 T
C pw dT Tw ,1 − Tw ,2 uma u u
(C pw ) ⎡⎢ +
KaV 1 1 1 1 ⎤ = 0.0125; mw = 0.014; T = 0.000195 (14)
= ∫ ≅ + + (7)
h 4 ⎥⎦
ma mw T
mw Tw ,1 hsw − ha 4 Δ
⎣ 1
h Δh 2 Δh 3 Δ
Using Eqs. (10)–(14), relative uncertainties involved in differ-
where: ent performance parameters based upon experimental results can
be calculated as below,
Δh1 = value of [hsw at {Tw ,2 + 0.1(Tw ,1 − Tw ,2 )} − {ha ,1 + 0.1(ha ,2 − ha ,1 )}]
Δh2 = value of [hsw at {Tw ,2 + 0.4 (Tw ,1 − Tw ,2 )} − {ha ,1 + 0.4 (ha ,2 − ha ,1 )}] uR u uQ w u
= 0.0021; M ev = 0.0125; = 0.0147; Q a = 0.0125;
Δh3 = value of [hsw at {Tw ,1 − 0.4 (Tw ,1 − Tw ,2 )} − {ha ,2 − 0.4 (ha ,2 − ha ,1 )}] R M ev Qw Qa
Δh 4 = value of [hsw at {Tw ,1 − 0.1(Tw ,1 − Tw ,2 )} − {ha ,2 − 0.1(ha ,2 − ha ,1 )}] u Me u
= 0.00198; e = 0.0019;
uKa
= 0.0194 (15)
Me e Ka
3.5. Effectiveness After proposing correlation for Me, the second objective is to solve
an inverse problem to predict feasible combinations of air and water
The effectiveness of a cooling tower is defined as the ratio of range flow rates satisfying a given Me. Another objective of solving an
to ideal range, i.e. [4], inverse problem is to verify the goodness of the proposed correla-
tion. For this, the following objective function is minimized,
(Tw ,1 − Tw ,2 )
e= (8)
(Tw ,1 − Twb ,1 )  − Me (m a, mw )⎤ 2
F = ⎡⎣Me (16)
⎦
Alternatively, the effectiveness is the ratio of the range to the sum  is the exact value of Merkel number and Me (m a, mw )
where Me
of range and approach.
is the one obtained by proposed correlation which is again a func-
tion of mass flow rates of air and water. To minimize the objective
3.6. Global heat and mass transfer coefficient function, F, DE is used which is discussed below.

The amount of heat and mass transfer inside the tower is rep-
5. Differential evolution based inverse method
resented by Ka as mentioned below [23],

mw
Tw ,2
C pw Storn and Prince [25] introduced DE which is a robust optimi-
Ka =
V ∫ h − ha
dT (9) zation method. Compared to other methods, DE has advantages such
Tw ,1 sw as fast convergence, simplicity and capability of efficiently working
 R.K. Singla et al./Applied Thermal Engineering 96 (2016) 240–249 243

4.0
5.5 mw= 0.108 (kg/s) 3.8 mw= 0.108 (kg/s)
mw= 0.125 (kg/s) 3.6 mw= 0.125 (kg/s)
5.0 3.4
mw= 0.145 (kg/s)
Range of cooling tower, R (K)

3.2

Heat transfer rate, Q (kW)

mw= 0.150 (kg/s) 3.0
4.5 2.8 Heat rejected by water
2.6
4.0 2.4
2.2
3.5 2.0
1.8
1.6
3.0 1.4
1.2
2.5 1.0
0.8
2.0 0.6 Heat gained by air
0.4
0.2
1.5 0.0
0.015 0.016 0.017 0.018 0.019 0.020 0.021 0.022 0.015 0.016 0.017 0.018 0.019 0.020 0.021 0.022
Mass flow rate of air, ma (kg/s) Mass flow rate of air, ma (kg/s)

Fig. 2. Variation of range of the cooling tower. Fig. 4. Variation of heat transfer rate.

