Performance characteristics of a trickle ﬁll in cross- and counter-ﬂow

7/22/2019 Performance characteristics of a trickle fill in cross- and counter-flow

1/10

Performance characteristics of a trickle fill in cross- and counter-flow

configuration in a wet-cooling tower

P.J. Grobbelaar 1, H.C.R. Reuter*, T.P. Bertrand 2

Department of Mechanical Engineering, Stellenbosch University, Private Bag X1, Matieland 7602, South Africa

h i g h l i g h t s

< Fill tests are performed on a trickle fill to accurately determine anisotropic fill behavior.

< Fill performance is found to have a strong dependency on fill configuration.

< High ratios of Gw/Ga favor crossflow, while low ratios favor counterflow.

< In counterflow trickle fill behaves similar to film fill.

< From test results, conclusions about heat transfer and fluid mechanics in the fill are made.

a r t i c l e i n f o

Article history:

Received 12 October 2011

Accepted 12 June 2012

Available online 21 June 2012

Keywords:

Cooling tower

Wet-coolingCrossflow

Counterflow

Anisotropic

Fill

a b s t r a c t

In cooling towers packed with trickle or splash fills, which have almost isotropic or anisotropic flow

resistance, the air flow through the fill is oblique or in cross-counterflow to the waterflow, particularly at

the cooling tower inlet when the fill loss coefficient is small or when the fill hangs down into the air inlet

region. This results that the fill Merkel number or transfer characteristic for cross-counterflow is between

that of purely counter- and cross-flow fills.

When using CFD to model natural draught wet-cooling tower performance for isotropic or anisotropic

fill resistance, two- or three-dimensional models and fill characteristics are therefore required to

determine overall fill performance.

In this paper, the test facility, measurement techniques and methods of analysis used to determine fill

performance characteristics in counter- and cross-flow configuration are presented and discussed.

Results obtained for a specific fill are presented and discussed which can be used for the evaluation of

cross-counterflow fill performance.

2012 Elsevier Ltd. All rights reserved.

1. Introduction

Wet-cooling towers are designed using performance models

that employ the e-NTU, Poppe and Merkel methods of analysis [1].

These are generally applied to either crossflow or counterflowcases. However, cross-counterflow applications of these models

have not been presented and verified in literature. The accuracy of

wet-cooling prediction and analysis may be significantly improved

if a method that can also be applied to cross-counterflow zones is

used.

The modeling of cross-counterflow in wet-cooling is compli-

cated by the fact that most types of fill are anisotropic, i.e. their

performance characteristics in crossflow are not the same as in

counterflow. In cross-counterflow their characteristics are some-

where between their crossflow and counterflow characteristics.In the last decade, numerous researchers [2e9] have published

work on using CFD to model coolingtowers,but this work is limited

to orthotropic fills, i.e. fillsthat are only porous in a single direction,

and can still not take anisotropic fill behavior into account.

Reuter [10] recently developed a method to analyze and predict

wet-cooling within such cross-counterflow zones that may be

integrated into CFD analyses. This method takes anisotropic fill

behavior into account by making use of a linear interpolation

between the cross- and counter-flow fill characteristics. However,

the use of this model (for cooling prediction purposes) requires

prior knowledge of the cross- and counter-flow fill characteristics.

* Corresponding author. Tel.: 27 21 808 4261 (O), 27 72 724 9819 (mobile);

fax: 27 21 808 4958.

E-mail addresses: [email protected] (P.J. Grobbelaar), [email protected]

(H.C.R. Reuter), [email protected] (T.P. Bertrand).1 Tel.: 27 72 2787 682.2 Tel.: 27 84 55 222 00 (mobile).

Contents lists available at SciVerse 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

1359-4311/$ e see front matter 2012 Elsevier Ltd. All rights reserved.

http://dx.doi.org/10.1016/j.applthermaleng.2012.06.026

Applied Thermal Engineering 50 (2013) 475e484
mailto:[email protected]:[email protected]:[email protected]://www.sciencedirect.com/science/journal/13594311http://www.elsevier.com/locate/apthermenghttp://dx.doi.org/10.1016/j.applthermaleng.2012.06.026http://dx.doi.org/10.1016/j.applthermaleng.2012.06.026http://dx.doi.org/10.1016/j.applthermaleng.2012.06.026http://dx.doi.org/10.1016/j.applthermaleng.2012.06.026http://dx.doi.org/10.1016/j.applthermaleng.2012.06.026http://dx.doi.org/10.1016/j.applthermaleng.2012.06.026http://www.elsevier.com/locate/apthermenghttp://www.sciencedirect.com/science/journal/13594311http://crossmark.dyndns.org/dialog/?doi=10.1016/j.applthermaleng.2012.06.026&domain=pdfmailto:[email protected]:[email protected]:[email protected]


2/10

Due to the complexity of the interaction between the fill

material, the water and the air that occurs within fill, fill perfor-

mance characteristics have up to date invariably been determined

through experimental testing.

