at SciVerse ScienceDirect

Applied Thermal Engineering 38 (2012) 160e167

Contents lists available

Applied Thermal Engineering

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

A simplified method on thermal performance capacity evaluation of counter flowcooling tower

Wanchai Asvapoositkul*, Supawat TreeutokDepartment of Mechanical Engineering, King Mongkut’s University of Technology Thonburi, Bang Mod, Thung Khru, Bangkok 10140, Thailand

a r t i c l e i n f o

Article history:Received 11 October 2010Accepted 11 January 2012Available online 21 January 2012

Keywords:Cooling tower thermal performancecapabilityCooling tower analysisMerkel theoryPredicting cooling tower performancePerformance curve

* Corresponding author. Tel.: þ662 470 9338; fax:E-mail address: wanchai.asv@kmutt.ac.th (W. Asv

1359-4311/$ e see front matter � 2012 Elsevier Ltd.doi:10.1016/j.applthermaleng.2012.01.025

a b s t r a c t

The thermal performance capacity of a wet cooling tower is dominated by weather conditions, partic-ularly ambient wet-bulb temperature. In this paper, the tower performance was predicted by a simplifiedmodel which was characterized by specification of a mass evaporation rate equation. The purpose of thisstudy was to present a calculation that was accurate and simple to implement, and could be applied toevaluate acceptance tests for new towers, to monitor changes in tower performance as an aid in planningmaintenance, and to predict tower performance under changed operating conditions. The results werevalidated and showed good agreement with experimental measurements. The results were also pre-sented in simple formats that were easy to use and understand. These allowed reduction of test data andcomparison of test results to design data. The method held a practical advantage for predicted towerthermal performance capability to which it was best suited when both flow rate and temperature of inletwater were near design conditions since it required neither the measurement of air flow rate nor thecalculation of tower characteristic ðhmassA=LÞ. The expected results of this study will make it possible toincorporate cooling tower design and simulation to evaluate and optimize the thermal performance ofpower plants for example.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Cooling towers have many applications in the fields of air-conditioning, refrigeration and power plants. In the case of powergeneration plants or sugar mill plants, the cooling tower require-ments are relatively large and it has been the practice in recentyears especially in Thailand to fabricate increasingly larger coolingtowers. For large towers or towers with special requirements thatare not Cooling Technology Institute (CTI) certified, in-situ testing isthe only way to guarantee that the towers will perform as required.For this purpose, it is quite common to use the Merkel theory suchas that of CTI [1] or ASME [2] for the computation of tower char-acteristic (hmassA=L) or Number of Transfer Units (NTU). The prob-lems usually encountered in analysis of cooling towers for largeprocess plants included measurement of many test data with highaccuracy instruments, analysis of test data and comparison of testresults to design point. This is an expensive and time-consumingprocess that should be undertaken only after due consideration.

The thermal capacity of a cooling tower is obtained by per-forming the test. The test data should be evaluated by comparing

þ662 470 9111.apoositkul).

All rights reserved.

correctly with the design conditions that were instructed accordingto the CTI cooling tower acceptance test code [1]. Incidentally, thesedata are not only useful in the determination of thermal capacity ofthe tower according to design conditions during the test run periodbut can also be used to determine the operating characteristics inthe change in atmospheric conditions, especially temperatures.Notable examples of techniques based on this approach are thework of Fujita and Tezuka [3], Peterson and Backer [4] and Lucaset al. [5]. They demonstrated that the cooling tower characteristiccurve predicted from the Merkel principle is simple in terms offormulation and can provide reliable estimate of cooling towerperformance at off-design. By this method, the tower operatingconditions are determined directly using the slope of the coolingtower characteristic curve.

Even though the method has been applied to predict the overallthermal evaluation of cooling towers, there are some concernsabout simplifying assumptions of the Merkel theory such as theneglecting of the reduction of waterflow rate by evaporation andthe saturated water vapor (or 100% relative humidity) of air at thetower exit. Themethod tends to underestimate the heat rejected bythe cooling tower but can be used if only the water outlettemperature is of importance [6]. Kloppers and KrÖger alsoproposed a technique to get accurate prediction by including thewater loss due to evaporation in the energy equation. The effect of