with real-coded numbers [26]. Initially DE generates a random pop- z k = Y k : Pc ≥ ℜ

ulation of vectors, Np. The population, Np lies within the solution (18)
z k = Z k : Pc < ℜ
space and also Np >> υ, where υ is the vector representing the number
of variables to be estimated. For the present work, a population size
of Np = 20 is considered. In successive steps, for a parent vector, Zk∈ Np where, Pc = 0.9 is the crossover probability and ℜ ∈[0, 1] is any
having number of variables equal to υ, three different vectors random number. The final retention of either child (z ) or parent
(τk, φk and χk∈ Np) are selected, each of these being of size υ, where (Z ) is dependent upon the following conditions,
τk, φk and χk∈ Np, but (τ , ϕ , χ )k ≠ Z k , where k is the index for gener-
ation of DE. Next, a mutant vector Yk is generated for Zk as follows z k = Z k : F (z k ) ≤ F ( Z k )
(19)
[25], Z k = z k : F (z k ) > F ( Z k )
Y k = τ k + S (ϕ k − χ k ) (17)
The above-mentioned optimization process is repeated until the
where S is the scaling factor which is considered to be 0.80 in the objective function, F becomes either zero or 100 generations are
present work [27]. Then, the child vector is generated as below, evaluated. Next, results and discussion are presented.

-4
8.0x10 0.42 mw= 0.108 (kg/s)
mw= 0.108 (kg/s)
mw= 0.125 (kg/s)
mw= 0.125 (kg/s)
-4
7.5x10 0.40
Water evaporation rate, Mev (kg/s)

mw= 0.145 (kg/s) mw= 0.145 (kg/s)

-4 0.38 mw= 0.150 (kg/s)
7.0x10 mw= 0.150 (kg/s)
Merkel number, Me

-4
0.36
6.5x10
0.34
-4
6.0x10
0.32
-4
5.5x10 0.30

5.0x10
-4 0.28

-4 0.26
4.5x10
0.24
-4
4.0x10
0.015 0.016 0.017 0.018 0.019 0.020 0.021 0.022 0.015 0.016 0.017 0.018 0.019 0.020 0.021 0.022
Mass flow rate of air, ma (kg/s) Mass flow rate of air, ma (kg/s)

Fig. 3. Variation of water evaporation rate. Fig. 5. Variation of Merkel number.

244 R.K. Singla et al./Applied Thermal Engineering 96 (2016) 240–249

0.52
mw= 0.108 (kg/s) 9.6
Heat-mass transfer coefficient, Ka (kg/s.m )

0.50 0.90
3

In
cr
ea
mw= 0.125 (kg/s)

s
9.4

in
0.48

g
Me
Approach
0.46 mw= 0.145 (kg/s) 0.85 9.2
0.44 mw= 0.150 (kg/s) 9.0

Merkel number, Me
0.42 0.80 mw/ma

Approach, (K)
0.40 8.8
0.38 0.75 Me 8.6
0.36 Approach
8.4
0.34
0.70 8.2
0.32
0.30 8.0
0.65
0.28 7.8
0.26
0.60 7.6
0.24
0.22 7.4
0.20 0.55
0.015 0.016 0.017 0.018 0.019 0.020 0.021 0.022 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3.0 3.1 3.2
Mass flow rate of air, ma (kg/s) Water to air ratio, mw/ma

Fig. 6. Variation of the heat and mass transfer coefficient. Fig. 8. Comparison of Merkel number and approach of present study with CTI per-
formance chart.

6. Results and discussion

it is observed that M ev increases with increase of ma, because with
As discussed previously, all experimental results of the cooling increase of ma, the evaporative heat transfer rate increases as dis-
tower are produced using five levels of ma and four levels of mw. cussed earlier. Alternatively, the interfacial area between the water
Fig. 2 depicts the range of the cooling tower, R against the mass flow and the air inside the tower increases which leads to higher M ev
rates of air, ma and water, mw. It is seen that the range, R increases with increase of ma. It is also found that the evaporation rate, M ev
with increase in ma. For a given water droplet, this increased ma increases with the increase of mw which is due to increased heat
stimulates more air to be in contact with water inside the tower, transfer coefficient at the water side [23].
which enhances the heat transfer rate and consequently de- The effect of different ma and mw on the heat rejected by water
creases the water outlet temperature. It is also observed that with and the heat gained by air is presented in Fig. 4. As expected, it is
increase in mw, the range, R decreases for all values of ma because, found that with increase in ma, the heat rejected by water in-
for a fixed mass of air flow rate, with increase in mw, more resis- creases because of increased contact area between air and water,
tance to the heat transfer comes into the picture, thereby decreasing which increases Ka. Due to this, the heat gained by air also in-
the range. This resistance is obvious because the specific humidi- creases. It may be noted that as ma increases, the difference between
ty, ω1 remains constant, whereas there is increase in mw. So, this the heat rejected by water and the heat gained by air increases due
resistance decreases the heat transfer rate between air and water to increased irreversibility of heat and mass transfer which further
and diminishes range, R. In Fig. 3, the effect of variation in ma and give rise to destruction of useful energy. Fig. 5 shows the variation
mw on the evaporation rate, M ev is presented. From the trend lines, of the Merkel number, Me with the variation of ma and mw. It is ob-
served that Me increases with increase in ma. Increased mass flow
rate of air, ma results in significant increase in the evaporative heat
25 transfer due to increased interfacial area, a between the air and the
mw= 0.108 (kg/s) water inside the tower which can also be understood from Fig. 3,
24
mw= 0.125 (kg/s) where the water evaporation rate, M ev increases with the
23
mw= 0.145 (kg/s)
22
mw= 0.150 (kg/s) Table 1
Effectiveness, e (%)