Cooling tower fills and their performance characteristics have

been the focus of researchers such as Khan, Qureshi and Zubair

[11,12] and Gharagheizi, Hayati and Fatemi [13]. Krger [[14]:256]

also lists several literature sources [15e20] on the performance

characteristics of cooling tower fills. However, none of the afore-mentioned studies provide an experimental comparison between

the crossflow and counterflow performance characteristics offills,

although De Villiers [21] performed such a comparison of cooling

tower rain zones.

To make such a comparison forfill materials, test facilities where

performance comparablefill tests could be done would be required.

Such facilities are available at the University of Stellenbosch and are

currently being used to determine 2-dimensional fill performance

characteristics.

This text describes the facilities and methodology with which

these tests are done and shows the results of a comparative fill test

done with a commercial trickle fill, shown in Fig. 1.

2. Fill performance characteristics

To evaluate and compare thermal performance of fills, Merkel

[22] derived a non-dimensional coefficient of performance or

transfer characteristic, now known as the Merkel number Me. The

Merkel number is defined as

Me hdafiAfrLfi

mw

hdafiLfiGw

ZTwi

Two

cpwdTwimasw ima

(1)

The air flow resistance is expressed by a loss coefficient Kfdm.

In this study, the Merkel number is calculated from fill perfor-

mance test data using the e-NTU method, presented in Krger[[14]:

274]. The Merkel number is calculated as a function of the number

of transferred (heat) units (NTU), the minimum evaporative

capacity rate Ce min and the water mass flow rate mw, as follows:

Me NTUeCe min=mw (2)

To determine NTU and Ce min the system of equations, presented

in Krger [4] needs to be solved simultaneously by means of an

iterative procedure. This system of equations is given below.

ee 1 exph

NTU0:22n

exp

CeNTU0:78 1o.

Cei

(3)

ee Q=Qmax (4)

Q maimao imai (5)

Qmax Ce minimaswi l imai (6)

Nomenclature

afi interfacial surface area between air and water per unit

volume offill zone, m1

Afr frontal area offill perpendicular to air flow direction,

m2

Ce evaporative capacity rate ratio, kg/s

cp specific heat at constant pressure, J/kg K

EB energy balance, %

g gravitational acceleration, m/s2

G mass velocity, kg m2 s1

hd mass transfer coefficient (mass base), m/s

i specific enthalpy, J/kg

ima specific enthalpy (per kg dry air), J/kg

Lfi fill length, m

m mass flow rate, kg/s

NTUe number of transfer units

patm atmospheric pressure, Pa

Dpfi pressure drop over fill, Pa

Q heat, W

T temperature, C

v velocity, m/s

Greek symbols

l correction factor (pure number)

r density, kg/m3

Dimensionless groups

ee effectiveness

Me Merkel number

Kfdm loss coefficient

Subscripts

a air

abs absolute

av airevapor mixture

avg average

i inlet

ma airevapor (per kg dry air)

max maximum

min minimum

o outlet

s saturated

w water

wb wetbulbx x-direction (along length)

z z-direction (vertical)

Fig. 1. Photo of the tricklefi

ll that was used in the comparative experimental tests.

P.J. Grobbelaar et al. / Applied Thermal Engineering 50 (2013) 475e484476


3/10

l imaswo imaswi 2imasw=4 (7)

Ce min min

mwcpw=

dimaswdTw

; ma

and Ce max

max

mwcpw=

dimaswdTw

; ma

(8)

dimaswdTw

z

imasw imaswoTwi Two

(9)

Ce Ce min=Ce max (10)

where the thermophysical properties are calculated according to

the property relations by Krger [4].

The loss coefficient is calculated as follows (Krger [[14]: 268]):

Kfdm 2hDpfi

ravov

2avo raviv

2avi

ravi

rav; avg

gLfi;z

irav; avgA

2fr=m

2av; avg (11)

For easy comparison to other fills, the fill Merkel number and

loss coefficient are divided by the fill length. The total Merkel

number is divided by the fill length in the water flow direction to

obtain the Merkel number per meter offill, while the loss coeffi-

cient is divided by fill length in the direction of the air flow to

obtain the loss coefficient per meter offill.

Fill characteristics data can be correlated by means of the

following functions:

Me=Lfi; z aGbwG

caT

dwi (12)

Kfdm=Lfi; x eGfwG

ga hG

iwG

ja (13)

An optimizer is used to minimize the least squares error

between the experimental data and the predictions made by

Equations (12) and (13) by changing the values of the constants a, b,

c, d, e, f, g, h, i and j.

