Está en la página 1de 6

G. F.

Halleit
Manager of Product Development,
Ceramic Cooling Tower Co.,
Performance Curves for Mechanical
Fort Worth, Texas

Draft Cooling Towers


This paper presents cooling tower theory and methods for analytical verification of man-
ufacturers' guaranteed performance curves for mechanical draft cooling towers. Both
ASME PTC-23 and Cooling Tower Institute Bulletin ATP-105 are being revised and
both test codes have historically used performance curves as a means of evaluating cool-
ing tower capacity. Techniques and methods are given for calculating performance
curves for both counterflow and crossflow type cooling towers. These procedures can be
used during bid evaluation to assess and predict tower performance at various operating
conditions other than the design point.

Introduction based on sound engineering principles. The cooling tower or other


Water cooling systems have been in use since before the turn of devices for this purpose can be considered as heat exchangers in
the century. These systems have included installations such as which water and air are in direct contact with one another. There
once-through cooling which used either natural or man-made lakes is no acceptable method for accurately calculating the total contact
or rivers. Other methods which were used and are still in operation surface between water and air. A "K" factor, or heat transfer coef-
for utility plants and other industrial applications include spray ficient, cannot be calculated directly from test data. Thermal per-
ponds and various spray devices. Due to the increasing lack of formance tests can be conducted for various specified cooling
available land for large cooling lakes and to government regulation tower designs and then evaluated and compared using accepted
of water usage, the need for cooling towers has become more preva- and proven methods. These same basic methods and theories can
lent over the years. Some of the first cooling towers were of the be used to predict performance at various operating conditions re-
natural draft type having wooden stacks for their air moving capa- lated to a design point.
bility. Other types of towers have included the atmospheric splash A generally accepted basic concept of cooling tower performance
deck tower which was dependent upon the ambient wind for air was developed by Merkel [l] 1 in 1925. His analysis and equations
movement. This type of tower almost defied analytical analysis include the sensible and latent heat transfer into an overall heat
and, therefore, was sold mostly by the foot of tower length rather and mass transfer process based on enthalpy difference as the
than by design specifications. basic driving force. A simple description of the transfer process is
More sophisticated means of analyzing and predicting perfor- as follows: each water surface is surrounded by a film of saturated
mance has become a necessity due to the size of our electrical gen- air at the bulk water temperature. The air is being heated and sat-
erating stations and the tremendous water quantities and associ- urated as it passes through the tower. The temperature and the as-
ated heat loads which must be dissipated. The size of our electrical sociated enthalpy of both water and air change in the tower ac-
generating sets has increased about 10 times over the last 10 years. cording to their relative positions. The heat and mass transfer is,
We now have stations of 1100 MW requiring in excess of 1,000,000 thus, related to the enthalpy difference at each point.
gpm circulating water flow. The sizes of the cooling tower installa- The Merkel theory further established that cooling tower per-
tions have increased proportionately and their costs are sometimes formance is basically related to and limited by the wet bulb tem-
in excess of $5,000,000 or as much as 5 percent of the total station perature which is an approximate measure of enthalpy. Other vari-
cost. Industrial and process towers have also increased proportion- ables which affect cooling tower performance are: the cooling
ately in size and cost. The expense of these installations can be range, the approach to the wet bulb, the entering wet bulb temper-
used to justify the many hours and dollars which the architects ature, and the total circulating water flow rate. Performance is also
and engineers spend in the job proposal and bid stage. affected by the tower and packing design and by the liquid-to-gas
Many theories have been developed since the early 1900s de- ratio required to meet the specified design point.
scribing the heat and mass transfer phenomenon which takes place Many of the towers which have been tested over the past several
in any atmospheric water cooling device. Most of these theories are years have shown a deficiency in their design levels unrelated to a
lack of engineering technology. Since the users of cooling towers
are becoming more knowledgeable, they are insisting that perfor-
mance acceptance tests be conducted on their installations. An ac-
Contributed by the Performance Test Codes Division, and presented at
the Winter Annual Meeting, New York, N. Y., November 17-22, 1974, of
THE AMERICAN SOCIETY MECHANICAL ENGINEERS. Manuscript
1
received at ASME Headquarters July 24,1974. Paper No. 74-WA/PTC-3. Numbers in brackets designate References at end of paper.