21
Details of the correlation at different ambient temperatures.
20
General correlation Ambient temperature
19 Me = c 1 + c 2 (mw ) + c 3 (m a ) + c 4 (mw2 )
27 °C 31.5 °C 36 °C
18 + c 5 (mw m a ) + c 6m a2 + c 7 (mw3 )
+ c 8 (mw2 m a ) + c 9 (mw m a2 ) + c 10 (m a3 )
17
16 Coefficient of determination, r 2 0.9205 0.9718 0.9954
Coefficients
15 c1 7.396 2.799 −1.404
14 c2 −99.45 −78.01 87.16
c3 −448.6 155.9 −345.7
13 c4 491.9 573.7 −406.1
c5 3631 161.1 −4289
12 1.19 × 104 −9.79 × 103 3.64 × 104
c6
0.015 0.016 0.017 0.018 0.019 0.020 0.021 0.022
c7 −1295 −1146 687.4
Mass flow rate of air, ma (kg/s) c8 2012 −6745 8142
c9 −1.17 × 105 4.70 × 104 6.09 × 104
c10 6.44 × 104 8.73 × 104 −8.30 × 105
Fig. 7. Variation of the effectiveness.
 R.K. Singla et al./Applied Thermal Engineering 96 (2016) 240–249 245

Table 2 an increase in mw, water held in the cells of the packing increases,
Raw data at 27 ± 1.5 °C used for calculation of performance parameters. which results in the relevant increase in interfacial area, a. The vari-
Air flow rate Water Inlet air Outlet air Range ation of effectiveness, e against ma and mw is shown in Fig. 7 which
flow rate is found to be increasing with increase in ma. This is somewhat
m a (kg s ) mw (kg s ) T a,1 (°C ) ω a,1 (kg kg dry air ) T a,2 (°C ) ω a,2 (kg kg dry air ) R (°C ) obvious because the effectiveness is directly proportional to range
0.0216 0.1083 26.34 0.0103 39.60 0.0425 5.32
and due to this fact, trend lines of effectiveness are similar to that
0.1250 26.75 0.0110 40.31 0.0436 4.51 of range, for all ma and mw. Further as expected, for any value of ma,
0.1417 26.97 0.0113 40.06 0.0448 3.60 the effectiveness decreases with an increase in mw. In order to check
0.1500 26.81 0.0116 40.19 0.0462 3.34 the reliability, in Fig. 8, present results of the cooling tower have
0.0206 0.1083 25.88 0.0115 40.19 0.0430 5.57
been compared with standard tower characteristic charts of Cooling
0.1250 26.38 0.0122 40.67 0.0446 4.65
0.1417 26.48 0.0127 40.76 0.0486 3.51 Technology Institute (CTI) Blue Book [28] that is widely used for in-
0.1500 26.52 0.0131 41.06 0.0501 3.37 dustrial and academic purpose [29,30] and any cooling tower obeys
0.0195 0.1083 27.96 0.0162 39.17 0.0422 3.92 trends in accordance with CTI chart. It is found that Merkel number
0.1250 28.47 0.0167 39.82 0.0448 3.33 decreases with increase in mw ma and this negative slope is repre-
0.1417 27.36 0.0151 38.44 0.0431 2.73
0.1500 28.42 0.0150 38.80 0.0441 2.51
sented by dotted line in tower characteristic curve that is in
0.0177 0.1083 26.91 0.0131 38.06 0.0401 3.72 accordance with CTI chart. It is also observed that the approach in-
0.1250 27.52 0.0134 38.43 0.0410 3.05 creases with decreasing Me, which is in line with standard
0.1417 27.73 0.0133 38.29 0.0433 2.65 performance curves. Hence, the results of the present study are well
0.1500 27.62 0.0135 38.46 0.0439 2.50
in agreement with available standards of CTI.
0.0149 0.1083 27.10 0.0138 39.15 0.0448 3.35
0.1250 27.86 0.0144 39.82 0.0465 3.09 Merkel number is an important parameter for evaluating the per-
0.1417 27.30 0.0141 40.01 0.0477 2.71 formance of a cooling tower. Depending upon the heating load on
0.1500 28.19 0.0148 40.56 0.0494 2.58 the cooling tower, from the designer and operator’s point of view,
it is important to operate a cooling tower to achieve a particular
value of Me. As pointed out earlier mass flow rates of water and air
increase of air flow rate, ma. On the other hand, Me decreases with are influencing parameters affecting Me which can be easily con-
the increase in water mass flow rate, mw due to increased resis- trolled by the operator. For expanded wire mesh packing, the
tance to heat transfer between water and air, thus reducing range unavailability of a theoretical relation accounting for the influ-
of the tower. From Eq. (7) it is also evident that any reduction in ence of mw and ma has motivated to propose a correlation. Using
the range lowers Me. experimental results at three different ambient temperatures, curves
The variation in ma and mw on Ka is investigated in Fig. 6 which are fitted using third order polynomial as mentioned in Eq. (20).
reveals that Ka increases with the increase of ma. It is already dem-
Me = c 1 + c 2 (mw ) + c 3 (m a ) + c 4 (mw2 ) + c 5 (mw m a )
onstrated that this increase of Ka is mainly due to the significant
increase of water evaporation rate inside the cooling tower with in- + c 6m a2 + c 7 (mw3 ) + c 8 (mw2 m a ) + c 9 (mw m a2 ) + c 10 (m a3 ) (20)
crease in ma, which in turn is due to increased interfacial area
between air and water. It is also noticed that with increase in mw, where c 1, c 2, c 3 … c 10 are coefficients which are presented in Table 1
Ka increases gradually which may be easily understood that with for various ambient temperatures.