3. Description of test facility

The experimental test facility is shown schematically in Fig. 2.

The facility is designed based on the standards of CTI [23].

Air is drawn into a 2 m 2 m wind tunnel inlet (1) by

a centrifugal fan (2). The fan is driven by a 50 kW electric motor (3),

which is connected to a variable frequency drive allowing for

controllable air speed through the tunnel. After entering the tunnel,

the air flows through the crossflow fill test section (4) and then

through mixing vanes (5) and a settling screen (6).

Four pairs of thermocouples in aspirated psychrometers (7)

measure the dry- and wet-bulb air temperatures downstream of

the settling screen. Insulation with a 100 mm thickness (8) and

a roof (9) shield the facility from convection and solar radiation

heat transfer to ensure accurate temperature measurements.

The air then flows through ASHRAE 51-75 [24] elliptical nozzles

located in a nozzle plate (10). The pressure drop measured over

these nozzles is used to calculate the airevapor massflow rate, mav,

through the wind tunnel. Turning vanes (11) guide the air around

90 bends into the counterflow fill test section (12).

The water that is used during a test is stored in an underground

storage tank that has a capacity of 45 m 3. Before a test, the water is

heated to the desired temperature (usually about 50 C) by circu-

lating it through a 100 kW diesel-fired boiler.

More detail of the respective test sections is provided in Sections

3.1 and 3.2.

3.1. Crossflow test section

The crossflow test section is shown schematically in Fig. 3.

Two supply pumps (1) pump hot water from the underground

storage tank to the crossflow test section, where it is distributed

evenly over the fill zone by a water distribution spray frame (2). On

its way to the test section, the water passes through a strainer (3)

and a cartridge filter (4). Any air that may have entered the water

supply line escapes through bleed valves (5) on the water distri-

bution spray frame.

In the water supply line, the water flow rate is controlled using

a valve (6) and measured using an electromagnetic flow meter (7).

This measurement is verified by measuring the pressure drop

across an orifice plate (8) installed in the supply line. The water

inlet temperature is measured by three thermocouples (9) just

before the water enters the water distribution spray frame.

Immediately upstream of the control valve at (6), there is a T-

junction through which water flow may be diverted to the coun-terflow section instead of the crossflow section by closing valve (6)

and opening valve (10).

After the water has passed through the crossflow test section fill

zone, it is collected in the two water catchment basins (11 and 12).

A pump (13) drains the collected water through a strainer (14) and

outlet pipe (15) and pumps it back to the storage tank. A pipe joins

with the crossflow outlet pipe at a T-junction immediately down-

stream of the drain pump (13). This pipe is the counterflow outlet

pipe and it may be sealed by a closing valve (16). The crossflow test

sections return waterflow rate is controlled by a valve (17). Vertical

plates (18) impede circumventive air flow through the main water

Fig. 2. Diagram of thefi

ll test facility the University of Stellenbosch.

P.J. Grobbelaar et al. / Applied Thermal Engineering 50 (2013) 475e484 477


4/10

catchment basin, forcing the air to pass through the fill zone

instead.

The water outlet temperature is measured at three stages of

mixing: in the mixing duct (19), the primary catchment basin (20)

and in a u-tube in the outlet pipe (21). Each of these stage

temperatures is measured by a set of two thermocouples.

Drift eliminators (22) are installed immediately downwind of

the fill zone to reduce the drift losses.

The airdry bulb and wet bulb temperatures at both the inlet and

outlet are measured by aspirated psychrometer stations (23 and 24

respectively). Such a station is visible on the foreground of Fig. 4.

Before reaching the outlet station, the air passes through air-mixing

vanes (25) and a settling screen (26).

Sensor rakes are used to measure the temperature profiles of

both the outlet air and water. The outlet air rake (27) is also an

aspirated psychrometer station consisting of 8 pairs of thermo-

couples, while the water outlet rake (28) consists of 13 thermo-

couples. All the three psychrometer stations are aspirated by

centrifugal fans that are installed underneath the tunnel (29, 30

and 31).

Pressure probes are fixed in front of (32) and behind (33) the fill

zone to measure the pressure drop over thefi

ll. The probes are

Fig. 3. Diagram of the crossflow test section, (a) side view section A-A, (b) top view section B-B.



5/10

connected by tubes to a Betz manometer inside the lab. Once test

conditions have stabilized, the reading from the manometer is

taken and recorded by hand. Atmospheric pressure is read off

a mercury barometer.

All thermocouples are of the coppereconstantan type and are

calibrated (at various temperatures) against a certified platinum

thermocouple.

