cti journal, vol. 29, no. 2 · 4 cti journal, vol. 29, no. 2 view from the tower the winter is gone...

CTI Journal, Vol. 29, No. 2 1

The CTI Journal(ISSN: 0273-3250)

PUBLISHED SEMI-ANNUALLYCopyright 2008 by The CoolingTechnology Institute, PO Box 73383,Houston, TX 77273. Periodicalspostage paid at Houston, Texas.

MISSION STATEMENTIt is CTI’s objective to: 1) Maintain andexpand a broad base membership ofindividuals and organizationsinterested in Evaporative HeatTransfer Systems (EHTS), 2) Identifyand address emerging and evolvingissues concerning EHTS, 3) Encour-age and support educationalprograms in various formats toenhance the capabilities andcompetence of the industry to realizethe maximum benefit of EHTS, 4)Encourge and support cooperativeresearch to improve EHTS Technologyand efficiency for the long-termbenefit of the environment, 5) Assureacceptable minimum quality levelsand performance of EHTS and theircomponents by establishing standardspecifications, guidelines, andcertification programs, 6) Establishstandard testing and performanceanalysis systems and prcedures forEHTS, 7) Communicate with andinfluence governmental entitiesregarding the environmentallyresponsible technologies, benefits,and issues associated with EHTS, and8) Encourage and support forums andmethods for exchanging technicalinformation on EHTS.

LETTERS/MANUSCRIPTSLetters to the editor and manuscriptsfor publication should be sent to: TheCooling Technology Institute, PO Box73383, Houston, TX 77273.

SUBSCRIPTIONSThe CTI Journal is published inJanuary and June. Complimentarysubscriptions mailed to individuals inthe USA. Library subscriptions $20/yr.Subscriptions mailed to individualsoutside the USA are $30/yr.

CHANGE OF ADDRESSRequest must be received atsubscription office eight weeks beforeeffective date. Send both old and newaddresses for the change. You mayfax your change to 281.537.1721 oremail: [email protected].

PUBLICATION DISCLAIMERCTI has compiled this publicationwith care, but CTI has not Investi-gated, and CTI expressly disclaimsany duty to investigate, any product,service process, procedure, design,or the like that may be describedherein. The appearance of anytechnical data, editorial material, oradvertisement in this publicationdoes not constitute endorsement,warranty, or guarantee by CTI of anyproduct, service process, procedure,design, or the like. CTI does notwarranty that the information in thispublication is free of errors, and CTIdoes not necessarily agree with anystatement or opinion in thispublication. The entire risk of the useof any information in this publicationis assumed by the user. Copyright2008 by the CTI Journal. All rightsreserved.

ContentsFeature Articles8 Tracking Molybdate in Cooling Water

Vadim B. Malkov

26 The Cost of NoiseRobert Giammaruti

34 Architectural Enclosure Influences on the Performance ofField Erected Counterflow Cooling TowersToby Daley

42 Improved Calcium Phosphate Control for StressedSystemsGary Geiger

54 Evaluate Your Cooling TowerRichard J. DesJardins

64 Seismic Qualification of Cooling Towers by Shake-TableTestingPanos G. Papavizas

Special Sections6 Data Sheet73 CTI Licensed Testing Agencies74 CTI ToolKit

Departments2 Meeting Calendar4 View From the Tower6 Editor’s Corner6 Multi Agency Press Release

see page........26

see page........8

see page........34

CTI Journal, Vol. 29, No. 22

CTI JournalThe Official Publication of The Cooling Technology Institute

Vol. 29 No.2 Summer 2008

Journal CommitteePaul Lindahl, Editor-in-ChiefArt Brunn, Sr. EditorVirginia Manser, Managing Editor/Adv. ManagerDonna Jones, Administrative AssistantGraphics by Sarita Graphics

Board of DirectorsDennis Shea, PresidentJess Seawell, Vice PresidentMark Shaw, SecretaryRandy White, TreasurerGary E. Geiger, DirectorRobert (Bob) J. Giammaruti, DirectorRichard (Rich) Harrison, DirectorChris Lazenby, DirectorFrank Michell, DirectorKen Mortensen, Director

FUTURE MEETING DATESCommittee AnnualWorkshop Conference

July 6-9, 2008 February 8-12, 2009Hyatt Regency, Orange County The Westin, Riverwalk

Garden Grove, CA San Antonio, TX

July 12-15, 2009 February 7-11, 2010Marriott Hotel The Westin Galleria

Colorado Springs, CO Houston, TX

Address all communications to:Virginia A. Manser, CTI AdministratorCooling Technology InstitutePO Box 73383Houston, Texas 77273

281.583.4087 - 281.537.1721 (Fax)

Internet Address: http://www.cti.org

E-mail: [email protected]


View From The TowerThe winter is gone and spring is here. If youbuild, inspect, test, maintain, treat or operate acooling tower, summer is fast approaching sothis is peak activity time preparing for full opera-tion. The CTI as well is gearing up for the JulyCommittee Workshop in California. The AnnualConference commitment deadlines for the Com-mittee Workshop are fast approaching. Com-mittee chairs are preparing meeting agendas andcatching up on details in preparation for the Com-mittee Workshop. It is a busy time for CTI Staff,Board of Directors and yours truly. It is time to

any organization it is time to look forward to thenext question. How can we meet our commit-ment to the Cooling Tower Industry better?The CTI standing and working committees areworking on the future but they need your help.The CTI is in constant need of new ideas and newdirections to venture into so we can grow. In theshort time since assuming the Presidency, I havehad several members offer ideas and suggestionsabout what they see as needs that the CTI can fill.We would also like to hear about areas where CTIis not fulfilling it obligations.

reflect for a moment on where CTI is at and where the orga-nization is headed.The CTI Annual Conference was very successful with excel-lent technical papers, technical panel presentations, owner/operator seminar, and our training seminar. A great deal ofwork was also completed during the committee meetings. Theaddition of working technical meetings at the Annual Confer-ence has increased the face to face time of all our committees.This means that the important work of writing, reviewing orbrainstorming new CTI Standards and Guidelines are progress-ing at a faster pace. The utilization of the new committeeorganization format has brought about increased participationfrom Manufacturers, Suppliers and Owner/Operators. As with

This past year we surveyed the CTI membership to begin theprocess of understanding the needs of our members. I wouldlike to invite anyone reading this Journal to help CTI by provid-ing input to CTI through your Company Representative, Boardof Directors, CTI Office, or myself. It is only through yourinput and ideas that CTI can meet the objectives of our organi-zation.I hope to see everyone in California for some fun and hardworkDenny SheaPresident 2008/2009

Denny Shea

REDWOOD

DOUGLAS FIR24 Hour Service on Your Lumber and Plywood Requirements

COMPLETE FABRICATION AND TREATINGSERVICE FROM OUR OPELOUSAS, LA PLANT

GAIENNIE LUMBERCOMPANY

BOX 1240 • OPELOUSAS, LA 70571-1240800-326-4050 • 337-948-3067 • 337-948-3069 (FAX)

Member


Editor’s Corner

Paul LindahlEditor-In-Chief

For Immediate ReleaseContact: Chairman, CTI Multi-Agency Testing

CommitteeHouston, Texas, 1 - May - 2008

The Cooling Technology Institute announces itsannual invitation for interested drift testing agenciesto apply for potential Licensing as CTI Drift TestingAgencies. CTI provides an independent third partydrift testing program to service the industry.Interested agencies are required to declare theirinterest by July 1, 2008, at the CTI address listed.

Please advise if you have any questions.

Introducing a new company, BOLToutlet.com.This company is dedicated to supplying a widerange of cooling tower related hardware, to thecooling tower manufacturer. The owner hasbeen in the hardware business for twenty-nine(29) years and is a second generation of buyingand selling industrial hardware. BOLToutlet.comhas a program that will bring the best quality andcompetitive pricing directly to you. They canoffer a complete line of cooling tower relatedbolts, nuts, washers, nails and other products.See their advertisement on page 67.

DaDaDaDaDattttta Sheeta Sheeta Sheeta Sheeta Sheet

The CTI program for thermal certification ofcooling towers, and soon for evaporative con-densers, is a remarkable program. The pro-gram is based on CTI STD-201.From the beginning of the existing program inthe early 1990’s, it has grown from a few com-panies in the United States to 16+ companieswith certified lines from Europe, Asia and theAmericas. There are 38+ lines certified. Moreare added on a regular basis, so the numbersquoted above are likely to understate the ac-tual by the time of printing this journal.The CTI certification program is now recognized by boththe State of California Energy Commission (in Title 24),and by the US Department of Energy (via its requirementfor states to adopt ASHRAE 90.1 or more stringent en-ergy efficiency standards). The CTI certification pro-gram is effectively required by law for open and closedcircuit cooling towers on air-conditioning applications inthe United States.The program is run by Tom Weast, the CTI CertificationAdministrator. Tom has done a magnificent job of apply-ing his personal integrity and strong work ethic, not to

mention his legendary skill in keeping costsdown, to build a credible and rapidly growingprogram. CTI owes Mr. Weast a substantialdebt of gratitude for the hard work and perse-verance applied to getting us where we are withthe program.We anticipate continued growth of the certifi-cation program as the green initiatives acrossthe new-building construction industry raise con-cerns for energy efficiency, and validation ofthat energy efficiency. This is the role thatSTD-201 plays in CEC Title 24 and ASHRAESTD 90.1. Carbon footprint concerns are di-

rectly related to energy consumption as well.If you have a chance to talk to Tom Weast, take the opportu-nity to thank him for a very fine job.

Respectfully,

Paul LindahlEditor-in-ChiefCTI Journal


Vadim B. Malkov (Hach Company),Blaine Nagao (ChemCal, Inc.),Steve Dumler (H2TrOnics),Phil Kiser (Hach Company)

AbstractMolybdate-based chemicals have been used as corro-sion inhibitors in cooling tower systems for severalyears. Although they provide superior performance, lev-els of molybdate have been reduced because of largeprice increases and certain local regulations. It has be-come more necessary to control molybdate levels to

A strategic partnership was developed with a companyproviding industrial clients with comprehensive servicefor cooling towers. The company was presented byH2TrOnics Inc. and ChemCal, Inc., which are parts of thesame organization. The present field testing was con-ducted utilizing their customer base, as well as their ex-isting electronic system allowing for remote monitoringof various parameters of cooling water. The system alsoallows direct interaction with eControllers over theinternet, which enables a degree of control of the treat-ment processes. The discussed process instrument wasnamed Hach MO42 analyzer.

Tracking Molybdate in Cooling Water

optimize performance versus cost and compliance in corrosion in-hibition. A new on-line analyzer has been developed that can mea-sure molybdate as molybdenum (Mo6+) with minimum maintenance.This analyzer can be used to monitor remotely when connected toa data acquisition system with web-based reporting component.Three of these complete systems have been evaluated for severalmonths at beta-testing sites in Texas. Comparisons have beenconducted versus bench tests. This paper will discuss results ofcurrent testing and some features of the web-based monitoringsystem with graphs and charts illustrating its performance.

IntroductionThe main purposes of this study were to validate performance of anewly developed process analyzer in real-life conditions when theinstrument is maintained solely by customers, and also to deter-mine suitability of the analyzer to application for monitoring Mo-lybdate tracer concentration in cooling water. A secondary goalwas to validate projected life time of the colorimeters equipped withnew polymeric light pipes.The method employed by the analyzer is a simple and well-knowncolorimetric method based on use of reaction between pyrocat-echol and molybdenum in aqueous solutions (A.I. Busev, “Ana-lytical chemistry of Molybdenum”, p.202, Ann Arbor – HumphreyScience Publishers, London 1969).Pilot version of this analyzer underwent vigorous field testing inspring of 2005 and met all advertised technical specifications (Table1).However, the testing revealed a short projected life time for thecolorimeter. In order to address the discovered issue, an extensivesearch for new materials and following laboratory testing was un-dertaken.New polymer implemented for molding light pipes serving to pro-vide unobstructed path between light source (an LED) and photodetector, showed very good stability during one-year-long test(Figure 1) and the colorimeter projected life time was calculated tobe minimum of 5-7 years. One limitation of the tested polymer wasits relatively high cost; therefore, it was decided to find less expen-sive materials suitable to the application.

ExperimentThe present β-test was started on 12/30/2006 at one industrial sitein Dallas, TX (Dallas-1). Second MO42 was installed at CoolingTower #1 at the power plant of University of Texas (UT) in Austin,TX on 01/23/2007, and the third instrument was installed at anotherindustrial facility in Dallas, TX on 04/17/2007 (Dallas-2). The instal-lations are shown in Figures 2-4 below.Initially colorimeters of the tested instruments were equipped withthe light pipes made from a cheaper polymer - Sylgard 184® (a reg-istered product of Dow Corning Corporation). However, this sili-cone rubber had some potential limitations for the application indi-cated in the data sheet – susceptibility to moisture, although it wassupposed to be reversible. Nevertheless, the testing was conductedin order to prove or disprove this material’s suitability. Eventually,in April of 2007, new colorimeters equipped with light pipes madefrom more expensive, but proven material (discussed in Introduc-tion) were installed on all three analyzers and the â-test was contin-ued and finished.

Reagents and StandardsExisting Hach MO42 reagent (p/n 28905-49) was used in the testand no standards were involved, because the analyzer did not re-quire calibration.

ProcedureThe analyzers were installed as parts of control panels includingseveral other sensors monitoring various parameters of coolingwater. The instruments were set to work unsupervised in the indus-trial environment. Two bottles of the reagent should provide onemonth of uninterrupted work and colorimetric cell manual cleaningwas to be done by the operators on either monthly basis or asrequired by their maintenance protocols.4-20 mA output was used to connect each analyzer to monitoringcontroller that transmitted the data to a secure server. Then pro-cessed data are available for end users through the serviceprovider’s web site. The web based application provides access todata generated by all sensors (eController), all logs (Operator andService), and all bench test results manually entered into the sys-tem. The system also has statistical tools to display the data in

Vadim B. Malkov


different formats (Trend Reports) and to compare various param-eters to each other. The system provides end customers with veryconvenient tools for monitoring and controlling their cooling tow-ers remotely from any computer connected to the internet usingtheir unique user ID and password. All data used in present reportwere collected using this system.