Table 3
Goodness of correlation for different experimental cases.

mw (kg s ) m a (kg s ) (m
w
ma ) (m a
mw ) Me (Experiment) Me (Correlation) Error

Within correlation (m a mw = 0.10 0.20)

1 0.1083 0.0197 5.4859 0.1823 0.3350 0.3542 5.75%
2 0.1250 0.0216 5.7784 0.1731 0.3210 0.3398 5.86%
3 0.1083 0.0216 5.0080 0.1997 0.3706 0.3974 7.22%
4 0.1250 0.0211 5.9211 0.1689 0.3238 0.3331 2.86%
5 0.1417 0.0211 6.7106 0.1490 0.3028 0.2977 1.70%
Outside correlation (0.20 < m a mw < 0.40)
1 0.0833 0.0177 4.7181 0.2120 0.4631 0.4936 6.17%
2 0.0833 0.0195 4.2816 0.2336 0.4952 0.5249 5.66%
3 0.0833 0.0206 4.0500 0.2469 0.5246 0.5635 6.90%
4 0.0833 0.0216 3.8523 0.2596 0.5199 0.6144 15.37%
5 0.0667 0.0177 3.7744 0.2649 0.5637 0.7351 23.32%
6 0.0667 0.0195 3.4253 0.2919 0.5975 0.7781 23.21%
7 0.0667 0.0216 3.0818 0.3245 0.6545 0.8985 27.15%
8 0.0617 0.0216 2.8507 0.3508 0.6385 1.0121 36.91%
9 0.0583 0.0216 2.6966 0.3708 0.7262 1.0959 33.74%
10 0.0542 0.0216 2.5040 0.3994 0.7788 1.2104 35.66%
Comparison of proposed correlation with literature
Me correlations from literature mw 0.1083 0.1250 0.1083
ma 0.0197 0.0216 0.0216
1 Me (present correlation) for wire mesh fill at 27 C o
0.3542 0.3398 0.3974
2 (
Me = 0.910464 (Gw G a )
−0.624555
)
L fi for trickle fill [15] 0.3773 0.3653 0.3995
Error 6.53% 7.52% 0.52%
3 Me = (1.907410Gw−0.599957G a0.649337Tw−0,1.209468 ) L fi for tricklle fill [15] 0.3539 0.3388 0.3690
Error 0.10% 0.27% 7.13%
Me = 0.79 (mw m a )
−0.48
4 for vertical grid apparatus fill [16] 0.3490 0.3404 0.3646
Error 1.48% 0.18% 8.25%
5 Me = (1.930306Gw−0.568230G a0.641400L−fi0.352377Tw−0,1.178670 ) L fi for trickle fill [20], where G = m A fi 0.3829 0.3690 0.4001
Error 8.10% 8.60% 0.68%
246 R.K. Singla et al./Applied Thermal Engineering 96 (2016) 240–249

