CHAPTER ONE

1. INTRODUCTION

Cooling tower are widely used in large air-conditioning installations to recycle

condenser cooling tower. The performance of cooling towers are often
expressed as Number of Transfer Unit (NTU).
A widely used method of calculating NTU is based on Merkels theory.
The theory is applicable to both counter flow and cross flow direct-contact
forced-draft air-water cooling towers as described by Stoecker and Jones
(1982).
Manual calculation of NTU is tedious and time-consuming. In this
research, computerized calculation are made to calculate the NTU of cross
flow cooling towers using Stoecker and Jones (1982), for a range of tower inlet
water temperature (tw1), outlet water temperature (tw2), ambient air wet bulb
temperature (twB) and ratio of water to air mass flow rate (L/G).
Multi-linear regression is then made to obtain correlations in the form
of :
NTU = a0. tw1a1.tw2a2.twba3(L/G)a4

Figure 1.1 A cross flow cooling tower by dividing it into sections.

1.1 Objective
To obtain theoretical correlations for cross flow cooling towers for the tower NTU as a
function of inlet water temperature, outlet water temperature, ambient air wet bulb
temperature, air flow rate and water flow rate.

1.2 Scope
Steady-state analysis of crossflow air-water cooling towers
Use of Merkel theory

CHAPTER TWO
2. LITERATURE REVIEW

Cooling towers are widely used in air-conditioning applications for reducing the
temperature of circulating water for reuse in condensers. A cooling tower is an
evaporative cooling device where a circulating water is cooled by direct contact
with atmospheric air, where some of water is evaporated. Direct-contact
forced-draft air-water cooling towers involve the transfer of mass and hear
between the hot water and ambient air flowing in counter flow and cross flow
configurations.
A simplified cooling tower analysis is often made using Merkels (1925)
theory. The model is based on the enthalpy difference between the saturated air
enthalpy or the air water interface and the enthalpy of the surrounding air as the
driving potential for energy transfer.
Baker and Shryrock (1961) have discussed Merkels theory and suggested
the use of on offset ratio refers to the ratio between the true enthalpy potential
and the difference in the air/water interface temperature and the bulk water
temperature. They also studied the effects of neglecting the evaporative water
loss and the liquid side thermal resistance on the cooling tower performance.
The liquid side thermal resistance was considered in a cross flow cooling
tower analysis, they have also shown that the true driving potential was lower
than that without the liquid side thermal resistance. By considering
evaporative water loss and the liquid side thermal resistance, the resulting
true NTUs were also shown to be greater than those calculated using
Merkels theory for an experimental cross-flow cooling tower.
Osterle (1991) presented a tower model for which the Merkel equations
were corrected to account for mass of water lost by evaporation. The corrected
NTUs were evaluated by solving a set of differential equations.

Majumdar et. al (1983) have developed a two dimensional model that

coupled the thermodynamics and

hydrodynamics for

simulating cooling

tower performance in either a counter-flow or cross-flow configuration. The

governing partial differential equations have been solved using finite
difference techniques in a computer code.
Bernier (1994) analysing the cooling tower heat rejection performance
must be obtained experimentally, where the NTU is evaluated using several
theories and models. The product of Ka in the NTU definition represents
average values of K and a in the tower, where K is the mass transfer
coefficient, and a is the air/water contact area per unit of tower volume. The
simplest and widely used theory for cooling tower analysis has been proposed
by Merkel in 1925.
Maiya (1995) analysing the performance of modified counter-flow
cooling tower. Bedekar (1998) investigation on the performance of a
counter-flow, packed-bed cooling tower.

CHAPTER THREE
3. METHODOLOGY
In a cooling tower, heat and mass are transferred from water to air. A cooling tower
with an effectiveness of 100% would exhaust saturated air from the tower at the
temperature of the entering water. In actual towers, the air leaves at a temperature
less than the entering water temperature and a relative humidity less than 100%
because the airflow is too high and the heat transfer area is too small. Effectiveness,

a, is defined10 in terms of the enthalpy of moist air as follows:

a = (hi,in - hi,out) / (t w2 t w1)

(3-1)

In Equation (1), hi,in and hi,out are the enthalpies of the air into and out of the tower, and

t w2 is the enthalpy of air saturated with water at the inlet water temperature.
Effectiveness is a dimensionless variable which can be determined from knowledge of
only two additional dimensionless variables, NTU.

