Sunteți pe pagina 1din 8

Applied Thermal Engineering 27 (2007) 520527 www.elsevier.

com/locate/apthermeng

Prediction of evaporation losses in evaporative uid coolers


Bilal Ahmed Qureshi, Syed M. Zubair
*
Mechanical Engineering Department, KFUPM Box 1474, King Fahd University of Petroleum and Minerals, Dhahran 31261, Saudi Arabia Received 28 February 2006; accepted 11 June 2006 Available online 17 August 2006

Abstract The accurate prediction of various aspects of thermal behavior of evaporative uid coolers is very important for both design and rating calculations. Exactly predicting evaporation losses is signicant since the process uid is cooled primarily by evaporation of a portion of the recirculating water that causes the concentration of dissolved solids and other impurities to increase. An empirical relation to predict evaporation losses is developed on the basis of the rule of thumb recommended by manufacturers, which is simple and accurate with a wide range of applicability. The predicted values are in good agreement with the numerical values obtained from the calibrated model where the maximum error was found to be approximately 4% but often less than 2%. 2006 Elsevier Ltd. All rights reserved.
Keywords: Evaporation losses; Fluid cooler; Numerical model

1. Introduction Cooling by evaporation is well known and it has found many industrial applications. A rational development of a combined cooling-tower heat-exchanger unit is typically dened as the evaporative uid cooler. In the evaporative uid cooler, the purpose of the cooling tower (to cool water) and that of the heat exchanger (to cool the process uid using the cooled water) are combined. It is important to note that, with the growth of the refrigeration and air conditioning industry, the evaporative cooler came into extensive use, principally, as a refrigerant condenser. Fig. 1 shows a schematic diagram of the evaporative uid cooler in which uid ows from the top to bottom. The modeling of an evaporative cooler is complicated by the fact that three uids, normally owing in dierent directions, interact with each other through heat and mass transfer processes. A constant spray water temperature was often assumed in the model but Parker and Treybal [1] realized that this assumption resulted in the governing equa-

Corresponding author. Tel.: +966 3 860 3135; fax: +966 3 860 2949. E-mail address: smzubair@kfupm.edu.sa (S.M. Zubair).

tions of the model to become inconsistent, thus, giving a futile answer. In the same reference, they also reported a detailed experimental study to dene the heat and mass transfer characteristics of evaporative uid coolers. Mizushina et al. [2] developed two dierent rating methods for evaporative coolers; one, a numerical procedure and the other, a straightforward analytical model based on the assumption of constant water temperature. Finlay and Grant [3] showed that this assumption might lead to errors in excess of 30%, for example, in large tube banks. A rating method, based on cooling tower procedures, was proposed by Tezuka et al. [4] but the assumptions made in this model were not as accurate as those used in the model of Parker and Treybal or of a simple model of Mizushina. In another study, Finlay and Grant [5] simplied the equations describing the mass transfer in an evaporative cooler by assuming that the vapor pressure of saturated moist air is a linear function of temperature. The model can be expected to be very accurate, as this is the only major assumption made in the analytical formulation. The nal design equations are somewhat complicated and therefore require a numerical solution procedure. Webb [6] performed a unied theoretical treatment for thermal analysis of cooling towers, evaporative condensers

1359-4311/$ - see front matter 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.applthermaleng.2006.06.008