(a)
1.4
Experiment
Correlation
1.2
Error (%)

1.0
Merkel number, Me

0.8

0.6
33.74 35.66
36.91

0.4 27.15
23.32 23.21
15.37
0.2
6.17 5.66 6.9

0.0
0.20 0.22 0.24 0.26 0.28 0.30 0.32 0.34 0.36 0.38 0.40
Air to water ratio, ma/mw
(b)

Fig. 9. (a) Proposed correlation curve for Me. (b) Percentage error in the correlation for different ma
mw ; ambient temperature at 27 ± 1.5 °C.

The coefficients of determination for correlations at different is satisfactory with the experimental Me up to m a mw ≈ 0.26
ambient temperatures of 27 ± 1.5°C, 31.5 ± 0.75°C and 36 ± 1°C are (Fig. 9b). However, beyond this, errors between the correlated and
0.9205, 0.9718 and 0.9954 , respectively. At 27 ± 1.5°C , the correla- experimental Me tend to increase (Table 3, Fig. 9b). Moreover, Table 3
tion curve and the raw data (used for evaluating the correlation) also presents a comparison of Me from the proposed correlation of
are presented in Fig. 9a and Table 2, respectively. To develop expanded wire mesh fill with other correlations available in literature
the correlation, twenty experiments are conducted within pertaining to different fills. To illustrate the application of inverse
0.10 < m a mw < 0.20 and to further check the goodness of the cor- methodology, the correlation at 27 °C is considered. In order to
relation, five more experiments are performed within the same range satisfy a given Merkel number, Me  , different combinations of
of m a mw . It is revealed that the maximum and the minimum error controllable parameters (m a and mw ) estimated using inverse anal-
are 7.22% and 1.70%, respectively (Table 3). To assess the goodness ysis are presented in Table 4. For predicting feasible combinations
of the correlation for m a mw > 0.20 , the study has been extended of these parameters, the objective function, F (Eq. 16) is minimized
and it is observed that Me obtained from the proposed correlation using DE by continuously updating the objective function using
 R.K. Singla et al./Applied Thermal Engineering 96 (2016) 240–249 247

Table 4
Comparison of experimental data and feasible combinations of controlling parameters satisfying a given Me

Sl. mw (kg s ) m a (kg s ) ma

mw Me Run mw (kg s ) m a (kg s ) ma
mw Me Objective function, F

Experimental Estimated

1 0.1083 0.0197 0.1823 0.3350 1 0.1253 0.0213 0.1701 0.3348 5.245 × 10−17
2 0.1153 0.0198 0.1718 0.3349 1.351 × 10−14
3 0.1209 0.0207 0.1709 0.3355 2.373 × 10−17
4 0.1259 0.0214 0.1701 0.3349 2.242 × 10−16
5 0.1297 0.0221 0.1706 0.3347 2.240 × 10−18
2 0.1250 0.0216 0.1731 0.3210 1 0.1259 0.0201 0.1593 0.3214 7.424 × 10−18
2 0.1360 0.0221 0.1625 0.3210 1.771 × 10−14
3 0.1250 0.0200 0.1596 0.3214 1.842 × 10−18
4 0.1368 0.0226 0.1651 0.3211 3.867 × 10−15
5 0.1246 0.0199 0.1597 0.3210 2.588 × 10−20
3 0.1083 0.0216 0.1997 0.3706 1 0.1300 0.0248 0.1909 0.3702 7.539 × 10−15
2 0.1155 0.0218 0.1887 0.3707 4.786 × 10−21
3 0.1276 0.0241 0.1889 0.3705 1.619 × 10−14
4 0.1126 0.0213 0.1894 0.3700 7.636 × 10−14
5 0.1233 0.0231 0.1876 0.3700 1.208 × 10−11
4 0.1250 0.0211 0.1689 0.3238 1 0.1245 0.0202 0.1621 0.3239 1.874 × 10−16
2 0.1245 0.0202 0.1622 0.3239 5.898 × 10−17
3 0.1205 0.0197 0.1632 0.3242 2.958 × 10−15
4 0.1187 0.0194 0.1635 0.3238 8.030 × 10−16
5 0.1130 0.0149 0.1320 0.3239 5.659 × 10−16
5 0.1500 0.0211 0.1407 0.3028 1 0.1349 0.0174 0.1289 0.3029 1.047 × 10−17
2 0.1447 0.0177 0.1223 0.3028 9.934 × 10−17
3 0.1429 0.0176 0.1231 0.3029 6.701 × 10−19
4 0.1379 0.0174 0.1265 0.3026 1.448 × 10−17
5 0.1369 0.0174 0.1273 0.3027 1.186 × 10−18