NTU = (hc A ) / (CpmL )

(3-2)

From Equation (3-7), NTU depends on the heat transfer coefficient, hc, the heat transfer
surface area, A , and the specific heat, Cpm. Energy efficient cooling towers with large heat
transfer surfaces and small fans might have nominal NTU values of 4.5 or more. Low initial
cost towers with small heat transfer surfaces and large fans might have nominal NTU
values of 1.5 or less.
Since hc depends on the airflow rate and the water flow rate, the NTU value of a tower

operating at off-design conditions will not be the same as the NTU value at design
conditions. An empirical equation useful for predicting NTU at off-design conditions is:

NTU =

(3-3)

= 4.19

The constants, a, m, and n, in Equation (3-3) are to be determined from published

performance data. In some cases, this data is unavailable.
The heat exchangers are usually characterized by the Number of Transfer Unit NTU.
Since the tower has similar hydrothermal operation, then it is possible to characterize
the tower performance using the NTU as following.

Since NTU are known for a particular cooling tower, the cooling tower
performance can be predicted at any operating condition using;
Multi Linear Regression ;
NTU = a0. tw1a1.tw2a2.twba3(L/G)a4

(3-4)

Estimate;
NTU = 1

-number of unit transfer

tw1 = 37oC

-entering water temperature

tw2 = 35oC

-leaving water temperature

twb = 25oC

-temperature wet bulb entering air

= 1.0
n x n = 10 x 10

-dividing section in cross flow cooling tower

calculate using eq.(3-9)

1= c.(37)c(35)c(25)c(1.0)c
a=9.2592 * 10-3 .into eq.(3.9) to get the actual NTU
So,

NTU actual = 0.0101

Although the counter flow cooling tower is widely used in industrial services, another
configuration is the cross flow tower, in which the air passes horizontally through the
falling water sprays. Cooling towers used for air-conditioning systems are often
located atop buildings, and cross flow cooling tower usually has a lower profile,
which lends itself better to architectural treatment.
The same principles of heat and mass transfer and balance of energy apply to a
cross flow tower and to the counter flow type, but the geometric treatment of the
cross flow tower is different. Figure 3-1 showns a cross flow tower divided into 12
sections for purposes of analysis. Water enters the top at a temperature tin while air
enters from the left with an enthalpy hin. In section 1 air enters at enthalpy hin and
leaves with enthalpy h1. Also in section 1 water entera at tin and leaves at t1. The

enthalpy of air entering section 2 is h1, and the water enters section 5 with a
temperature t1. The temperature of water leaving sections 9, 10, 11, and 12 at the
bottom of the tower are t9, t10, t11, and t12, respectively. These temperatures are all
different, and the streams combine to form one stream of temperature tout.

If the value of

hcA/Cpm is known for the entire cooling tower, the outlet water

temperature can be predicted when the inlet water temperature, inlet air enthalpy,
and the flow rates of water and air are known. The tower can be divided into a
number of small increments (12 increment, for example, in Fig 2-1) and (hcA/Cpm )/12
assigned to each increment. Let

Figure 2.1 Analysis a cross flow cooling tower by dividing it into sections.

Divide tower into 3 rows, 4 columns. Total cells = 12.

Given;
tin

o
: 37 C

twb

o
: 25 C

Total cells

: 3 x 4 = 12

air flow, L

: 24 kg/s

water flow, G : 15 kg/s

hin

: 80.8

1. Divide whole tower NTU into cellular NTUe

Guess NTU = 10

[kW/kJ.K]

[kW/kJ.K]
NTUe = 20

2.

[kg/s]

[kg/s]
Ge = 5 [kg/s]

3.

[kg/s]

[kg/s]
Le = 6 [kg/s]

4. Calculate saturated air specific enthalpy at inlet water temperature tin

hi,in = 4792.6 + 2568tin 29.834tin2 + 1.6657 tin3

[J/kg]

hi,in = 4792.6 + 2568(37) 29.834(372) + 1.6657(373)

hi,in = 143338.5561

[J/kg]

5. Let
t1 = 37 + 0.1 = 37.1oC

6. Calculate saturated air specific enthalpy at outlet water temperature t1

hi,out = 4792.6 + 2568tin 29.834tin2 + 1.6657 tin3

[J/kg]

hi,out = 4792.6 + 2568(37.1) 29.834(37.12) + 1.6657(37.13)

hi,out = 144060.2397 [J/kg]
[oC]

Calculate specific heat of outlet air

[J/kg]

h1 = 153938.8348

[J/kg]

Heat gain by air

[W]

dq = 5 (162-80.8) [W]

dq = 438833.0088 [W]

equals heat loss from falling water:

[oC]

[oC]
t1 = 16.0333
7. If

[oC]
repeat step 6.