Nomenclature

A exposed surface-area (air/water interface area), m2

Appr approach (Tw,o � Twb), �Ccp,a specific heat of dry air at constant pressure, kJ/kgKcp,w specific heat of water at constant pressure, kJ/kgKG dry air mass flow rate, kg/sh specific enthalpy, or specific total air enthalpy, kJ/kghconv convective heat transfer coefficient, W/m2Khfw enthalpy of saturated liquid water evaluated at

Tw, kJ/kghgw enthalpy of saturated water vapor evaluated at

Tw, kJ/kghmass convective mass transfer coefficient, kgair/m2shmassA=L or NTU tower characteristicsL water mass flow rate, kg/sP fan power, wattr2 correlation coefficientR range (Tw,i � Tw,o), �CRH relative humidityT temperature, �C

Twb wet-bulb temperature, �C

Greek symbolsw specific volume, m3/kgr density, kg/m3

u humidity ratio, kgw/kgda

Subscriptsa aird design valueda dry airfw saturated liquid of watergw saturated vapor of wateri inleto outletsw saturated watert test valuev vaporw waterz z coordinates

Fig. 1. Control volume for cooling tower.

W. Asvapoositkul, S. Treeutok / Applied Thermal Engineering 38 (2012) 160e167 161

evaporation causes the water flow rate to decrease from inlet tooutlet; as a result, the ratio of water-to-air (L/G) varies through thetower. These two effects (evaporation loss and variable L/G) wereinvestigated by Baker and Shryock [7]. For calculation of counterflow cooling tower with evaporation loss, a constant L/G ratioresults in a 4.4% increase in NTU at a 22 �C range. And evaporationloss and varied L/G ratios result in a 1.34% increase in NTU at thedegree range.

The purpose of this study was to apply the cooling tower perfor-mance characteristics to determine the operating characteristics forthecooling towerbeing considered. Thestudywasalsotodetermine ifthe performance curves could be used to evaluate thermal perfor-manceof the cooling towerwithoutmeasurement of airflow rate andthe calculation of hmassA=L. And tower capacity was more accuratelyexpressed in terms of the ratio of water-to-air loading (L/G) thatincluded water evaporation and unsaturated air leaving the tower.

2. Theoretical analysis (basic equation)

The analysis considers an increment of a cooling process as incontrol volume dz of Fig. 1 wherewater mass flow rate L and dry airmass flow rate G flow uniformly of plane area. All horizontalsections through the tower are assumed to be the same, in whichboth streams move in an opposite and vertical direction (watermoves downward while air moves upward).

A mass balance and an energy balance for a steady water-sprayflow with total exposed surface-area (air/water interface area)element dA, as in flow path dZ (assuming negligible kinetic andpotential energies and work).

Mass balance for dry air

dG ¼ 0 (1)

Mass balance for water

dL ¼ G dua (2)

Energy balance

Gdha ¼ Ldhfw þ hfwdL ¼ Ldhfw þ hfwGdua (3)

Heat is removed from the water by a transfer of sensible heatdue to a difference in temperature levels, and by latent heat

equivalent of a mass transfer resulting from the evaporation ofa portion of the circulating water. The energy balance on the waterside in terms of heat and mass transfer coefficients, hconv and hmassrespectively, is

Ldhfw ¼ hconvðTsw � TaÞdAþ hmassðusw � uaÞhgwdA (4)

Themass balance on the air side of the evaporated water mass is

Gdua ¼ hmassðusw � uaÞdA (5)

The simultaneous heat and mass transfer takes place and can beexpressed by substituting (4) and (5) into (3) and through rear-rangement we get,

Gdha ¼ hconvðTsw � TaÞdAþ hmassðusw � uaÞhgwdA (6)

Fig. 3. Cooling tower demand curve.

W. Asvapoositkul, S. Treeutok / Applied Thermal Engineering 38 (2012) 160e167162

By applying and replacing the Lewis factor (hconv=hmasscp;a orratio of overall heat transfer to overall mass transfer), and ther-modynamic properties of airewater, the simplified equation (6) is

Gdha ¼hmassdA�ðhsw�haÞþ

�hconv

hmasscp;a�1

��ðhsw�haÞ

�ðusw�uaÞhgw�� ð7Þ

If the Lewis factor is equal to 1, we get

Gdha ¼ hmassdA½ðhsw � haÞ� (8)

And if the reduction of water flow rate by evaporation isneglected in the energy balance, we get

Ldhfw ¼ hmassdA½ðhsw � haÞ� (9)

hmassdAL

¼ d hfwðhsw � haÞ (10)

Integrating

hmass AL

¼Zh2

h1

dhfwðhsw � haÞ ¼

ZT2

T1

cdTfwðhsw � haÞ (11)

This is known as the Merkel equation. Integration for equation(11) is done by using Tchebyshev’s method which gives a highdegree of accuracy in the case of large cooling ranges as suggestedby CTI [1].