Results and DiscussionLight Pipes βββββ-testThe test of two analyzers (Dallas-1 and UT sites) containing lightpipes made from the first tested silicone material lasted for severalmonths (January – April) and revealed in general positive results.Main parameter that the customers were interested in was the accu-racy, which was expressed in tracking of the bench test results.Bench analysis for molybdate (as Mo6+) was conducted by theoperators using Hach Method #8169 according to their laboratoryprotocols and procedures. Results of the tracking for both sites/analyzers are presented in Figures 5 through 7 below. Some addi-tional explanations are provided in the titles for corresponding fig-ures.Blue lines on the provided charts represent trends built based uponthe bench test results, and as seen from the Figures 5 and 6 thetracking was very good and it made the customers very confidentin the instrument. The third instrument started working at Dallas-2facility just a few days before the colorimeter was replaced, there-fore the collected data was considered not representative and wasnot taken into account.Based on previously acquired knowledge about the instrumentperformance (discussed in the Introduction), Reference Counts datawere collected and analyzed in order to prove acceptable light pipeslife expectancy. The reference counts were found in the instrument’smenu, manually recorded by the operators on a regular basis (dailythrough weekly), and then entered into the system. Several follow-ing figures below show the reference counts trends reflecting clean-liness of the sample and/or the colorimetric cell, and most impor-tantly the downward trend of the counts values during prolongedtime.The deep drop in the counts (circled on Fig. 7) was explained by adirty cell and following maintenance, which brought the countsback to normal.Based on the slope (Fig. 8) and assuming that the decay rate re-mained the same regardless of the LED intensity level, the pro-jected life time for these light pipes would be approximately 650-750days, which is not acceptable. Removing the outliers makes theslope even steeper. Similar behavior was observed at another β-site(Fig. 9, 10).According to the displayed slope (Fig.10), life expectancy for thiscolorimeter would be even worse than at the Dallas-1 site.Observed trends for the reference counts basically disprovedSylgard 184® as the material suitable for the application. Appar-ently, increased sensitivity to moisture/humidity indicated in themanufacturer’s data caused some irreversible changes in the lightpipes and the observed downward trend has confirmed that.The β-test was continued with new colorimeters equipped withlight pipes made out of the more expensive polymer (further calledOE4500), which had been tested in Environmental Chamber at Hach

Company and proved itself very well. This part of the β-test startedon 4/20/07 at Dallas-1 and Dallas-2 sites and on 4/25/07 at UT inAustin, TX. The procedure of the test did not change, thus maininformation was collected from comparison between bench analy-sis and MO42 readings as well as the reference counts trends.Since this silicone material had been extensively tested in the laband environmental chamber, the decision was made to run a short-ened β-test. The test results are presented in two groups similarlyto the first test results – comparison between MO42 readings vs.bench test results and analysis of the reference counts trends (Fig.11 – 16).It is necessary to mention that the customers consciously useddata averaging feature of the analyzer, and therefore, MO42 read-ings sometimes showed smoother curves than the bench test (Fig.12). Nevertheless, as seen from Figures 11 – 13, tracking betweenbench analyses and MO42 readings was very good and the cus-tomers expressed very high level of confidence in the analyzer.At the time of testing the instruments at Dallas-2 and UT sites wereused for monitoring only, while the MO42 at Dallas-1 was initiallyused for partial control of the corrosion inhibitor feed pump. Thepartial control was accomplished by allowing the controller to baseload the tower with corrosion inhibitor at a rate of approximately 80percent of the theoretical demand. In May 2007, the analyzer wasallowed a greater portion of control. The increased level of controlresulted in a greater than 20 percent increase in the number of in-range test results as recorded by the on-site operators and thechemical service representatives. The greater control was accom-plished by reducing the base loading of corrosion inhibitor andallowing the controller to use the instrument data for a much largerportion of the feed. Feed was on/off control based with a dead-band of approximately 0.01 ppm.Growing confidence in the instrument has led to the customers’desire to use the other instruments for control of the feed pumps,too. The feedback was received through both the strategic partnerand directly from the end users during visiting of the β-sites.When asked about spikes or drops in MO42 readings, customersnormally explained it by some maintenance performed on eithercooling towers or the analyzers. As seen from all data (Fig. 5, 6 and11 – 13) there were not many such situations during the entire test,and apparently, it never affected operation of the towers, thereforethe customers never complained. Analysis of possible interferingfactors, such as Temperature, pH, Oxidation-Reduction Potential(ORP), Alkalinity, Hardness, iron and copper ions concentration,which might have caused the spikes, or drops, or otherwise a dis-crepancy between the results, did not reveal any significant oreven observable interference. Some examples of such analysis arepresented and discussed in the end of this report (InterferenceAnalysis).Next several figures (Fig. 14 – 16) show the trends of the referencecounts and comparison with MO42 readings. Drops in the refer-ence counts visible on the charts usually occur due to either dirtysample or the colorimetric cell and do not cause adverse effect tothe concentration readings. This conclusion comes clear from across-comparison between graphs for each β-site, because it isnecessary to see all data: reference counts, MO42 concentrationreadings and bench test results simultaneously (Fig. 11 – 16).


Separate analysis of the reference counts trends presented in fig-ures 14 – 16 and comparison with previously accumulated data onthe OE4500 light pipes life expectancy (discussed in Introduction)confirmed previously established 5-7 years time span, which wasconsidered sufficient for this product (see Figure 17).

Interference AnalysisSeveral examples of trend analyses for potential interference fromother monitored/controlled parameters of the cooling water are pre-sented below. Main parameters analyzed for possible interferencewere identified to be the Alkalinity, Hardness, Temperature, pH,and ORP of the treated water. Concentrations of disinfectants aswell as iron and copper ions present in the sample have also beenanalyzed from the interference point of view. Explanations for con-ducted analyses are mostly provided in the titles for correspondingfigures below. All data have been obtained from the monitoringsystem over the internet.Bromine (generated from reaction between sodium bromite andchlorine) is used as disinfectant at this cooling tower (Dallas-1).The disinfectant may not be strong enough to prevent bio-growth(see Fig. 2 – overflow pipe).Chlorine in form of hypochlorite is used as disinfectant at Dallas-2site and there was no bio-film found in the instrument sample prepa-ration system (see Fig. 4 – overflow pipe).It is interesting to discuss disinfection of cooling water in somemore details. As seen from figures 22 and 23 above as well as fromFigures 2 and 4, different disinfectants might change oxidation-reduction potential (ORP) of the sample in a different manner andsome disinfectants might not be strong enough to prevent bio-growth in the system. However, other factors such as increasedcorrosion ability of chlorine should be taken into account, there-fore application of one or another disinfection agent at specificsites should be thoroughly considered. At UT in Austin, TX chlo-rine dioxide is used for disinfection (Fig. 24) and as expected itprevents formation of the bio-growth very well, however, it mightcause slight interference to the Molybdate analysis (MO42) asshown in figures below.As seen in Figure 25, measured molybdate concentration droppedevery time the disinfectant was injected in system that reflected inthe spikes of ORP, however, the injections did not affect benchanalysis results. According to the customer, the observed drops infirst MO42 readings right after the injections were ~0.05 – 0.07 ppm(outside the specified repeatability), while recorded values werelower (0.02 – 0.03 ppm) due to engaged signal averaging feature.Because pH of the water is always alkaline (Fig. 21), it is assumedthat ClO2 is converted into chlorite ions fairly fast, therefore theinterference might be caused by molecular chlorine dioxide existingin the system in concentration ~ 0.2 ppm (customer’s bench testresults – DPD bleaching method) during the injections (Fig. 26).It is necessary to mention that after change of the main PCB (soft-ware) and the colorimeter, the momentary drops in measured Mo-lybdate concentration (MO42) disappeared (Fig. 27). This observa-tion leads to conclusion that the momentary drops in MO42 con-centration readings outside of the specified precision (±0.03 or 3%,whichever is greater) was possibly caused not by chemical interfer-ence, but either electronic or optical response to the presence ofvolatile compounds such as chlorine dioxide. This phenomenon

has not been seen at the other β-sites using chlorine or bromine fordisinfection.There was no direct correlation found between MO42 readings andsample conductivity (Fig. 28, 29).This actually indicated that the Molybdate tracer concentrationcannot be simply monitored by conductivity of the sample, althoughconductivity provides valuable feedback about other parametersof cooling water.Some correlation was observed between molybdate concentrationand concentrations of iron and copper ions in the sample deter-mined intermittently by the bench testing, however, no interferencewas found (Fig. 30, 31). In the same time, no correlation was ob-served between MO42 and data obtained from newly implementedcorrosion meters for Mild Steel (MS) and Copper corrosion (Fig. 32– 35). Corrosion rate of the metals is expressed in Mills per Year –MPY and reflects presence of the corresponding ions in the sample.The observed inconsistency may be attributed to the trending pro-vided for bench test results by the software and therefore shouldbe disregarded.

Follow Up StudyPerformance of the beta-units was recently checked and it wasfound that in onc case the instrument helped to identify and repaira mechanical break down of one of the protected systems at aDallas site. The instrument started showing multiple spikes in mo-lybdate concentration that was not captured by regular bench test-ing – Figure 36. By a coincidence the bench testing was performedat times when the system was working properly (see the markedareas on the chart, Fig. 36) and the malfunctioning went unnoticedfor some time. After the problem was discovered and mechanicalissue causing chemical to back up in the header was fixed, thesystem returned to normal operation.

Conclusions• Analysis of all data accumulated on MO42 performance dur-

ing two β-tests along with extensive in-lab and environmen-tal chamber testing has shown that the analyzer will performto the advertised specifications in cooling water applica-tions.

• The instrument will be suitable for both Molybdate Tracer(~0.5 ppm) and Molybdate Inhibitor (~3 ppm) applications(concentrations as Mo6+).

• The study has found no interfering factors in cooling waterMolybdate Tracer application.

• The electronic system providing easy access to all data al-lows for efficient remote data monitoring and analysis, aswell as allows controlling the cooling water parameters fromany computer over the internet.

AcknowledgementThe authors wish to thank the operators and managers of the facili-ties where the study was conducted for their help and patience.The special thanks go to Kevan Decker (University of Texas), andDavid McDougall (Dow Corning) for their ongoing understandingand support.


Figures and Tables

Fig. 1 – Reference counts showing rate of decay ofthe colorimeter. Average slope = -9.6 ± 7.4 (based

on calculations for each month of testing andthe LED intensity level).

Table 1. Molybdate Process AnalyzerTechnical Specifications

Fig. 2 – MO42installed at Dallas-1

industrial site.

Fig. 3 – MO42installed at UTPower Plant.

Fig. 4 – MO42installed at

Dallas-2industrial site.

Fig. 5 – Dallas-1: Results of tracking between MO42 andMolybdate bench testing (blue) from the βββββ-test start through4/20/07 when the colorimeter and main PCB were replaced

Fig. 6 – UT: Results of tracking between MO42 and Molybdatebench testing (blue) from the βββββ-test start through 4/25/07 when

the colorimeter and PCB were replaced.

Fig. 7 – Dallas-1: Trend analysis for Reference Counts (blue)vs. Molybdate concentration (MO42).


Whether your project requiresnew construction or retrofit,standard products or customsolutions, Shepherd TowerComponents are a perfect fit.

· PVC Coated Hanger Grids· Stainless Steel Hanger Grids· Gull Wing Splash Fill Slats· V-Bar Splash Fill Slats· Film Pack· Drift Reduction Units· Nozzles & Accessories

C. E. Shepherd Company, L.P.2221 Canada Dry Street

Houston, TX 77023Telephone: 713.924.4300

Fax: 713.928.2324www.ceshepherd.com

[email protected]

Shepherd Standard high quality products for cooling towers include:

Since 1957, our primary business has been innovation!We encourage inquiries for custom product solutions!


Fig. 8 – Dallas-1: Reference counts trend for colorimeterequipped with first tested silicone light pipes (LED level 1).

Fig. 9 – UT: Trend analysis for Reference Counts (blue) vs.Molybdate concentration. Deep drops in the counts

were explained by a dirty cell. Following maintenancebrought it back to normal.

Fig. 10 – UT: Reference counts trend for colorimeter equippedwith first tested silicone light pipes (LED level 1).

Fig. 11 – Dallas-1: Results of tracking between MO42and Molybdate bench analyses (blue) since the new

colorimeter was installed on 4/20/07.

Fig. 12 – Dallas-2: Results of tracking between MO42and Molybdate bench analyses (blue) since the new


Fig. 13 – UT: Results of tracking between MO42 andMolybdate bench analyses (blue) since the new



Fig. 14 – Dallas-1: Reference Counts trend for colorimeterequipped with OE4500 light pipes.

Fig. 15 – Dallas-2: Reference Counts trend for colorimeterequipped with OE4500 light pipes.

Fig. 16 – UT: Reference counts trend for colorimeterequipped with OE4500 light pipes.

Fig. 17 – Reference counts and life expectancy analysisfor all three analyzers equipped with OE4500 light pipes.

Average slope = -6.5 ± 16.5 (both negative andpositive outliers removed).

Fig. 18 – Dallas-1: parallel trends for MO42 (blue) andAlkalinity readings obtained from the analyzer and bench

testing (Alkalinity). No interference observed.

Fig. 19 – Dallas-1: parallel trends for MO42 (blue) and TotalHardness readings obtained from the analyzer and bench

testing (Hardness). No interference observed.


Fig. 20 – Dallas-1: parallel trends for MO42 (blue) andTemperature (°F) readings obtained from the analyzer and

temperature sensor. The temperature has fluctuated within68 - 82°F (20 - 28°C) window that fit into the ideal sampletemperature range (17-27°C). No interference observed.

Fig. 21 – UT: parallel trends for MO42 (blue) and pH readings.No interference observed.

Fig. 22 – Dallas-1: parallel trends for MO42 (blue) andOxidation-Reduction Potential (ORP) readings. No

interference observed.

Fig. 23 – Dallas-2: parallel trends for MO42 readings (blue)and ORP readings. No interference observed.

Fig. 24 – UT: MO42 (blue) and ORP readings obtained througheController. Chlorine dioxide is used as disinfectant at this

cooling tower.

Fig. 25 – UT: an example of momentary negative interferenceof ClO2 injections (blue) to MO42 readings (circled areas).


Fig. 26 – UT: Molybdate (MO42, blue) and Chlorine Dioxideconcentration trends. No direct interference observed.

Fig. 27 – UT: No interference to MO42 readings (red) after thereplacement of the PCB and colorimeter.

Fig. 28 – Dallas-1: Molybdate concentration (blue) vs.Conductivity. No correlation observed.

Fig. 29 – UT: Molybdate concentration (blue) vs. Conductivity.No correlation observed.

Fig. 30 – UT: Comparison between MO42 readings (blue) andbench test results for Iron concentration

Fig. 31 – UT: Comparison between MO42 readings (blue) andbench test results for Copper concentration.


Fig. 32 – UT: Comparison between MO42 readings (blue)and Mild Steel corrosion trend (MPY).

Fig. 33 – Dallas-2: Comparison between MO42 readings(blue) and Mild Steel corrosion trend (MPY).

Fig. 34 – UT: Comparison between MO42 readings (blue)and Copper corrosion trend (MPY).

Fig. 35 – Dallas-2: Comparison between MO42 readings (blue)and Copper corrosion trend (MPY).

Fig. 36 – Mechanical failure captured by the analyzerat one of the Dallas sites


Robert GiammarutiHudson Products CorporationJess SeawellComposite Cooling Solutions, LLC

ABSTRACTToday, owner/operators, OEM’s and suppli-ers are facing lower and lower near and farfield noise limits with respect to their equip-ment. However, lost in this race to see whocan out quiet who is the impact of cost. Spe-cifically – the cost of noise with respect notonly to fans, but the fan mechanical/struc-tural parts as well

· Tolerance on near and far field SoundPressure Level (SPL) and Sound PowerLevel (PWL) noise predictions are +/- 2dB(A).

· Tolerance on cost estimates is +/-10%to 15%.

Additionally the authors wish to emphasizethat, with respect to cooling towers, only fannoise spectrum and reduction was consideredhere and this paper does not attempt to ad-dress the higher frequency water noise. While

The Cost of Noise

Robert Giammaruti Jess Seawell

This paper will look at two specific applications, one a bank ofinduced draft air-cooled heat exchangers and the other being a setof field erected cooling tower cells. In both case studies, the costof lower and lower near and far field noise will be evaluated withrespect to the fan and mechanical and structural components.

INTRODUCTIONWe will look at two case studies here, the first being a 11 bay bankof Air-Cooled Heat Exchangers (ACHEs) and the second being a 8cell counter flow Cooling Tower (CT). Before getting into the de-scriptions of the equipment, we will review the assumptions thatwent into this analysis. The following assumptions apply to bothsystems unless otherwise noted:

· The total airflow and static pressure delivered by the fans ismaintained as noise is reduced.

· Noise reduction is achieved solely by speed reduction ofthe fans and modification of the fans to different blade countsand blade types. No other noise abatement devices wereconsidered and standard motors, gears or drives were em-ployed.

· Near and far field noise predictions are for fans only. Noattempt was made to assess noise generated by the drive/gear systems, motors or waterfall (cooling tower only).

· Inlet conditions and tip clearances of the fans remained con-stant.

· ACHE near field noise were predicted one meter below theACHE bank center with all the fans running

· CT near field noise were predicted between cells 4 and 5along the centerline of the towers, two meters above thedeck level, with all the fans running.

· Far field noise for both the ACHE and CT were predicted at100 meters perpendicular to the long side of the units, 2meters above the ground.

· The noise correlations used were the same for both the ACHEand CT systems.

· No additional bays or cells have been added – plot arearemains constant.

· Heat transfer surface remained constant.· CT water flow remained constant.

fan noise can be analyzed without changing the overall systemresistance, water noise reduction/suppression requires adding at-tenuators to the inlet and/or outlets of the cooling tower. Theseattenuators increase the overall system resistance (i.e. increasesmotor power draw) and thus add a level of complexity that, whileimportant, was outside the scope and purpose of this paper. How-ever, the authors do agree that the subject of cooling tower waternoise should be addressed in the future as a separate paper. Fi-nally, it is acknowledged that other design options are available fornoise reduction such as ACHE/CT redesign or other low noisetechnologies. But as with water noise, our scope here is solelylimited to fan noise reduction.