B.A. Qureshi, S.M. Zubair / Applied Thermal Engineering 27 (2007) 520527

521

Nomenclature a A b cp d E ErrB1 ErrB2 ErrG Err24 h hc hc,w hD hf,w hg,w h0 g hfg,w hs,w L Le mratio interfacial area per unit volume of a tube bundle, m2 m3 outside surface area of cooling tubes, m2 empirical constant (See Eq. (21)) specic heat at constant pressure, kJ kg1 K1 diameter, m percentage of water evaporated, % percentage error between numerical data and Baltimore Aircoils rule of thumb, % percentage error between numerical data and Baltimore Aircoils procedure, % percentage error between the numerical data and Grahams rule of thumb, % percentage error between the numerical data and Eq. (24), % 1 specic enthalpy of moist air, kJ kg a convective heat-transfer coecient of air, kW m2 K1 lm heat-transfer coecient, kW m2 K1 convective mass transfer coecient, kgw m2 s1 specic enthalpy of water evaluated at tw, 1 kJ kg w specic enthalpy of saturated water vapor eval1 uated at tw, kJ kg w specic enthalpy of saturated water vapor eval1 uated at 0 C, kJ kg w change-of-phase enthalpy (hfg,w = hg,w hf,w), 1 kJ kg w specic enthalpy of saturated moist air evalu1 ated at tw, kJ kg w Length of tube, m Lewis factor (Le = hc/hDcp,a) _ w;i =m _ a mass ow ratio m _ m ntr NTU
_

P _ Q Rf R2 Re Rint t U W Ws,w / C Dtp

mass ow rate, kga s1 number of tube rows _ a number of transfer units hD A=m pitch, m heat transfer, kW fouling, m2 C kW1 coecient of determination Reynolds number interface resistance, m2 C kW1 temperature, C overall heat transfer coecient, kW m2 K1 1 humidity ratio of moist air, kgw kg a humidity ratio of saturated moist air evaluated 1 at, kgw kg a relative humidity water ow rate per unit tube length, kg m1 s1 cooling range (tp,i tp,o), C

Subscripts a air db dry-bulb g,w vapor at water temperature i inlet int at airwater interface o outlet os outside p process uid s,w saturated moist air at water temperature t tube tot total v vapor w water wb wet-bulb

and evaporative uid coolers. In this paper, a specic calculation procedure is outlined for sizing and rating each type of evaporative heat exchanger. In another paper, Webb and Villacres [7] presented computer algorithms that have been developed to perform rating calculations of evaporatively cooled heat exchangers. The algorithms are particularly useful for rating commercially available heat exchangers at o-design conditions. Dreyer [8] presented various mathematical models for thermal evaluation of evaporative coolers and condensers. These models ranged from the exact model based on Poppe and Ro gener [9] to the simplied models of Mizushina et al. [2,10]. However, these authors did not discuss the impact of fouling on thermal performance of these evaporative heat exchangers, which is recently presented by Qureshi and Zubair [11,12]. The objective of this paper is to obtain an equation that would enable users of evaporative uid coolers to predict evaporation loss more accurately but which is also simple enough to avoid the use of tables or charts.

2. Evaporative uid cooler model Fig. 2 illustrates an innitesimal control volume of the basic model for evaporative uid coolers in which the air (subsystem I) is owing in an upward direction whereas the water (subsystem II) is sprayed in a downward direction. It is considered that the process uid (subsystem III) ows in the upward direction in the evaporative cooler. The major assumptions that are used to derive the basic modeling equations may be summarized as [2,6,8,12]: The system is in a steady state. The apparatus and the cooling water recirculating circuit are insulated from the surroundings. Radiation heat transfer is ignored. Water lost by drift is negligible. The heat and mass transfer coecients are constant within the tube bundle. Complete surface wetting of the tube bundle is assumed.

522

B.A. Qureshi, S.M. Zubair / Applied Thermal Engineering 27 (2007) 520527

. Air Out (ma , ho )


Drift Eliminators

Spray Distribution

External . Water mw,i , t w,i

. m p , t p ,o , h p ,o
Closed-Circuit Heat Exchanger Coil Air In Water . m w , t w,o Air . ma , t wb ,i

_w oW 1 om 2 _ a oA oA m The mass ow of recirculating water evaporating into air, in terms of the mass transfer coecient, hD, can be written as   _w om _w _ w hD W s;int W dA m 3 dA m oA After simplication, we get _ w hD W s;int W dA dm 4

. m p , t p ,i , h p ,i
Air In

Eqs. (2) and (4) indicate that the mass ow rate does not remain constant as some of the water evaporates into the atmosphere. At the airwater interface simultaneous heat and mass transfer takes place, which can be expressed as   ! oh _a h _ a h hc tint ta dA m dA m oA hD W s;int W hfg;int dA simplifying, we get _ a dh hc tint ta dA hD W s;int W hfg;int dA m 6 5

Pump
Fig. 1. Schematic of a counter-ow evaporative uid cooler.