Journal of Engineering for Power OCTOBER 1975 / 503


Copyright 1975 by ASME

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


curate performance test conducted in accordance with an accept- The methods used for counterflow towers cannot be applied to
able test code such as ASME PTC-23 or the Cooling Tower Insti- crossflow towers with the same degree of accuracy, although there
tute Bulletin ATP-105 is the only assurance that the user has, in are certain similarities in the methods. The heat and mass transfer
fact, received a tower capable of 100 percent capacity. 2 processes in both types of towers are based on the same potential
Many improvements have been made by the individual cooling for cooling and can be represented as being equal to an overall
tower companies on the basic theories used for analysis. Because transfer or diffusion coefficient K, to the interfacial area between
the industry is a very competitive market, these refinements have the media, and to the potential for cooling. The total heat transfer
been considered proprietary and closely guarded. Several papers in the tower is the sum of the heat transferred at each differential
have been written by the companies describing the theory but few surface. The process of integration is dependent upon the relative
publications are available concerning the practical application. flow patterns of water and air in each type of tower.
The purpose of this paper is not to derive equations describing The counterflow tower can be resolved into a one-dimensional
basic theories but to apply known and proven theories in a practi- problem with the assumption that the flow pattern is vertical with
cal manner so that the engineer may verify performance curves the water falling downward through the tower and the air being
which have been submitted to him. This paper presents methods forced upward. Since enthalpy is a function of temperature, each
which can be used for calculating or verifying "guaranteed perfor- incremental change in water temperature results in a correspond-
mance curves" for both counterflow and crossflow mechanical ing change in the enthalpy of the main air stream. The change in
draft cooling towers. potential is in one direction and is linear for a fixed water-to-air
ratio. It is also assumed that the counterflow tower can be divided
Counterflow Tower Theory and Demand Curves into unit towers of 1 sq ft of ground area, and that the unit towers
Experience and research conducted by the cooling tower indus- are working in parallel. The following equation was developed and
try and other organizations have shown that the method of correla- presented by Lichtenstein for determining the counterflow tower
tion developed by Lichtenstein approximates the performance of characteristic:
counterflow cooling towers [2]. This method which can be used for
design, evaluating, and predicting counterflow performance has KaV dT
gained wide acceptance by the industry. L STt7 1 . (1)
Lichtenstein pointed out that his theory does not completely ex-
plain the effect of air flow rate on the counterflow characteristic Where
curve. Other research reported by various investigators has shown K overall mass transfer coefficient, lb/(hr)(sq ft of contact
that performance varies more significantly with the inlet water area) (lb water/lb dry air);
temperature to the tower [3, 4, 5]. The method of correlation de- a = interfacial contact surface, sq ft/cu ft of tower volume;
veloped by Lichtenstein does not include the effect of these vari- V = tower volume, cu ft/sq ft of plan area;
ables; however, the effect of inlet water temperature can be incor- L = water flow rate, lb/hr sq ft plan area;
porated as shown on Cooling Tower Institute form ATP-107R. The 7i = water temperature entering tower, F;
effect of air flow rate on the characteristic curve is very minimal T2 = water temperature leaving tower, F;
for spaced packings and can be neglected over the narrow operat- dT = temperature differential, F;
ing range of air loadings for counterflow towers. The characteris- h" = enthalpy of saturated air at water temperature, Btu/lb
tics of some film packings do vary with air flow rate, and the varia- dry air;
tion in performance is determined only by tests. h = enthalpy of main air stream, Btu/lb dry air.