2.1. Tower characteristics

The tower characteristics (hmassA/L) are a dimensionless variablewhich can be determined by integrated value of equation (11) atdesign condition. This value is based on the equipment’s designrequirements or is a measure of the difficulty of the task [8]. Incooling tower design practice, it is referred to as an acceptedconcept of cooling tower performance [1,2]. For the given coolingtower, its value depends on the ratio of water-to-air loading (L/G).And the dimensionless variable, (L/G), can be determined from theknownwaterflow and known air flow. The hmassA/L value of a toweroperating at off-design conditions will not be the same as the

Fig. 2. Cooling tower characteristic curve with design point and test point.

hmassA/L value at design conditions. An empirical equation usefulfor predicting hmassA/L at off-design conditions is [8]:

hmassAL

¼ c�LG

��n

(12)

Values of c and n are determined from performance dataprovided by manufacturers. Typical values of n are in the range of0.4< n< 0.6 [9]. If a typical value of n is assumed, the value of c canbe determined from L and G at nominal design conditions. Once cand n are known for a particular cooling tower, the cooling towerperformance can be predicted at any operating condition given thewater inlet temperature Tw,i, the ambient air wet-bulb temperatureTwb, and the flow rates L and G. The tower characteristic hmassA=Lcan then be plotted against varying (L/G) ratio, and this givesa measure of the ability of the tower to effect the transfer such asshown in Fig. 2.

2.2. Cooling demand curves

The water temperature and air temperature or enthalpy arebeing changed along the tower and the Merkel relation can only beapplied to a small element of the heat transfer surface. Referring to

Fig. 4. Tower demand and characteristic curve.

Fig. 6. Cooling tower test rig configuration.

W. Asvapoositkul, S. Treeutok / Applied Thermal Engineering 38 (2012) 160e167 163

the right-hand side of equation (11), (hsw � ha) is the differencebetween the enthalpy of saturated air at the water temperaturesand the enthalpy of air temperature at each location in the tower.This equation is used to calculate thermal demand based on thedesign temperatures and selected L/G.

Fig. 3 is an example of a curve, on which the required hmassA/L,for a given inlet air wet-bulb temperature and range, is plottedversus L/G with the approach as a parameter. This is known asa demand curve.

Now, it is possible to superimpose the tower characteristic curve(Fig. 2) over the demand curve (Fig. 3), the intersect being theoperating point for the tower being considered for the duty such asshown in Fig. 4.

2.3. Simulation calculation

Cooling towers operate most of the time at conditions differentthan their design conditions therefore the data extracted from Fig. 4would be important information to have for plant thermal opti-mization. For a given cooling tower, its characteristics are describedby equation (12) which (hmassA/L) will remain unchanged as long asthe ratio of water-to-air loading (L/G) is constant. Weather condi-tions, particularly ambient wet-bulb temperature, will affect therange and the approach of the cooling tower. The cooling watertemperatures relate to the range and the approach, as follows.

Tw;i ¼ Tw;o þ R (13)

Tw;o ¼ Twb þ Appr (14)

A procedure for simulating the performance of a cooling toweris the simultaneous solution of equations (11) and (12). The

Fig. 5. Flow diagram for the cooling tower simulation calculation.

sequence of the calculation is shown by the flow diagram in Fig. 5.Starting with trial values of Appr for an ambient Twb and R, thevalue of (hmassA/L) can be obtained from equations (11) and (12). Inpractice, the equations are solved iteratively with the updatedvalues of Appr until the specified hmassA/L from equation (12) issatisfied.

2.4. Modifications

The calculating of hmassA/L is computed using either equation(11) or (12) which is obtained once L/G is determined. A charac-teristic point is experimentally determined by first measuring anambient dry bulb and wet-bulb temperatures, air discharge drybulb andwet-bulb temperatures, and cooling water inlet and outlettemperatures. The L/G ratio is then calculated as follows;

LG

¼�ha;o � ha;i

� hw;o�ua;o � ua;i

�hw;i � hw;o

¼�ha;o � ha;i

� hw;o�ua;o � ua;i

cp;w

�Tw;i � Tw;o

(15)

Once the value of L/G is known, the procedure for calculatinghmassA/L is computed using the enthalpy values at the measuredtemperatures. This provides the evaluation of tower characteristics(hmassA/L) on the basis of the true L/G.