The 53 thermocouples, the electromagnetic waterflowmeter and

the nozzle- and orifice-plate pressure transducers, are all connected

toan Agilentdata logger, which is in turn connectedto a PCthroughan

USBcable.On thePC, LabViewsoftwareis installed through which the

sensors are monitored in real-time. Once test conditions have stabi-

lized, all monitored test data are written and saved to disc. The saved

data files are processed using Python 2.7 scripts and MS Excel.

To check the quality of the results, an energy balance EB is

calculated as

EB Qa Qw

Qw 100% (14)

where Qw is the heat extracted from the water stream and Qa is the

heatabsorbed by air stream, both calculated using the measured inlet

and outlet temperatures and mass flows. According to the conserva-

tion of energy, the energy balance should ideally be equal to zero.

Table 1 shows a summary of the recorded data (excluding air

and water outlet temperature profiles), after averaging, from

a single crossfl

ow test. The derived quantities from the same testare shown in Table 2.

3.2. Counterflow test section

A photo of the counterflow test section (12) is shown in Fig. 5

and the details of the counterflow test section are shown sche-

matically in Fig. 6.

The counterflow fill test section has a 1.5 m 1.5 m cross-

section. The fill height can be varied up to 5 m. After the air,coming from the wind tunnel as shown in Fig. 2, has entered the

counterflow test section (12), the dry- and wet-bulb temperatures

of the air just below the fill (13) are measured. The air flows

through the water extraction troughs (22), the fill (14) and ulti-

mately the drift eliminators (15). The drift eliminators are installed

to reducethe amountof waterlost by drift losses. Thepressure drop

across the troughs and the fill is measured using two independent,

calibrated, electronic pressure transducers (16).

The first part of the water supply line to the counterflow test

section is described in Section 3.1.

The same electromagnetic flow meter that is used for crossflow

tests is used to measure the flow rate of the hot water before it is

pumped into the counterflow test section (17). Thermocouples (19)

measure the temperature of the water before it is sprayed onto thetop of the fill via a water distribution spray frame (20) through

a spray zone (21). A cartridge filter (18) is installed in-line imme-

diately upstream of the aforementioned thermocouples. The

temperature of the water, after it has traveled through the spray

zone, is measured again to ensurethat the temperature of the water

(22) entering the fill is known. The water falls through the fill (14),

is cooled, and the temperature of the water leaving the fill is

measured (23). The water leaving the fill is collected by two layers

of water extraction troughs rotated 90 to each other (24).

The outlet water temperature is again measured in the top and

bottom troughs (25) as well as in the pipe work draining the top

(26) and bottom(27)troughs. The water is collected in a sump from

where it is pumped back to the underground storage tank (28).

Fig. 4. Photo of the empty fill zone in the crossflow test section.

Table 1

Measured inputs during a crossflow test.

TaiC Twbi

C Twi C Two

C mavo kg/s mwkg/s

Dpfi Pa TwboC patm Pa

13.1 10.5 34.0 17.4 11.4 4.3 74 20.1 100 900

Table 2

Derived quantities from a crossflow test.

EB % Gw kg/m2 s Ga kg/m2 s Me/Lfi m1 Kfdm/Lfi m

1

0.1 1.4 2.8 1.16 10.9

Fig. 5. Photo of the counterfl

ow test section.



6/10

The Merkel number for the fill and troughs is calculated using

the inlet and outlet water temperatures (22) and (25) respectively.

The Merkel number for the troughs, calculated using the inlet (23)

and outlet (25) water temperatures, is subtracted from the total

Merkel number to obtain the reported Merkel number which

represents the Merkel number of the fill only.

An Agilent data logger is connected to a PC and is used to

monitor test conditions in real-time through a LabVIEW program

on the PC. Once the test conditions have stabilized, all the moni-

tored test data is written and saved to file.

Table 3 shows a summary of the recorded data from a single

counterflow test, while the derived quantities from the same test

are shown in Table 4.

3.3. Instrumentation errors

In order to determine the accuracy and reliability of the

measured data, both the crossflow and counterflow test sections

were critically evaluated by respectively Grobbelaar [25] and Ber-

trand [26]. The measurement uncertainties that were determined

for both the crossflow and the counterflow test sections are given

Table 5.

Note that the uncertainties reported in the Table 5 tend to vary

slightly with water and air mass velocity. The standard deviation of

the outlet water temperature in the counterflow test facility alsohas a strong dependence on the measure of redistribution of water

by the fill the given value of 0.75 K applies (approximately) to the

tested trickle fill.

The uncertainty in the Twbi for the counterflow facility has not

been determined, but since it is measured in a way identical to the

Twbi for the crossflow facility, the uncertainty in it is expected to be

of similar magnitude.

