Calculating Realistic PMroEmissions from Cooling TowersJoelReismon ond Gordon FrisbieGreystone Environmental Consultants, Inc,, 10470 Old Placerville Road, Suite 110, Sacramento, CA95827

Emissions of particulate matter less tban 10 micrometers indiameter (PM r) from wet cooling towers rnay be calculatedusing tbe methodologlt presented in EPA's AP-42 [1], wbicbassumes tbat all total dissolued solids (TDS) emitted in "drift"particles (liquid water entra.ined in the air stream and can'iedout of tbe tower tbrougb tbe induced drafifan stack) are PMro.Howner, tbis assurnption is ouerly conservatiue because it doesnot consider that, upon euaporation of tbe drift, many of tbesolid particles tbat remain are larger tban PMro. Particleslarger than 10 micrometers do not represent a bealtb bazardand are not regulated under current air quality regulations.For example, for wet cooling towers witb medium to higb TDSleuels, tbe AP-42 metbodologjt predicts significantly higberPM ,o emissions tban would actually occur, euen for towersequipped witb uery bigb fficiency drift eliminators (e.g.,0.00060/o drift rate). Sucb ouer-prediction may result in unre-alktically bigb PMromodeled concentra.tions and./or tbe needto purcbase expensiue Ernission Reduction Credits (ERCs) inPMronon-attainmmt areas, Since tbese towers hauefairly lowemissionpoints (10 to 15 m aboue ground), ouer-predictingPMro emission rates can also result in exceedingfederal Pre-uention of Signfficant Dete4oration (PSD) significance leuelsat a project'sfenceline. Tltis paper presents a metbodfor com-puting realistic PM ,o emissions from cooling towers with medi-um to bigb TDS leuels by enabling tbe engineer to determinetbe PMromassfraction of the total amount of particulate emit-ted by a coolingtower,

INTRODUCTIONCooling towers are heat exchangers used to dissi-

pate Iarge heat loads to the atmosphere. 'Wet, or evap-orative, cooling towers rely on the latent heat of waterevaporation to exchange thermal energy between theprocess and the air passing through the cooling tower.The cool ing water may be an integral part of theprocess or may provide cooling viaheat exchangers,for example, steam condensers. Wet cooling towersprovide direct contact befween the cooling water and

Environmental Progress (Vo1.21, No.2)

air passing through the tower, and as part of normaloperat ion, a very smal l amount of the c i rculat ingwater may be entrained in the air stream and be carried out of the tower as "drift" droplets. Because thedrift droplets contain the same chemical impurities asthe water circulating through the tower, the particulatematter within the drift droplets may be classified as anemission. The magnitude of the drift loss is influencedby the number and size of droplets produced withinthe tower, which are determined by fill design, towerdesign, the air and water patterns, and design of thedrift eliminators.

AP.42 METHOD OF CATCUTATING DRIFI PARTICUTATEEPA's AP-421. provides available particulate emis-

sion factors for wet cooling towers, however, thesevalues only have an emission factor rating of "E" (thelowest level of confidence acceptable). They are alsorather high, compared to typical present-day manufac-turers' guaranteed drift rates, whiqh are on the orderof 0.0006%. (Drift emissions are typically expressed asa percentage of the cooling tower water circulationnte). AP-42 states that "a consen)Atiuely bigb PMroemission factor can be obtained by: (a) multiplyingthe total liquid drift factor by the TDS fraction in thecirculat ing water, and (b) assuming that once thewater evaporates, all remaining solid particles arewithin the PM10 range." (Italics per EPA).

If TDS data for the cooling tower are not available,a source-specific content can be estimated by obtain-ing the TDS for the make-up water and multiplying itby the cooling tower cycles of concentration. (Thecycles of concentration is the ratio of a measuredparametet for the cooling tower water [such as con-ductivity, calcium, chlorides, or phosphatel to thatpatametet for the make-up water.)