AIR-COOLED HEAT EXCHANGERDESCRIPTIONThe base design air-cooled heat exchanger (ACHE) described inthis paper (Figures 1 and 2) is a grade mounted, carbon steel in-duced draft item built to the API-661 Standard (Reference 1). TheACHE has a thermal duty of 29.3MW (100 Million Btu/hr) coolinglight gasoline from 60.3C (141F) to 37.8C (100F) at an ambient de-sign temperature of 32.2C (90F). The item consists of 11 individualbays with 12.2 m (40.0ft) long 25.4 mm (1.0 in) OD carbon steel tubeswith extended surface. The extended surface consists of extrudedaluminum fins 15.9 mm (0.625 in) high fins spaced at 10 fins per inch.The tubes are spaced in an equilateral tube pitch of 63.5 mm (2.5 in).The individual bays are 4.98 m (16.34 ft) wide with an overall itemwith of 54.8 m (179.8 ft). Height from the bottom of the tube bundleframe to grade is 2.74 m (9.0 ft).The base mechanical fan drive systems consists of 25 HP, 60 HZ,single speed motors, synchronous belt speed reduction, and 3.96m (13 ft) fiberglass reinforced plastic (FRP) fans with 4 blades.Each bay has two mechanical drive systems with the entire itemcontaining twenty-two such systems.

Figure 1. Induced air-cooled heat exchanger front (inlet) view


Figure 2. Induced air-cooled heat exchanger side view

COOLING TOWER DESCRIPTIONThe base CT described in this paper is an 8 cell, in line, induced,mechanical draft cooling tower (Figures 3, 4 and 5) designed to theapplicable CTI standards and guidelines (References 2,3). Thetower structure is constructed with fire retardant FRP structuralcomponents and incorporates 304 SS hardware for all structuraland mechanical connections. The roof deck is FRP with a non-skidsurface applied. The interior and exterior casing is 12 oz – fire retar-dant (FR), FRP casing. The tower includes two – FR, FRP stair-ways, one at each end of the tower and one FR-FRP ladder andcage, located in the center of the tower. Cell size is 14.63 m x 14.63m (48 ft x 48 ft), with 1.83 m (6 ft) of low fouling PVC film fill that isbottom supported. The drift eliminators are PVC, cellular type, 0.40mm (0.015 in) thick; with a maximum allowable drift rate of 0.0015 %of design flow. The tower design flow is 401,254 L/min (106,000GPM) total with entering hot water of 37.8 C (100 F), exiting coldwater of 29.4 C (85 F) and a design wet bulb temperature of 25.6 C(78 F).The base mechanical fan drive systems consists of 250 HP, 480VAC, 60 HZ, single speed motors, double reduction gear reducers,composite drive shafts and 10 m (32.8 ft), FRP fans with 9 blades.The FR-FRP fan stacks are velocity recover type design, 10 m (32.8ft) in diameter and 3.05 m (10 ft) in height. Controls consist of lowoil level cut off switches and vibration switches with manual andremote reset features mounted on the gear reducer. The tower isinstalled on a customer supplied concrete basin with a basin depthof 1.22 m (4.0 ft).

Figure 3. Cooling tower front view

Figure 4. Cooling tower top view

Figure 5. Cooling tower end view

CASE 1 – AIR-COOLED HEAT EXCHANGERTable 1 lists the common operating parameters of all the ACHE fansstudied. Table 2 lists the base case fan condition along with thelower noise options listed in order of decreasing noise.

Table 1. ACHE fan-operating parameters.

Table 2. ACHE fan-operating conditions in order ofdecreasing noise.

As one can see from Table 2, the fan speed was lowered to approxi-mately 50% of the base design while the fan shaft power remainedfairly constant. Fan speed was varied in approximately 10% incre-ments by a combination of drive ratios and motor speeds to achievethe required noise reductions. Multiple fan selections were per-formed for the STD, VLN and ULN fans types as applicable toachieve the overall most cost effective solution at that particular


speed. Cost increases were estimated for fan Cases 1 through 8taking into account changes (if any) in fan, motor, drives, mechani-cal structure and fan rings. The cost increase estimates were thenplotted against the near and far field SPL noise predictions as shownin Figure 6 below.

Figure 6. ACHE bay cost increase vs. near and far field SPLThe overall ACHE cost per bay significantly increases as noise isdecreased with the lowest noise option increasing the per bay costby approximately 35%. Given the non-linearity of the data (a gen-eral 3rd order polynomial was used to trendline the data) one cansee that the cost increase accelerates faster for a relative steadydecrease in noise. Put another way, the first 10 dB(A) of far fieldnoise reduction increased the overall cost per ACHE bay by about4% from the base design. However the next 10 dB(A) increased thecost 20% from the base design.The majority of this increase is related to the fan and depending onthe type of fan (VLN or ULN – See Figure 7 for a general sizecomparison) will range from 70 to 90% of the total increase. This isshown below in Figure 8 where the costs increases are plotted as afunction of fan sound power level.

Figure 7. Size comparison of standard noise (foreground),very low noise (middle) and ultra low noise (back) fan blades

The variance shown in Figure 8 is simply a function of the designpoints chosen. For example, at the 92.7 dB(A) SPL point, nearly30% of the cost increase is due to non-fan components due to thefact that the fan, in this instance, did not change as the speed wasreduced. However, when looked at in total, this combination of fan,motor, drives and so forth provided the least amount of overall costincrease. Another factor is the fan material. Aluminum fans tend tobe less costly than FRP fans and will move the cost split moretowards the non-fan components. A good rule of thumb is that thefan will account for approximately 70% (Aluminum) to 85% (FRP) of

the total cost increase as noise is reduced. However, every situa-tion is unique and the owner/operator is encouraged to investigateall possibilities with the OEM/Supplier.

CASE 2 – COOLING TOWERTable 3 lists the common operating parameters of all the coolingtower (CT) fans studied. Table 4 lists the base case fan conditionalong with the lower noise options listed in order of decreasingnoise.

Table 3. Cooling Tower fan-operating parameters.

Table 4. Cooling Tower fan-operating conditions in order ofdecreasing noise.

As one can see from Table 4, the fan speed was lowered to approxi-mately 60% of the base design while the fan shaft power remainedfairly constant until Cases 6 and 7. Here the lower efficiencies ofthe ULN fans significantly increased the required fan shaft power.This is not uncommon for these types of fans as the goal is nor-mally lowest possible noise, not power optimization. It should benoted here that it is highly unusual for a client to accept 300HPmotors and in most cases, additional cells or other methods will beemployed by the CT OEM to stay at or below 250 HP motors.However, for the purposes of this analysis we have assumed thatthe client will accept the larger motors.

Figure 8. Percent cost increase split between fans andmechanicals


Fan speed was varied in approximately 10% increments (Base Caseto Case 4) by a combination of gear ratios and motor speeds toachieve the required noise reductions. Cases 5, 6 and 7 were thelowest speeds possible for the given CT operating conditions.Multiple fan selections were performed for the STD, LN, VLN andULN fans types as applicable to achieve the overall most costeffecting solution at that particular speed. Cost increases wereestimated for fan Cases 1 through 7 taking into account changes in(if any) fan, motor, gears, mechanical structure and fan stacks. Thiscost increase estimates were then plotted against the near and farfield noise predictions as shown in Figure 9.The overall CT cost per cell significantly increases as noise is de-creased with the lowest noise option increasing the per cell cost byapproximately 30%. Given the non-linearity of the data (a general3rd order polynomial was used to trendline the data) one can seethat the cost increase accelerates faster for a relative steady de-crease in noise. Put another way, the first 10 dB(A) of far field noisereduction increased the overall cost per CT cell by about 9% fromthe base design. However the next 3.4 dB(A) increased the cost28% from the base design.

Figure 9. CT cell cost vs. near and far field SPL

The majority of this increase is usually related to the fan and, de-pending on the type of fan, will range from 20 to 80% of the totalincrease. This is shown below in Figure 10 where the costs in-creases are plotted as a function of fan sound power level.

The variance shown in Figure 10 is simply a function of the designpoints chosen. For example, at the 101.5 dB(A) SPL point, nearly80% of the cost increase is due to non-fan components due to thefact that the fan cost, in this instance, did not increase that muchrelative to the speed reduction. In this case, it was the gear drivingthe cost increase, as it was necessary to move up to the next boxsize. However, when looked at in total, this combination of fan,motor, gears and so forth provided the least amount of overall costincrease. A good rule of thumb is that the fan will account forapproximately 60% of the total cost increase as noise is reduced.However as mentioned in Case 1, every situation is unique and theowner/operator is encouraged to investigate all possibilities withthe OEM/Supplier.

SUMMARYAs presented in the two cases herein, the cost of noise, in this caselower noise, significantly impacts the cost of an ACHE or CT de-pending on how low and where the noise guarantee points arelocated. And while significant near and far field noise reductions isachievable (in these cases without adding ACHE Bays or CT Cells),as the curves show, the increase in cost is not a linear function butrather a 3rd order polynomial that accelerates quickly as one re-duces the noise levels.While this analysis is far from exhaustive, the authors hope thispaper provides the reader with some perspective in the realm ofACHE and CT fan noise. As the old adage goes “there are no freelunches” with the same being true here. However, in this instance,the cost of “lunch” will probably increase rapidly compared towhat you receive in return.

REFERENCES1. American Petroleum Institute, “Air-Cooled Heat Exchang-

ers for General Refinery Service,” API Standard 661, SixthEdition, February 2006.

2. Cooling Technology Institute Standards 111,131, 136, 137and 153.

3. Cooling Technology Institute ESG 152

Figure 10. Percent cost increase split betweenfans and mechanicals


Toby DaleyComposite Cooling Solutions, L.P.

AbstractArchitectural enclosures for cooling towers are notnew phenomena. Ideal clearances are provided bymanufacturers to achieve the rated performance.However predicting the tower capability when theclearances are less than ideal can be complex. Thispaper will present a study of an architectural louverenclosure and its influence on the performance ofthe tower when less than ideal clearances are achiev-able.

Introduction

stantly shrinking. The end result is the tower isplaced in a location that is usually not free from in-fluences caused by obstructions.This paper will present some fundamental consider-ations involved which should be addressed whenplacing a cooling tower in an enclosure and the po-tential influences on performance if not properly ad-dressed.

Types of EnclosuresDepending upon the type of facility there are twobasic types of enclosures or a combination of thetwo.

Basic Types of Enclosures

Architectural Enclosure Influences onthe Performance of Field ErectedCounterflow Cooling Towers

Locating a cooling tower inside of an enclosure is not a new idea. Ithas been the goal to hide the cooling tower from sight especially onarchitecturally sensitive projects. As a result, architectural con-crete cooling towers were introduced. Utilization of these towershas created some very remarkable and impressive architecturalstructures that blend in with the building. As construction costsincreased for concrete towers in the 1970’s and 1980’s a new type oftower constructed of Pultruded fiberglass was introduced. It wasdesigned to replace concrete construction. The Pultruded fiber-glass tower provided the durability and longevity of concrete butsubstantially less in weight. The additional benefit was the abilityto create an architecturally pleasing tower which would not requirea screen wall or enclosure. Thousands of concrete towers havebeen constructed since the 1940’s and architectural fiberglass tow-ers from 1980. However, the vast majority for HVAC and architec-tural applications are still enclosed in some type of a surround orenclosure wall. The types of towers range from factory assembledto field erected and crossflow to counterflow and metal to wood tofiberglass. The purpose of the enclosure is to make the tower dis-appear and/or provide sound attenuation. However, in most casesthe pursuit of hiding the tower overrides the basic need of a cool-ing tower which is to receive and discharge the required air volumein a uniform pattern.In the past ten to fifteen years HVAC cooling tower applicationshave grown in size in response to chiller manufacturers supplyinglarger chillers. Thus, it is not uncommon today to see massivecentral district cooling plants approaching 650,000 tons in capacityand utilize large field erected counterflow cooling towers. They arelimited in available real estate and located in populated areas andsubsequently the towers are located on the roof of the plant andmust be enclosed. Sound is an issue in many locations and soundattenuators are often added to the inlet and discharge which addsadditional air blockage to the tower. In most cases the roof is alsolimited in space and thus the space for the cooling tower is con-

· Solid Wall – The tower is completely surrounded by a solidwall.

· Porous Wall – The tower is completely surrounded by awall that air can pass through.

Solid Wall Enclosure -A solid wall enclosure can only receive air from above. Thus, thecooling tower is discharging air into the same air space that thetower is trying to pull air into the air inlets. The obvious problemwith this type of installation is the tendency for the discharge air torecirculate into the air inlets and thus degrading the thermal perfor-mance. If the enclosure wall is higher than the top of the fan stackthen the recirculation influence is much greater. This would bemost prevalent when cross winds and/ or inversions are occurring.Porous Wall Enclosure –This enclosure can be constructed of limitless materials and shapes.When a cooling tower is placed in this enclosure the air is drawnboth through the louver wall and vertically from above the tower.The following lists the most basic types.Basic Types of Porous Wall Enclosures –

· Architectural Concrete Wall – This type is usually a geo-metric pattern of openings built into a wall. The entire wallcan be porous or it can be a combination of porous andsolid.

· Architectural Louver Panel Wall – This type is usuallyconstructed of commercially available panels comprised ofblade type louvers. The panels can be oriented in an exhaustor intake configuration. The intake configuration is usedwhen air is entering the enclosure. The exhaust configura-tion is used when air is discharged or venting an enclosure.Depending upon the size of the project and the local struc-tural design parameters the panels can require a substantialamount structural steel framing for support.

Toby Daley


· Perforated Screen Wall – This type of panel is constructedof a perforated material which may be combined with freeopen areas to create the architectural effect. This type ofenclosure will also require structural framing to provide sup-port for the screen.

Solid Wall Enclosure

Architectural Louver Wall

Perforated Metal Screen Enclosure

Free Open Area, Blockage and ClearancesThe following sketch presents a typical installation where the toweris located in a restricted environment.

When rating or selecting a tower for placement within a restrictedenclosure there are three basic influences that must be considered.Basic Influences on Performance –

· Net Free Open Area – The net free open area of the airpathway to the air inlets of the tower must be taken intoaccount.

· Blockage – Resistance in air flow can exist due to enclosurestructural framing, external walkways, common header pip-ing, valves, and pumps. These block the air path and in-crease the inlet velocity, pressure, and cause mal-distribu-tion of air.

· Clearances – The clearance between the tower air inlet andthe enclosure wall determines how the air volume will bebalanced between the air approaching from above the towerand / or the air approaching through the porous wall.

The Impact on Thermal Performance can occur in two ways.Type of Influence –

· Increased Pressure Drop – The tower as located could pos-sibly not have an ambient air inlet pressure as a startingpoint. Thus, additional system pressure prior to the air inletwas not included in the tower design. This would translateto not being able to achieve the required air flow for thetower.


· Mal-distribution of Air – If the tower does not have properclearances or is influenced by blockage then the air enteringthe tower may not be as uniform as a free field environmentunder which the tower was rated. The mal – distribution cantranslate into areas of the fill material being bypassed orunderutilized. This can cause two problems 1) increasedstatic pressure as a result of high localized velocities and 2)hot water bypass in areas of the fill that are being bypassed.

· Recirculation – A major impact on performance is the recir-culation of the discharge air back into the fresh air inlets.Recirculation increases the wet bulb temperature of the airentering the tower. If the tower was designed for a 78° F InletWet Bulb temperature (IWBT) and the recirculation is in-creasing the IWBT to 80° F then the tower performance ca-pability will still be 100 % based upon the manufacturer’sperformance curves. However, colder water could be achievedif the recirculation was not occurring. Thus, the energy effi-ciency of the total system suffers. Recirculation can occurdue to wind forcing the discharge air stream to turn backinto the enclosure and the tower air inlets. If the clearancebetween the tower and the enclosure walls is not sufficient,then the greater the downward air velocity next to the wall.The higher this downward velocity compared to the towerdischarge velocity, the greater the recirculation tendency.Secondly, the height of the enclosure wall also influencesthe magnitude of recirculation. If the wall exceeds the heightof the stack, then the greater the tendency of mild windconditions to cause recirculation of the hot discharge airback into the fresh air intakes.

