Вы находитесь на странице: 1из 88

The CTI Journal

(ISSN: 0273-3250)
published semi-annually
Copyright 2015 by The Cooling Tech-
nology Institute, PO Box #681807
Houston, Texas 77268. Periodicals
Contents
postage paid at Houston, Texas.

mission statement
It is CTI’s objective to: 1) Maintain
and expand a broad base member-
Feature Articles
ship of individuals and organizations
interested in Evaporative Heat Transfer
8 Hope Creek Circulating 144-inch Water Pipeline Carbon Fiber
Systems (EHTS), 2) Identify and ad- Upgrade
dress emerging and evolving issues
concerning EHTS, 3) Encourage and
Anna Pridmore, Ph.D. & Jim Melchionna
support educational programs in vari-
ous formats to enhance the capabili- 18 Proposed Methodology For CTI ATC-128 Sound Certificaiton of
ties and competence of the industry to
realize the maximum benefit of EHTS, Factory Assembled Towers
4) Encourge and support cooperative John Dalton & Larry Burdick
research to improve EHTS Technol-
ogy and efficiency for the long-term 30 Safety In Cooling Tower Maintenance
benefit of the environment, 5) Assure
acceptable minimum quality levels and Magose Abraham Eju
performance of EHTS and their compo-
nents by establishing standard speci- 36 Advancements in Cleaning and Passivation of Cooling Water Systems
fications, guidelines, and certification
programs, 6) Establish standard test- Raymond M. Post, P.E., Prasad Kalakodimi, Ph.D., Je"rey O’Brien & Richard H. Tribble
ing and performance analysis systems
and prcedures for EHTS, 7) Communi- 48 Mechanical Behavior of Polymer Fills
cate with and influence governmental
entities regarding the environmentally
Nina Woicke, Ph.D., & Daniel Dierenfeld
responsible technologies, benefits,
and issues associated with EHTS, and 58 ASHRAE Legionella Standard 188: Evidence-Based Interpretation
8) Encourage and support forums and
methods for exchanging technical
and Application
information on EHTS. Janet E. Stout, Ph.D.
letters/manuscripts
Letters to the editor and manuscripts 60 New York Legionella Regulations: Are They Missing The Boat?
for publication should be sent to: The Sarah Ferrari
Cooling Technology Institute, PO Box
# 681807 Houston, TX 77268.

subscriptions
The CTI Journal is published in
January and June. Complimentary
Special Sections
subscriptions mailed to individuals in
the USA. Library subscriptions $45/
74 CTI Certified Towers
yr. Subscriptions mailed to individuals
outside the USA are $45/yr. 80 CTI Licensed Testing Agencies
change of address 82 CTI ToolKit
Request must be received at subscrip-

Departments
tion office eight weeks before effective
date. Send both old and new ad-
dresses for the change. You may fax
your change to 281.537.1721 or email:
vmanser@cti.org.

publication disclaimer
2 Multi Agency Press Release ...see page 17
CTI has compiled this publication with
care, but CTI has not Investigated,
2 Meeting Calendar
and CTI expressly disclaims any duty
to investigate, any product, service 4 View From the Tower
process, procedure, design, or the
like that may be described herein. 6 Editor’s Corner
The appearance of any technical data,
editorial material, or advertisement in
this publication does not constitute en-
dorsement, warranty, or guarantee by
CTI of any product, service process,
procedure, design, or the like. CTI ...see page 34
does not warranty that the information
in this publication is free of errors, and
CTI does not necessarily agree with
any statement or opinion in this pub-
lication. The entire risk of the use of
any information in this publication is
assumed by the user. Copyright 2016
by the CTI Journal. All rights reserved.
...see page 54
...see page 26
...see page 44
CTI Journal, Vol. 37, No. 2 1
CTI Journal
The Official Publication of The Cooling Technology Institute