initially and randomly-generated values of two parameters selecting appropriate combinations of water and air mass flow rate
(m a and mw ) . For calculating the updated Me, the proposed corre- to satisfy a given value of Me.
lation (Eq. 20) is employed.
The iterative variation of ma and mw along with the objective func-
tion, F corresponding to the best individual of DE population is
Acknowledgements
revealed in Figs. 10 and 11, respectively. The variation corre-
sponds to the first run of Table 4. From the variations of estimated
The ISIRD grant received from IIT Ropar for this work is duly
parameters (Fig. 10) along with variation of the objective function
acknowledged.
(Fig. 11), it is seen that unknown parameters (m a and mw ) undergo
continuous variation during the iterative process. But, the objec-
tive function, F attains negligible value F < O (10−6 ) beyond 20
Appendix
iterations (approx.). This suggests that different combinations of un-
knowns (m a and mw ) satisfy a particular value of Me as presented
Thermo-physical parameters of air and water vapor mixture
in Table 4. However, for a given value of Me, the ratio of the two
[1,20]
parameters m a mw remains approximately the same. The success-
ful solution of inverse analysis again confirms the goodness of the Dry air: 220 K < T < 380 K
proposed correlation. The conclusion based upon the present work
is discussed next. Saturated water vapor and liquid: 273.15 K < T < 380 K

The enthalpy of the air water vapor mixture is given below,

ha = C pa (T − 273.15) + ω [hfgwo + C pv (T − 273.15)]
7. Conclusion
where C pa and C pv are specific heats of air and water calculated at
(T + 273.15)
In this work, performance of a forced draft cooling tower with 2 as below,
expanded wire mesh packing is investigated. The experiments are
C pa = 1.045356 × 103 − 3.161783 × 10−1 (T )
conducted with the variation of two controlling parameters such
as mass flow rates of water and air. Variation of controlling param- + 7.083814 × 10−4 (T 2 ) − 2.705209 × 10−7 (T 3 )
eters on the performance parameters has been investigated in detail.
In this investigation it is pointed out that Merkel number (Me) is a C pv = 1.3605 × 103 + 2.31334 × (T )
critical parameter that is sensitive to mass flow rates of water and − 2.46784 × 10−10 (T 5 ) + 5.91332 × 10−13 (T 6 )
air. Using experimental results, a correlation has been proposed for
Me as a function of water and air mass flow rates. Secondly, an op- The latent heat of water is
timization analysis has been conducted to simultaneously estimate
multiple and feasible combinations of water and air mass flow rates
hfgwo = 3.4831814 × 106 − 5.8627703 × 103 (273.15) + 12.139568
× (273.15) − 1.40290431 × 10−2 (271.15)
2 3
yielding a fixed value of Me. The study is proposed to be useful in
 248 R.K. Singla et al./Applied Thermal Engineering 96 (2016) 240–249

4
0.210 10
Me = 0.3350 2
0.195 10
Me = 0.3210
0.180 Me = 0.3706 10
0
Mass flow rate of air, ma (kg/s)

0.165 Me = 0.3238 10
-2

Me = 0.3028

Objective function, F
0.150
-4
0.135 10
-6
0.120 10
0.105 10
-8

0.090 -10
10
0.075
10
-12
Me = 0.3350
0.060
10
-14 Me = 0.3210
0.045
-16
Me = 0.3706
0.030 10 Me = 0.3238
0.015 10
-18
Me = 0.3028
0.000
1 10 100 1 10 100
Generation Generation
(a)
Fig. 11. Variation of the objective function with generation of differential evolu-
0.32 tion (DE).