Structural Framing for support of the louver wall can present majorblockage for the tower air inlet.

Air Inlet Blockage Due to Structural Framing for LouverEnclosure

Roof Top Installation with Louver Wall Enclosure

Architectural Louver Selection-The architectural louver should be selected with consideration ofthe tower type and available clearance. If the tower is located lessthan the manufacturers recommended distance from the louver wallthen the louver can directly impact the tower performance. It isoften the misconception that if you select a louver with a reason-able percent free open area then the tower will be okay. However, anarchitectural louver can have an acceptable percent free open areabut still have an unacceptable pressure drop.The direction of the louver can also impact how the air enters thetower. Improper orientation can cause the air to take multiple turnscausing an increase in pressure and potential air mal-distribution.It is important to understand the basis of the louver manufacturer’spressure drop curves. If it is based upon net velocity then thepressure drop should be calculated using the velocity associatedwith the predicted louver air flow volume and the net free openarea. The net free open area should also include reduction for struc-tural framing which blocks the tower air inlet. Additionally, there isa pressure drop curve for an intake or an exhaust orientation.The following is an example of three architectural louvers and therange of variability that can occur when making a selection basedupon free open area only.

Louver Model B and C are very close to the same free open area butLouver C has a 34% higher static pressure due to its more restric-tive shape. Louver A has a velocity pressure coefficient less thanlouver B but only 37.6% free open area. It has the highest staticpressure because of the higher net velocity.Air Balance Equilibrium and Effective Louver Height -When a tower is located in a louvered enclosure the air flow throughthe air paths will reach an equilibrium point. At equilibrium theactual amount of air that is delivered through the louvers andthrough the vertical well (the area between the tower and the louverwall) will be established.


At equilibrium the effective height of the louver wall will not be thefull height of the wall. If, for example, the height of the louver wall is30 feet only a portion of the wall would be utilized by the air path.This effective louver height is different for each installation and isdependent upon the parameters previously discussed. If you wereto expect all of the required air to pass through the louver wall andchose a louver with a free open area of 50% then you would needan effective louver area equal to twice the tower air inlet area tosimulate a free field environment. Or the net velocity of the louverwall equals tower air inlet velocity.

Selecting the Proper Distance -Selecting the correct distance to locate the tower from the louverwall influences the tower performance. Whether in a solid or po-rous wall arrangement the velocity and trajectory of the air ap-proaching the air inlet will determine the uniformity of the air flow inthe tower. If the tower has louvers the type of louvers will alsoinfluence the trajectory. Tower manufacturers have recommendedvelocity limits for the well and louver velocity which ultimatelydefines the required distance. If these distances are achievablethen the tower rating is applicable. The following table presentssome example of published velocity limits.

The following presents net velocity limits based upon experiencewith architectural louver installations and large counterflow cool-ing towers and practical clearances that can be achieved.

A modeling system was developed which solves for the equilib-rium point of the air flow streams for specified distances away fromthe tower. The tower, enclosure and louver characteristics are en-tered into the model. The analysis results in a recommended dis-tance from the enclosure wall. The following is the model for theexample tower, installation and Louver A.

\

In this example only 20% to 30% of the air is passing through thelouvers. Thus, 70% to 80% of the air is coming from the verticalwell. Even if the tower was located closer to the louver wall the airflow would equalize at only 38% to 48% through the louvers andthe well velocity would be on the order of 950 to 1100 FPM. Thus,the velocity limits not only influence the tower performance byimproved air flow distribution but also by minimizing recirculationresulting from high vertical well velocities.


However, if the recommended distance is not achievable then amore complex study involving computational fluid dynamics mod-eling is required to predict the deficiency in performance. It is notuncommon for the tower capability to be reduced to 85% when itnot properly placed in the enclosure. The graph below presents thepotential deficiency if the tower and louver are not optimized forthe installed clearances.

Enclosure Checklist for Field Erected CoolingTowersThe following is a recommended checklist when designing an en-closure for a field erected counterflow cooling tower.

· Select the type of enclosure such as solid, louvered, porousor a combination.

· Define the space available for the tower and walls that willbe solid, louvered or porous.

· Select the architectural louver type and obtain the % freeopen area and static pressure curves.

· Determine the structural framing required for both horizon-tal and vertical support of the louver wall that needs to befactored into the net free open area.

· Determine in general the plan for the tower supply and re-turn piping that needs to be factored into the net free openarea.

· Determine in general the plan for access to and around thetower that needs to be factored into the net free open area.

· Contact the cooling tower manufacturer and provide theabove information along with the thermal duty requirementsand any special requirements such a specific number of cells,horsepower limits, etc.

IF YOU BLOCK THE AIR TO THECOOLING TOWER

THE THERMAL PERFORMANCEWILL BE IMPAIRED

· Request that the tower manufacturer provideyou with recommended –o Size and quantity of cells for the thermal

duty.o Clear distance required from the tower to

the enclosure walls.o The predicted capability of the tower if the

available space cannot accommodate therequired distance.

ConclusionsIn concluding, the performance related influ-ences can often be overlooked when placingcooling towers in restrictive enclosures. Thekey variables to address are air pathways, ve-locities, trajectories, clearances, obstructionsand recirculation. The prediction of these influ-ences is complex but is predictable using cus-tom analysis techniques. A successful result isdependent upon properly selecting the louversand tower for optimum performance in the in-stalled environment. If an analysis is not per-formed then at least consideration should begiven to increasing the tower size in an attemptto compensate for these influences.


Gary Geiger, Caroline SuiWater and Process TechnologiesGE Infrastructure4636 Somerton Road, Trevose, PA 19053

ABSTRACTInorganic phosphate is the most widely used mildsteel corrosion inhibitor for open recirculating cool-ing water systems. However, effective control of cal-cium phosphate precipitation must be maintainedboth in the recirculating cooling water and at heatedsurfaces, if corrosion is to be controlled without aloss of heat transfer efficiency. Over the past 30 yearsnotable advances have been made in polymeric dis-persant technology that have improved calcium phos-

Improved Calcium Phosphate ControlFor Stressed Systems

Inorganic phosphate is the most widely used steelcorrosion inhibitor for open recirculating coolingsystems. Orthophosphate suppresses both the an-odic and cathodic corrosion reactions. High dos-ages of orthophosphate (12-20 ppm) catalyze theformation of a protective oxide film, that retards theanodic corrosion reaction.2.3 Orthophosphate re-stricts electron transfer at the cathode through theformation of an insoluble calcium phosphate film.Orthophosphate-based corrosion programs may in-clude the addition of polyphosphate (pyrophos-phate or hexametaphosphate) and/or zinc to enhancecathodic corrosion inhibition and guard against lo-

phate control. This paper discusses the performance of a recentlydeveloped polymer that has shown excellent performance understressed conditions.Keywords: deposit control, corrosion control, stabilized phosphate,polymer, phosphate, azole, cooling water, cooling tower, high cycle,and stressed water condition

INTRODUCTIONCooling water treatment programs must control corrosion, deposi-tion and microbial activity to maintain heat transfer efficiency andavoid premature corrosion failures of process equipment. All threeareas of concern are interrelated and must be simultaneously ad-dressed. Excessive steel corrosion will result in flow restrictionsand the accumulation of iron corrosion products that impede heattransfer. Soluble iron released at the corroding surface can poisonpolymeric dispersants, resulting in diminished efficacy and the for-mation of iron phosphate deposits. Deposits not only impede heattransfer, but also are responsible for premature corrosion failuresdue to under-deposit corrosion mechanisms. Deposits provide asafe haven for anaerobic bacteria that are responsible for microbio-logically influenced corrosion (MIC). Both sessile and plaktonicbacteria control are necessary to prevent slime forming bacteriafrom establishing biofilms that restrict heat transfer and provide aprotective environment for the colonization of anaerobic bacteria.Open recirculating cooling water circuits are typically mixed metal-lurgy systems containing both carbon steel and copper-based al-loys. Electrochemical corrosion of these metals involves both an-odic and cathodic corrosion reactions. Metal loss occurs at theanode and oxygen reduction occurs at the cathode. (Oxygen re-duction at the cathode is the primary reaction with air-saturatedwaters operating at neutral to alkaline pH.) For ferrous metals,soluble iron is released at the anode and hydroxide anions aregenerated at the cathode. For corrosion to proceed, the anodic andcathodic reactions must occur simultaneously and at the same rate.Corrosion protection is achieved by inhibiting either or both reac-tions.1

calized corrosion. The use of polyphosphate is favored wheredischarge restrictions limit or preclude the use of zinc. Like ortho-phosphate, polyphosphates form insoluble calcium salts at the cath-ode due to the localized high pH.Corrosion inhibitor programs for mixed metallurgy systems con-taining both copper and iron alloys will include an azole. Azolesform a protective chemisorbed film on the copper surface and in-hibit the cathodic corrosion reaction, although there is some evi-dence that they also retard the anodic reaction.4,5 If copper corro-sion cannot be adequately controlled, steel corrosion will suffer.As copper corrodes, copper ions enter the bulk cooling water andwill plate on carbon steel surfaces forming galvanic corrosion cells.This results in severe localized corrosion (pitting-type) of the steelsurface. The commercially available azoles are benzotriazole (BZT)and tolyltriazole (TTA). A proprietary halogen resistant azole (HRA)is also available.6,7,8

The use of inorganic phosphate for steel corrosion protection re-quires an effective calcium phosphate precipitation inhibitor tomaintain the phosphate soluble in the bulk cooling water and pre-vent deposition at heat transfer surfaces. Effective polymeric in-hibitors/dispersants for calcium phosphate were first developed inthe late 1970’s. Since that time, a wide variety of copolymers andterpolymers have been introduced that have expanded the rolefrom calcium phosphate inhibition to particulate fouling control.However, the primary role of the polymeric dispersant in an inor-ganic phosphate-based program is to prevent calcium phosphateformation. Calcium phosphate demonstrates retrograde solubilitywith both pH and temperature. At any given level of calcium hard-ness and phosphate, the dosage of the dispersant is dictated bythe temperature of the hottest process equipment and the operat-ing pH range, if scaling is to be avoided. Precise pH control isrequired to minimize high pH swings and avoid exceeding the con-trol capabilities of the inhibitor.Microbiological (MB) control is extremely important in maintainingefficient and reliable equipment operation. If unchecked, microor-ganisms can accumulate and form slime deposits on heat transfer

Gary Geiger


equipment, transfer piping, tower fill and decking, etc., causing aloss of heat transfer efficiency as well as promoting under-depositcorrosion. Some microbes also pose significant health concernswhile others can degrade water treatment chemicals. There arenumerous approaches for MB control including the use of halo-gens (chlorine, bromine), organic biocides, (both oxidizing and non-oxidizing), ionization, and UV (electromagnetic radiation). Of these,chlorine is by far the most cost-effective, being far less expensivethan non-oxidizing biocides and certain stabilized bromine treat-ments. Chlorine gas, when dissolved in water, hydrolyzes quicklyto form both hypochlorous and hydrochloric acids. The former isalso generated when using hypochlorite solutions (bleach) insteadof dissolving chlorine gas in water. Hypochlorous acid is a weakacid, which can undergo further dissociation to a hydrogen ion anda hypochlorite ion. While both hypochlorous acid and the hy-pochlorite ion are oxidizing agents, the former is more germicidal.Its effectiveness is attributed to the ease in which the molecule canpenetrate the cell wall of a microorganism. Hypochlorous acid hasa similar structure to that of water and can thereby permeate the cellwall. Both hypochlorous acid and hypochlorite ion can degradeorganic deposit control agents and corrosion inhibitors. Chlorinetolerant treatment additives can allow the unrestricted use of chlo-rine for MB control, reducing the costs associated with non-oxidiz-ing biocides. If the organic treatment components (polymeric dis-persant, azole) are susceptible to chlorine oxidation, the free chlo-rine residual must be closely control to avoid excessive treatmentdegradation.This paper focuses on laboratory generated application data for arecently developed stress tolerant polymer (STP) for neutral pH,high phosphate treatment programs. STP represents the culmina-tion of over 30 years of GE Water & Process Technologies’ polymerresearch.

LABORATORY SCREENING & POLYMEREVALUATIONThe advanced scale control chemistry, referred as stress tolerantpolymer (STP), was evaluated in the laboratory and compared to asulfonated acrylic acid copolymer (SAA) and other commercialcopolymers and terpolymers, designated polymer A, B, C, and D.The majority of the comparative studies were performed with theSAA polymer since its performance was comparable to that of theother commercially available polymeric dispersants in calcium phos-phate beaker studies. The SAA polymer is one of the most widelyused commercial calcium phosphate inhibitors.Evaluations were conducted using a static beaker test protocol anda dynamic Bench Top Unit (BTU) protocol. Static beaker tests aredesigned to be a screening tool that can quantify the relative effi-cacy of calcium phosphate inhibitors. The BTU is a recirculatingsystem incorporating a heat transfer surface to assess deposition.Both testing methods have been described in prior papers.9 BTUtests are designed to simulate a severe system condition. Bulkwater and heat transfer surface temperatures are controlled, pH ismonitored and controlled, and corrosion rates of low carbon steel(LCS) and Admiralty brass (ADM) are measured with test coupons.The BTU is not an evaporative system, so it does not simulate anopen recirculating system. The BTU simulates a heat exchangerthat is constantly recirculating hot water (120OF) for cooling. Theunit uses a heat exchanger tube fitted with an electrical heater to

adjust the surface temperature and a cooling coil to remove excessheat and maintain the temperature of the recirculating water. Thesystem retention time (holding time) is controlled by adjusting therate of makeup to the water reservoir. Test results using the twomethods are discussed in detail in the paper.

1. Static beaker testsA. Calcium Phosphate Inhibition TestingCalcium phosphate inhibition studies were conducted under stan-dard water conditions. The standard test water contained 400 ppm(mg/L) Ca, 100 ppm (mg/L) Mg, 35 ppm (mg/L) M-alkalinity (all asCaCO3), 10 ppm (mg/L) orthophosphate, and the polymer underevaluation. Each solution was adjusted to pH 8.2 and placed in awater bath controlled at 158OF (70OC) for a period of 18 hours. Thesolutions were then filtered through a 0.2 micron filter and the fil-trate analyzed for soluble phosphate. Percent inhibition of calciumphosphate was then calculated for each polymer relative to a con-trol sample without treatment. The polymers were evaluated at dos-ages of 5, 10 and 15 ppm (mg/L). The stress tolerant polymer (STP)was compared to a sulfonated acrylic acid copolymer (SAA) andfour other commercially available polymers (A, B, C & D).As shown in Figure 1, the STP polymer was highly effective even ata dose rate as low as 5 ppm. SAA provided greater than 95%inhibition at 10 ppm, while the other commercial polymers gave lessthan 10% PO4 inhibition at dose rates equal to or less than 10 ppm.These results clearly demonstrated the superior performance of theSTP compared to SAA and other commercial polymers.B. Calcium Pyrophosphate Inhibition TestingPyrophosphate is included in many neutral pH programs to fortifycathodic corrosion inhibition. The use of pyrophosphate is limitedby calcium pyrophosphate solubility concerns, similar to ortho-phosphate. To exploit its properties as a corrosion inhibitor, thepyrophosphate dosage must be adjusted based on the calciumhardness of the cooling water and take into consideration its rever-sion to orthophosphate. Polyphosphates revert (hydrolyze) toorthophosphate and the degree of reversion is a function of tem-perature, pH and system holding time.Calcium pyrophosphate inhibition efficacy was evaluated at pH 8.2with a 400 ppm (mg/L) calcium hardness water containing 10 ppmpyrophosphate (as PO4). The solution was placed in a water bathcontrolled at 158OF (70OC) for a period of 18 hours. The solutionwas then filtered through a 0.2 micron filter and analyzed for totalinorganic phosphate. The percent inhibition was calculated for theSTP and SAA polymers relative to a control sample without treat-ment. The polymers were evaluated at dosages of 5, 7.5, 10, and 15ppm (mg/L). The degree of reversion of pyrophosphate was deter-mined by measuring the amount of orthophosphate at the end ofthe test period. Reversion was less than 20% (2 ppm).A comparison of polymer performance demonstrates that the STPpolymer provides a greater degree of stabilization than the SAAcopolymer. The SAA copolymer was ineffective below a dosage of10 ppm, while the STP polymer provided nearly complete inhibitionat the lowest dosage tested (5 ppm). Figure 2 summarizes the dose-response data.