Vol. 37 No. 2 Summer 2016

Journal Committee
Future Meeting Dates
Committee Annual
Paul Lindahl, Editor-in-Chief
Virginia Manser, Managing Director/Advertising Manager Workshop Conference
Donna Jones, Administrative Assistant
Andrew Manser, Administrative Assistant July 10-16, 2016 February 5-9, 2017
Angie Montes, Administrative Assistant
Graphics by Sarita Graphics Pointe Hilton Tapatio Cliffs Sheraton New Orleans
Phoenix, AZ New Orleans, LA
Board of Directors
Bill Howard, President July 16-19, 2017 February 4-8, 2018
Anthony DePalma, Vice President
Brandon Rees, Secretary Hilton Orlando – Lake Buena Vista Hilton Houston North
Frank Michell, Treasurer Lake Buena Vista, FL Houston, TX
Jon Cohen, Director
James W. (Jim) Cuchens, Director July 15-18, 2018 February 10-14, 2019
Narendra Gosain, Director
Brian Hanel, Director La Cantera Sheraton New Orleans
Jean-Pierre Libert, Director San Antonio, TX New Orleans, LA
Helene Troncin, Directorr

Address all communications to:


Virginia A. Manser, CTI Administrator
Cooling Technology Institute
PO Box #681807
Houston, Texas 77268
281.583.4087
281.537.1721 (Fax)

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

E-mail:
vmanser@cti.org

For Immediate Release


Contact: Chairman, CTI
Multi-Agency Testing Committee
Houston, Texas
2-September-2016
Cooling Technology Institute, PO Box
681807, Houston, Texas 77268 – The
Cooling Technology Institute announces
its annual invitation for interested ther-
mal testing agencies to apply for poten-
tial Licensing as CTI Thermal Testing
Agencies. CTI provides an independent
third party thermal testing program to
service the industry. Interested agencies
are required to declare their interest by
March 1, 2017, at the CTI address listed.

2 CTI Journal, Vol. 37, No. 2


CTI Journal, Vol. 37, No. 2 3
View From The Tower
Sixty six years of education. Sixty six years of These are what creates CTI’s solid foundation. So
growth. Sixty six years of commitment. Sixty while at the summer workshop meeting in beautiful
six years of setting the industry standards. Sixty Phoenix AZ, please join in, and take an ACTIVE role
six years of fun. All this due to, sixty six years of in ensuring that CTI remains the fore-font leading
CTI. Without the participation by all the attendees technical organization when it comes to all types
over the many years, CTI would never be what it is of cooling technologies. It will be a time for a ton
today. But let’s face it, without ACTIVE participa- of fun in the hot hot sun! It will be a time to golf
tion, CTI would be even less than it is today. I would (Sunday only though, time to roll up those sleeves
like to thank the many ACTIVE volunteers of CTI for the balance of the meeting! ), a time to see old
including all our board members over the years, all friends, meet new friends and have a drink or two
the presidents over the years and all the committee in the hospitality suite after a long days work.It
chairs and task group chairs and certainly even all will be a time when owner operators can network
the task group participants over the years that have with other owner operators on trends and cooling
helped create and update all our codes, standards and technology specific to their plants and applications.
Bill Howard
guidelines. By taking an ACTIVE role, both you and It will be a time when direct competitors can join
your company are providing a valuable service to CTI, but equally hands (certainly not literally!) and be friends in the quest for a better
important is the impact we are having on the world. CTI has numer- organization mostly it will be a time to:
ous international members which we embrace and welcome to our “Ask not what CTI can do for you, rather ask what you can do
CTI family. Over the past several years, this number has continued for CTI.”
to grow and as can be seen by the attendance at our meetings, these
members are committing their time and efforts in helping to make CTI, “Join in, get involved, learn, educate others, par-
CTI the world class technical organization that it is. The CTI staff ticipate and thrive!”
continues to deliver the excellence that we all have come to ap- Sincerely and respectfully,
preciate and a special thanks goes out to Vicki and her entire staff. Bill W. Howard, P.E.
A technical organization is only as good as the documents and in- CTI President
formation they contain as well as the people within the organization.