t1 - t1,old = 16.0333 - 37.1

t1 - t1,old = -21.0667

8. Repeat calculations (steps 46) for the other cells.

9. At the bottom of tower calculate the average exit water temperature tout

CHAPTER FOUR

4. DATA ANALYSIS AND RESULTS

A computer program has been written to calculate the NTU of cross flow cooling
towers using Merkels theory as described by Stocker and Jones (1982). The inputs to
the program are inlet water temperature (tw1), outlet water temperature (tw2),

ambient air wet bulb temperature (twB) and the ratio of water to air mass flow rate
(L/G). The program calculates the NTU as described in Chapter 3.
Correlation are developed for the following conditions;
(a) Atmospheric pressure, 101 325 Pa.
(b) Tower divided into 10 rows and 10 columns (i.e. 100 cells)
(c) 34.5 oC < tw1 < 35.5 oC
(d) 29 oC < tw2 < 30 oC
(e) 25 oC < twB < 26 oC
(f) 0.5 < L/G < 1.5

4.1

L/G between 0.5 and 0.7

Table 4.1 shows the true NTU as calculated using Merkels theory and

NTU estimated from the correlation:

NTU = 6.800115. tw15.403319 .tw2-10.30583 .twb2.785535 (L/G)0.4895199

The maximum single-point error when using the correlation is -4.5%. Figure
4.1 shows that NTUcorrelation is in good agreement with the true NTU. Figure
4.2 gives the percentage error distribution for all NTUcorrelation where the
errors range from -4.5% to 3.6%.

Table 4.1 NTU for 0.5 < L/G < 0.7

tw1
34.5
34.5
34.5
34.5

tw2
29
29
29
29

twB
25
25
25
25.5

L/G
0.5
0.6
0.7
0.5

NTUtru
e
0.94
0.99
1.046
1.039

NTUregressio
n
0.9306
0.9853
1.034
1.0334

% Error
-1
-0.48
-1.15
-0.54

34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
tw1
35
35.5

29
29
29
29
29
29.5
29.5
29.5
29.5
29.5
29.5
29.5
29.5
29.5
30
30
30
30
30
30
30
30
30
29
29
29
29
29
29
29
29
29
29.5
29.5
29.5
29.5
29.5
29.5
29.5
29.5
29.5
30
30
30
30
30
30
30
tw2

25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
twB

0.6
0.7
0.5
0.6
0.7
0.5
0.6
0.7
0.5
0.6
0.7
0.5
0.6
0.7
0.5
0.6
0.7
0.5
0.6
0.7
0.5
0.6
0.7
0.5
0.6
0.7
0.5
0.6
0.7
0.5
0.6
0.7
0.5
0.6
0.7
0.5
0.6
0.7
0.5
0.6
0.7
0.5
0.6
0.7
0.5
0.6
0.7
0.5
L/G

1.1
1.17
1.166
1.243
1.332
0.777
0.811
0.849
0.849
0.89
0.935
0.94
0.989
1.046
0.643
0.665
0.69
0.696
0.723
0.753
0.761
0.794
0.83
1.004
1.061
1.125
1.108
1.177
1.257
1.242
1.329
1.43
0.837
0.876
0.92
0.914
0.961
1.014
1.009
1.067
1.132
0.699
0.726
0.756
0.756
0.788
0.824
0.826

1.0941
1.1483
1.1453
1.2126
1.2726
0.7583
0.8029
0.8426
0.8421
0.8916
0.9357
0.9333
0.9881
1.037
0.6201
0.6565
0.689
0.6886
0.7291
0.7652
0.7632
0.808
0.848
1.0103
1.0697
1.1226
1.122
1.1879
1.2467
1.2434
1.3165
1.3816
0.8233
0.8717
0.9148
0.9143
0.968
1.0159
1.0133
1.0728
1.1259
0.6732
0.7128
0.7481
0.7476
0.7916
0.8307
0.8286
NTUregressio
n
0.9207
1.0956

30

26

0.7

NTUtru
e
0.907

29

25

0.5

1.065

-0.53
-1.86
-1.78
-2.45
-4.46
-2.4
-1
-0.75
-0.81
0.18
0.08
-0.71
-0.09
-0.86
-3.56
-1.27
-0.14
-1.06
0.84
1.61
0.29
1.77
2.17
0.63
0.82
-0.21
1.26
0.93
-0.82
0.12
-0.94
-3.38
-1.64
-0.49
-0.56
0.03
0.73
0.19
0.42
0.55
-0.54
-3.69
-1.82
-1.05
-1.11
0.45
0.82
0.31
% Error
1.51
2.87

35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5

29
29
29
29
29
29
29
29
29.5
29.5
29.5
29.5
29.5
29.5
29.5
29.5
29.5
30
30
30
30
30
30
30
30
30

25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26

0.6
0.7
0.5
0.6
0.7
0.5
0.6
0.7
0.5
0.6
0.7
0.5
0.6
0.7
0.5
0.6
0.7
0.5
0.6
0.7
0.5
0.6
0.7
0.5
0.6
0.7