2 The left-hand side of equation (1) is a dimensionless group which


An average of more than 15 towers per year have been tested by Midwest relates the variables and capabilities of a particular tower design.
Research Institute since 1970 in accordance with CTI Bulletin ATP-105.
The average performance capabilities for 1970-1973 were 85, 82, 87, and 75 The equation may also be used to determine the order of difficulty
percent respectively. or the demand for a specified cooling design.

Nomenclature ,

Counterflow
K overall mass transfer coefficient, ing design but also with water ture to calculate interior points
lb/(hr) (sq ft of contact area) temperature entering tower Ka = the unit-volume characteristic or
(lb water/lb dry air) (7i) a unit-volume coefficient for a
interfacial contact surface, sq ft/ n = exponent related to packing de- specified crossflow packing,
cu ft of tower volume sign determined from test data Btu
tower volume, cu ft/sq ft of plan hr cu ft Btu/lb dry air
area Crossflow C = constant determined by test
water flow rate, lb/hr sq ft plan Ha = enthalpy of saturated air, Btu/lb varies with packing design and
area _ of dry air water temperature entering
Tt water temperature entering X = mesh size, dimensionlessuse a tower (Ti)
tower, F value of 0.06 and retain each L = liquid loading, lb/hr sq ft of hori-
water temperature leaving tower, fifth value of air enthalpy to zontal water area
F calculate interior points G' air loading, lb dry air/hr sq ft of
dT temperature differential, F Hw - enthalpy of saturated air at water vertical air inlet area
h" enthalpy of saturated air at water temperature, Btu/lb of dry air Ml, 2 exponents determined by per-
temperature, Btu/lb dry air i, j_ = array notation for point being formance tests
enthalpy of main air stream, considered, dimensionless packing depth, in the direction of
Btu/lb dry air Tw = temperature of water at Z co- air flow, ft
C = constant related to packing de- ordinate, F packing height, ft.
sign, or the intercept of the Z = mesh size, dimensionlessuse a
characteristic curve at L/G = value of 0.06 and retain each total liquid-to-gas ratio, (lb
1.0varies not only with pack- fifth value of water tempera- VGA- water/hr)/(lb dry air/hr)

504 / OCTOBER 1975 Transactions of the ASME

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


TOWER CHARACTERISTIC CURVE SHEET tion since this variable is a function of tower and packing design
for use with CTI Bulletin ATP-105
Manufacturer determined by tests. The effect is greatest for a spray filled tower
and at a minimum for a pure film-type packing. The reduction in
Ivlodel no. Job location KaV/L at an L/G of 1.0 can be from 0 to 10 percent per 10 in-
Proposal no. dated Inquiry no.
Other data crease in inlet water temperature with about 5 percent as a good
A. Show L/G design point and guaranteed . Show design approach curve at design
performance curve extended to 10% range and design WBT, intersecting average for industrial packings.
above and 10% below design point. curve A at design L/G.
KaV/L = C(L/G)"-6
The manufacturer is usually required to supply a set of guaran-
. Plot test curve paralleling curve A.
teed performance curves covering operating conditions such as
water flow, cooling range, cold water temperature, and wet bulb
temperature. A typical set of curves is shown in Fig. 1. These
curves show cold water temperature as a function of wet bulb tem-
perature for various cooling ranges. The design value of L/G and
^ ^::::::::^:;:HWT - F::::::::::::::::::::::::::
KaV/L is presumed to remain constant for the design gpm. The
> " - ^ : - " : : : = = - uo values for range, cold water temperature, and wet bulb tempera-
ture may be read from points on the curves and the value of KaV/
L may be calculated for the design L/G. The value of KaV/L is
considered constant within the limits given by CTI and ASME test
codes.
The curves may be calculated using analytical methods similar
to those used by the Cooling Tower Institute to determine perfor-
mance levels from field test data [6]. The procedure for calculating
1A
zi'zzz::i'-ZZ~z :":~:::::""~::::~::" : : - - : : - : : : : performance curves for any counterflow tower is as follows:
1 Estimate a cold water temperature or approach for the wet
bulb temperature being considered. Using equation (1), the value
of Ka V/L may be calculated for the estimated temperatures.
1.6 1.7
2 If the calculated value of Ka V/L does not agree within a rea-
WATER-TO-AIR RATIO, l/G
Design Test Test Test
sonable tolerance with the design or adjusted Ka V/L, then
Conditions Conditions (1) Conditions (2) Conditions (3) 3 Use a trial and error solution and a refined estimate for the
Water flow 100,000
Hot water temp. Fm 120 cold water temperature until tolerances on Ka V/L are met.
Cold water temp. F The CTI form ATP-127 is included to show a format and calcu-
Wet bulb temp. F 80
Total fan horsepower hp 75 lation procedure for Ka V/L based on the Tchebycheff method for
L/G " 1.7309 numerically evaluating the integral given in equation (1). This
KaV/L. = 1.1181
method giyes consistent results over a wide variety of cooling rang-
_Bj_ es and wet bulb temperatures. The method also lends itself to pro-
Revised January 1960 CTI form ATP-107R gramming on a digital computer.
The ASME and CTI test codes are being revised and both have
historically used performance curves for the determination of
tower capacity. It has been suggested that the performance curves
should be plotted based on constant air volume, and provision is
Counterflow Performance Curves also being made to include the effect of water temperatures. The
Usually the quoted bid data will include enough information so mass flow rate through the tower would then become a function of
that the design KaV/L may be calculated or verified if desired. inlet air and water temperatures; therefore, appropriate adjust-
The design point is the intersection of the demand or approach ments should be made to the L/G and the available KaV/L. A
curve and the cooling tower characteristic curve as shown in CTI lower L/G and higher value of KaV/L will be available for points
form ATP-107R. The equation for the tower characteristic curve
can be written in the general form as follows:

^ = C(L/GY" (2)
Where: COUNTERFLOW TOWER PERFORMANCE CURVES
DESIGN CAPACITY 100% GPM
KaV/L = tower characteristic; DESIGN CONDITIONS'. WATER FLOW 100,000 GPM
C = constant related to packing design, or the intercept HWT"I20F CWT"95F WBT=80F
L/G 1.7309 KaV/L'1.1181
of the characteristic curve at L/G = 1.0;
n = exponent related to packing design determined from
COOLING RANGE-F
test data.
The tower characteristic curve for a particular cooling tower design
is usually determined from accurate test data and performance
tests conducted in a research facility. The research data are then
related to field performance tests for further substantiation. The
values of " C " and "n" in equation (2) are a function of packing de-
sign. The value of n is the slope of the characteristic curve for the
packing design. Values can vary from as low as -0.25 to 1.0. The
lower values are generally characteristic of spray fill towers, and
the upper limits are usually associated with high transfer, film-
type packings. The average value of n for industrial type packings
is from 0.5 to 0.6. In accordance with CTI requirements, a slope
of - 0 . 6 can be used within the test code limits without appreciable
error.
60 62 64 66 68 70 72 74 76 78 80 82 84 86
If the tower characteristic curve includes the effect of hot water
temperature, then the value of KaV/L varies with temperature. WET BULB TEMPERATURE- F
The cooling tower manufacturer must supply the degree of varia- Fig. 1 Counterflow tower performance curves