The distribution of air and water is uniform at the inlets and this uniformity is maintained. Thus, the temperatures in the unit will only depend on the vertical position in the unit, which implies the model is one-dimensional. The lm temperature at the airwater interface is equal to the bulk lm temperature. Water temperature at the inlet and outlet is same. The airwater interface area is approximately equal to the outer surface area of dry tubes i.e. the water lms on the tubes are very thin. The water mass balance (refer to Fig. 2), yields   !   _w om oW _ aW m _w _a W _wm m dA dA m oA oA simplifying, we get

The above equation can be simplied by introducing Lewis factor, Le = hc/hDcp,a, and assuming that hfg % hg. This gives [8]: _ a dh hD dAhs;int h Le 1fhs;int h m W s;int W hg;int g 7

where cp,a is the specic heat of the mixture and hg,int is specic enthalpy of water vapor evaluated at the interface temperature, tint. If Lewis factor is taken as unity, we get _ a dh hD hs;int h dA m 8 The overall energy balance on the process uid can be expressed as:   ! ohp _ p hp _ p hp m 9 m dA U os tp tint dA oA

Fig. 2. Mass and energy balance of an evaporative uid cooler.

B.A. Qureshi, S.M. Zubair / Applied Thermal Engineering 27 (2007) 520527

523

If the enthalpies of the water and process uid are written as     ohf ;w ot w ohp ot p c p ;w ; c p ;p 10 oA oA oA oA we can write on substituting Eq. (10) in (9) as _ p cp;p dtp U os tp tint dA m 11

form by integrating both sides. We then use EES to solve initial value dierential equations as detailed below [12,13]. For the process uid in the uid cooler, we write Z A  dtp tp tp;in dA 15 dA Ai where the term in brackets is known from Eq. (11). Similarly, for air and water, the other relevant equations are written as Z A  dh h hin dA 16 dA Ai Z A  dtw tw tw;in dA 17 dA Ai  Z A dW dA 18 W W in dA Ai  Z A _w dm _ w;out _w m m dA 19 dA Ai where the term in brackets for Eqs. (16)(19) are known from Eqs. (8), (13), (2) and (4), respectively. In a uid cooler, the process uid can either be considered to be owing from bottom-to-top or top-to-bottom. In the present study, the former conguration was considered as this is chosen for comparatively easier calculation, since the process uid temperature at the bottom (tp,in) is known. The number of transfer units of the evaporative uid cooler is calculated by Z hout hD A dh NTU 20 _a m h int h hin It is important to note that the NTU is a measure of the airwater interface area required to aect the required heat transfer duty. The calculations regarding the uid cooler have been validated from the experimental data provided by Jang and Wang [14] shown in Fig. 3 and the results were found

where Uos is the time-dependent (due to fouling) overall heat transfer coecient. Therefore, the overall energy balance on the control volume for the evaporative cooler results in   !   ! _w om ohf ;w _ ah m _w _ p hp m dA h f ; w dA m oA oA   ! oh _ _ w hf ;w ma h dA m oA   ! ohp _ p hp m dA 12 oA Simplifying and using Eq. (10), we get _ a d h c p ;w t w d m _wm _ p cp;p dtp _ w cp;w dtw m m 13