1.129
1.202
1.174
1.252
1.342
1.314
1.412
1.525
0.894
0.939
0.99
0.975
1.029
1.089
1.076
1.141
1.216
0.753
0.785
0.82
0.814
0.851
0.892
0.888
0.933
0.982

1.16
1.2174
1.2167
1.2882
1.3519
1.3484
1.4277
1.4983
0.8928
0.9453
0.9921
0.9915
1.0498
1.1017
1.0988
1.1634
1.221
0.7301
0.773
0.8112
0.8108
0.8584
0.9009
0.8985
0.9513
0.9984

2.75
1.28
3.64
2.89
0.74
2.62
1.11
-1.75
-0.13
0.67
0.21
1.69
2.02
1.17
2.12
1.96
0.41
-3.04
-1.53
-1.07
-0.4
0.87
0.99
1.19
1.96
1.67

Figure 4.1 NTU_correlation versus NTU_true for L/G 0.5 0.7

Figure 4.2 Percentage error in NTUcorrelation for L/G 0.5 0.7

4.2

L/G between 0.7 and 0.9

Table 4.2 shows the true NTU as calculated using Merkels theory and

NTU estimated from the correlation:

NTU = 6.661408. tw13.968963 .tw2-8.622578 .twb2.638664 (L/G)0.3987166

The maximum single-point error when using the correlation is -5.5%. Figure 4.
3 shows that NTUcorrelation is in good agreement with the true NTU. Figure 4.
3 gives the percentage error distribution for all NTUcorrelation where the
errors range from -5.3% to 4.1%.
Table 4.2 NTU for 0.7 < L/G < 0.9
tw1
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
35
35
35
35

tw2
29
29
29
29
29
29
29
29
29
29.5
29.5
29.5
29.5
29.5
29.5
29.5
29.5
29.5
30
30
30
30
30
30
30
30
30
29
29
29
29

twB
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5

L/G
0.7
0.8
0.9
0.7
0.8
0.9
0.7
0.8
0.9
0.7
0.8
0.9
0.7
0.8
0.9
0.7
0.8
0.9
0.7
0.8
0.9
0.7
0.8
0.9
0.7
0.8
0.9
0.7
0.8
0.9
0.7

NTUtru
e
1.046
1.11
1.183
1.17
1.25
1.343
1.332
1.436
1.558
0.849
0.89
0.937
0.935
0.986
1.044
1.046
1.109
1.183
0.69
0.718
0.748
0.753
0.785
0.821
0.83
0.869
0.914
1.125
1.199
1.284
1.257

NTUregressio
n
1.0367
1.1056
1.1702
1.1654
1.2429
1.3156
1.3071
1.3941
1.4756
0.8251
0.88
0.9314
0.9276
0.9893
1.0471
1.0404
1.1096
1.1745
0.6593
0.7031
0.7442
0.7411
0.7904
0.8366
0.8313
0.8866
0.9384
1.1358
1.2114
1.2822
1.2769

% Error
-0.89
-0.4
-1.08
-0.39
-0.57
-2.04
-1.87
-2.92
-5.29
-2.81
-1.12
-0.59
-0.79
0.33
0.3
-0.53
0.06
-0.72
-4.45
-2.07
-0.5
-1.57
0.69
1.91
0.16
2.02
2.67
0.96
1.03
-0.14
1.58

35
35
35
35
35
tw1
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5

29
29
29
29
29
tw2

25.5
25.5
26
26
26
twB

0.8
0.9
0.7
0.8
0.9
L/G

1.349
1.457
1.43
1.549
1.69

1.3618
1.4414
1.4322
1.5274
1.6167
NTUregressio
n
0.9642
1.0205
1.0163
1.0839
1.1473
1.1399
1.2157
1.2868
0.7223
0.7704
0.8154
0.812
0.866
0.9167
0.9108
0.9714
1.0282
1.2428
1.3255
1.403
1.3972
1.4901
1.5772
1.5671
1.6713
1.769
0.9892
1.055
1.1167
1.1121
1.186
1.2554
1.2473
1.3303
1.408
0.7904
0.843
0.8922
0.8885
0.9477
1.003
0.9966
1.0629
1.125

29.5

25

0.8

NTUtru
e
0.969

29.5
29.5
29.5
29.5
29.5
29.5
29.5
30
30
30
30
30
30
30
30
30
29
29
29
29
29
29
29
29
29
29.5
29.5
29.5
29.5
29.5
29.5
29.5
29.5
29.5
30
30
30
30
30
30
30
30
30