Journal of Engineering for Power OCTOBER 1975 / 505

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


below the design wet bulb temperature and for lower inlet water gration are dependent upon the relative flow patterns of water and
temperatures. These variables are generally not included in manu- air in each type of tower. The crossflow tower involves a two-di-
facturers' performance curves since the procedures will result in a mensional flow pattern in which water falls downward through the
lower predicted or guaranteed cold water temperature. Most tow- tower and air is drawn horizontally through the packing. The en-
ers are tested below the design wet bulb temperature. thalpy of the air changes not only in the vertical direction but also
Counterflow type curves may be calculated graphically by using in the horizontal.
the CTI Cooling Towers Performance Curves, 1967 [7], which is In 1956, Zivi and Brand presented a method for crossflow tower
commonly referred to as the "Blue Book" or Foster Wheeler Bulle- analysis based on the principle of enthalpy difference as the poten-
tin CT-43-2, also known as the "Black Book" [8]. The CTI bulletin tial [11]. Their method consists of a point-by-point determination
is available from their headquarters in Houston, Texas. The Foster of the water temperature and air enthalpy distribution within an
Wheeler bulletin was printed in limited supply in the early 1940s actual tower. The temperature contours of the actual tower are
and is no longer available except for existing copies. Both volumes then related to the contours of a dimensionless X and Z coordinate
of "demand" performance curves present the required KaV/L as a grid or matrix for specified boundary conditions. Equations are de-
function of L/G for various approaches for a constant cooling range rived for relating the X and Z coordinates to the depth and height,
and wet bulb temperature. A similar discussion of theory and a respectively, for an actual unit-width tower. The dimensionless
limited span of demand curves was published by J. F. Pritchard curves can be used to determine the characteristic required for a
and Company in 1957 [9]. crossflow tower to meet a given cooling specification. The method
may also be used to evaluate test data and to predict performance
Crossflow Tower Theory and Demand Curves of a particular tower design. Other methods for crossflow analysis
References on the crossflow principle of water cooling originated have been presented by authors such as Baker and Mart [3], Baker
in the early 1920s [10]. Some of the first water cooling devices such and Shryock [12], and Snyder [5]. Referenced methods are based
as spray ponds and towers incorporated the crossflow principle in on the same enthalpy difference as a basis for cooling and produce
part, but the process was not analyzed separately. The double-flow results of acceptable accuracy, although the mathematical ap-
crossflow cooling tower concept for industrial applications was in- proach for solving the double integration varies.