4 CTI Journal, Vol. 37, No. 2


CTI loses two of their best in 2016
Arthur “Art” Brunn Jr.
1934 - 2016 Arthur Frederick Brunn, Jr., 81, of Beaumont, died Monday, March 21, 2016, at Harbor
Hospice, Beaumont. He was born June 27, 1934, in Richmond Heights, Missouri, to Arthur Frederick
Brunn, Sr. and Aileen Dorothy Rummel Brunn. They lived in Webster Groves, Missouri until 1939,
when they moved to Kansas City, Missouri. The Brunn family moved to Memphis, Tennessee in 1950.
Art graduated from Treadwell High School, where he was a member and president of the Key Club and
a member of the National Honor Society. Art went to the University of Tennessee in Knoxville where
he met Clarice Marie Tallent. They were married December 18, 1955. Art graduated with a Bachelor of
Science degree in chemical engineering in August, 1956. Art and Marie moved to Milwaukee, Wisconsin
where Art worked for Allis-Chalmers Manufacturing Company. Their first son, David Arthur, was born
December 15, 1957. In February, 1958, Art, Marie and David moved to Houston, Texas, where Art did
sales and service work for Allis-Chalmers in the field of water treatment. Their second son, Michael Charles was born April 13,
1960 in Houston. In 1967 the family moved to Beaumont, Texas when Art started working for DuPont as a consultant in the
field of industrial water treatment. Art was active with N.A.C.E., an international technical organization covering all phases of
corrosion. He became chairman of the Technical Practices Committee of N.A.C.E., a member of their Board of Directors, and
received the Brannon Award for service to the organization. He was also active in the Cooling Tower Institute, where he was
on the Board of Directors, treasurer, and president. Art was named an Honorary Life Member of CTI and remained active after
retirement from DuPont. Art was an active member of Redeemer Lutheran Church. He held the office of treasurer, chairman of
the Voter’s Assembly, Chairman of the Children’s Center Board, and was active in many other phases of the church. Survivors
include his wife, Clarice Marie Tallent Brunn of Beaumont; sons, David Brunn and his wife, Arlene; Michael Brunn and his
wife, Stacey, both of Houston; sister, Julie Mochnak of Livonia, Michigan; five grandchildren; and four great-grandchildren.

Bruce Byron Brand


Bruce Byron Brand, 88, passed away peacefully at home on Sunday, May 15, 2016, in Fort Worth.
Bruce was born Sept. 27, 1927, in Kansas City, Missouri to Ella and Carl Brand. He graduated from
high school at the age of 16, attended Virginia Military Institute for one year, spent time in the U.S.
Army teaching cadets at West Point, then went on to the University of Missouri and graduated with a
Bachelor of Science in Chemical I, Engineering. He was a lifetime member of Sigma Alpha Epsilon
Fraternity. He married Ellie, his neighbor and sweetheart, in 1953. The couple began their life in Kan-
sas City, Kansas and Prairie Village, Kansas. Over the next 11 years, the couple welcomed their four
children into their home. His children remember him as a kind, gentle, smart but stubborn father who
encouraged and aided them to pursue their goals. Bruce was an accomplished cooling tower engineer.
He spent his early career with Havens Cooling Towers in Kansas City, Missouri. In 1973, he and Ellie picked up their four
children and moved to Fort Worth were he started his new job at Ceramic Cooling Tower and remained there until he retired.
Retirement did not slow him down, though; he became a consultant with Tower Engineering in Fort Worth and continued
to “work” until just a few months before his passing. Outside of his work, he, Ellie and his children were avid sailors and
members of the Fort Worth Boat club for many, many years. During this time, he amassed a room full of sailing trophies in-
cluding some from the Lightning and Catalina 22 Nationals. When his children (crew) grew up and moved away, Bruce took
up golf in his 50s and was an active member of the Ridglea Men’s Golf Association where he played up to three rounds per
week until he was age 87 and continued to play, when he could, until just a few months before his passing. He was an active
member of Ridglea Country Club since 1975, and he and Ellie could be seen there almost every night enjoying dinner and
friends. Bruce was preceded in death by his parents, Ella and Carl Brand of Kansas City, Missouri; and his daughter, Mary
Elizabeth Tracy of Syracuse, Kansas. Survivors: Loving wife of 62 years, Eleanor Arnold Brand; children, Melanie Brand
Tracy and her husband, Timothy Allen Tracy of Syracuse, Kansas, Thomas Marshall Brand of Forth Worth, Jeffrey Douglas
Brand and his wife, Lindy Denise Brand of Austin, Texas; grandchildren, Liz and husband, Josh Porter of Ambridge, Penn-
sylvania, James Tracy of Wichita, Kansas, Whitney Tracy of Syracuse, Kansas, Jeffrey Brand Jr., Haley Rojas, Harper Brand
and Matt Brand of Austin, Texas; and two great-grandchildren, Gabe Hrustic and Laney Porter of Ambridge, Pennsylvania.