Grobbelaar [25] also did tests to determine the uniformity of the

water distribution at the inlet of the crossflow test facility, and it

was found that, depending on the total water mass flow, the

standard deviation of the local water mass velocity in the test area

was equal to 9e12% of the average mass velocity.

Bertrand [26] did similar tests on the counterflow test facility

and found that the water distribution had, on average a Christian-

sen coefficient of 0.95.

Grobbelaar [25] performs an uncertainty analyses for the

crossflow facility, in which he shows that the variance in (cross-

flow) Me/Lfi and Kfdm/Lfi as a function of the variances in the

experimental measurements. These sensitivities are given in

Table 6.

Grobbelaar [25] shows that, when the sensitivities and the

measurement uncertainties in the crossflow test section are

considered together, the (conservative) uncertainties in the deter-

mined Me/Lfi and Kfdm/Lfi are respectively 5.4% and 5.2%.

4. Results

4.1. Fill configurations tested

The trickle fill that was used in the comparative experimental

tests has cross-fluted channels to facilitate air flow in a specific

direction. Normally, the fill is installed so that these channels align

with the expected direction of air flow, i.e. when installed in

counterflow, these channels will be vertically orientated.

Table 3

Measured inputs during a counterflow test.

Tai C Twbi

C TwiC Two

C mavo kg/s mw kg/s Dpfi Pa patm Pa

18.2 15.0 48.8 31.7 2.3 3.4 9.5 100 940

Table 4

Derived quantities from a counterflow test.

Gw kg/m2 s Ga kg/m

2 s Mee-NTU/Lfi m1 Kfdm/Lfi m

1

1.5 1.0 0.78 12.7

Fig. 6. Diagram of 1.5 m 1.5 m counterflow test section.

Table 6

Me/Lfi and Kfdm/Lfi sensitivities to variances in the experimentally measured

quantities.

Measurement Twbi Twi Two mavo mw Dpfi Twbo

Unit of Sensitivity %/K %/K %/K %/% %/% %/Pa %/K

Me/Lfi Sen sit ivi ty 13.4 16.9 47.8 1.6 1.6 0.0 0.3

Kfdm/Lfi Sensitivity 0.1 1.2 1.2 1.9 0.1 2.5 0.0

Table 5

Measurement uncertainties, in terms of standard deviation, determined for the

crossflow test section.

Twbi Twi Two mavo mw Dpfi Twbo

Crossflow:

avg abs dev

0.047 K 0.071 K 0.048 K e e e 0.104 K

Crossflow:

uncertainty

interval

e e e 2.70% of

measured

massfl

ow

0.97% of

measured

massfl

ow

0.5 Pa e

Counterflow:

standard

deviation

e 0.02 K 0.75 K e e e e

Counterflow:

uncertainty

interval

e e e 2.70% of

measured

mass flow

1.25% of

measured

mass flow

0.63 Pa e



7/10

In the comparative experimental tests, three fill configurations

were tested:

1. In the counterflow test facility with the fill orientated as it

would normally be in counterflow (channels vertical).

2. In the crossflow test facility with the fill orientated as it would

normally be in crossflow (channels horizontal).3. In the crossflow test facility with fill orientated as it would nor-

mally be in counterflow (channels vertical). This configuration is

henceforth referred to as Crossflow (counterflow config.).

These three configurations are shown schematically in Fig. 7.

4.2. Tests done

The characteristic equations that are determined for the fill are

based on data from multiple fill tests. This subsection briefly

describes the scope of the data on which the presented fill char-

acteristic equations are based.

A single testis defined as a period of about 1 min during which

all data from the experimental facility was recorded. During thistime, the water mass velocity Gw, air mass velocity Ga and water

inlet temperature Twi remained virtually constant and there was

almost no change in the outlet conditions.

A test series is defined as a group of tests that were done one

shortly after the other. Within a test series all possible combina-

tions of several predefined air- and water-mass velocities are tested

together in respective tests, e.g. 3 predefined water mass velocities

and 4 predefined air mass velocities make for 3 4 12 tests

within the test series. Water inlet temperature decreases

throughout the tests in the same order that the tests were done in.

Table 7 shows data recorded from such a test series of 12 tests.

The amount of experimental tests on which the various fill

performance characteristic equations are based is given below:

Crossflow:

3 Test series with 4 water mass velocities and 4 air mass

velocities

Total 48 testsCounterflow:


velocities

Total 84 tests

Crossflow (counterflow config.):


velocities 48 tests


velocities 16 tests

Total 64 tests

4.3. Fill performance characteristics

The fill performance characteristic relations that were deter-

mined for the three fill configurations, using the method described

in Section 2, are given in Tables 8 and 9.

Fig. 7. The three fill configurations that were tested.

Table 7

Example of experimental data that is recorded

during a test series in the crossflow test facility.