vented in the early 1940s. This type of tower has gained wide usage The remainder of this paper presents computer methods and
for cooling requirements having large water flow rates and for me- modifications for the Zivi and Brand method and shows the effect
dium to less difficult design temperature conditions. The greatest of variables which have been discussed but not completely defined
advantages of the industrial crossflow tower are its design capabili- in previous publications. Fig. 2 shows the format for a two-dimen-
ty of water loadings to 20 gpm/sq ft of packing area and its air ve-, sional array used for the incremental analysis of a crossflow tower.
locities to 600 fpm. The hot water temperature is considered constant along the X
As previously discussed, the heat and mass transfer process in boundary in the direction of air flow and the entering wet bulb
both counterflow and crossflow towers are based upon the same temperature is considered constant along the vertical Z boundary.
potential. However, the method of analysis and the process of inte- Equations given below have been developed relating the change of
air enthalpy and water temperatures along each boundary. The
COOUNG TOWER INSTITUTE temperatures and physical properties of water and air at each of
PERFORMANCE ANALYSIS
Calculation StiMts
the interior points may be calculated using a trial and error solu-
tion of two equations which relate the air enthalpy and the water
Sheet 3 of _ temperature at the point being considered. The dimensionless X
Date
and Z coordinates, or the demand required for an average cold
VI. Design Approach Curve: (Note trial plot. Select val s of L/G to obtain approach line that crosses curve rhru
test point. For first point. sume approach curve slopes upward to right at 45.)
water temperature, may be determined by plotting lines of con-
stant, average cold water temperatures for specified entering hot
0.1 (T, - T , ) _
0.4 (T, - T , ) _ water and wet bulb temperatures. The average cold water temper-
h, (T, - T,) L/G - _ ature at an interior point is the average of the individual tempera-
0.1 (T, - T,)L/G - "
0.4 (T, - T,> L/G -
h
T,F(orhw *h . <hw - h> 5h
T,
TWO-DIMENSIONAL ARRAY FOR TYPICAL I N C R E M E N T A L
T, * 0.1 IT, - T . ) - + 0.1 (T, - T,) L/G - Ah, UNIT-VOLUME ANALYSIS OF CROSSFLOW PACKING
T. t 0.4(T.--T., - h, * 0.4 (T, - T,) L/G - Ah.-
T, - 0.4 (T, - T.) - - 0.4 (T, - T , ) L/G - Ah, - CONSTANT HWT, T w
X
T, - 0.1 (T, - T . ) - 0.1 (T, - T , ) L/G - Ah, -
(1,1) (2,1) (3,1) (4,1)
T. -
AT,"
*Samt u.valuti in lablr npagr
o
;; IN
t-> 4h
(Plot n scratch plot to determine next value of L/G.)
L/G- h, - b, (T, - Tr)L/G (1,2) (2,2) (3,2) (4,2)
0.1 (T, - T.) L/G " 6.1 * 1- *
0.4 (T, - T.) L/G 0.4 *
03
T , F for h . h. h. ( h . - h.) 5
4h
T, - h. H
T. * 0.1 (T, - T.) - h, * 0.1 (T, - T.) L/G - Ah,-- Z (1,3) (2,3) (3,3) (4,3)
CO
T, 0.4 (I. - T.) - h. + 0.4 (T, - T.) L/G - 4h. - Z
o
Ah.- o
T, - 0 . 1 (T, - T . ) - h. - 0 . H T . - T , ) L/G - Ah. -