CTI Journal, Vol. 37, No. 2 5


Editor’s Corner
Dear Journal Reader, China Cooling Tower Institute: This organization
A new year for CTI started after the CTI An- has been growing in China, and has reached out to
nual Meeting in February, 2016. Bill Howard of CTI to explore ways the two organizations can work
CTD,Inc., has started his two year term as President together. Several meetings have occurred between
of CTI. Several outside influences have had an im- the organizations, and we continue to work toward
pact on CTI, and some new activities are moving, future cooperation.
so the next few years should be interesting. Emerging Activities:
Outside Influences: CTI Research Update: A research grant was award-
DOE Fan Rule: My last Editor’s Corner made ed to CleanAir to complete Pitot tip investigations to
CTI Members aware that a Department of Energy find a replacement for the (no longer commercially
rulemaking is in progress with regard to fans, which available) Simplex tip, which has been the CTI water
could include any of the products of interest to CTI flow measurement standard. CleanAir is currently
that move air with a fan or blower. CTI has asked Paul Lindahl proceeding with this, the first research project de-
for exemption of heat rejection equipment from the Editor-In-Chief veloped under the relatively new CTI Research and
rule, and have been successful in gaining the exemption in the term Development Committee.Projects are proposed via
sheet agreed upon by the working group. The DOE is writing the the standing technical committees, and proceed through a process
final rule, which is expected in 2016, so we will need to continue to administered by the R&D committee.Funding is raised and admin-
watch this closely. CTI members who produce fans for any purpose istered via the CTI Finance Committee. Other projects are under
should be aware that they could be affected by this rulemaking, consideration, and new ones may be proposed within the standing
which will set minimum fan efficiencies by equipment types. For technical committees at any time.
more information, contact Larry Burdick, Frank Morrison, or me. Licensed Sound Testing: Licensed sound testing is in its first full
Legionnaires Disease: Outbreaks in New York City and Flint, MI, year of operation in 2016. We look forward to growth in this test-
have brought media attention to the issue, and in the case of NYC/NYS ing service by CTI.
have focused on cooling towers. Since then, the CDC has become There is much happening with our very active CTI volunteer tech-
vocal about the need for a whole building approach to Legionnaires nical organization; we again encourage you to get involved in the
Disease risk, and has published a nicely done Toolkit to supplement CTI technical committees.
ASHRAE STD-188 (2015) Legionella: Risk Management for Build- Respectfully,
ing Water Systems. CTI’s GDL-159 is expected to be a useful tool
with this standard when completed, hopefully, this year. The CDC
toolkit and the ASHRAE standard are available on their websites. Paul Lindahl, CTI Journal Editor