Gwa

kg/m2 s

Gaa

kg/m2 s

TwiC Tai

C Twbi C Two

C Dpfi Pa TwboC patm Pa

1.5 1.0 52.01 19.21 17.18 29.72 27 39.04 100 860

1.5 1.5 51.01 19.14 17.15 26.26 55 35.53 100 860

1.5 2.0 49.93 19.07 17.10 23.37 95.5 32.50 100 860

1.5 2.5 49.43 19.01 17.07 22.45 163 30.31 100 860

3.0 1.0 48.24 19.02 17.04 34.34 31 42.30 100 860

3.0 1.5 47.50 19.18 17.13 30.97 60 39.65 100 860

3.0 2.0 46.47 18.99 17.05 28.14 106 37.08 100 860

3.0 2.5 45.62 18.78 16.94 25.86 174 34.74 100 860

4.5 1.0 43.89 19.08 17.04 35.47 36 40.85 100 860

4.5 1.5 42.94 18.96 17.00 32.76 66 38.82 100 860

4.5 2.0 41.34 18.77 16.92 29.71 121 36.13 100 860

4.5 2.5 40.25 18.63 16.85 27.61 192 34.08 100 860

a

Rounded.

Table 8

Fill characteristic equations for Merkel number per meter of fill that were deter-

mined for different cross- and counter-flow fill configurations.

Fill configuration Determined fill

characteristic equations

for Me/Lfi, z

Correlation

coefficient

Crossflow Me/Lfi, z 1.2330 Gw0.7550

Ga0.3450 Twi

0.02790.987

Counterflow Me/Lfi, z 1.6293 Gw0.9250

Ga0.7760Twi0.0986

0.994

Crossflow

(counterflow config.)

Me/Lfi, z 1.5258 Gw0.7754

Ga0.7996 Twi

0.07300.983

Table 9

Fillcharacteristicequations forloss coefficientper meteroffill thatweredetermined

for different cross- and counterflow fill configurations.

Fill configurati on De termi ned fill characteristic

equation for Kfdm/Lfi

Correlation

coefficient

Crossflow Kfdm/Lfi, x 11.007 Gw0.2458

Ga0.0974 3.4886 107

Gw5.6876 Ga

6.5011

0.994

Counterflow Kfdm/Lfi, z 3.1980 Gw0.4920

Ga1.4110 7.6960 Gw

0.1100 Ga0.0910

0.982

Crossflow

(counterflow config.)

Kfdm/Lfi, x 29.0167 Gw0.1332

Ga0.0774 2.9590 107 Gw

8.9749 Ga2.0027

0.939

http://-/?-http://-/?-http://-/?-http://-/?-


8/10

The correlation coefficients, given in the last column ofTables 8

and 9, show the goodness of thefi

t. A correlation coeffi

cient of 1will indicate that the equation fits perfectly through the data.

Fig. 8 shows how the fill characteristic equations approximate

the experimental data.

The fill characteristic equations are plotted on comparative

graphs in Figs. 9 through 11. It is clear from the characteristic

equations in Table 8 that the Merkel number is a very weak function

of water inlet temperature. All graphs are therefore plotted only for

a single water inlet temperature, namely Twi 45C.

Fig. 12 shows comparisons between the relative performance of

the three fill configurations, where the relative fill performance is

the Merkel number per meter offill divided by the loss coefficient

per meter offill.

5. Discussion of results

The obtained Me/Lfi and Kfdm/Lfi are of comparable magnitude to

those obtained by Krger [[14]:267] for an expanded metal

(stainless steel) fill, which has a Me/Lfi of 0.3646 m1 and Kfdm/Lfi of

1.8365 m1. Since Krgers fill is significantly more porous than the

tested tricklefill, it is expected to have both a lower Me/Lfi and Kfdm/

Lfi than the tested trickle fill. This is indeed the case.

From the characteristic equations it can be seen the Merkel

number is a very weak function of water inlet temperature for all

the tested fill configurations.

Figs. 9 through 12 reveal that the fills behavior is highly

dependent on the relative angle between the water and the airthrough it.

When the cross-fluted channels arenot aligned with the airflow

direction, the loss coefficient per meter of fill doubles (approxi-

mately) and a 25% (approximately) increase in Merkel number is

observed. The higher loss coefficient is expected because of the

aforementioned misalignment and it is speculated that the increase

in heat transfer is due to higher turbulence that is associated with

the steeper pressure gradient across the fill.