T. - h, -_.
y AT
(1,4) (2,4) (3,4) (4,4)

' Ah

Calculate new points in V as required, fteplot on CTI form /MP-107R.


'. Summary: L/G t intetsection of curves from V & VI - .
Capacity - 100 x L / G ' intersection . 1 0 0 x (Tear curve above
L / G . design Fig. 2 Two-dimensional array for typical incremental unit-volume analysis
i 1963 CTf Form ATT-127 of crossflow packing

506 / OCTOBER 1975 Transactions of the ASME

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


tures of all points to the left. The equations and procedure for the 10 M 12

point-by-point analysis are as follows:


For X Boundary. Entering hot water temperature, constant.
Calculate air enthalpy for points on X boundary.
For 2 Boundary. Entering wet bulb temperature, constant. Cal-
culate water temperature for points on Z Boundary.

S t a r t i = 1.0, j = 1.0, Ai = 1.0, Aj = 0.0

H
AX
-
HaU-i.J) n l # w < M , J ) + HwU.j) ~Hai.i-i,l)\
Ha{i,j)
AX
1.0 +
(3)

Where:
Ha = enthalpy of saturated air, Btu/lb of dry air;
AX = mesh size, dimensionless. Use a value of 0.06 and re- Fig. 3 Typical crossffow demand curves
tain each fifth value of air enthalpy to calculate
interior points;
Hw - enthalpy of saturated air at water temperature,
Btu/lb of dry air; 3 Solve for Ha, Tw at (i,j) using estimated enthalpy difference.
", " / ' = array notation for point being considered, dimen- 4 Determine Hwaj) for Tw(ij) in equation (7).
sionless. 5 Calculate Hwaj) Haaj).
6 Compare with estimated value.
Start i = 1. 0, j = 1. 0, Aj = 1. 0, Ai = 0. 0 If not within tolerance the second try could be the average of the
estimated and calculated values.
vUJ)
7 Repeat until tolerance is met.
The foregoing equations and array notation have been given in a
general form for use in computer programming. Although the
AZ equations and procedure may be used for manual calculations, it is
<t>U,j-D L-^iu,j-i> +
H<mt, H,
highly desirable that a digital computer program be available for
(4) the computations. A new set of curves is required for each change
in either boundary conditions. A computer program may be writ-
Where: ten in a matter of hours and is well worth the effort.
Tw = temperature of water at 2 coordinate, F; A typical set of crossflow demand curves are shown in Fig. 3 for
AZ = mesh size, dimensionless. Use a value of 0.06 and retain a hot water temperature of 120F and an inlet wet bulb tempera-
each fifth value of water temperature to calculate ture of 80F. The demand curves are dimensionless and are not re-
interior points. lated to a particular tower design. The curves show the order of
1 Estimate Tw(U). First try could be = [ 7 V , j _ y - (Tw(ij-i) difficulty of the temperature conditions. The above procedure and
- 7W1/2. (5) equations can be used to calculate demand curves for various
2 Substitute Hwaj) for estimated Tui(ij) above in equation (4). crossflow conditions similar to the CTI "Blue Book" for counter-
3 Calculate T ^ j ; , equation (4). flow duty.
4 If Tw estimated = Tw calculated (within tolerance), proceed.
5 If not, second try could be average of estimated and calculat- Crossflow Performance Curves
ed values. The thermal characteristics for crossflow packings must be de-
6 Repeat Step 5 until tolerance is satisfied. termined by performance tests. The crossflow heat and mass
Note: Calculate H and T values to tolerance of 0.01. transfer process for a crossflow tower is best described as a unit-
Interior Points. Known values of Tw, Hw, Ha from preceding volume coefficient or characteristic. The general form of the equa-
points. tion describing the thermal characteristics can be written as fol-
lows:
Start i = 2. 0, j = 2. 0, A i = 1.0, Aj = 1.0
H
Ka CG'"iL"2 (8)
aU,S)
Where:
AX
HaU-l,j) + ~2~ l^w{i-i,j) ~ Haii-l,j) + HwU.j) ~ Haii,j)\ Ka the unit-volume characteristic or a unit-volume
coefficient for a specified crossflow packing,
(6) Btu
u(i,J> hr cu ft Btu/lb dry air
C = constant, determined by test. Varies with packing de-
AZ sign and water temperature entering tower (Ti);
L = 'liquid loading, lb/hr sq ft of horizontal water area;
(7) G' - air loading, lb dry air/hrsq ft of vertical air inlet area;
i i , rt2 = exponents determined by performance tests.
Note: The values of A X and A 2 can be changed to 0.3 without The unit-volume coefficient for a specified packing may be used to
loss of accuracy for interior points. calculate the characteristic curves for a line of towers having vari-
1 Estimate Hwaj) - Haaj). First try could be = Hwa-ij) - ous cell and tower dimensions. The X and 2 coordinates for points
Ha(i-lj) on a characteristic curve may be calculated by solving the fol-
2 Substitute in equations (6) and (7). lowing equations given for various assumed values of G' and L.