It is important to note from Eq. (13) that some of the heat removed from the process uid goes to heating (or cooling) the water lm. Based on outside surface area of the tubes, the overall heat transfer coecient Uos, as a function of actual progress of fouling process, can be written as !   ! 1 1 d t;os d t;os d t;os 1 Rint Rf ln U os hc;is d t;is h c; w 2k t d t;is 14 Now, if the temperature of the interface lm is considered the same as the bulk water temperature, then all the terms with the subscripts (s, int) will be replaced by (s,w) i.e., interface thermal resistance, Rint ! 0. This approach was used in the current work. 3. Numerical solution The full system of ve dierential equations describing the operation of the evaporative uid cooler is given by Eqs. (2), (4), (8), (11) and (13). It should be noted that the correlations for convective and condensation heat transfer coecients inside the tubes as well as the mass transfer and the lm heat transfer coecients outside the tubes are used from various sources which are described in Qureshi and Zubair [11,12], and are also summarized in Appendix A. A computer program is written in Engineering Equation Solver (EES) for solving these equations for the uid cooler. In this program, properties of airwater vapor mixture are needed at each step of numerical calculation, which are obtained from the built in functions provided in EES. For example, in this program any rst-order dierential equation is rst transformed into an appropriate

Fig. 3. Verication of evaporative cooler model from the data of Jang and Wang [14].

524

B.A. Qureshi, S.M. Zubair / Applied Thermal Engineering 27 (2007) 520527

Table 1 Percentage error in calculated values of outlet process uid temperature _ a (kg/s) m 1.880 0.166 2.070 2.913
a b c

_ w;i (kg/s) m 2.667 0.458 1.845 2.500

_ p (kg/s) m 15.00 0.325 2.670 6.000

tdb,i (C) 25.0 17.5 10.0 25.0

twb,i (C) 19.50 13.43 8.450 18.00

tp,i (C) 50.0 44.8 15.6 50.0

tref p;o (C) 48.32a 37.50b 13.55c 44.15d

tcalc p;o (C) 48.11 37.58 13.55 45.14

   calc  tp;o tref p;o  (C) 0.21 0.08 0 0.01

Dreyer [8]. Mizushina and Miyashita [10]. Finlay and Harris [15].

to be in good agreement (less than 4% error in prediction of total heat transfer). Furthermore, numerical examples provided by other authors [8,10,15,16] were also compared with the values obtained by our model, and the results are summarized in Table 1. The absolute values of the temperature dierence in outlet process uid temperature (tp,out) from the literature and the present model are given in the last column and were found to be less than 0.25 C. 4. Empirical equation Evaporation occurs as heat is transferred from the process uid where the lowest achievable temperature is the inlet wet-bulb temperature, which is basically governed by the inlet dry-bulb temperature and relative humidity. Therefore, the maximum potential for evaporation is the dierence between the inlet process uid and inlet wet-bulb temperatures (tp,i twb,i). For any xed value of the relative humidity, a higher dry-bulb temperature yields a higher wet-bulb temperature, which indicates a smaller potential for evaporation. It is important to note that commercial companies in the eld of evaporative cooling devices have determined their own rule of thumb to calculate losses due to evaporation. In this regard, we will be considering, for example, Baltimore Aircoil [17] and Graham Manufacturing Company [18]; of which the former also uses a simplied calculation procedure to achieve a rened prediction. The rule of thumb recommended by Baltimore Aircoil uses the fact that the evaporation rate is approximately 1.8 liters per 1000 kcal (or 4180 kJ) of heat rejection. Thus, mathematically, the percentage of water lost due to evaporation can be written as _ p D t p =m _ w;i 100 E bm where b 0:0018 21