From Fig. 9 it is observed that this particular fill displays a higher

Merkel number in counterflow than in crossflow for the majority of

air- and water-mass velocities. It is believed that this is because, in

counterflow, the water flow is retarded by the air flow, with the

result that a given water particle spends a longer time in the cooling

zone than it would in crossflow. This theory is further supported by

the observation that crossflow performance relative to counterflow

increases as the ratio of water mass velocity to air mass velocity

increases, with crossflow even slightly outperforming counterflow

at the highest (Gw/Ga) ratios. This can be observed best on Fig. 12.

In this particular fill, the loss coefficient for counterflow is

always lower than for crossflow. This is not due to the natural draft

effect that is only present in counterflow: that is already taken into

account when the loss coefficient is calculated (the term

(ravi ravm) g Lfi). Rather, it is speculated that the cause for this

effect lies in the microflow pattern of the water: In counterflow, the

Fig. 8. Example of experimental data from crossflow tests compared to presented fill characteristic equations.

Fig. 9. Me/Lfi predicted byfi

ll characteristic equations, as a function of Ga, for different cross- and counter-fl

owfi

ll confi

gurations and Gw

s.



9/10

water flow is close to aligned to the cross-fluted channels in the fill,

and it is therefore expected that the water would mostly stick to

the perimeter of the channels, causing little obstruction of the

channel. In crossflow, the water falls closer to perpendicular to the

channels, which should mean that more of the water drips through

the channels. These falling drops would partly obstruct flow

through the channels, and cause a slightly higher loss coefficient.

The characteristic equations in Table 8 show that, in the coun-

terflow characteristic equation for the Merkel number, the expo-

nent of Gw is close to 1, i.e. the counterflow Merkel number is

Fig. 10. Me/Lfi predicted by fill characteristic equations, multiplied by Gw, i.e. effectively hd Afi/Lfi, as a function ofGa for different cross- and counter-flow fill configurations and Gws.

Fig. 11. Kfdm/Lfi predicted by fill characteristic equations, as a function of Ga for different cross- and counter-flow fill configurations and Gws.

Fig. 12. Relative fill performance predicted by fill characteristic equations, as a function of the ratio between Ga and Gw, for different cross- and counter-flow fill configurations and

water mass velocities.



10/10

almost inversely proportional to the water mass velocity. When we

substitute the definition of the Merkel number (per meter offill)

hd afi/(Gw Lfi) into the characteristic equation, Gw appears under the

line on both sides of the equation and may be canceled to reveal

that, in counterflow, the term hd afi is practically independent of

water mass velocity. This is illustrated in Fig.10, where the lines for

different water mass velocities in counterflow lie very close to each

other.

The independence of the term hd afi from water mass velocity is

what one would expect from a film fill, where the water flows

mainly in a film. When the water mass velocity increases, this film

merely becomes thicker and there is little change in afi. Therefore

the term hd afi remains almost constant. Evidently the tested trickle

fill, when in counterflow, behaves similar to a film fill. This supports

the previous assertion with regard to the water sticking to the

channel sides when in counterflow.

The fact that the crossflow configurations do show some

dependence, albeit weak, on water mass velocity, support the

current understanding of the flow mechanics within the fill: When

in crossflow, significant dripping does take place within the fill and

a higher water mass velocity will lead to more (and bigger) drops,

increasing afi and consequently increasing hd afi/Lfi, as is seen in

Fig. 10.The amount of heat transferred in a fill zone depends mainly on

the term hd afi, rather than on the Merkel number. Therefore, it may

be better to describe fill performance in terms of hd afi rather than

in terms of the Merkel number, especially if hd afi is given as n

characteristic equation in terms ofGa and Gw. Such equations would

plainly reveal how somefills, such as film fills, do not transfer more

heat when Gw increases, while other fills, such as splash fills, can

take some advantage out of higher water mass flow rates.

Conclusions

The observations made during this experimental study prove

that this fill is anisotropic. It also highlights the necessity to develop

models that can accurately model anisotropic fill.

The characteristic equations determined duringfill testingcapture

and reveal the anisotropicfill behavior and may therefore be used in

fill models that can take anisotropicfill behavior into account.

The experimental results and the high correlation coefficients of

the empirical equations also demonstrate the test facilities

capacity to accurately determine fill performance characteristics.

From the recorded data, interesting observations are made.

These observations increase our understanding of the heat transfer

and flow mechanics within the fill and of the variation in fill

performance characteristics between crossflow and counterflow.

Acknowledgements

This study was supported by Eskom TESP (Tertiary Education

Support Program) and by IWC (Industrial Water Cooling).

References

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

[2] R. Al-Waked, M. Behnia, CFD simulation of wet cooling towers, AppliedThermal Engineering 26 (4) (2006) 382e395.

[3] R. Al-Waked, M. Behnia, Enhancing performance of wet cooling towers,Energy Conversion and Management 48 (10) (2007) 2638e2648.