25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26

0.9
0.7
0.8
0.9
0.7
0.8
0.9
0.7
0.8
0.9
0.7
0.8
0.9
0.7
0.8
0.9
0.7
0.8
0.9
0.7
0.8
0.9
0.7
0.8
0.9
0.7
0.8
0.9
0.7
0.8
0.9
0.7
0.8
0.9
0.7
0.8
0.9
0.7
0.8
0.9
0.7
0.8
0.9

1.025
1.014
1.073
1.142
1.132
1.207
1.293
0.756
0.789
0.825
0.824
0.863
0.906
0.907
0.955
1.008
1.202
1.286
1.384
1.342
1.446
1.57
1.525
1.659
1.819
0.99
1.047
1.111
1.089
1.158
1.238
1.216
1.302
1.402
0.82
0.858
0.902
0.892
0.939
0.99
0.982
1.038
1.102

0.95
-1.07
0.15
-1.39
-4.34
% Error
-0.5
-0.44
0.23
1.02
0.46
0.7
0.72
-0.48
-4.45
-2.36
-1.16
-1.45
0.35
1.18
0.42
1.72
2
3.4
3.07
1.37
4.11
3.05
0.46
2.76
0.74
-2.75
-0.08
0.77
0.51
2.12
2.42
1.4
2.58
2.17
0.43
-3.61
-1.75
-1.08
-0.39
0.92
1.32
1.49
2.4
2.09

Figure 4.3 NTU_correlation versus NTU_true for L/G 0.7 0.9

Figure 4.4 Percentage error in NTUcorrelation for for L/G 0.7 0.9

4.3

L/G between 0.9 and 1.1

Table 4.3 shows the true NTU as calculated using Merkels theory and

NTU estimated from the correlation:

NTU = 7.26992. tw14.627645 .tw2-9.369511 .twb2.502056 (L/G)0.5178051

The maximum single-point error when using the correlation is -6.43%. Figure
4.3 shows that NTUcorrelation is in good agreement with the true NTU. Figure
4.3 gives the percentage error distribution for all NTUcorrelation where the
errors range from -6.43% to 4.74%.
Table 4.3 NTU for 0.9 < L/G < 1.1
tw1
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
35
35
35
35
35
35

tw2
29
29
29
29
29
29
29
29
29
29.5
29.5
29.5
29.5
29.5
29.5
29.5
29.5
29.5
30
30
30
30
30
30
30
30
30
29
29
29
29
29
29

twB
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5

L/G
0.9
1
1.1
0.9
1
1.1
0.9
1
1.1
0.9
1
1.1
0.9
1
1.1
0.9
1
1.1
0.9
1
1.1
0.9
1
1.1
0.9
1
1.1
0.9
1
1.1
0.9
1
1.1

NTUtru
e
1.183
1.268
1.368
1.343
1.452
1.582
1.558
1.704
1.882
0.937
0.99
1.05
1.044
1.11
1.186
1.183
1.267
1.366
0.748
0.781
0.818
0.821
0.861
0.907
0.914
0.964
1.021
1.284
1.384
1.502
1.457
1.585
1.739

NTUregressio
n
1.1741
1.2654
1.3541
1.3406
1.4449
1.5462
1.5268
1.6456
1.7609
0.9067
0.9772
1.0457
1.0353
1.1158
1.194
1.179
1.2707
1.3598
0.7032
0.7579
0.811
0.8029
0.8654
0.926
0.9144
0.9856
1.0547
1.3013
1.4025
1.5008
1.4858
1.6014
1.7137

% Error
-0.75
-0.2
-1.02
-0.18
-0.49
-2.27
-2
-3.43
-6.43
-3.24
-1.3
-0.41
-0.84
0.52
0.67
-0.33
0.3
-0.45
-5.99
-2.96
-0.86
-2.2
0.51
2.1
0.05
2.24
3.3
1.35
1.34
-0.08
1.98
1.03
-1.46

35
35
35
tw1
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5

29
29
29
tw2

26
26
26
twB

0.9
1
1.1
L/G

1.69
1.86
2.069

1.6922
1.8238
1.9517
NTUregressio
n
1.083
1.1589
1.1474
1.2366
1.3233
1.3068
1.4084
1.5071
0.7794
0.84
0.8989
0.8899
0.9591
1.0263
1.0135
1.0923
1.1689
1.4401
1.5521
1.6609
1.6444
1.7723
1.8965
1.8728
2.0184
2.1599
1.1121
1.1986
1.2826
1.2698
1.3686
1.4645
1.4462
1.5587
1.668
0.8625
0.9296
0.9948
0.9849
1.0615
1.1359
1.1217
1.2089
1.2936