Similarly, Berkeley [19] explained that evaporation loss is approximately 0.75 pounds (0.34 kg) for every 1000 Btu (or 1055 kJ) of cooling duty. Therefore, in this case, the percentage evaporation loss can be expressed as _ w;i 100 _ p cp;p Dtp =1055=m E 0:34m 22

the ambient condition, the process temperatures, the unit specic data etc. Therefore, only a simplied calculation (demonstrated by means of an example in the appendix) can be used. For this purpose, a psychometric diagram is needed to determine air conditions and assumptions have to be made for the unknown variables. Even though this is still a simplied calculation, it gives a more accurate result than the rule of thumb [21]. In the present work, it is proposed that b in Eq. (21) is a function of ambient air conditions (i.e., b = f(/i, tdb,i)). For this purpose, numerical data regarding evaporation loss was generated for the following set of input data for an evaporative uid cooler given in Mizushina and Miyashita _ a 0:1666 kga =s [10]: tp,i = 50 C, A = 1.915 m2, Le = 1, m _ p 0:325 kgw =s. The water-to-air mass ow ratio is and m an important factor and it aects all aspects of the cooler performance. In this particular case, the data was generated for a ow ratio of 2. Also, depending on the location, the ambient air can vary from very dry to humid conditions. Therefore, the relative humidity of the incoming air was varied from 0.1 to 0.7 for an inlet dry-bulb temperature range of 2050 C. After the evaporation loss data had been generated, instead of keeping b constant, as in Eq. (21), it was chosen in each case so that the error in prediction of evaporation loss did not exceed 0.03%. Then the possibility of accurate prediction of evaporation loss using only one factor was considered taking into account the inlet dry-bulb temperature and inlet relative humidity one by one where the equations were tted in the form of a polynomial of the second degree. To generate the polynomial based on only the inlet dry-bulb temperature, the average value of b, for the investigated range of the inlet relative humidity (i.e. 0.10.7), was used at each dry-bulb temperature considered. The polynomial based on only the inlet relative humidity was found in a similar manner. Both polynomial regression equations are given below:
6 b 2:85 107 t2 db;i 3:97 10 t db;i 0:00141

23a 23b

b 5:28 10

/2 i

4:19 10 /i 0:00169

It is explained in the operating and maintenance manual of Baltimore Aircoil [20] that the rule of thumb (refer to Eq. (21)) is a rule which can be used worldwide, but only for indicative purposes. Furthermore, calculating it with greater accuracy depends on various parameters such as

The coecient of determination (R2) was found to be 99.83% and 100%, respectively. Furthermore, using the same data, a regression equation based on both factors was also found and is given below: b 113 8:417/i 1:6147tdb;i 105 24

B.A. Qureshi, S.M. Zubair / Applied Thermal Engineering 27 (2007) 520527

525

The coecient of determination (R2), in this case, was 96.18%. To ascertain the validity of the above equations, three random sample problems were selected from the literature [8,15,16] and their results compared. The error range in evaporation loss prediction for Eq. (23b) was approximately (10.235.8%). Still, it was noticed that Eq. (23a) and (24) were equally accurate in two cases (percentage error less than 1%) but for the sample problem of Finlay and Harris [15], where a combination of very low dry-bulb temperature (10 C) and high relative humidity (82%) was experienced at the inlet, Eqs. (23a) and (24) produced an error of 17.4% and 2.4%, respectively. Therefore, it was determined that both factors were required for an accurate evaluation of b and it was used for comparison with the other simple methods as detailed above. Figs. 4a4f are drawn for the same set of input data described above with the addition of data generated for water to air ow ratio of 0.5 and 1 as well. Here, Eq. (24) was reused to determine b and the results were compared with Eqs. (21) and (22) as well as the simple calculation procedure. It was seen that the error range in evaporation loss prediction found was approximately (7.7 to 27%), (4.9 to 31%), (3.7 to 4.1%), and (12 to 6.7%) for Eqs. (21), (22), (24) and the simple calculation procedure, respectively. For Eq. (21), the percentage error is greater than 22% at low inlet dry-bulb temperatures but decreases at higher values of the temperature. On the other hand, for Eq. (22), the percentage error is found to be approximately 30% at high dry-bulb temperatures but decreases at lower values. However, Eq. (24) shows greater error for a combination of high dry-bulb temperature and high relative humidities with a maximum error of 4.1% but it is less than 2% in most cases. Furthermore, the simple calculation procedure exhibited a maximum error of 12% for some extreme cases (combination of high

mratio = 2, Range = 5.1 - 9.37 C

tdb,i = 20 C
1. 5

E (%)