[4] N. Williamson, M. Behnia, S.W. Armfield, Thermal optimization of a naturaldraft wet cooling tower, International Journal of Energy Research 32 (14)(2008) 1349e1361.

[5] N. Williamson, M. Behnia, S. Armfield, Comparison of a 2D axisymmetric CFDmodel of a natural draft wet cooling tower and a 1D model, International

Journal of Heat and Mass Transfer 51 (9e10) (2008) 2227e2236.[6] N. Williamson, S. Armfield, M. Behnia, Numerical simulation of flow in

a natural draft wet cooling tower e the effect of radial thermofluid fields,Applied Thermal Engineering 28 (2e3) (2008) 178e189.

[7] A. Klimanek, A. Ryszard, R.A. Bia1eckia, Z. Ostrowskia, CFD two-scale model ofa wet natural draft cooling tower, Numerical Heat Transfer 57 (2) (2010).

[8] A. Klimanek, R.A. Bia1ecki, Solution of Heat and Mass Transfer in CounterflowWet-cooling Tower Fills. Institute of Thermal Technology, Silesian Universityof Technology, Konarskiego 22, 44e100 Gliwice, Poland (2009).

[9] A. Klimanek, R.A. Bia1ecki, On a numerical model of a natural draught wet-cooling tower, Archives of Thermodynamics 29 (4) (2008) 63e72.

[10] H.C.R Reuter, Performance evaluation of natural draught cooling towers withanisotropic fills, PhD thesis, University of Stellenbosch, South Africa (2010).

[11] J. Khan, B.A. Qureshi, S.M. Zubair, A Comprehensive Design and Performance

Evaluation Study of Counter Flow Wet Cooling Towers, Department ofMechanical Engineering, King Fahd University of Petroleum and Minerals,Saudi Arabia, 2004.

[12] B.A. Qureshi, S.M. Zubair, A Complete Model of Wet Cooling Towers withFouling in Fills, Department of Mechanical Engineering, King Fahd Universityof Petroleum and Minerals, Saudi Arabia, 2006.

[13] F. Gharagheizi, R. Hayati, S. Fatemi, Experimental Study on the Performance ofMechanical Cooling Tower with Two Types of Film Packing, Department ofChemical Engineering, Faculty of Engineering, University of Tehran, Iran,2006.

[14] D.G. Krger, Air-cooled Heat Exchangers and Cooling Towers e Thermal-flowPerformance Evaluation and Design, PennWell Corporation, Oklahoma, 2004.

[15] H.J. Lowe, D.G. Christie, Heat Transfer and Pressure Drop Data on CoolingTower Packings and Model Studies of Resistance of Natural Draft Towers toAir Flow, International Heat Transfer Conference, Boulder, Colorado, 1961.

[16] Cooling Tower Institute, Cooling Tower Institute Performance Curves, CTI,Houston, Texas, 1967.

[17] N.W. Kelly, Kellys Handbook of Crossflow Cooling Tower Performance, N.W.Kelly and Assoc., Kansas City, MO, 1976.

[18] S.A. Cale, Development of Evaporative Cooling Packing. Report EUR 7709 EN,Commission of European Communities, Luxembourg, 1982.

[19] R.D. Fulkerson, Comparative evaluation of counterflow cooling tower fills,Journal of the Cooling Tower Institute 9 (2) (1988).

[20] B.M. Johnson, Cooling Tower Performance Prediction and Improvement.Volume 1, Applications Guide, EPRI Report GS-6370, Volume 2, KnowledgeBase, Electric Power Research Institute, Palo Alto, CA, 1989.

[21] E. De Villiers, Heat, mass and momentum transfer in the rain zone of a coolingtower, Stellenbosch University, MSc thesis, South Africa (1999).

[22] F. Merkel, Verdunstungskhlung, Zeitschrift des Vereines Deutscher Ingen-ieure 70 (1925) 123e128.

[23] CTI (Cooling technology institute), Acceptance Test Code for Water CoolingTowers. CTI Code ACT-105 (2000). Houston, TX.

[24] American Society of Heating, Refrigeration and Air Conditioning, LaboratoryMethods of Testing Fans for Rating. Standard 51e75, ASHRAE, New York,1975.

[25] P.J. Grobbelaar, Experimental and numerical evaluation of anisotropic fill incross- and counterflow, MSc thesis, University of Stellenbosch, South Africa(2010).

[26] T.P. Bertrand, Evaluation of a 1.5 1.5 m2 counter-flow fill performance testfacility with a view to contributing to a fill performance standard, MSc thesis,Stellenbosch University, South Africa (2011).


Performance characteristics of a trickle ﬁll in cross- and counter-ﬂow

Documents

Transcript of Performance characteristics of a trickle ﬁll in cross- and counter-ﬂow