2. Bench Top Unit EvaluationsThe STP and SAA polymers were further evaluated at neutral pHunder recirculating, heat transfer conditions in the BTUs. The BTUs


were conducted under standard and stressed cooling water condi-tions. The standard water composition and operating parametersare shown in Table 1. The cooling treatment contained 15 ppm (mg/L) orthophosphate, 3 ppm (mg./L) pyrophosphate (as PO4), and 1.2ppm (mg/L) Halogen Resistant Azole (HRA), unless specified oth-erwise. The treatment program is designed for a near neutral pHrange (6.8-7.4). High levels of inorganic phosphate were used toensure excellent steel corrosion protection. This was necessary topreclude the possibility of iron “poisoning” of the polymer due toexcessive corrosion. Polymer dosages were varied based on testconditions and the polymers used. The test duration was 7 days.Samples were taken from the sump daily for water chemistry char-acterization including turbidity, ortho-PO4 and ICP analysis of bothfiltered (F) and un-filtered (UF) samples. At the end of each BTUevaluation, the corrosion coupons and heat exchanger tube werevisually inspected for deposits. Coupon corrosion rates were de-termined by a weight loss method.A. Polymer Dosage Requirements As A Function of pH And Tem-peratureTesting was directed at determining the minimum polymer dosagerequired to control calcium phosphate precipitation in the recircu-lating water and prevent deposition on the heat transfer surface.Maintaining phosphate solubility is necessary to ensure steel cor-rosion protection. Table 2 provides a summary of the data.Studies performed at standard conditions, over a pH range of 7.2-7.8 (Tests1-8) demonstrate the superior dose-response performanceof the STP polymer over that of the SAA copolymer. Although theSAA copolymer provided excellent scale control, the STP polymerachieved the same level of performance at a much lower dosage(33-50%). At pH 7.2, 4 ppm (mg/L) SAA was required while only 2ppm STP ppm (mg/L) was needed. As the pH increased from 7.2 to7.8 the polymer requirement increased significantly. Calcium phos-phate deposition control was achieved with 18 ppm (mg/L) SAAvs.12 ppm (mg/L) STP, at pH 7.8. The effect of pH on calciumphosphate solubility in relation to the polymer requirement is clearlyevident in these studies. As the calcium phosphate super-satura-tion increases with pH, higher levels of polymer are required tocontrol scaling. Good pH control is necessary to minimize pHswings that lead to scaling. In the absence of good control, thepolymer dosage must be increased when the pH exceeds the upperlevel of the pH control range. Alternatively, the phosphate dosagemust be reduced or the cycles of concentration decreased.The relative performance of the two polymers was further investi-gated at the pH 7.2 water condition with heat exchanger tube sur-face temperatures from 130OF (54.4OC) to 160OF (71.1OC). Increas-ing the surface temperature from 130 OF required a slight increase ofthe STP dosage from 2 to 3 ppm (mg/L) at 140OF (60OC), and to 4ppm (mg/L) at 160OF to maintain calcium phosphate scale control(Test 1, 10 & 12). The SAA copolymer controlled phosphate depo-sition at the 140 OF condition at a dosage of 6 ppm, but was unableto maintain the heat transfer surface free of deposits at 160OF at 8ppm.B. Impact of pH Upset On Calcium Phosphate Deposition ControlThe ability of the STP and SAA polymers to recover from a loss ofacid feed was evaluated under standard conditions. Testing wasinitiated at a pH of 7.2 and the pH incrementally increased to 7.9 tosimulate the loss of acid feed. The pH was then decrease back to

7.2. The pH excursion was limited to pH 7.9 to avoid calcium car-bonate precipitation/deposition. The soluble orthophosphate (0.2micron filtrate) was monitored to track calcium phosphate inhibi-tion efficacy. The STP and SAA polymers were applied at 4 and 8ppm (mg/L), respectively. Previous testing demonstrated that thesedosages were sufficient to prevent both bulk water precipitationand deposition of calcium phosphate at pH 7.4. Figures 3 & 4provide a summary of the data.Precipitation of calcium phosphate occurred when the pH exceed7.4. This was not unexpected based on the previous pH dose-response testing (Table 2). The loss of phosphate from the recircu-lating water was nearly identical for both polymers when the pHwas increased to 7.9. Soluble orthophosphate levels decreasedfrom an average of 16 ppm to 8 ppm, indicating a complete loss ofprecipitation control. However, as the pH was decreased from 7.9back to 7.4, the soluble orthophosphate increased to its initial valuewith the STP treated system, but not with the SAA treated system.Soluble phosphate levels remained at 14 ppm with the SAA poly-mer and did not increase with time, indicating continued loss ofphosphate. All of the phosphate (16 ppm) was recovered with theSAA polymer when the pH was decreased to 7.2 and an additional8 ppm polymer was added.Figure 5 provides a visual record of the comparative depositionperformance of the two polymers. As can be seen, calcium phos-phate deposition is significantly less with the STP copolymer.Calcium phosphate precipitation and deposition resulted in a sig-nificant decrease in the active polymer. The active (free) SAAcopolymer level decreased from an initial value of 8 ppm to 2 ppm asthe pH increased from 7.4 to 7.9, while the STP dosage decreasedfrom its initial value of 4 ppm to 1 ppm. Polymer loss is attributed toadsorption on calcium phosphate particulate in the circulating wa-ter and on surface deposits.Under actual field operating conditions the appropriate responseto an alkaline pH excursion would be to increase the polymer dos-age. However, the STP polymer offers a significant degree of pro-tection compared to the SAA polymer when the pH drifts out of thecontrol range.The necessity of increasing the polymeric dispersant dosage whenthe calcium phosphate super-saturation is exceeded was evaluatedat standard condition with the SAA copolymer. Testing was per-formed over a pH from 7.2 to 7.8. Automatic polymer control wasused to continuously maintain the active SAA dosage at a con-stant 4 ppm, i.e., the dosage necessary for the pH 7.2 condition.Calcium phosphate control was monitored by measuring the solublephosphate maintained in the circulating water. As can be seen forFigure 6, the soluble phosphate level decreased with increasingpH, even though the SAA copolymer dosage was held constant.The heat transfer surface was covered with a uniform calcium phos-phate deposit. This study clearly demonstrates that the polymerdosage must be adjusted as the calcium phosphate super-satura-tion increases. Merely maintaining a constant active polymer re-sidual without regard for the stresses that affect calcium phos-phate solubility will result in deposition.C. Comparison of Halogen Stable and Conventional TreatmentsTesting in the presence of a free chlorine residual compared thecorrosion and deposit control efficacy of two treatment programsunder standard water conditions, at pH 7.2. The treatments only


differed in the polymer and azole used for copper alloy control. TheSAA containing program contained 4 ppm SAA and 3 ppmbenzotriazole (BZT). The STP program used 2 ppm STP and 1.2ppm halogen resistant azole (HRA). Both programs utilized the 15ppm ortho-PO4 / 3 ppm pyro-PO4 blend for steel corrosion protec-tion. Systems were continuously chlorinated to maintain a freechlorine residual of 0.5 to 1.0 ppm. Chlorination was initiated 24hours after the start of the test to ensure initial passivation of boththe steel and brass metallurgy.Performance of the STP program was excellent with coupon corro-sion rates averaging <0.5 mpy for low carbon steel and <0.2 mpy forAdmiralty brass. The difference between filtered and unfilteredorthophosphate levels (delta PO4) averaged 0.3 ppm. The cou-pons and heat exchanger tube at the end of the study were free ofdeposits and corrosion. See Figure 7.The conventional program using SAA and BZT experienced slight,uniform calcium phosphate deposition on both the corrosion cou-pons and heat exchanger tube. The difference between filter andunfiltered orthophosphate levels averaged 0.6 ppm, indicating thatthe polymer was preventing precipitation from the recirculatingwater. Corrosion rates average 0.6 mpy for low carbon steel and 0.2mpy for Admiralty brass. Slight pitting corrosion occurred on theheat exchanger tube. The increase in deposition experience duringchlorination is attributed to the effects of corrosion (i.e., solubleiron generation) and not degradation of the polymeric dispersant.Chlorine is a strong oxidizing agent that accelerates the cathodiccorrosion reaction.The SAA/BZT treatment was modified by increasing the SAA dos-age to 8 ppm to address the deposition experienced at the 4 ppmlevel. At the higher polymer dosage the calcium phosphate scalewas completely eliminated, but localized corrosion of steel surfacesincreased significantly. Both the low carbon steel coupons andheat exchanger tube experienced severe pitting-type corrosion. Thecoupon corrosion rates averaged 1.4 mpy for low carbon steel and0.6 mpy for Admiralty brass. Elimination of the calcium phosphatedeposits left the metal surfaces more susceptible to chlorine-in-duced corrosion (Figure 8).The studies performed with the SAA copolymer demonstrate thatcalcium phosphate films can provide some corrosion protection.However, this barrier film technology is not practical since the rateof deposition cannot be controlled in systems having equipmentwith a wide range of heat loads (surface temperatures).D. Evaluation of a Pyrophosphate-Based Corrosion ProgramEvaluation of STP as a calcium pyrophosphate scale inhibitor wasconducted at standard conditions with a modified corrosion pro-gram. The 15 ppm orthophosphate/3 ppm pyrophosphate blendwas replaced with 18 ppm pyrophosphate (as ppm PO4). STP wasapplied at 10 ppm. Low polymer dosages were not evaluated, sincethe purpose of this study was to verify the precipitation inhibitionstudies.The pyrophosphate program provided excellent corrosion perfor-mance. Steel coupon corrosion rates averaged 0.5 mpy and heatexchanger tube was free of localized corrosion. STP preventedboth calcium pyrophosphate precipitation in the circulating cool-ing water and scaling of the heat exchanger tube. The differencebetween filtered (0.2 micron) and unfiltered total inorganic phos-phate levels was less than 0.2 ppm.

CONCLUSIONS1. The stress tolerant polymer (STP) is an excellent inhibitor

for both calcium orthophosphate and calcium pyrophos-phate.

2. STP provides superior calcium phosphate control withnear-neutral pH, inorganic phosphate-based programsunder heat transfer conditions.

3. The polymer dosage must be adjusted for increases incalcium phosphate super-saturations due to alkaline pHexcursions.

4. STP provides a significant margin of safety during highpH excursions

5. The excellent corrosion and scale control performance ofthe phosphate-based program employing STP and a halo-gen resistant azole (HRA) is attributed to the chlorinestability of the additives and their superior inhibition prop-erties.

ACKNOWLEDGEMENTSThe authors would like to acknowledge the support of GE Waterand Process Technologies’ Cooling Research and Technical Mar-keting teams and to acknowledge the technical assistance of LisaLarks, Julie Davis, John Mahurter, Gloria Tafel and Debby Arnoldfor their involvement with laboratory testing of the STP polymerdevelopment.

REFERENCES(1) Betz Handbook of Industrial Water Conditioning, Ninth Edition,1991, pp. 176-177(2) R. C. May, G. E. Geiger, D. A. Bauer, “A new Non-ChromateCooling Water Treatment Utilizes High Orthophosphate LevelsWithout Calcium Phosphate Fouling,” Corrosion/80, Paper No. 196,1980(3) W. F. Beer, J. F. Ertel, “Experience With High Phosphate CoolingWater Treatment Programs,” Corrosion/85, Paper No. 125, 1985(4) Hollander, O.; May, R. C., Corrosion, 1985, Vol. 41, pp. 39-45(5) T. Notoya, G. W. Poling, Boshuku Gijutsu, 1981, Vol. 30, p. 381(6) D. W. Reichgott, et.al., U.S. Patent #5,772,919, “Methods ofInhibiting Corrosion Using Halo-Benzotriazoles,” June 30, 1998(7) L. Cheng, R. C. May, K. M. Given, “A New Environmentally-Preferred Copper Corrosion Inhibitor,” Corrosion/99, Paper #101,1999(8) R. C. May, L. Cheng, K. M. Given, P. R. Higginbotham, “Appli-cation of a New Corrosion Inhibitor for Copper Alloys at the HarrisNuclear Plant,” Power Conference, December, 1997(9) S. M. Kessler, N. T. Le, “Performance of a New Paper MillSupply Treatment Program,” Materials Performance, Vol. 36, No. 8,1997, pp. 35-41


a chemical or refinery project, often requires a major adjustment inconceptualizing for both project engineers and managers. For ahigh rise office building the cost of energy is necessary to retainthe tenants. In a refinery electrical consumption is often consid-ered a cost of producing the product. In the power industry elec-tricity is the product. It may be a surprise that cost of operating thecentral plant in a high rise building complex can be one-third of thetotal cost of operation of the entire building. Current politics re-garding “greenhouse gasses” has stirred considerable pressure toreduce our energy requirements. More than ever it is necessary tolook at all of the equipment in the plant or building cooling loop totry to make improvements.Power consumed in operating a power plant is obviously non-productive. You can either use it, or you can sell it. For somecogeneration projects the power consumed is purchased from thelocal utility and the power generated is sold to your client. The twocan be at different rates. There are significant variations on themethod of establish-ing a power contract, and it is not the intent ofthis paper to evaluate the various contracts. We are more con-cerned with the effect on selection of equipment and plant layoutand how they might affect the project.

What is energy worth?Energy rates can be found on the internet. Energy InformationAdministration (EIA) posts rates for residential, commercial, indus-trial and transportation categories monthly for every state. April2007 kilowatt-hour rates (cents per kW-hr) for a few locations aretabulated below:

State Commercial Industrial All SectorsConnecticut 15.75 13.02 16.58

Hawaii 19.91 16.36 19.26Missouri 5.57 4.29 5.78

Idaho 4.82 3.32 5.15U. S. average 9.28 6.16 8.77

In addition to the $/KW charges there can be demand charges tooffset the peak power requirements the plant may need for hot daysor other conditions.However, we are not talking pennies. Here are examples of powerconsumption over time using the U. S. average cost of industrialpower (unadjusted for inflation-8,000 hours per year).

ByRichard J. DesJardinsCooling Tower Consultant

It is the purpose of this paper to show a few simplemethods to determine the best way to arrive at the mosteconomical cooling tower selection for a project.Low first cost may not be the best method for selectinga cooling tower. Power consumption for pumps andfans, the cost of the basin, piping and electrical equip-ment, and the choice of tower layout should often bethe deciding factors. Optimization of tower design con-

fall in a size where many different types of cooling tow-ers are available. These can range from crossflow tocounterflow, from packaged factory assembled towersto the small or large field erected towers. A great varietyof products are available for your consideration: almosttoo many. Larger power plants can be air cooled, eitherby direct or indirect steam condensation, or water cooledwith mechanical or natural draft cooling towers, and toa lesser extent, rivers and the oceans. Refineries, highrise air conditioning and other projects usually use cool-ing towers.Evaluating a power project, as compared to evaluating

Evaluate Your Cooling Tower

ditions related to other equipment such as heat exchangers andcondensers is discussed.This paper evaluates options of present worth value, annual costand capitalized costs of revenue streams, projected life span, returnon investment, depreciation, taxes, general administrative expenses,insurance requirements and the cost of power, and it provides com-ments on the proper choice of decision making formulas.It is well known that only a few specific items are needed to selecta cooling tower, such as the water flow rate, the hot water tempera-ture, the cold water temperature, and the wet bulb temperature.Generally, when we are called for a selection, the engineers andpurchasing agents contacting us are sophisticated enough to knowthese basic requirements.One can only hope they have considered what one degree colderwater would do to the overall plant capability and profitability.Also, it is not always the most economical to choose the highestrecorded wet bulb temperature. It is usually desirable to investi-gate the optimum balance between the cost of the cooling towerand the heat exchangers or other equipment it is cooling. Theconcept of alternate design temperatures is presented here just toprovoke thought on the possibilities and the effect they may haveon the overall plant economics. Although these are important con-cepts they will not be detailed in this paper.In reality we often really need more than just the temperatures andflow rate. We need to know what power is worth, and this becomesa major stumbling block since the engineer and purchasing agentprobably have not discussed it with management or even men-tioned to management that power will be consumed as well as gen-erated or it is needed to make a product. The reactions we get arefrom denial to “what you are talking about? Why do you need toknow that?” We need to know it because it may make a big differ-ence on the price or the plot space required.There is a natural tendency to take bids and pick the low price.That may or may not be the best approach for your project. Theremay be a very significant long term savings for the project if theeconomics of operation and cost of capital are investigated.