Using AP-42 guidance, the total particulate emis-sions (PM), after the pure water has evaporated, canbe expressed as:

July 2002 1,27

PM : \Water Circulation Rare x Drift Rate x TDS (1)

For example, for a typical power plant wet coolingtower with a water circulation rate of 746,000 gallonsper minute (gpm), drift rate of 0.0006%, and TDS of7,700 parts per million by weight (ppmw),

PM : 145,000 gptn x 8.34lb water/gal x 0.0006/100 x7,700lb solids/10o lb water x 60 min/hr = 3.38Lb/hr

On an annual basis, this is equivalent to almost 1.5tons per year (tpy). Even for a state-of-the-art drift elimi-nator system, this is not a small number, especially ifassumed to all be equal to PM10, a regulated criteria poflutant. However, as the following analysis demonstrates,only a very small fraction is actually PMro.

coMPUT|NG THE PMro FMCT|oNBased on a representative drift droplet size distribu-

tion and TDS in the water, the amount of solid mass ineach drop size can be calculated. That is, for a given ini-tial droplet size, assuming that the mass of dissolvedsolids condenses to a spherical particle after all the waterevaporates, and assuming the density of the TDS isequivalent to a representative salt (e.g., sodium chlo-ride), the diameter of the final solid particle can be calcu-Iated. Thus, using the drift droplet size distribution, thepercentage of drift mass containing particles smallenough to produce PMro can be calculated. This methodis conservative as the final particle is assumed to be per-fectly spherical, hence, as small aparticle as can exist.

The droplet size distribution of the drift emited fromthe tower is critical to performing the analysis. Brent-wood Industries, a drift eliminator manufacturer, wascontacted and agreed to provide drift eliminator test datafrom a test conducted by Environmental Systems Corpo-ration (ESC) at the Electric Power Research Institute(EPRI) test facility in Houston, Texas, in 1988. The parti-cle size distribution is included in the first and lastcolumns of Tables 1, and 2. The data consist of waterdroplet size distributions for a drift eliminator thatachieved a tested drift rate of 0.00030/0. As we are using a0.00060/o drift rate, it is reasonable to expect that the0.00030/o drift rate would produce smaller droplets, there-fore, this size distribution data can be assumed to beconservatiue for predicting the fraction of PMro in thetotal cooling tower PM emissions.

In calculat ing PMro emissions, the fo l lowingassumptions were made:r Each water droplet was assumed to evaporate

shortly after being emitted into ambient air, into asingle, solid, spherical particle.

. Drift water droplets have a density (p) of water;1.0 g/cmj or 1.0 * 10-6 pglpm3.

. The solid particles were assumed to have the samedensity (pror) as sodium chloride, (i.e.,2.2 g/c^3).

Using the formula for the volume of a sphere,V=4n^rJ/3, and the densi ty of pure water, p*=1.0g/.*3, the following equations can be used to derivethe solid particulate diameter, Do, as a function of the

t28 July}oA2

TDS, the density of the solids, and the init ial driftdroplet diameter, Do:

Volume of drift droplet = (4/S;Sn(OutZ)t Q)

Mass of solids in drift droplet : (TDSXp,(Volume of drift droplet) (3)

substituting,

Mass of solids in drifr : (TDS)(p)(4/Dn(Dut2)3 (4)

Assuming the solids remain and coalesce after thewater evaporates, the mass of sol ids can also beexDressed as:

Q)

Mass of solids = (pros) (solid particle volume) :

QroJ G/)n(Dr/z): (5)

Equations 4 and 5 are equivalent:

(pros) (4/)n(Do/2)3 = (TDS) (p)(4/in(Du/2)3 (6)

Solving for Do:

Dp = Da t(TDsxp,/pror)lt6

rilflhere:

TDS is in units of ppmwDo = diameter of solid particle, micrometers (pm)D6 = diameter of drift droplet, pm