29.5

25

NTUtru
e
1.088

29.5
29.5
29.5
29.5
29.5
29.5
29.5
30
30
30
30
30
30
30
30
30
29
29
29
29
29
29
29
29
29
29.5
29.5
29.5
29.5
29.5
29.5
29.5
29.5
29.5
30
30
30
30
30
30
30
30
30

25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26

1.1
0.9
1
1.1
0.9
1
1.1
0.9
1
1.1
0.9
1
1.1
0.9
1
1.1
0.9
1
1.1
0.9
1
1.1
0.9
1
1.1
0.9
1
1.1
0.9
1
1.1
0.9
1
1.1
0.9
1
1.1
0.9
1
1.1
0.9
1
1.1

1.161
1.142
1.22
1.312
1.293
1.394
1.513
0.825
0.866
0.911
0.906
0.956
1.011
1.008
1.07
1.14
1.384
1.499
1.637
1.57
1.717
1.895
1.819
2.014
2.255
1.111
1.186
1.272
1.238
1.33
1.438
1.402
1.519
1.66
0.902
0.95
1.005
0.99
1.049
1.116
1.102
1.175
1.259

0.13
-1.94
-5.67
% Error
-0.46
-0.18
0.47
1.36
0.86
1.06
1.03
-0.39
-5.53
-3.01
-1.33
-1.78
0.33
1.52
0.55
2.09
2.54
4.06
3.54
1.46
4.74
3.22
0.08
2.96
0.22
-4.22
0.1
1.06
0.83
2.57
2.9
1.85
3.15
2.61
0.48
-4.38
-2.15
-1.02
-0.52
1.19
1.78
1.78
2.88
2.75

Figure 4.5 NTU_correlation versus NTU_true for L/G 0.9 1.1

Figure 4.6 Percentage error in NTUcorrelation for for L/G 0.9 1.1

4.4

L/G between 1.1 and 1.3

Table 4.3 shows the true NTU as calculated using Merkels theory and
NTU estimated from the correlation:
NTU = 8.420994. tw14.314386 .tw2-9.973997 .twb3.167216 (L/G)0.7143654

The maximum single-point error when using the correlation is -8.12%. Figure
4.3 shows that NTUcorrelation is in good agreement with the true NTU. Figure
4.3 gives the percentage error distribution for all NTUcorrelation where the
errors range from -8.12% to 5.71%.
Table 4.4 NTU for 1.1 < L/G < 1.3
tw1
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
35
35
35
35

tw2
29
29
29
29
29
29
29
29
29
29.5
29.5
29.5
29.5
29.5
29.5
29.5
29.5
29.5
30
30
30
30
30
30
30
30
30
29
29
29
29

twB
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5

L/G
1.1
1.2
1.3
1.1
1.2
1.3
1.1
1.2
1.3
1.1
1.2
1.3
1.1
1.2
1.3
1.1
1.2
1.3
1.1
1.2
1.3
1.1
1.2
1.3
1.1
1.2
1.3
1.1
1.2
1.3
1.1

NTUtru
e
1.368
1.486
1.629
1.582
1.74
1.934
1.882
2.101
2.381
1.05
1.119
1.199
1.186
1.274
1.378
1.366
1.483
1.625
0.818
0.859
0.905
0.907
0.958
1.016
1.021
1.086
1.161
1.502
1.644
1.817
1.739

NTUregressio
n
1.3598
1.4872
1.6148
1.5851
1.7335
1.8823
1.8423
2.0148
2.1877
1.0087
1.1032
1.1978
1.1758
1.2859
1.3963
1.3666
1.4945
1.6228
0.752
0.8224
0.893
0.8766
0.9587
1.041
1.0188
1.1142
1.2098
1.53
1.6732
1.8168
1.7834

% Error
-0.6
0.08
-0.87
0.2
-0.37
-2.67
-2.11
-4.1
-8.12
-3.93
-1.42
-0.1
-0.86
0.94
1.33
0.04
0.78
-0.14
-8.07
-4.26
-1.32
-3.35
0.07
2.46
-0.21
2.6
4.21
1.86
1.78
-0.01
2.56

35
35
35

tw1
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5

29
29
29

tw2
29
29.5
29.5
29.5
29.5
29.5
29.5
29.5
29.5
29.5
30
30
30
30
30
30
30
30
30
29
29
29
29
29
29
29
29
29
29.5
29.5
29.5
29.5
29.5
29.5
29.5
29.5
29.5
30
30
30
30
30

25.5
25.5
26

twB
26
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5

1.2
1.3
1.1

L/G
1.3
1.1
1.2
1.3
1.1
1.2
1.3
1.1
1.2
1.3
1.1
1.2
1.3
1.1
1.2
1.3
1.1
1.2
1.3
1.1
1.2
1.3
1.1
1.2
1.3
1.1
1.2
1.3
1.1
1.2
1.3
1.1
1.2
1.3
1.1
1.2
1.3
1.1
1.2
1.3
1.1
1.2