Why evaluate?First, the need for the analysis and the effect that it can have shouldbe demonstrated. Many of the projects proposed for cogenera-tion, ethanol or other chemical plants and refineries just happen to

Richard J.DesJardins


Do you select a big tower or a little tower?The cooling tower doesn’t set the heat load. The plant sets the heatload, and in order to do the cooling duty you can be provided witha big tower that does not use much power or a little tower that usesa lot of power. The tower can be tall with little plot space or shortwith a lot of plot space with resulting variations in pumping costs.The fans can be efficient or cheap. The variations can be dramatic.However, they can all be designed to do the same required cool-ing duty. So, which one should be selected?Selecting the evaluation toolsIt is not believed necessary to be sophisticated, at least in the earlystages of the project. You need to establish the cost of powerconsumption which may well be the same as the selling price foryour locality or the value in your power contract. Next, it is neces-sary to consider the cost of capital, which can be considered youroverall corporate cost of capital, the cost of borrowing money forthis project, the inflation rate plus a percent or so, or the marginalreturn that you could get on your money if you did not invest it inthe project at all. Third you need to establish the duration of theanalysis. It is also necessary to consider variations on how theequipment will be used.There are several methods of analysis which can be used. Themethod used (present worth, annual cost, or capitalized cost, dis-counted cash flow rates of return, payout period, or present worthcombined with a capital recover, etc.) does not make much differ-ence in the final analysis. Annual cost and capital cost methodsmake use of levelizing techniques, which employ present worthanalysis either directly or indirectly. Management sometimes pre-fers the annual cost method because they can relate it to flow offunds,As will be demonstrated later it is necessary to adjust the initialcost of capital to include a return on the investment you are makingand to account for annual fixed expenses. The adjustments need tobe brought to “present worth” values and compared to the energyconsumption to get a true comparison of one alternative to another.The method used for analysis probably has more to do with thepersonality of management than the actual method used. What isimportant is to get the understanding that some consideration mustbe made for the perpetual conflict between capital cost and powerconsumption. Once the necessary information is obtained the easi-est methods to use to evaluate equipment is a combination of “capi-tal recovery factor” and “present worth”. This paper will explainthe process to obtain that information and reduce it to a few simplefactors that can be used to facilitate the analysis of which coolingtower or tower component is best for your plant.

What is the evaluation period?Are you going to operate the plant for a full 30 years, and you willbe the owner? Have you signed a power agreement for 30 years butyou’re really an investor looking to make a quick buck and takeadvantage of the current tax laws. Maybe you really plan to unloadthe plant to somebody else in about 5 years so they can write it upagain and start the depreciation all over. Are you a tax paying

entity or are you part of a public institution that has no tax advan-tages? How many hours a year do you expect the project to operateconsidering downtime for maintenance? That new high rise build-ing in downtown is supposed to last well over 100 years (I saw onein great shape in Brussels Belgium with a corner stone dated 1492),so how long are you going to design yours for? Casinos in LasVegas are never designed for over 15 years – they tear them downand build something bigger.Historically power projects are evaluated over a period of 25 to 30years with the normal being 20-25-30 years of operation. Petroleumand refinery projects are generally evaluated on a 7 or 8 year periodwith typical economic evaluations being made on 3 to 5 year payouts.Why is this? I guess I wondered why for a long time, and I finallydetermined in a conversation with one of my banker friends that thebanks would only loan monies to a chemical or refinery project for7 or 8 years, and since it takes 3 or 4 years to build the plant they areonly allowed 3 to 5 years to pay it out. If you cannot pay it back inthe required payout period, the project is not viable. The banks willnot loan you the money. Recent research of projects on the internetand talks with bankers revealed loan durations for power plantsand large buildings are about 30 years. Cogeneration, ethanolplants, geothermal and general industry loans appear to be about10 to 15 years, but there are few solid guidelines. Bankers arelooking at the power contract, life of equipment, projection of longterm demand for the product and the general business climate be-fore making a decision on loan duration. They want to be sure theyget their money back.Projects with long duration evaluations increase the magnitude ofthe before and after tax affect of capital usage and the cost of powerconsumption. Your management wants to get a return on theirinvestment, and that may be the deciding factor for evaluation.More will be discussed about this later.

Present worth calculationThe present worth of a single or multiple future payments (knownas cash flows) is the nominal amounts of money to change handsat some future date, discounted to account for the time value ofmoney, and other factors such as investment risk. A given amountof money is always more valuable sooner than later since thisenables one to take advantage of investment opportunities.Present values are therefore smaller than corresponding futurevalues. (Wikipedia)Calculating the present worth of a stream of income or expenses issimple.PW = C [1-[1+(r/100)] -n]/(r/100)PW = Present worth of a stream of payments or expenses. r = annual interest rate in percent C = Capital expenditure. n = number of yearsThe choice of interest rate (r) can vary depending on the consider-ation for risk. It could be the prime rate, loan rate, corporate weightedaverage cost of capital, minimum return on capital, inflation rateplus a premium, or other risk adjusted value.For example the present worth factor for $1 for 20 years at 7%interest is:PW = 1 [1-[1.07] -20] / (0.07) = 10.6

What is the cost of capital?The first cost is not the only consideration. Many adjustmentsmay be necessary. A few considerations are listed below:


Adjustments to Capital:A Minimum Return on Capital is required if the project is to besuccessful. Usually this is set as the overall cost of corporatecapital which might be from stock, bonds, or loans, and it is often acombination of the weighted average of all three.Income Tax on Return of Capital for the purpose of evaluation ofdifferent investments is not the same as for the Internal RevenueService. The only concern is for the tax on the capital with respectto the minimum acceptable rate of return. If you make more you willpay more, but that should not be part of the decision making pro-cess.

T = R/1-R ( i+Mid-d’) C’

T = Tax rate in % of CapitalR= Income tax rate in decimal format (0.40 +/-)i = minimum acceptable rate of return on investment (0.075)id = depreciation/amortization reserve in % (decimal equivalent) onsinking fund basisd’= bank depreciation for IRS in % (say 3.3% straight line- decimalequivalent)C’ = 1-(B*b/i) [taxable capital]B = Fraction of Capital related to bondsb = interest on bond debtAd Valorem Taxes (Local or state property taxes) are generallybased on an assessed value set as a percentage of the initial plantand equipment total capital cost (say $20 per $100 of total cost).Then the assessed value is taxed at a rate per $100 of the “assessedvalue”. The final result might be 1.5 to 2.5% of the initial totalcapital cost.General and Administrative expenses might be a small percentageof the initial capital cost of a large power plant (maybe 1%) and alarger percentage of a smaller plant where the costs of salariesmanagers, legal services, advertising, etc. make up a larger percent-age of the initial plant cost (maybe 2% or higher)Insurance expenses are often 0.1% to 1% of the initial capital costof the plant.Depreciation and Amortization (or capital recovery) is the lossdue to wear and tear, obsolescence, and other factors that causethe ultimate retirement of the equipment. The “minimum acceptablerate of return” is the return on investors’ capital, and “depreciation/amortization” is the return of investors’ capital. This does not needto be the same as the allowable tax rate of depreciation. Think of itas a sinking fund where the amount of a periodic payment increasesto the value of the initial cost.A = F [i/(1+i)n-1]A = periodic paymentF = accumulated moneyi = interest rate (%/100)Percentage depreciation = A/capital costCost of Capital Example

% of CapitalMinimum Return on Capital 7.0Income Tax on Return 2.5Depreciation/Amortization Reserve 1.0Ad Valorem Taxes 2.5Administration General Expenses 1.0Insurance 0.2

=========Total annual fixed charge 14.2%

Years of evaluation 0Required return on capital 7%Annual fixed charge 14.7%

Present worth factor for $1 for 20 years at 7%PW = 1 [1-[1.07] -20]/(0.07) = 10.6Present Worth of Capital = (0.142)(10.6) * C = 1.50 CEach dollar of capital expenditure today should be considered as$1.50 because it requires having $1.50 on hand today to pay off theequal annual obligation of $0.142 fixed charges.Therefore, the simple adjustment to the first cost of the product orcomponent is to increase the price by the Present Worth of Capitalfactor.

What is the cost of energy?National Average cents per kW-hr = 8.77Hours per year = 8,000 (there are 8760 hours in a year, but there

may be downtime for maintenance or lackof demand for the product)

Present worth Factor 20 years at 7% = 10.6Present worth cost of power = 0.0877 * 8,000 * 10.6 = $7,437/kW

= 7,437 * 0.746 = $5,548 / horsepowerIf power consumption is tax deductible (estimate 40% total rateState and Federal):Power Cost = 5,548 * 0.6 = $3,330 / HP

What components should be evaluated?The cooling tower fans are not the only source of power consump-tion. The pumping loop through the condenser interconnectingpiping and the tower distribution system is often the most criticalcontrolling factor. The fans may not run all the time, but at leastone pump probably does. Especially on large projects with flowrates in excess of 30,000 GPM we find the pumping head controlsthe basic configuration of tower used, such as crossflow versuscounterflow and film fill versus splash fill in many instances.Part of the design process might include an investigation to see if itpays to use a taller fan stack. If there are no evaluation factors theonly tool to use is an attempt to adjust the height to fully load amotor, and that may not be the correct decision. If evaluationfactors are available the cost of the taller fan stack can be comparedto the savings in fan power for getting a greater velocity recoveryeffect. The same can be true for investigating different fill heights,types of fill and drift eliminators, air inlet heights, and various towerconfigurations. For these reasons evaluation factors should beprovided with the request for proposals.

Modifications for variations in operatingmodes, weather and heat loadsA review of the actual plant operation for an entire year mightindicate that the cooling tower fans will not run at full speed every-day, all day. Winter heat loads for a building will be 30% of summerload while the heat load for a refinery might be the same all year.Cooling too much in the winter might causes problems for a steamturbine. A geothermal plant may be able to use the coldest water itcan get all year around. It might pay to run a chiller with colderwater down to the point just prior to sucking oil in order to reducethe compressor back pressure, but at that point it is necessary tostart shutting down fans, number of cells, number of pumps and


run fewer chillers. Each project should be reviewed to determinethe possibility of variations for optimum usage. 33% or 50% pump-ing capacity will reduce the evaluated pump power, and using ½speed or VFDs on the fans will reduce the evaluated fan power.Another major consideration is the method of operation to be used.I know of some installations that are operating only in the after-noons on a hot summer day to keep the peaking load down andthrough this method they have great penalty costs for exceedingmaximum allowable meter readings as allocated by the power com-panies. Maybe this is starting a diesel engine, a gas turbine ordiverting steam from an operation, but in many cases it is a justifi-able application.For example fans running at half speed will use about 1/5 or 1/6 ofthe full speed power for half the hours of a year and the cost of fanpower can be reduced accordingly. Using the overall power costscalculated above the actual fan power costs for the year can befurther reduced:[(4,000 * $3,330) + (4,000 * $3,330 * 1/5)] / 8,000 = $1,980 / HPweighted average fan power cost.Some condenser and heat exchanger pumps are always running,even if the cooling tower fans are turned down for cold weatheroperation, and the pumping cost must certainly be considered aprimary expense of operating the plant. There may be practicallimits on turn-down of pumps in cold weather because the coolingtower nozzles may not provide adequate water distribution to pre-vent freezing in the fill, fouling of the fill or excessive drift due tomal-distribution of the air or water. Maybe only half of the cells orpumps run in the winter.Pumping power (English Units) at 85% pump efficiency can becalculated as:BHP/ft = 0.000297 * GPM (simplified 0.0003 * GPM )BHP = brake horsepowerFt = foot of pump headGPM = flow rate in gallons per minutePumping Cost = 50,000 * 0.0003 * $3,330 = $49,950/ft pump headChances are that the pumps run all the time and fans run 50% of thetime. To simplify the presentation the example below uses thisassumption. Remember, the object is not trying to evaluate theentire project: rather it is to make rational decisions regarding whichcooling tower design options are best.I don’t know of any manufacturer who is really thrilled with thepossibility of doing the full optimization study for you until youhave gone through the preliminary steps. If you go far enough togive them some basic tools to work with they will generally makeseveral selections to help you confirm your optimization. They justdon’t want to make a hundred selections when you are only goingto choose one. At the time of requesting bids you should give thevendor guide lines: cost of fan power – cost per foot of pump head– concrete basin cost – demand cost for electrical connections andcontrols – cost of piping.Although it is not the purpose of this paper to design the tower, thepurchaser should realize that sometimes all of the power evaluationcannot be used. There are good practice design standards thatshould be considered. Some examples are: 1) high fan power evalu-ation leads to low fan discharge velocities with increased possibil-ity of recirculation, 2) high pump head evaluation leads to low inletheight with high inlet air velocities to the extent air by-passes por-tions of the fill and the tower does not perform properly, 3) high fanand pump power costs lead to large tower plot areas with lightwater loadings that cause poor distribution with inadequate wet-ting of all of the fill surfaces, winter freezing problems, and poor

nozzle distribution when one pump is shut off. It is recommendedyou review the design and your anticipated operating modes withyour cooling tower consultant.

Analyzing the BidsAfter the bids are received the first task should be to make sureeach selection will perform as required. Equalize the bids for com-pliance with the specification and then compare power consump-tion. Premature evaluation is not advised because adjustments tothe design may be necessary to assure compliance with the tech-nical specifications.The selections shown in the table below are examples of the varia-tions that can exist. All selections were made for the same thermalduty of flow rate, hot water, cold water, and wet bulb temperatures.The exact conditions are not important for the purpose of this pre-sentation. The object is to show the relative magnitude of costs ofenergy versus capital expenditure. If you budget too low on capitalcost initially funding may not be adequate when you go to buy theequipment. If you budget too high it may kill the project before itgets started.Pump heads shown are for the tower only, and they do not includepiping losses or pressure drops of other equipment. For a moredetailed analysis you may want to add in the cost of various num-bers and sizes of risers and other piping, the cost of VFDs for oneor more of the fans, 2-speed motors, wiring and control costs, etc.Evaluation factors are also helpful in considering various towercomponents. Please note that in the examples given above the fanstack height varies. The energy cost can be used to decide if thepresent worth savings in fan power due to better velocity recoveryfrom a tall fan stack saves enough to offset the extra cost of thetaller fan stack. Changes to motor and starter sizes can also beincluded. The number of cells may change the piping costs.The ratio of the savings in operating costs and the additional fixedcharges should be at least one to one if the client is to break even.Some clients, though, for business reasons may not value thepresent worth of future savings as much as current capital expendi-tures. If the capital is not available to start the project it will neverget built. Maybe management will apply an additional 1.2 factor onextra capital expenditures for this reason. Decisions like this can bejustified if there is a possibility the equipment could be modified inthe future. For example, maybe another cell could be added in thefuture to reduce plant cold water temperatures for increased pro-duction after the project has been proven profitable.All of these economic studies will assist in making a selection re-garding which type of product to purchase and what its powerevaluation should be. However, ultimately the final decision mustalso include the reliability of the supplier, including reputation orexpectation that the product will meet its thermal performance re-quirements and that the equipment will be suitable for the expectedmaintenance life of the components of the tower and the life of thetower itself. It is suggested that you condition each vendor’s offer-ing to assure that it will meet the design performance requirementsbefore you buy, and that the people evaluating the bids have theexpertise and database to tell the difference. After all of that, it isrecommended that you test the final product to assure that youreceived what you asked for.As you consider the concepts of this paper please keep in mindthat it can be used just as easily to justify modifications to existingcooling towers as it can for evaluating new towers. Some Engineer/Constructors have been advising that their clients have been modi-fying their cooling towers in an effort to reduce energy consump-tion in order to do what they can to reduce their contribution to


Evaluation factors are also helpful in considering various towercomponents. Please note that in the examples given above the fanstack height varies. The energy cost can be used to decide if thepresent worth savings in fan power due to better velocity recoveryfrom a tall fan stack saves enough to offset the extra cost of thetaller fan stack. Changes to motor and starter sizes can also beincluded. The number of cells may change the piping costs.The ratio of the savings in operating costs and the additional fixedcharges should be at least one to one if the client is to break even.Some clients, though, for business reasons may not value thepresent worth of future savings as much as current capital expendi-tures. If the capital is not available to start the project it will neverget built. Maybe management will apply an additional 1.2 factor onextra capital expenditures for this reason. Decisions like this can bejustified if there is a possibility the equipment could be modified inthe future. For example, maybe another cell could be added in thefuture to reduce plant cold water temperatures for increased pro-duction after the project has been proven profitable.All of these economic studies will assist in making a selection re-garding which type of product to purchase and what its powerevaluation should be. However, ultimately the final decision must

also include the reliability of the supplier, including reputation orexpectation that the product will meet its thermal performance re-quirements and that the equipment will be suitable for the expectedmaintenance life of the components of the tower and the life of thetower itself. It is suggested that you condition each vendor’s offer-ing to assure that it will meet the design performance requirementsbefore you buy, and that the people evaluating the bids have theexpertise and database to tell the difference. After all of that, it isrecommended that you test the final product to assure that youreceived what you asked for.As you consider the concepts of this paper please keep in mindthat it can be used just as easily to justify modifications to existingcooling towers as it can for evaluating new towers. Some Engineer/Constructors have been advising that their clients have been modi-fying their cooling towers in an effort to reduce energy consump-tion in order to do what they can to reduce their contribution toglobal warming.The simplified concept of a present worth of capital factor alongwith the present worth of fan power and pump power is all youneed to make an economic analysis of your cooling tower bids.