Model 0. 5 Ours BAC Graham BAC(ROT) 0 0. 1 0.2 0. 3 0.4 0. 5 0.6 0. 7

tdb,i = 40 C

Relative humidity, i
Fig. 4b. For mass ow ratio of 2 and dry-bulb temperature of 40 and 20 C.
3

mratio = 1, Range = 2.26 - 8.24 C


2. 5

tdb,i = 30 C

E (%)

1. 5 Model Ours 1 BAC Graham BAC(ROT) 0. 5 0. 1 0.2 0. 3 0.4 0. 5 0.6 0. 7

tdb,i = 50 C

Relative humidity, i
Fig. 4c. For mass ow ratio of 1 and dry-bulb temperature of 50 and 30 C.
3. 5

mratio = 2, Range = 2.39 - 8.72 C


o

mratio = 1, Range = 4.82 - 8.86 C

1. 5

tdb,i = 30 C
2. 5

tdb,i = 20 C

E (%)

Model 0. 5 Ours BAC Graham BAC(ROT) 0 0. 1 0.2 0. 3 0.4 0. 5 0.6 0. 7


0. 5 0. 1
o

E (%)

1. 5

Model Ours BAC

tdb,i = 50 C

Graham BAC(ROT)

tdb,i = 40 C

Relative humidity, i

0.2

0. 3

0.4

0. 5

0.6

0. 7

Relative humidity, i
Fig. 4d. For mass ow ratio of 1 and dry-bulb temperature of 40 and 20 C.

Fig. 4a. Comparison of evaporation prediction for dierent operating conditions. For mass ow ratio of 2 and dry-bulb temperature of 50 and 30 C.

526

B.A. Qureshi, S.M. Zubair / Applied Thermal Engineering 27 (2007) 520527

5. 5

mratio = 0.5, Range = 2.12 - 7.74 C

tdb,i = 30 C
4. 5

E (%)

3. 5

not only does Eq. (24) predict much better than the rules of thumb of Baltimore Aircoil and Graham Manufacturing Company, it is just as good as the simple calculation procedure (See Table 2) giving it the added advantage of requiring only a calculator as well as a very small amount of time. In general, we nd that Eq. (24) oers an overall advantage compared to other methods currently used. 5. Concluding remarks Based on the rule of thumb provided by Baltimore Aircoil, the empirical relation to predict evaporation losses is further improved, by inserting a variable constant in a functional form. It is very simple and reliable as it only requires the use of a calculator at the site and is valid for a wide range of operating conditions. The predicted values obtained from this relation are compared with numerical data calculated from an accurate model of evaporative uid coolers that is solved by using Engineering Equation Solver (EES) program. We nd that the predictions are in excellent agreement. For accurate prediction, the range of application should be limited so that the smallest cooling range is greater than 2 C, the maximum value of the inlet air relative humidity is 0.7, the dry-bulb temperature of the incoming air is between 10 and 50 C and the water-to-air mass ow ratio is between 0.5 and 2. Appendix A. Heat and mass transfer correlations

2. 5

Model Ours BAC Graham BAC(ROT)

1. 5

tdb,i = 50 C

0. 5 0. 1 0.2 0. 3 0.4 0. 5 0.6 0. 7

Relative humidity, i
Fig. 4e. For mass ow ratio of 0.5 and dry-bulb temperature of 50 and 30 C.