0.30 Me = 0.3350
Me = 0.3210
M ass flow rate of water, mw (kg/s)

0.28
Me = 0.3706
0.26 Me = 0.3238 Specific humidity is mentioned below,
0.24 Me = 0.3028 ⎡ 2501.6 − 2.3263 × (Twb − 273.15) ⎤
ω=⎢ ⎥
0.22 ⎣ 2501.6 + 1.8577 × (T − 273.15) − 4.184 × (Twb − 273.15) ⎦
⎡ (0.62509 × pvwb ) ⎤
0.20 ⎢ p − 1.005 × p ⎥
⎣ ( abs vwb ) ⎦
0.18 ⎡ 1.00416 × (T − Twb ) ⎤
−⎢ ⎥
0.16 ⎣ 2501.6 + 1.8
8577 × (T − 273.15) − 4.184 × (T wb − 273.15) ⎦
0.14
Nomenclature
0.12
0.10 a Interfacial area (m2 m3 )
Cp Specific heat at constant pressure ( kgkJ⋅K )
0.08 e Effectiveness
F Objective function
1 10 100
Generation h Specific enthalpy ( kg kJ
)
(b)
K ( )
Mass transfer coefficient mkg2⋅s
Me Merkel number
M ev Water evaporation rate ( kgs )
Mass flow rate ( kgs )
Fig. 10. Variation of estimated parameters (a) ma, (b) mw with generation of dif-
ferential evolution (DE).
m
Np Population size in DE
p ( )
Pressure mN2
Pc Crossover probability
Q Heat transfer rate (kW )
R Range (K )
The vapor pressure of saturated water vapor is given by r2 Coefficient of determination
pv = 10z N m2 S Scaling factor
T Temperature (K )
where u Uncertainty
V Volume of exchange core (m3 )
⎡ ⎛ 273.15 ⎞ ⎤ ⎛ 273.15 ⎞
z = 10.79586 × ⎢1 − ⎜ ⎟⎠ ⎥ + 5.02808 × log10 ⎜⎝ ⎟ Y Mutation vector
⎣ ⎝ T wb ⎦ Twb ⎠ Z Parent vector
⎡
+ 1.50474 × 10−4 ⎢1 − 10
{
−8.29692×⎛⎜ wb −1⎞⎟ ⎤
T
⎝ 273.15 ⎠
⎥
} z Child vector
⎣ ⎦
Greek symbols
⎡ ⎧ ⎛ 273 .15 ⎞⎫ ⎤
+ 4.2873 × 10−4 ⎢ 4.76955 × ⎨1 − ⎜ ⎟ ⎬ − 1 + 2.786118312 χ, τ, φ Random vectors in DE
⎣ ⎩ ⎝ Twb ⎠ ⎭ ⎥⎦ ω Specific humidity (kg kg of dry air )
 R.K. Singla et al./Applied Thermal Engineering 96 (2016) 240–249 249