Acknowledgement:1. Several concepts contained in this paper come from

notes and private conversations with Millard Cherry andPaul Leung of Bechtel Corporation during the spring of1974 in. preparation for a presentation to the PacificEnergy Association in Los Angeles, CA.

2. Cost of energy in various USA locations can beobtained at:www.eia.doe.gov/cneaf/electricity/epm/tables5_6_b.htm

3. Federal Corporation Taxes can be obtained at:www.infoplease.com/ipa/A0005946.html

4. State Corporate Income Tax Rates can be obtained at:www.taxadmin.org/fta/rate/corp_inc.html


Panos G. Papavizas, P.E.Baltimore Aircoil Company

AbstractThe design and qualification requirements defined inbuilding codes for active mechanical equipment to resistseismic forces are dependent on the desired level of earth-quake safety. For the basic level of safety where theintent is to reduce the hazard to life posed by equipmentbecoming detached or toppling during an earthquake,the seismic design provisions focus on equipment sup-

could constitute a hazard to life. The design and quali-fication of mechanical equipment for this higher level ofearthquake safety (i.e., functionality) must reliably as-sure functionality in addition to position retention.Building codes address this higher level of earthquakesafety by requiring the design of equipment itself, inaddition to the supports and attachments, to resist seis-mic forces that are greater than those required for posi-tion-retention considerations alone. Though the struc-tural stability and integrity of equipment is directly af-

Seismic Qualification of Cooling TowersBy Shake-Table Testing

ports and attachments and the basis of qualification can either beby analysis, testing, or experience data. For the higher level ofsafety where equipment functionality is vital for the continued op-eration of critical facilities after earthquakes, the design provisionsfocus on the equipment itself, as well as the supports and attach-ments. The basis for qualification in this case is limited to shake-table testing or experience data. Due to the rarity of strong-motionearthquakes and the lack of substantiated seismic experience datafor the current generation of equipment, shake-table testing be-comes the most reliable method of qualifying equipment function-ality.The code-recognized shake-table test protocol, ICC-ES AC156, pro-vides a generic methodology for verifying post-seismic test func-tionality. It is incumbent upon equipment manufacturers to definethe specific functional verification activities that must be conductedas part of the seismic qualification program. The unique functionalcharacteristics of cooling towers necessitate special considerationin developing a test plan that can reasonably and reliably assurepost-earthquake functionality. Baltimore Aircoil Company has de-veloped a functional verification methodology as part of a compre-hensive seismic qualification program for its products. This meth-odology and the basis for its development is the subject of thispaper.

IntroductionThe seismic design of mechanical equipment is focused primarilyon equipment supports and attachments. The intent of the seismicdesign provisions included in building codes is to reduce the haz-ard to life posed by the sliding, toppling, or falling of equipmentduring earthquakes. Typically, the basis for qualifying equipmentsupports and attachments for this level of earthquake safety (i.e.,position retention) is static analysis. This qualification approach,though satisfactory and acceptable for verification of equipmentrestraint in most applications, is not sufficient for all applications.Mechanical systems often serve vital functions in critical buildingfacilities such as emergency response centers, communication cen-ters, and hospitals. The continued operation of these facilitiesafter an earthquake is partly dependent on the ability of the vitalsystems, and the equipment within these systems, to function asintended. Failure of equipment to function in these applications

fected using this approach, functional reliability is only indirectlyimpacted. It is recognized that this approach may not be sufficientfor all types of equipment and for all critical applications. Accord-ingly, the qualification requirements for active equipment (i.e., equip-ment with moving or rotating parts) that must function following anearthquake go beyond the requirements for position retention. Func-tional qualification must be based on experience data or approvedshake-table testing. With the rarity of strong-motion earthquakesand the lack of substantiated seismic performance data for the lat-est generations of equipment, shake-table testing becomes the mostreliable way of assuring equipment functionality.The code-recognized shake-table test protocol, Acceptance Crite-ria for Seismic Qualification by Shake-Table Testing ofNonstructural Components and Systems (AC156), issued by ICCEvaluation Service, Inc., provides a framework for verification ofequipment functionality. The test protocol is applicable to all typesof equipment including mechanical and electrical equipment. Assuch, it does not define specific functional compliance tests andactivities. It does require the test specifier, which is typically theequipment manufacturer, to provide a detailed description of theequipment functional characteristics. The test specifier must alsodefine all the pre- and post-seismic test functional verification ac-tivities to be performed as part of the overall seismic qualificationtest plan.Cooling towers have unique functional and structural characteris-tics that must be carefully considered in the development of func-tional qualification requirements. The challenge for cooling towermanufacturers, and perhaps for the cooling technology industry asa whole, is defining the specific functional verification tests andactivities to be conducted as part of an AC156-compliant test pro-gram that can reasonably and reliably assure equipment function-ality after an earthquake. Baltimore Aircoil Company (BAC) hastaken up this challenge in developing a comprehensive seismicqualification program for its products. The functional verificationmethodology developed by BAC and the basis for its developmentis the subject of this paper.

BackgroundThe seismic design provisions for mechanical equipment attachedto buildings and other structures have evolved substantially fromtheir initial inclusion in the 1976 edition of the Uniform Building

Panos G.Papavizas


Code™ (UBC). The early provisions focused on the attachment ofequipment and were largely based on observed equipment damagein prior earthquakes. In later editions of the UBC, though the pri-mary focus of the code was still equipment attachment, the need forgreater assurance of equipment functionality in critical applica-tions was recognized and addressed, albeit indirectly. The coderequired equipment itself to be designed for seismic forces with“the assumption that if structural integrity and stability of the equip-ment are maintained, function and operability are ‘reasonably’ pro-vided for, although by no means assured” (Porush, 1992, p. 27).Shake-table testing as a means of seismically qualifying equipmentfor building applications first appeared in national seismic provi-sions in 1985 with the initial publication by the Federal EmergencyManagement Agency (FEMA) of the National Earthquake Haz-ards Reduction Program (NEHRP) Recommended Provisions forthe Development of Seismic Regulations for New Buildings. How-ever, shake-table testing for functional qualification of mechanicalequipment has not been employed to any significant extent outsidethe nuclear industry. This can be explained in part by the absenceof a recognized shake-table test methodology suitable fornonstructural building components. Where tests have been per-formed for non-nuclear applications, they have been hampered byinconsistent interpretation and translation of the building codeseismic provisions into actual test procedures (Gatscher et al., 2003,p.74). Tests so conducted result in uncertain equipment reliability.The test methodology void was filled by the issuance of AC156 in2000. This protocol was developed through collaboration betweenSchneider Electric, the Building Seismic Safety Council, and theInternational Conference of Building Officials Evaluation Service,which is now the ICC Evaluation Service, Inc. (Caldwell et al., 2003,p. 449). The most recent edition of AC156 became effective inJanuary, 2007. It is suitable for equipment qualification in accor-dance with the 1997 edition of the UBC and the 2006 edition of theInternational Building Code® (IBC). The AC156 protocol is refer-enced in the 2003 NEHRP provisions and the American Society ofCivil Engineers standard ASCE/SEI 7-05, Minimum Design Loadsfor Buildings and Other Structures.

Current Seismic Qualification Requirementsfor Critical EquipmentThe earthquake regulations found in most state and local buildingcodes are based on the International Building Code, the modelbuilding code published by the International Code Council (ICC)®.The NFPA 5000®, Building Construction and Safety Code®, whichis published by the National Fire Protection Association, is used asthe basis for earthquake regulations to a much lesser extent and inonly a handful of code jurisdictions. Both of these model codesrefer to the consensus standard ASCE/SEI 7-05 for seismic designcriteria.The seismic design requirements for nonstructural componentsincluding mechanical equipment are contained in Chapter 13 ofASCE/SEI 7-05. All manufactured equipment (i.e., packaged equip-ment that typically are not custom engineered for specific projects)that fall within the scope of Chapter 13 and are not exempt from theseismic design provisions must be certified and qualified by one ofthree methods defined in Section 13.2.1. These methods are:

a. Analysisb. Testingc. Experience data.

Critical mechanical components that are required to function afteran earthquake are classified as “designated seismic system” com-ponents. These components are assigned a component impor-tance factor, Ip, of 1.5 and are themselves required to be designedfor seismic forces. Non-critical components are assigned an impor-tance factor of 1.0.The more restrictive qualification requirements for designated seis-mic system components are contained in Section 13.2.2. Specifi-cally, paragraph 13.2.2.a. states:

“Active mechanical and electrical equipment that must re-main operable following the design earthquake shall becertified by the supplier as operable based on approvedshake table testing in accordance with Section 13.2.5 orexperience data in accordance with Section 13.2.6. Evi-dence demonstrating compliance of this requirement shallbe submitted to the authority having jurisdiction after re-view and approval by the registered design professional.”

The requirements for seismic testing are defined in Section 13.2.5.This section states in part:

“…testing shall be deemed as an acceptable method todetermine the seismic capacity of components and theirsupports and attachments. Seismic qualification by test-ing based upon a nationally recognized testing standardprocedure, such as ICC-ES AC 156, acceptable to the au-thority having jurisdiction shall be deemed to satisfy thedesign and evaluation requirements provided that the sub-stantiated seismic capacities equal or exceed the seismicdemands determined in accordance with Section 13.3.1and 13.3.2.”

The experience data requirements are provided in Section 13.2.6,which states:

“As an alternative to the analytical requirements of Sec-tions 13.2 through 13.6, use of experience data shall bedeemed as an acceptable method to determine the seismiccapacity of components and their supports and attachments.Seismic qualification by experience data based upon na-tionally recognized procedures acceptable to the author-ity having jurisdiction shall be deemed to satisfy the de-sign and evaluation requirements provided that the sub-stantiated seismic capacities equal or exceed the seismicdemands determined in accordance with Sections 13.3.1and 13.3.2.”

Clearly, of the three qualification methods only shake-table testingand experience data are recognized and acceptable for functionalqualification of critical equipment. There are limitations, however,in the use of experience data.

Limitations of Qualification by ExperienceObservations of buildings and equipment in the aftermath of earth-quakes when correlated with recorded ground-motion and build-ing-motion data are an invaluable source of information that can beused to gain an understanding of actual performance during seis-mic events. This type of reconnaissance information, or experiencedata, has guided code-writers in developing national seismic provi-sions to mitigate hazards from future earthquakes, and has revealedto equipment manufacturers the strengths and weaknesses of theirequipment designs. Experience data is important for the ongoingdevelopment of codes and seismically resistive equipment, but islimited when used as the basis for qualifying the seismic capacity


of the current generation of equipment designs for new buildingapplications.The most obvious limitation is that data is relatively rare due to theinfrequency of strong-motion earthquakes. Furthermore, experi-ence data that is available likely pertains to equipment designs thatare obsolete or that are several generations old. Considering thatmany equipment manufacturers frequently upgrade and modify theirproducts to gain competitive advantage through differentiation,experience data can become outdated very quickly.Experience data is also relatively difficult to obtain due to the re-strictions on access in earthquake damaged areas and the con-cerns about legal liability. Where access is permitted, ground-motion or building-motion data may not be available making it dif-ficult to determine the actual seismic demand experienced by theequipment. And assuming all the information is available, the sub-stantiated seismic capacity may be so low that it is virtually uselessfor new applications.To be acceptable as a basis for seismic qualification of equipment,ASCE/SEI 7-05 requires the use of experience data to be based on anationally recognized procedure. This requirement is necessary toensure the validity and proper application of the data. One proce-dure that exists and is used on a regular basis is the Seismic Quali-fication Utilities Group (SQUG) Generic Implementation Procedure(Eder, 2007). However, this procedure is proprietary to SQUG and isused primarily for qualification of new and replacement equipmentin nuclear power plants. No comparable procedure exists for build-ing applications.These limitations by no means preclude the use of experience data,but suggest that for most building applications it is not a viable

method for equipment qualification. The limitations also dictatethat if this method becomes viable in the future it should be usedwith extreme care.The more reliable alternative to using experience data for functionalqualification is to perform shake-table testing in accordance withAC156.

Scope and Purpose of AC156The stated scope and purpose of the AC156 test protocol is toestablish the “minimum requirements for the issuance of ICC Evalu-ation Service, Inc., evaluation reports on seismic qualification shake-table testing of nonstructural components and systems…” (ICC-ES, AC156, p. 2). ICC-ES is a nonprofit subsidiary of ICC thatperforms technical evaluations of products to determine code com-pliance. Evaluation reports are issued by ICC-ES following thereview and approval of test data submitted by an applicant to sub-stantiate code compliance. The reports are made available to thepublic and can be used by engineers, contractors, specifiers, andothers, but the primary aim is to aid building officials in the enforce-ment of code regulations. Building officials can accept or rejectICC-ES evaluation reports as proof of code compliance given thatthe reports are only advisory.To fulfill the stated scope and purpose, the AC156 test protocol notonly clearly defines the test procedure, but also specifies the ac-creditation requirements for test laboratories and the format re-quirements for test reports. These requirements must be satisfiedfor ICC-ES to consider the testing and to issue an evaluation re-port. However, an evaluation report does not have to be the finaloutcome of qualification testing in accordance with AC156.


Equipment manufacturers must carefully consider the advantagesand disadvantages of pursuing ICC-ES evaluation reports for theirtested products before proceeding with a testing program. Gener-ally, for test reports to be considered they must come from accred-ited laboratories. In special cases ICC-ES will consider reports fromnon-accredited laboratories as long as on-site assessments areperformed to its satisfaction (see ICC-ES, Acceptance Criteria forTest Reports, AC85, for accreditation requirements). Though nec-essary for the issuance of evaluation reports, these requirementsmay limit a manufacturer’s choice of test facilities and may be pro-hibitively expensive and burdensome.The evaluation report requirements of AC156 should not dissuademanufacturers from pursuing seismic qualification test programs.The benefits of qualification by testing outweigh the limitationsresulting from the absence of an evaluation report. To satisfy theintent of the accreditation requirements, it is advisable and prudentfor manufacturers to use test laboratories that are reputable andthat have experience performing tests in accordance with AC156.

Functional Verification Requirements ofAC156The functional verification requirements of AC156 are intended tobe universally applicable to all types of equipment, including me-chanical and electrical equipment. The protocol essentially pro-vides a framework within which equipment manufacturers can de-scribe the functional characteristics of their products and definethe specific pre- and post-seismic test functional verification ac-tivities to be performed as part of the qualification test program.The protocol also provides acceptance criteria for functional verifi-cation.Section 4.0 of AC156 details all the information that must be pro-vided by the equipment manufacturer for the unit to be tested,defined as the Unit Under Test or UUT. Specifically regarding thefunctional requirements, Section 4.4 states:

“A listing and detailed description shall be provided ofthe functional and operability equipment requirements and/or tests used to verify pre- and post-seismic-testing func-tional compliance.”