Journal of Engineering for Power OCTOBER 1975 / 507

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


tion determines the liquid-to-gas ratio at which the tower must be
operated to meet the design temperature conditions. The design
point may be used in a similar manner for the counterflow tower in
that it may be superimposed on the demand curve for any other
boundary conditions. Appropriate adjustments should be made in
the liquid-to-gas ratio and the cooling potential available for
changes in hot water temperature and air flow rate. A typical vari-
ation attributable to the hot water temperature effect in Ka is 5
percent reduction per 10F increase in hot water temperature.
Test values have shown that the variation can be considered linear
between 90F and 150F. The variation of performance with air
flow is inherent in the method of analysis.
The characteristic curves shown in Fig. 4 were used to calculate
the crossflow performance for a 60F inlet wet bulb temperature
and a cooling range of 25F. The predicted cold water temperature
for the crossflow tower without corrections to the L/G or perfor-
mance levels is approximately 0.6F lower than the counterflow
performance shown in Fig. 1 at 60F wet bulb temperature. The
crossflow performance curve will have a steeper slope than the
Fig. 4 Crossflow cooling tower characteristic curves
counterflow curve. Crossflow performance curves submitted for
proposal and bid purposes may be calculated using a trial and
error solution for the design point and predicted cold water tem-
The equations relate the packing dimensions and thermal charac- perature on a variety of demand curves which cover the guaran-
teristics to the dimensionless grid shown in Fig. 3. teed zone of cooling tower operation. The procedure is similar to
that given for the counterflow tower with certain exceptions. Per-
formance is predicted for various operating and boundary condi-
tions by solving for the cold water temperature, or the approach to
Where: the wet bulb, for constant cooling ranges.
X = X coordinate, dimensionless;
x = packing depth, in the direction of air flow, ft. Conclusions
The size and cost of our present day cooling towers have in-
creased the need for analyzing competitive bids and proposals.
This paper presents discussions of known and accepted methods
Where: which may be used by the engineer prior to bid awarding to ana-
Z = Z coordinate, dimensionless; lyze and predict counterflow and crossflow performance. The dis-
z = packing height, ft. cussions are limited to mechanical draft type towers and to the op-
Equations (8), (9), and (10) are used to calculate the characteristic erating conditions covered by existing test codes. The theories and
curves shown in Fig. 4. The curves are calculated for towers having techniques may also be used to analyze and predict performance
air travel depths of 12 and 15 ft and packing heights of 20 and 45 for natural draft towers if appropriate adjustments are made for
ft, respectively, for a base hot water temperature of 100F and G' the air moving capabilities and for the resulting change in thermal
of 2000 Ib/hr sq ft. The curves also indicate the effect on perfor- performance levels.
mance level for variations in hot water temperature and air load-
ing. Fig. 4 also shows lines for the total liquid-to-gas ratio to be References
used with the characteristic curves. The following equations ob- 1 Merkel, F., "Verdunstungskuehlung," VDI Forschungsarbeiten No.
tained directly from equations (9) and (10) in the foregoing can be 275, Berlin, 1925.
,-,,aMsed to relate this ratio to the dimensionless grid. 2 Lichtenstein, J., "Performance and Selection of Mechanical-Draft
Cooling Towers," TRANS. ASME, Vol. 65,1943, p. 779.
3 Baker, D. R., and Mart, L. T., "Cooling Tower Characteristics as De-
termined by the Unit-Volume Coefficient," Refrigerating Engineering,
L/G b a s e d on unit a r e a i s -, - * - (n) 1952.
G' Z x 4 Mickley, H. S., "Design of Forced Draft Air Conditioning Equip-
L L x ment," Chemical Engineering Progress, Vol. 45, 1949, p. 739.
L/G b a s e d on t o t a l m a s s flow r a t e s () r = 7 - (12) 5 Snyder, N. W., "Effect of Air Rate, Water Rate, Temperature, and
G G z Packing Density in a Cross-Flow Cooling Tower," American Institute of
Chemical Engineers, Heat Transfer Symposium, Mar. 20,1955.
Equation (11) will plot as a series of straight lines on the grid for 6 Cooling Tower Institute Bulletin ATP-105.
various values of L and G' based on a unit area liquid-to-gas ratio 7 Cooling Tower Institute Cooling Tower Performance Curves, 1967.
for a tower having specified dimensions of x and 2. The L/G' is 8 Bulletin CT-43-2, Foster Wheeler Corporation, New York, N. Y.,
usually preferred for tower design and selection. Equation (12) will 1943.
also plot as a family of straight lines based on the total liquid-to- 9 "Counterflow Cooling -Tower Performance," J. F. Pritchard, and Co.
of California, 1957.
gas ratio. The (L/G) T has advantages for evaluating field test data 10 "Cross-Current Flow Type Water-Cooling Tower," Mechanical Engi-
and for the prediction of tower performance. Fig. 4 is based on neering, Vol. 12, Apr. 29,1921, p. 227.
(L/G)T- Fig. 4 can be plotted as an overlay for use with Fig. 3. The 11 Zivi, S. M., and Brand, Bruce B., "An Analysis of the Cross-Flow
design point for a specified tower to meet a given cooling require- Cooling Tower," Refrigerating Engineering, 1956.
12 Baker, Donald R., and Shryock, Howard A., "A Comprehensive Ap-
ment is the intersection of the tower characteristic curve and the proach to the Analysis of Cooling Tower Performance," Journal of Heat
given average cold water temperature line. The point of intersec- Transfer, Aug. 1961, pp. 339-348.

508 / OCTOBER 1975 Transactions of the ASME

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

También podría gustarte