1.927
2.162
2.069

1.9504
2.1178
2.0728

NTUtru
e
2.668
1.161
1.245
1.345
1.312
1.42
1.549
1.513
1.656
1.83
0.911
0.963
1.021
1.011
1.075
1.148
1.14
1.221
1.316
1.637
1.803
2.009
1.895
2.115
2.395
2.255
2.561
2.96
1.272
1.373
1.493
1.438
1.567
1.723
1.66
1.83
2.04
1.005
1.068
1.14
1.116
1.194

NTUregressio
n
2.4614
1.1349
1.2412
1.3477
1.3229
1.4468
1.571
1.5376
1.6815
1.8258
0.8461
0.9253
1.0047
0.9863
1.0786
1.1712
1.1463
1.2536
1.3612
1.7185
1.8794
2.0407
2.0032
2.1908
2.3788
2.3282
2.5462
2.7647
1.2748
1.3941
1.5138
1.486
1.6251
1.7646
1.727
1.8887
2.0509
0.9504
1.0394
1.1286
1.1078
1.2116

1.22
-2.04
0.18

% Error
-7.74
-2.25
-0.31
0.2
0.83
1.89
1.42
1.62
1.54
-0.23
-7.12
-3.91
-1.59
-2.44
0.34
2.02
0.55
2.67
3.44
4.98
4.24
1.58
5.71
3.58
-0.67
3.25
-0.58
-6.6
0.22
1.54
1.39
3.34
3.71
2.41
4.04
3.21
0.53
-5.44
-2.68
-1
-0.73
1.47

35.5
35.5
35.5
35.5

30
30
30
30

25.5
26
26
26

1.3
1.1
1.2
1.3

1.284
1.259
1.357
1.475

1.3155
1.2875
1.4081
1.529

2.46
2.27
3.77
3.66

Figure 4.7 NTU_correlation versus NTU_true for L/G 1.1 1.3

Figure 4.8 Percentage error in NTUcorrelation for for L/G 1.1 1.3

4.5

L/G between 1.3 and 1.5

Table 4.3 shows the true NTU as calculated using Merkels theory and
NTU estimated from the correlation:
NTU = 9.458658. tw14.383413 .tw2-11.27517 .twb4.164945 (L/G)0.8327527

The maximum single-point error when using the correlation is -11.47%. Figure
4.3 shows that NTUcorrelation is in good agreement with the true NTU. Figure
4.3 gives the percentage error distribution for all NTUcorrelation where the
errors range from -11.47% to 7.22%.
Table 4.5 NTU for 1.3 < L/G < 1.5
tw1
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
34.5
35
35
35
35
35
35
35
35
35
35

tw2
29
29
29
29
29
29
29
29
29
29.5
29.5
29.5
29.5
29.5
29.5
29.5
29.5
29.5
30
30
30
30
30
30
30
30
30
29
29
29
29
29
29
29
29
29
29.5

twB
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26
25

L/G
1.3
1.4
1.5
1.3
1.4
1.5
1.3
1.4
1.5
1.3
1.4
1.5
1.3
1.4
1.5
1.3
1.4
1.5
1.3
1.4
1.5
1.3
1.4
1.5
1.3
1.4
1.5
1.3
1.4
1.5
1.3
1.4
1.5
1.3
1.4
1.5
1.3

NTUtru
e
1.629
1.804
2.026
1.934
2.18
2.501
2.381
2.748
3.253
1.199
1.294
1.407
1.378
1.504
1.657
1.625
1.799
2.018
0.905
0.958
1.018
1.016
1.083
1.161
1.161
1.249
1.354
1.817
2.033
2.312
2.162
2.465
2.87
2.668
3.12
3.758
1.345

NTUregressio
n
1.6242
1.817
2.017
1.9498
2.1812
2.4213
2.3324
2.6092
2.8964
1.1381
1.2732
1.4133
1.3663
1.5284
1.6967
1.6344
1.8283
2.0296
0.8023
0.8975
0.9963
0.9631
1.0774
1.196
1.1521
1.2888
1.4307
1.8658
2.0872
2.3169
2.2398
2.5056
2.7814
2.6793
2.9973
3.3271
1.3074

% Error
-0.29
0.72
-0.45
0.82
0.06
-3.19
-2.04
-5.05
-10.96
-5.08
-1.61
0.45
-0.85
1.62
2.39
0.58
1.63
0.57
-11.35
-6.31
-2.13
-5.2
-0.51
3.02
-0.77
3.19
5.66
2.68
2.67
0.21
3.6
1.65
-3.09
0.42
-3.93
-11.47
-2.8