As part of the required information for Section 4.0, the importancefactor, Ip, must be specified by the manufacturer. This factor estab-lishes the performance level of the equipment. An importance fac-tor of 1.5 indicates that the functionality requirements of the proto-col apply.The pre-seismic test functional compliance verification activitiesmust be performed in accordance with Section 6.3. Specifically,Section 6.3 states:

“Functional and operability requirements and/or tests, asspecified in Section 4.4, shall be performed by an accred-ited testing laboratory to verify pre-test functional perfor-mance. Functional testing could be performed at eitherthe test facility or at the UUT manufacturing facility. Testdescription and results shall be documented in accordancewith Section 5.2 (Test Reports).”

As indicated in the discussion of the scope and purpose of AC156,the accreditation requirements are pertinent when an evaluationreport will be sought from ICC-ES. Otherwise, the functional testsshould be professionally and competently performed and recordedin the test report.

The post-seismic test functional compliance verification require-ments are defined in Section 6.7. Specifically, this section states inpart:

“Based upon the specified UUT importance factor in Sec-tion 4.3, equipment being qualified must be capable of per-forming its intended functions after the seismic event. Func-tionality and operability requirements and/or tests, asspecified in Section 4.4, shall be performed on the UUT toverify post-test functional and operational compliance.”

The acceptance criteria for verification of functionality are definedin paragraph 6.7.2 and require the post-test results to be “equiva-lent” to the pre-test results. To ensure that the results can bereadily interpreted and confirmed, the manufacturer should pro-vide objective pass/fail criteria in the test plan.Paragraph 6.7.2 also includes acceptance criteria for basic struc-tural integrity. It is required that the UUT structural system andanchorage are not compromised during the test, though some struc-tural damage such as local yielding is acceptable as long as it doesnot affect the functionality of the equipment. Repairs to the equip-ment are allowed as long as they are relatively minor. The exampleprovided in AC156 to illustrate what constitutes a minor repair is“replacing a bulb” (ICC-ES, AC156, p. 8).

Functional Characteristics of Cooling Towersand Qualification ConsiderationsThe function of cooling towers, including both open and closedcircuit towers, is to reject waste heat into the atmosphere by evapo-rative cooling. This is accomplished through a complex interactionof the various sub-systems within towers. These sub-systemsinclude the air moving system or mechanical system (i.e., fan, mo-tor, and drive system), the water distribution system (i.e., integralspray pumps, internal piping, nozzles, distribution basins, and col-lection basins), the heat transfer system (i.e., fill media and/or inte-gral heat exchangers) and the structural system (i.e., support mem-bers, bracing, enclosures, and anchorage). These tower sub-sys-tems must remain largely intact after an earthquake for coolingtowers to perform their intended function. Therefore, the integrityof these sub-systems must be verified in a comprehensive test andinspection program in order to assure functionality.The most recognized and accepted method of determining the ther-mal capability of cooling towers is testing in accordance with ATC-105, Acceptance Test Code for Water Cooling Towers, and its supple-ment for closed circuit towers, ATC-105S, both of which are pub-lished by the Cooling Technology Institute (CTI). Though thermaltesting in accordance with ATC-105 could be used to verify coolingtower functionality, it is unnecessary and excessive within the con-text of seismic qualification. Additionally, when thermal testing isused as the sole means of verification, it may be inadequate toaddress all the functional characteristics of cooling towers.The primary purpose of thermal performance testing in accordancewith ATC-105 is to determine the absolute thermal capability ofcooling towers. The AC156 functionality requirements, however,focus on the relative performance of the equipment. As long as thepre-test and post-test functional results are equivalent, the func-tional compliance requirements of AC156 are satisfied. As an alter-native to ATC-105 testing, the functional tests and activities shoulddemonstrate that the various cooling tower sub-systems remainsubstantially intact and can deliver equivalent pre- and post-seis-mic test functional results. This approach can reasonably assure


cooling towers will function after an earthquake as well as theywould prior to an earthquake.Some of the functional compliance tests and activities must focuson the characteristics of cooling tower sub-systems that an ATC-105 thermal performance test would not necessarily address, suchas mechanical system integrity. Though the ATC-105 code requiresan overall condition assessment of mechanical equipment prior toperforming a thermal test, this assessment may not assure the con-tinued safe operation of the mechanical system following an earth-quake. Other aspects of cooling tower sub-systems that warrantadditional consideration are the water containing function of distri-bution/collection basins and internal piping as well as the struc-tural integrity of the fill media and/or integral heat exchanger in thecase of a closed circuit tower.The functional compliance verification methodology developed byBAC includes specific tests and activities that address all aspectsof cooling tower function. This methodology is described in thefollowing section.

Functional Verification MethodologyThe functionality of cooling towers after shake-table testing can bereasonably and reliably assured through a comprehensive inspec-tion and test program that focuses on the various internal sub-systems of towers. The tests and inspections must verify that:

· The air moving system is substantially intact and deliversequivalent pre- and post-seismic test air flow.

· The integrity of the mechanical system is not compromisedand that the mechanical components can be operated safely.

· The water distribution system is substantially intact anddelivers equivalent pre- and post-seismic test water flowwithout significant leaks or drift.

· The heat transfer system is substantially intact and is notstructurally or thermally compromised.

· The structural system and anchorage are substantially in-tact with only minor yielding or distortion that does notimpact functionality.

These criteria can be satisfied by conducting thorough inspec-tions and production-type tests that establish fan/motor perfor-mance, demonstrate water distribution, measure vibration charac-teristics, and assess structural integrity. Specific inspections andtests to be conducted both before and after shake-table testing foropen cooling towers and closed circuit cooling towers are pro-vided in Tables 1 and 2, respectively, found at the end of this paper.A similar matrix or outline should be included in the seismic testplan, along with pass/fail criteria suited to the type of tower, toclearly establish the verification requirements.The results of all functional verification tests and inspections shouldbe included in the test report. Test and inspection forms can bedeveloped to facilitate the collection of test data and the interpreta-tion of the test results. The report should also include before andafter photographs, including close-up pictures of any post-testdamage.The production-type tests listed in Tables 1 and 2 typically wouldbe conducted at the UUT manufacturing facility or research facilitysince it is unlikely that the shake-table test laboratory has the capa-bility (i.e., hardware, instrumentation, and expertise) to conductthese tests. Therefore, it is critical that the fragility level of theUUT is not exceeded during shake testing so that the UUT can be

transported to the functional test facility. This requirement shouldbe communicated to the shake-table lab personnel so that all par-ties understand the performance expectations. Minor repairs to the UUT are acceptable as long as they are on thesame order of magnitude as the example provided in the AC156protocol. A partial listing of analogous cooling tower-type repairsthat should be acceptable to a reviewing authority is providedbelow:

· Sealing of minor leaks.· Replacement of dislodged water distribution branches or

nozzles.· Minor retightening of pipe joints.· Minor retightening of drive belts.· Minor realignment of mechanical components.· Replacement of a few dislodged fill packs or drift elimina-

tors.· Replacement of a few dislodged louvers or air inlet screens.

All minor repairs that occur during testing must be documented inthe test report.When conducted as part of a comprehensive seismic qualificationprogram in accordance with AC156, the functional verification testsand inspections listed in Tables 1 & 2 and described herein canreliably assure cooling tower functionality.

Certification RequirementsEquipment that has been seismically qualified by shake-table test-ing must be certified by the manufacturer as specified in paragraph13.2.2.a. of ASCE/SEI 7-05. Currently, there are no specific guide-lines in the code for preparing a certificate of compliance, but it isrecommended that the following information be included as a mini-mum for generic building applications (i.e., where the building dy-namic characteristics are not known).

1. Name of the manufacturer.2. Product line covered by the certificate.3. The code for which compliance was evaluated (e.g., 2006

IBC).4. Reference to AC156 as the test protocol.5. Performance level (i.e., Ip = 1.5).6. Certified seismic capacity, defined in terms of the design

spectral acceleration parameter at short period, SDS.7. Installation restrictions, if any (e.g., outdoor, grade level).8. Product restrictions, if any (e.g., accessories not covered

by the certificate)Manufacturers should include the certificate of compliance in theirproposal packages. With the seismic qualification basis and thecertified seismic capacity clearly defined in the certificate of com-pliance, equipment specifiers and purchasers can readily determinewhether or not the equipment is suitable for their specific project.A certificate of compliance in itself, however, is not sufficient evi-dence of testing and seismic qualification. All supporting docu-mentation, including the seismic qualification test plan, the testreport, and any supporting analyses, must be available for reviewand approval by the registered design professional and the build-ing official.

System Considerations


It has been established that cooling towers in designated seismicsystems must be designed and qualified to the building code re-quirements. However, qualification of cooling towers alone doeslittle to assure post-earthquake functionality of the systems withinwhich they operate. All components within these systems and theway that they interact must receive equivalent attention. This sys-tem approach to qualification should be directed by the registeredprofessional responsible for the system design.Cooling tower manufacturers can do their part to assure the func-tional level of earthquake performance in critical systems by satis-fying the requirements of the building codes, qualifying their equip-ment using a methodology similar to the one described herein, andproviding sufficient information to the registered professional tofacilitate cooling system design.

ConclusionsCooling towers classified as designated seismic system compo-nents that must operate and function as intended following anearthquake are required to be qualified using either experience dataor shake-table testing. Due to various limitations of experiencedata, the most reliable method of qualification is shake-table test-ing in accordance with the code-recognized test protocol, ICC-ESAC156.Seismic qualification programs for cooling towers must take intoconsideration their unique functional characteristics and includespecific pre- and post-shake test functional verification activitiesto reasonably and reliably assure post-earthquake functionality.The methodology developed by BAC and presented herein ad-dresses all aspects of cooling tower function and satisfies the veri-fication requirements of AC156. In the absence of specific func-tional verification requirements for cooling towers in the AC156protocol, a companion guideline or standard developed by the cool-ing technology industry could help ensure that all manufacturersapproach functional qualification with equal rigor.Cooling towers qualified by shake-table testing must be certifiedby the manufacturer. Certificates of compliance must contain suffi-cient information for specifiers and purchasers to determine thesuitability of the equipment for a specific project. Additionally, allsupporting documentation must be readily available for review andapproval by the registered design professional and the buildingofficial.Though functional qualification of cooling towers in critical appli-cations is necessary and important, a focus on the towers alone isinsufficient. Cooling tower qualification is just one part of a systemapproach to assuring the functional reliability of critical systems.

ReferencesAmerican Society of Civil Engineers. (2006). Minimum Design Loads

for Buildings and Other Structures. ASCE/SEI 7-05. Reston,Virginia: ASCE.

American Society of Heating, Refrigerating and Air-ConditioningEngineers, Inc. (2007). Heating, Ventilating, and Air-Condi-tioning Applications. 2007 ASHRAE Handbook. Chapter 54.Atlanta, Georgia: ASHRAE.

Building Seismic Safety Council. (1986). NEHRP RecommendedProvisions for Seismic Regulations for New Buildings. Part1: Provisions (1985 edition). FEMA 95. Washington, D.C.: Fed-eral Emergency Management Agency.

Building Seismic Safety Council. (2004). NEHRP RecommendedProvisions for Seismic Regulations for New Buildings andOther Structures. Part 1: Provisions (2003 edition). FEMA450-1. Washington, D.C.: Federal Emergency ManagementAgency.

Building Seismic Safety Council. (2004). NEHRP RecommendedProvisions for Seismic Regulations for New Buildings andOther Structures. Part 2: Commentary (2003 edition). FEMA450-2. Washington, D.C.: Federal Emergency ManagementAgency.

Caldwell, P. J., Gatscher, J. A. (2003). Equipment Qualification forProduct Line Families Using a Shaker Table Type Testing Cam-paign. Proceedings of Seminar on Seismic Design, Perfor-mance, and Retrofit of Nonstructural Components in CriticalFacilities (pp. 445-455). ATC-29-2. Redwood City, California:Applied Technology Council.

Cooling Technology Institute. (2000). Acceptance Test Code forWater Cooling Towers. CTI Code ATC-105. Houston, Texas:CTI

Cooling Technology Institute. (1996). Acceptance Test Code forClosed Circuit Cooling Towers. CTI Code ATC-105S. Hous-ton, Texas: CTI

Eder, S. J. (2007). Seismic Qualification of Equipment by Analysis.Presented at Symposium on Seismic Regulations and Chal-lenges for Protecting Building Equipment, Components &Operations. October 12, 2007. University at Buffalo.

Gatscher, J. A., Caldwell, P. J., & Bachman, R. E. (2003). NonstructuralSeismic Qualification: Development of a Rational Shake-TableTesting Protocol Based on Model Building Code Requirements.Proceedings of Seminar on Seismic Design, Performance, andRetrofit of Nonstructural Components in Critical Facilities(pp. 63-75). ATC-29-2. Redwood City, California: Applied Tech-nology Council.

ICC Evaluation Service, Inc. (2003). Acceptance Criteria for TestReports. AC85. Whittier, California: ICC-ES.

ICC Evaluation Service, Inc. (2007). Acceptance Criteria for Seis-mic Qualification by Shake-Table Testing of NonstructuralComponents and Systems. AC156. Whittier, California: ICC-ES.

International Code Council, Inc. (2006). International BuildingCode (2006 edition). Washington, D.C.: ICC.

International Conference of Building Officials (1976). Uniform Build-ing Code (1976 edition). Whittier, California: ICBO.

National Fire Protection Association (2005). NFPA 5000, BuildingConstruction and Safety Code (2006 edition). Quincy, Massa-chusetts: NFPA.

Porush, A. R. (1992). An Overview of the Current Building CodeSeismic Requirements for Nonstructural Elements. Proceed-ings of ATC-29 Seminar and Workshop on Seismic Designand Performance of Equipment and Nonstructural Elementsin Buildings and Industrial Structures. (pp. 17-31). ATC-29.Redwood City, California: Applied Technology Council.


Table 2: Closed Circuit Cooling TowersReferencesTable 1: Open Cooling Towers


Advance Grp Cooling Tower Pvt Ltd ....... 55

AHR Expo ..................................................... 63

Aggreko Cooling Tower Service ........... 38-39

Amarillo Gear Company ............................ IBC

Amcot Cooling Tower .................................... 3

American Coalition for Ethanol ................ 29

American Cooling Tower, Inc ....................... 9

AMSA, Inc .............................................. 13, 47

Bailsco Blades & Casting, Inc ..................... 62

Bedford Reinforced Plastics, Inc ................ 65

BOLToutlet.com ........................................... 67

Brentwood Industries .................................... 17

ChemTreat, Inc ............................................. 21

CleanAir Performance Group ..................... 25

CTI License Testing Agencies .................... 73

CTI ToolKit ............................................ 74-75

Composite Cooling Solutions, LP .............. 41

Cooling Tower Resources ............................ 33

Dominion ....................................................... 32

Dynamic Fabricators .................................... 43

Emerson ......................................................... 61

Fibergrate Composite Structures ................. 49

Gaiennie Lumber Company ............................ 4

Hesiler Green ................................................. 53

Howden Cooling Fans ...................................... 7

Hudson Products Corporation .................... 59

Industrial Cooling Towers .................. IFC, 40

Liang Chi Industries ...................................... 57

Metrix ............................................................. 19

Midwest Towers, Inc .................................... 31

Paharpur Cooling Towers &Equipment LTD ............................................ 23

Power-Gen ..................................................... 69

ProvibTech .................................................... 27

Rexnord Industries ........................................ 11

C.E. Shepherd Company, LP ...................... 15

Spraying Services, Inc. ................................. 45

SPX Cooling Technologies ...................... OBC

Strongwell .......................................................... 5

Swan Secure Products, Inc. ............................. 2

Tower Performance, Inc. ............................ 76

Index of Advertisers

cti journal, vol. 29, no. 2 · 4 cti journal, vol. 29, no. 2 view from the tower the winter is gone...

Documents