6. 5

mratio = 0.5, Range = 4.53 - 8.33 C


5. 5

tdb,i = 20 C

4. 5

E (%)
3. 5 Model Ours 2. 5 BAC Graham BAC(ROT) 1. 5 0. 1 0.2 0. 3 0.4 0. 5 0.6 0. 7

tdb,i = 40 C

The lm heat transfer coecient from outside of the tubes in a counter ow horizontal tube evaporative cooler are obtained from Mizushina et al. [10], which is also recommended by Dreyer [8]. It is given by hc;w 2102:9C=d t;os with 0:195 < C=d t;os < 5:556 where _ w;i d t;os =2ntr P t L C m
_ 1=3

A:1 A:2

Relative humidity, i
Fig. 4f. For mass ow ratio of 0.5 and dry-bulb temperature of 40 and 20 C.

dry-bulb temperature and low water-to-air ow ratio); otherwise, it is much less in most cases. We notice from Figs. 4a4f that the investigated cooling range is between 2.1 and 9.4 C. It is important to note that the cooling range given in these gures is a result of our analysis, i.e. it depends on the state of ambient air at the inlet to the evaporative cooler. For further comparison, the same three sample problems mentioned above were used and it was noted that

A:3

It is important to note that the correlation given by Mizushina et al. [10] was obtained from test data using tube diameters of 12.7 mm, 19.05 mm and 40.0 mm. The following volumetric correlation for the mass transfer coecient was determined to t the test data of Mizushina et al. [10]: hD a 5:027 108 Rea Rew
0:9 0:15

d t;os

2:6

A:4

Table 2 Examples of numerical and predicted values of percentage evaporation loss Ref. 8 15 16 Model 1.578 0.354 1.732 BAC 1.904 0.534 2.100 ErrB1 (%) 20.60 50.831 20.857 BAC-rened 1.492 0.353 1.743 ErrB2 (%) 5.493 0.335 0.346 Graham 1.426 0.276 1.474 ErrG (%) 9.671 21.79 15.13 Eq. (24) 1.569 0.363 1.739 Err24 (%) 0.633 2.448 0.111

B.A. Qureshi, S.M. Zubair / Applied Thermal Engineering 27 (2007) 520527

527

The interfacial area per unit volume of a tube bundle in a (2 dt,os) array can be expressed as a pd t;os p 0:9069 p p d t;os 2 3d t;os 2d t;os 3d t;os A:5

Step 4: Calculate evaporation: (46.182 30.037) * 31.7 * 1.042 = 533.3 g/s.