If the effect of evaporation is ignored, equation (15) may bewritten as

LG

¼�ha;o � ha;i

�hw;i � hw;o

¼�ha;o � ha;i

cp;w

�Tw;i � Tw;o

(16)

Table 1Cooling tower design condition.

Design condition

Water loading 60 L/min-m2

Hot water temp. (Tw,i) 38.5 �CCold water temp. (Tw,o) 33.5 �CInlet wet-bulb temp. (Twb,i) 29 �CInlet dry bulb temp. (Tdb,i) 36 �CTotal fan driver power 185 wBarometric press. 1.0013 barLiquid to gas ratio (L/G)d 1.163

Twb,o exper. ( oC )

30.5 31.0 31.5 32.0 32.5 33.0

Tw

b,o

pred

icte

d (

o C )

30.5

31.0

31.5

32.0

32.5

33.0

-2%

+2%

Fig. 8. Comparison of wet-bulb temp. at cooling tower exit between the experimentdata and the predicted value from equation (16).

Table 2Specifications of the measuring devices.

Measurement Instrument Accuracy Resolution

Water flow rate Rotameter �2% 5 L/minWater temp. RTD temperature probe �2% 0.1 �CAmbient wet/dry Temp. RTD temperature probe �0.8 �C 0.1 �CInlet/outlet air velocity Vane Anemometer �2% 0.1 m/sFan power Multi-meter �2% 1 V, 0.1 A

W. Asvapoositkul, S. Treeutok / Applied Thermal Engineering 38 (2012) 160e167164

It may be assumed that the air discharge is saturated. There-fore, the air discharge is at its wet-bulb. This is based on MerkelModel.

The CTI code determines the test value of (L/G)t from the averagewater flow rate and fan driver output power at the time of test. Itsvalue is calculated from [1]:

�LG

�t¼

�LG

�d

�LtLd

��PdPt

�13�rtrd

�13�wt

wd

�(17)

3. Experiments

3.1. Apparatus

The experiment was performed in the induced draft counterflow cooling tower test rig (see Fig. 6). The tower was made ofstandard industrial cooling tower equipment andmaterial with oneexception e one side of its walls was made of clear high strengthpolycarbonate material that allowed direct observation of the drifteliminator, the fill under test, the spray pattern of the nozzle andthe interaction of the air and water. The tower’s inside dimensionswere 1000 mm � 1000 mm with a total height of 3350 mm andcould accommodate up to 1500mmof fill. Design conditions for thetower are summarized in Table 1.

The spray nozzle was attached to a movable frame that enabledaccurate placement of the nozzle spray, which allowed for full spraycoverage of the fill under test. The test section’s rain zone (fallingwater below the fill) was adjusted to 400 mm for all tests.

Water was circulated by a centrifugal pump. The flowwas variedmanually by means of a flow control valve and measured by

Fig. 7. Schematic diagram of c

a rotameter. A 70 kW gas burner supplied the heat load to thecirculated water. The water was then delivered to an insulated tankwhere its temperature was maintained at a constant value duringtesting with two supplemental electrical heaters each of 9 kW.

Induced air was circulated counter flow by an axial flow fan. Thefan speed could be varied by variable frequency drives. Air velocitywas measured by a vane anemometer. Inlet and exit air wet and drybulb temperatures were measured with a Resistance TemperatureDetector (RTD) temperature probe which calibrated to mercury-in-glass thermometers.

The specifications of the measuring devices are shown inTable 2. And the test rig schematic is shown in Fig. 7.

3.2. Procedure

In this experiment, the inlet hot water temperature of the towerwas kept constant while the flows of water and air were varied. Thetower test was conducted in accordance with the Cooling

ooling tower test facility.

Fig. 9. Comparison of L/G between the experiment data and the predicted value fromtest values of discharged air properties (TDA), test values of discharged air propertieswithout evaporation (TWE) and test fan driver output power (TFD).

wet-bulbe temp. ( o

C )

18 20 22 24 26 28 30 32 34

Ran

ge ( o

C )

3

4

5

6

7

8

9

+5%CAPACITY

-5%CAPACITY

100%CAPACITY

Fig. 11. Cooling tower evaluations at inlet water temp. of 38.5 �C.