Subscripts [12] M. Gao, F.Z. Sun, K. Wang, Y.T. Shi, Y. Bin Zhao, Experimental research of heat
transfer performance on natural draft counter flow wet cooling tower under
a Air
cross-wind conditions, Int. J. Therm. Sci. 47 (2008) 935–941.
abs Absolute [13] P.K. Mondal, S. Mukherjee, B. Kundu, S. Wongwises, Investigation of the
fg Fluid–gas mixture crosswind-influenced thermal performance of a natural draft counterflow
k Index for iteration cooling tower, Int. J. Heat Mass Transf. 85 (2015) 1049–1057.
[14] E. Elsarrag, Experimental study and predictions of an induced draft ceramic
sw Saturated air at bulk water temperature tile packing cooling tower, Energy Convers. Manag. 47 (2006) 2034–2043.
wb Wet bulb [15] J.C. Kloppers, D.G. Kröger, Refinement of the transfer characteristic correlation
v Vapor of wet-cooling tower fills, Heat Transf. Eng. 26 (2005) 35–41.
[16] M. Lemouari, M. Boumaza, I.M. Mujtaba, Thermal performances investigation
w Water of a wet cooling tower, Appl. Therm. Eng. 27 (2007) 902–909.
1 Inlet [17] A.H. Copeland, R.S. Segall, C.D. Ringo, B. Moore, Mathematical modelling of
2 Outlet inverse problems for oceans, Appl. Math. Model. 15 (1991) 586–595.
[18] R.K. Singla, R. Das, Application of decomposition method and inverse prediction
of parameters in a moving fin, Energy Convers. Manag. 84 (2014) 268–281.
Superscripts [19] R. Ramkumar, A. Ragupathy, Thermal performance of forced draft counter flow
~ Exact value of Me wet cooling tower with expanded wire mesh packing, (2011) 19–23.
[20] J.C. Kloppers, A Critical Evaluation and Refinement of the Performance Prediction
of Wet-Cooling Towers, University of Stellenbosch, Stellenbosch, 2003.
References [21] A.M. Lavasani, Z.N. Baboli, M. Zamanizadeh, M. Zareh, Experimental study on
the thermal performance of mechanical cooling tower with rotational splash
type packing, Energy Convers. Manag. 87 (2014) 530–538.
[1] D.G. Kröger, Air-cooled Heat Exchangers and Cooling Towers; Thermal-Flow
[22] R. Perry, D. Green, J. Maloney, Perry’s chemical engineers’ handbook, 1997.
Performance Evaluation and Design, vol. 1, PennWell Books, Tulsa, Oklahoma,
[23] M. Lemouari, M. Boumaza, A. Kaabi, Experimental analysis of heat and mass
2004.
transfer phenomena in a direct contact evaporative cooling tower, Energy
[2] F. Merkel, Verdunstungskühlung, VDI-Verlag, 1925.
Convers. Manag. 50 (2009) 1610–1617.
[3] M. Poppe, H. Rögener, Berechnung Von Rückkühlwerken, VDI-Wärmeatlas, 1991,
[24] R.J. Moffat, Contributions to the Theory of Single-Sample Uncertainty Analysis,
Mi 1–Mi 15.
J. Fluids Eng. 104 (1982) 250.
[4] H. Jaber, R.L. Webb, Design of Cooling towers by the effectiveness-NTU method,
[25] R. Storn, K. Price, Differential evolution–a simple and efficient heuristic for global
J. Heat Transfer 111 (1989) 837–843.
optimization over continuous spaces, J. Glob. Optim. 11 (1997) 341–359.
[5] J.C. Kloppers, D.G. Kröger, A critical investigation into the heat and mass transfer
[26] A.K. Parwani, P. Talukdar, P.M.V. Subbarao, Performance evaluation of hybrid
analysis of counterflow wet-cooling towers, Int. J. Heat Mass Transf. 48 (2005)
differential evolution approach for estimation of the strength of a heat source
765–777.
in a radiatively participating medium, Int. J. Heat Mass Transf. 56 (2013)
[6] H. Inazumi, S. Kageyama, A successive graphical method of design of a cross-flow
552–560.
cooling tower, Chem. Eng. Sci. 30 (1975) 717–721.
[27] A. Bhowmik, R.K. Singla, P.K. Roy, D.K. Prasad, R. Das, R. Repaka, Predicting
[7] A.A. Dreyer, P.J. Erens, Modeling of cooling tower splash pack, Exp. Therm. Fluid
geometry of rectangular and hyperbolic fin profiles with temperature-
Sci. 7 (1993) 130.
dependent thermal properties using decomposition and evolutionary methods,
[8] H.R. Goshayshi, J.F. Missenden, Investigation of cooling tower packing in various
Energy Convers. Manag. 74 (2013) 535–547.
arrangements, Appl. Therm. Eng. 20 (2000) 69–80.
[28] Cooling Tower Performance Curves: Blue Book, The Cooling Tower Institute,
[9] Y.J. Dai, K. Sumathy, Theoretical study on a cross-flow direct evaporative cooler
Houston, Texas, US, 1967.
using honeycomb paper as packing material, Appl. Therm. Eng. 22 (2002)
[29] G.F. Hallett, Performance curves for mechanical draft cooling towers, J. Eng.
1417–1430.
Power. 97 (1975) 503.
[10] F. Gharagheizi, R. Hayati, S. Fatemi, Experimental study on the performance of
[30] R. Rosaler, Standard Handbook of Plant Engineering, McGraw-Hill, 2002.
mechanical cooling tower with two types of film packing, Energy Convers.
Manag. 48 (2007) 277–280.
[11] J. Facão, A. Oliveira, Heat and mass transfer correlations for the design of small
indirect contact cooling towers, Appl. Therm. Eng. 24 (2004) 1969–1978.