35
35
tw1
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5
35.5

29.5
29.5
tw2
29.5
29.5
29.5
29.5
29.5
30
30
30
30
30
30
30
30
30
29
29
29
29
29
29
29
29
29
29.5
29.5
29.5
29.5
29.5
29.5
29.5
29.5
29.5
30
30
30
30
30
30
30
30
30

25
25
twB
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26
25
25
25
25.5
25.5
25.5
26
26
26

1.4
1.5
L/G
1.4
1.5
1.3
1.4
1.5
1.3
1.4
1.5
1.3
1.4
1.5
1.3
1.4
1.5
1.3
1.4
1.5
1.3
1.4
1.5
1.3
1.4
1.5
1.3
1.4
1.5
1.3
1.4
1.5
1.3
1.4
1.5
1.3
1.4
1.5
1.3
1.4
1.5
1.3
1.4
1.5

1.463
1.607

1.4626
1.6235

NTUtru
e
1.706
1.902
1.83
2.048
2.329
1.021
1.089
1.167
1.148
1.234
1.336
1.316
1.429
1.565
2.009
2.27
2.613
2.395
2.76
3.259
2.96
3.505
4.291
1.493
1.639
1.818
1.723
1.916
2.16
2.04
2.308
2.658
1.14
1.224
1.324
1.284
1.392
1.521
1.475
1.616
1.789

NTUregressio
n
1.7557
1.949
1.8774
2.1002
2.3314
0.9216
1.031
1.1445
1.1064
1.2377
1.3739
1.3234
1.4805
1.6434
2.139
2.3929
2.6563
2.5678
2.8726
3.1888
3.0717
3.4362
3.8144
1.4989
1.6768
1.8613
1.7993
2.0129
2.2344
2.1524
2.4078
2.6728
1.0566
1.182
1.3121
1.2684
1.4189
1.5751
1.5173
1.6973
1.8841

-0.03
1.03
% Error
2.92
2.47
2.59
2.55
0.1
-9.74
-5.33
-1.93
-3.63
0.3
2.84
0.56
3.6
5.01
6.47
5.41
1.66
7.22
4.08
-2.16
3.77
-1.96
-11.11
0.39
2.3
2.38
4.43
5.06
3.45
5.51
4.33
0.56
-7.32
-3.43
-0.9
-1.22
1.93
3.56
2.86
5.03
5.32

Figure 4.9 NTU_correlation versus NTU_true for L/G 1.3 1.5

Figure 4.10 Percentage error in NTUcorrelation for for L/G 1.3 1.5

CHAPTER FIVE
5. CONCLUSION

Merkerls theory was used to obtain theoretical correlation for predicting the
NTU of cross-flow cooling towers in the form of power equation,
NTU = a0. tw1a1.tw2a2.twba3(L/G)a4

Where the coefficients a0, a0, a0 and a0 are summarized in table 5.1.
Table 5.1 Coefficients of NTU Correlation
L/G

a0

a1

a2

a3

a4

0.5 - 0.7

6.8001
15
6.6614
08
7.2699
2
8.4209
94
9.4586
58

5.4033
19
3.9689
63
4.6276
45
4.3143
86
4.3834
13

-10.305
83
-8.6225
8
-9.3695
1
-9.9740
0
-11.275
2

2.7855
35
2.6386
64
2.5020
56
3.1672
16
4.1649
45

0.4895
2
0.3987
17
0.5178
05
0.7143
65
0.8327
53

0.7 - 0.9
0.9 - 1.1
1.1 - 1.3
1.3 - 1.5

Maximu
m error
-4.5%
-5.3%
-6.4%
-8.1%
-11.5%

REFERENCES

1.W.F.Stoecker & J.W.Jones, Refrigeration and Air Conditioning Second Edition,

McGraw-Hill
2.Dickey, J. B., Evaporative Cooling Towers, Marley Cooling Towers Co., 1978.
3. ASHRAE, Cooling Towers, in ASHRAE HandbookHVAC Systems and
Equipment, New York: ASHRAE, 2004.
4. Wistrom, G., Cold Weather Operation of Crossflow Cooling Towers, Plant
Engineering, September 8, 1975.
5. Liptk, B. G. (ed.), Cooling Towers: Their Design and Application, in
Environmental Engineers Handbook, Lewis Publishers, 1997.
6.Optimal Design of Cooling Towers, Heat and Mass Transfer - Modeling and
Simulation, Prof. Md Monwar Hossain (Ed.), ISBN: 978-953-307-604-1, InTech,
Available from: http://www.intechopen.com/books/heat-andmasstransfer-modeling-and-simulation/optimal-design-of-cooling-towers