Using Formulas 2 through 7 and the particle sizedistribution test data, TabIe l can be constructed fordrift from a wet cooling tower having the same char-acteristics as our example: 7,700 ppmw TDS and a0.00060/o drift rate. The first and last columns of thistable are the particle size distribution derived from testresul ts provided by Brentwood Industr ies. Usingstraight-line interpolation for a solid particle size 10 pmin diameteq we conclude that approximately I4.9o7o o1the mass emissions are equal to, or smal ler than,PMro. The balance of the solid mriterial are particu-lates greater than 10 pm. Hence, PMro emissions fromthis tower would be equal to PM emissions x 0.149, or3,38lb/hr x 0.1.49 = 0.50 lblhr. The process is repeat-ed in Table 2, with alI parameters equal except thatthe TDS is 11,000 ppmw. The result is that approxi-mately 5.t10/o are smaller at 11,000 ppm. Thus, whiletotal PM emissions are Iarger by virtue of a higher TDS,overall PMro emissions are actually louer, because moreof the solid particles arc larger than 10 pm.

The percentage of PMI. /PM was calculated forcooling tower TDS values from 1,000 to 12,000 ppmwand the results are plotted in Figure 1.. Using thesedata, Figure 2 presents predicted PMro emission ratesfor the 745,000 gpm example tower. As shown in thisFigure, the PM em,ission rate increases in a straightl ine as TDS increases, however, the PMro emissionrate increases to a maximum at around a TDS of 4,000ppmw, and then begins ta decline. The reason is thatat higher TDS, the drift droplets contain more solids

Environmental Progress (Vo1.21, No.2)

Ioble L Resultant solid particulate size distribution (TDS = 7,700 ppmw).

Tobfe 2. Resultant solid particulate size distribution (TDS = 11,000 ppmw),

Environmental Progress (Vo1.21, No.2) Iulv 2002 I29

90

80

70

60

850b40o.

30

10

0

20

1000 2000 3000

Figure l. Percentage of drift PM that evaporates to PMro.

4000 5000 6000 7000 8000 9000 10000 11000 12000

Girculating Water TDS (ppmw)

circulation rate of 146,000 GPM and 0.0006% drift rate,

PM Emission Rat\; z

PM16 Emission

Ell

oGtC

.9o.9,EuJ(f

=o

6.0

5.0

4.0

3.0

2.0

1.0

0.01000 2000 3000 4000 5000 6000 7000 8000 9000 10000 11000 12000

Circulating Water TDS (ppmw)

Figule 2. PM10 emission rate vs. TDS.

and therefore, upon evaporation, result in larger solidparticles for any given initial droplet size.

c0NcLusl0NThe emission factors and methodology given in

EPA's AP-42, Chapter 1.3.4 Wet Cooling Towers [1], donot account for the droolet size distribution of thedrift exiting the tower This is a critical factor, as morethan 850/o of the mass of particulate in the drift frommost cooling towers will result in solid particles largerthan PMro once the water has evaporated. Particleslarger than PMro are no longer a regulated air pollu-tant, because their impact on human health has beenshown to be insignificant. Using reasonable, conserva-tive assumptions and a realistic drift droplet size distri-bution, a method is now available for calculating real-

730 July 2002

istic PMro emission rates from wet mechanical draftcooling towers equipped with modern, high-efficiencydrift eliminators and operating at medium to high lev-els of TDS in the circulating water.

TITEMTURE CITED1. "Compilation of Air Pollutant Emission Factors,"

AP-42, Fifth Edition, Yolume I: Stationary Pointand Area Sources, Chapter I3.4 \/et Cooling Tow-ers, http:/ /www.epa.gov/ttn/chief /ap42/ , rJ.S.Environmental Protection Agency, Office of AirQuality Planning and Standards, Research TrianglePark, NC, January 1.995.

Environmental Progress (Vo1.21, No.2)