References
[1] R.O. Parker, R.E. Treybal, The heat, mass transfer characteristics of evaporative coolers, AIChE Chemical Engineering Progress Symposium Series 57 (32) (1961) 138149. [2] T. Mizushina, R. Ito, H. Miyashita, Characteristics and methods of thermal design of evaporative coolers, International Chemical Engineering 8 (3) (1968) 532538. [3] I.C. Finlay, W.D. Grant, The accuracy of some simple methods of rating evaporative coolers, Report No. 584, National Engineering Laboratory, East Kilbride, Glasgow, 1974. [4] S. Tezuka, T. Takada, S. Kasai, Performance of an evaporative cooler, Heat Transfer-Japanese Research 6 (1) (1976) 118. [5] I.C. Finlay, W.D. Grant, Air coolers, cooling towers and evaporative coolers, Report No. 534, National Engineering Laboratory East Kilbride, Glasgow, 1972, pp. 165328. [6] R.L. Webb, A unied theoretical treatment for thermal analysis of cooling towers, evaporative condensers, and uid coolers, ASHRAE Transactions 90 (Part 2B) (1984) 398415. [7] R.L. Webb, A. Villacres, Algorithms for performance simulation of cooling towers, evaporative condensers, and uid coolers, ASHRAE Transactions 90 (Part 2B) (1984) 416458. [8] A.A. Dreyer, Analysis of evaporative coolers and condenser, MSc Thesis, University of the Stellenbosch, Rep. of South Africa, 1988. [9] M. Poppe, H. Ro gener, Evaporative Cooling Systems, VDI Warmeatlas, Section Mh, 1984. [10] T. Mizushina, R. Ito, H. Miyashita, Experimental study of an evaporative cooler, International Chemical Engineering 7 (4) (1967) 727732. [11] B.A. Qureshi, S.M. Zubair, The impact of fouling on performance evaluation of evaporative coolers and condensers, International Journal of Energy Research 29 (14) (2005) 13131330. [12] B.A. Qureshi, S.M. Zubair, A comprehensive design and rating analysis of evaporative coolers and condensers: Part I; performance evaluation, International Journal of Refrigeration 29 (4) (2006) 645 658. [13] S.A. Klein, F.L. Alvardo, EES-Engineering Equation Solver. Users manual for Microsoft Windows Operating Systems, version 7.155, FChart Software, Madison, WI, 2001. [14] J.Y. Jang, Z.J. Wang, Heat and mass transfer performances of closedtype cooling towers, in: Proceedings of the International Conference on Advanced in Computational Heat Transfer, Palm Cove, Queensland, Australia, May 2025th, 2001, pp. 269276. [15] I.C. Finlay, D. Harris, Evaporative cooling of tube banks, International Journal of Refrigeration 7 (4) (1984) 214224. [16] P.J. Erens, Comparison of some design choices for evaporative cooler cores, Heat Transfer Engineering 9 (2) (1988) 2935. [17] Available from: <www.BaltimoreAircoil.com>. [18] Available from: <www.graham-mfg.com>. [19] F.D. Berkeley, Now Consider Evaporative Cooling, Graham Manufacturing Co., Inc., Batavia, N.Y., 1961. [20] Operating and Maintenance Instructions, Baltimore Aircoil, Bulletin M308/39. [21] Personal correspondence with Georges Hoeterickx of Baltimore Aircoil. [22] W.M. Kays, Numerical solution for laminar ow heat transfer in circular tubes, Transactions of ASME 77 (1955) 12651274.

The mass transfer coecient can now be written as hD 5:544 108 Rea Rew
0 :9 0:15

d t;os

1:6

A:6

Kays [22] determined the heat transfer coecient during laminar ow (Re < 2300) inside a duct with a constant wall temperature. Nup 3:66 0:104Rep Prp d t;in =L 1 0:016Rep Prp d t;in =L
0:8

A:7

Dreyer [8] used the equation below for the heat transfer coecient for a ow inside of a tube in a turbulent ow regime. Nup fD =8Rep 1000Prp 1 d t;in =L
:67 1 1 12:7fD =80:5 Pr0 p 0:67

A:8

where the friction factor fD for smooth tubes was dened by fD 1:82 log10 Rep 1:64
2

A:9

Eq. (A.8) is valid for the following ranges: 2300 < Rep < 106 ; 0:5 < Prp < 104 ; 0 < d t;in =L < 1 Appendix B. Simplied calculation procedure Consider Baltimore unit FXV 542-L with airow of 31.7 m3/s. It is selected to cool 1000 kW at an ambient temperature of 50 C dry bulb and 35 C wet bulb. Step 1: Determine inlet air parameters: Air inlet temperature (tdb,i/twb,i) (C): 50/35. Air inlet density (kg/m3): 1.042. Air inlet enthalpy (kJ/kg): 128.13. Air inlet water content (g/kg): 30.037. Step 2: Calculate enthalpy pick-up: Pick-up = kW/(airow * density). Pick-up = 1000/(31.7 * 1.042) = 30.274 kJ/kg. Step 3: Determine discharge air condition: Discharge air enthalpy = inlet enthalpy + pickup = 158.4 kJ/kg. Assume 100% RH. Determine water content = 46.182 g/kg.

S-ar putea să vă placă și