W. Asvapoositkul, S. Treeutok / Applied Thermal Engineering 38 (2012) 160e167 165

Technology Institute (CTI) Acceptance Test Code for Water-CoolingTowers ATC-105 [1].

According to the experiments testing data were Tw,i ¼ 38.5 �Cand L/G was 0.90, 1.06, 1.10, 1.14, 1.2 and 1.24. The fill of 600 mmheight was chosen for the experiments. The test was conductedwithin the following variations from the test conditions. Circulatingwater flow, heat load and range were not varied by more than 2%.Instantaneous air temperature readings fluctuated during the test,but variations in average readings during the test period did notexceed 1 �C per hour for wet and dry bulb temperatures.

After reaching steady state conditions, the inlet and exit watertemperature was taken at every 5-min interval. A total of 12readings were taken, and then an average was calculated. Inlet andexit air wet and dry bulb temperatures weremeasured at the centerof each side of the louvers and that of the fan stack exit. A readingwas taken every 5-min, and then the average was computed. Thewater flow rate was measured at every 20-min interval. A total of 3readings were taken and then the average was computed. Fanpower consumption was measured by using a multi-meter. Areading was taken at every 30-min interval. A total of 2 readingswere taken and then the average was computed.

Fig. 10. Comparison of hmassA/L calculated from experiment data L/G and predicted L/Gfrom test values of discharged air properties (TDA), test values of discharged airproperties without evaporation (TWE) and test fan driver output power (TFD).

4. Application and comparison with Merkel model

To calculate tower characteristics (hmassA/L) from the Merkelequation, it is necessary to know the air wet-bulb temperature atthe inlet, hmassA/L, and the water temperature at the inlet and theexit. The enthalpy of air at the exit is approximated from equa-tion (16) where saturated air at the inlet and the exit as well asno water evaporation are assumed. The saturated air temperatureat the exit can be determined easily from a Psychometric chart orfrom a computer program [1]. Fig. 8 shows the comparison of theexperimental and numerical values of discharged air wet-bulbtemperature Twb,o. It can be seen that the outlet wet-bulbtemperatures from the experimental and predicted values arein good agreement. The maximum errors were found to be lessthan 2%.

If the properties of air at the inlet and the exit are known, theeffect of evaporation and the true air properties can be used tocalculated the ratio of water-to-air loading (L/G) from equation (15).The values of L/G base on equations (15)e(17) were calculated andcompared with that of experimental data as shown in Fig. 9. Itshould be noted that the result from equation (15) were

Fig. 12. Predicted inlet water temp. at 29 �C entering wet-bulb temp. and 60% RH.

Fig. 13. Predicted inlet water temp. for 105% of design water circulation. Fig. 15. Predicted inlet water temp. for 95% of design water circulation.

W. Asvapoositkul, S. Treeutok / Applied Thermal Engineering 38 (2012) 160e167166

determined from test values of discharged air properties (TDA)while that from equation (16) were determined from test values ofdischarged air properties without evaporation (TWE) and that fromequation (17) were determined from test fan driver output power(TFD). The linear regression (r2) of the predicted values was rankedin ascending order as follows: TWE, TFD and TDAwith the values of0.657, 0.893 and 0.936, respectively.

Fig. 10 presented the calculated values of tower characteristics(hmassA/L) obtained with L/G from experiments, TDA, TWE and TFD.The difference among the data indicated the influence of L/Gwhereresult from TDA showed excellent agreement with hmassA/L ob-tained from the experiment value of L/G. The linear regression (r2)of the predicted values from TDA, TFD and TWE was 0.980, 0.974and 0.876, respectively. The deviation was less than 2% thereforethe methods were understood to be suitable since percent errors ofabout 5%e10% always go with the heat balance in performancetests [3].

The values from TDA were plotted for the tower characteristiccurve, shown in Fig. 2, at the design condition of Twb ¼ 29 �C,and L/G ¼ 1.163. The line was fit to the model given in equation

Fig. 14. Predicted inlet water temp. at design water circulation.

(12). The values of n and c were found to be 0.531 and 0.764,respectively.

5. Evaluation of tower performance

In evaluating cooling tower thermal capacity, the designconditions of the tower must be available (either from the manu-facturer or test data). In this illustration, the tower characteristic isshown in Fig. 2. Its operating characteristics were predicted asdescribed in section 2.3.

With properly selected demand curves (preferably with givenconstant inlet water temperature) and subject to certain L/G (pref-erably with given tower capacity �5%) the method could beemployed to meet a wide range of service requirements. Fig. 11illustrates evaluation cooling tower capacity curves for inlet watertemperature, Tw,i ¼ 38.5 �C; this could be expanded to other inletwater temperature ifdesired. Therefore, theoperatorof sucha coolingtower can determine the tower capability from the graph as shown inFig.11. In considering other inlet water temperaturewith varying Twbat water flow rate of �5% of the design flow rate, see Figs. 12e15.

The cooling tower capacity illustration in Fig.11was based on theassumption that the test conditions of the water flow rate and inletwater temperaturewere near design conditions. Themethod has anadvantage in that neither the measurement of air flow rate nor thecalculation of hmassA/L was required. This method was proposed byFujita and Tezuka [3]. For practical use, inlet water temperatures arewithin �2 �C, water flow rates are within �5% and inlet wet-bulbtemperatures arewithinþ3 �C/-17 �C fromthedesign conditions [3].

6. Conclusions

A calculation method for predicting the behavior of induceddraft wet cooling tower has been developed, with a new methodthat included water evaporation and unsaturated air leaving thetower. In the case where those two conditions are ignored(TWE), the maximum error from the predicted outlet wet-bulbtemperatures was less than 2%. While those two conditionswere considered (TDA), the predicted values of (L/G) were foundto be best suited with those from the experiment with the linearregression (r2) of 0.936. The other predicted values of (L/G) weredetermined from the test fan driver output power (TFD) thatgave r2 of 0.893 and those from TWE that gave r2 of 0.657. Hence

W. Asvapoositkul, S. Treeutok / Applied Thermal Engineering 38 (2012) 160e167 167

the results of the calculated values of tower characteristics(hmassA/L) where the r2 of the predicted values from TDA, TFDand TWE were 0.980, 0.974 and 0.876, respectively.

With the available data either from the design conditions or thetest data of the cooling tower, the prediction operating conditionscan be presented in simple formats. The method is also applied topredict the cooling tower thermal performance capability whenboth flow rate and temperature of inlet water near design condi-tions without the measurement of air flow rate and the calculationof hmassA/L. Subsequently, the results can be used to determine oroptimize counter flow wet cooling tower design for a given set ofoperating conditions.

Acknowledgements

This research has been supported by the Thailand Research Fundthrough the MAGWindow I Program (Grant No. MRG-WI525E078),and the Thai Cooling Tower Company.

References

[1] Cooling Technology Institute, Acceptance Test Code for Water-Cooling TowersATC-105, Cooling Technology Institute, Houston, TX, 2000.

[2] The American Society of Mechanical Engineers, Atmospheric Water CoolingEquipment PTC 23-2003, ASME, New York, 2003.

[3] T. Fujita, S. Tezuka, Calculations on thermal performance capability ofmechanicaldraft cooling towers, Bulletin of JSEM 27 (225) (1984) 490e497.

[4] N. Peterson, Luc De Backer, A simplified method to evaluate cooling tower andcondenser performance using the CTI toolkit, CTI Journal 30 (1) (2009).

[5] M. Lucas, P.J. Martinez, A. Viedma, Experimental study on the thermal perfor-mance of a mechanical cooling tower with different drift eliminators, EnergyConversion and Management 50 (2009) 490e497.

[6] J.C. Kloppers, D.G. KrÖger, Cooling tower performance evaluation: Merkel,Poppe, and e-NTU methods of analysis, Journal of Engineering for Gas Turbineand Power 127/1 (2005).

[7] D.R. Baker, H.A. Shryock, A comprehensive approach to the analysis of coolingtower performance, Journal of Heat Transfer ASME Technical Bulletin (August1961) R-61-P-13.

[8] Stephen A. Leeper, Wet Cooling Tower: ‘Rule-of-Thumb’ Design and Simulation,U.S. Department of Energy Assistant Secretary for Resource Application, 1981,Office of Geothermal, under DOE Contract No. DE-AC07e76ID01570.

[9] D. Baker, Cooling Tower Performance, Chemical Publishing Co. Inc., New York,1984.