6 CTI Journal, Vol. 37, No. 2


CTI Journal, Vol. 37, No. 2 7
Hope Creek Circulating 144-inch
Water Pipeline Carbon Fiber Upgrade
Jim Melchionna, Corporate Program Manager - Engi- from 5 to 20 lineal feet. A common failure mode
neering Services, Piping, PSEG Nuclear –Salem & Hope
Creek Generating Stations
of PCCP is breakage of the prestressing wires
Anna Pridmore, Ph.D., Vice President-Pipeline Solu- within individual PCCP sections.
tions, Structural Technologies
Once enough prestressing wires break on an
individual segment of PCCP, the concrete core
Abstract in the region near the broken wires is no longer
in compression and can crack, exposing the steel
Hope Creek Nuclear Generating Station has a
cylinder to ground water, causing corrosion. De-
comprehensive asset management program to
pending on the nature of the wire break, number
address their cooling tower in addition to their cir-
of broken wires and other factors, the process of
culating water pipelines. Results from condition
pipe failure can take from days or weeks to years
assessment of their circulating water pipelines
to complete the process.
in recent years indicated the need for repair of
seven (7) distressed segments. The plant elected Because of the segmental construction of PCCP
to repair the segments of 144-inch pre-stressed Anna Pridmore and the predictability of failure following a large
concrete cylinder pipe (PCCP) of the circulating number of wire breaks, technology was invented
water pipelines during an outage taking place in October 2013. to inspect it with an objective of identifying broken wires. Elec-
Carbon fiber-reinforced polymer (CFRP), a trenchless structural tromagnetic inspection technologies assist owners by allowing
repair system, was selected to strengthen the damaged PCCP. CFRP them to locate, with precision and accuracy, individual sections in
is a repair material used to provide targeted structural upgrades need of replacement or repair. An electromagnetic inspection was
within distressed or damaged pipeline sections. This paper will performed and led to identifying the areas of broken wires – and
discuss all aspects of design and implementation of the repair. ultimately the distressed segments – which Hope Creek selected
for upgrade during the fall 2013 refueling outage.
Background On Hope Creek Generating Sta-
tion Options Analysis
In 1974, construction began on the Hope Creek Nuclear Generating Once the segments were identified, Hope Creek then had to pri-
Station. In operation since 1986, Hope Creek Nuclear Generat- oritize and discuss options for how to best proactively address
ing Station is owned and operated by PSEG Nuclear LLC. It is a the pipeline system. To immediately mitigate further distress to
single unit boiling water reactor with a total generating capacity the above-ground portions of the circulating water pipelines, an
of 1,178 megawatts net. As one of four licensed nuclear power external concrete encasement was installed. Beyond this stopgap
reactors in New Jersey, whose combined output supplies over one measure the pipes either needed to be replaced or undergo a full
half the electrical power to the state, Hope Creek generates enough structural repair in place. Due to their size, location and outage
electricity to power approximately one million homes each day. duration, replacement of the pipe segments was ruled out as an
Hope Creek shares a 700 acre man-made island with its sister plant, option. Hope Creek opted for a trenchless, fully structural repair
Salem Nuclear Generating Station. utilizing a CFRP liner to address the distressed sections.
Several years ago external distress was observed on the 144-inch The benefits of this option included completion of the repairs with
PCCP circulating water system. Due the observed distress and due no excavation required, the ability to perform the repair within
to its location directly adjacent to the natural draft cooling tower, the outage window and nearly no hydraulic loss because of the
Hope Creek management implemented a condition assessment pro- low-profile finish of CFRP liners. Unlike most lining systems,
gram for the large diameter pipelines. It was through this program CFRP liners, which are hand laid into place, can be engineered and
that seven (7) segments were identified for repair in 2013, and a installed on the specific pipe segments requiring structural upgrade.
proactive upgrade project was launched, budgeted and executed. Design Of The Cfrp System
Background On Circulating Water System Given that the CFRP repair method needed to provide a fully
And Condition Assessment Program structural repair, the design process was an important phase of the
project. PSEG worked directly with the CFRP designer, Simpson
The 144-inch diameter sections of the circulating water system
Gumpertz & Heger, to make sure all design requirements were met.
for Hope Creek are comprised of embedded-cylinder type (ECP-
The CFRP lining system was designed as a standalone structural
type) PCCP, which is composed of an inner concrete core, a steel
system where the CFRP system resists all internal and external
cylinder, an outer core, high strength wires under tension that are
structural demands without reliance on the host pipe.
wound over the outer core, and a protective mortar coating (Figure
1). PCCP is generally a segmental pipe install in sections ranging

8 CTI Journal, Vol. 37, No. 2


The CFRP design process involves collection of the design infor- were an important element of successful project completion. In
mation, including as-built drawings, lay schedules, pipe specifica- all CFRP pipeline projects, the environment within the interior of
tions, and pipe inspection reports. Typical parameters accounted the pipe repair area needs to have the temperature and humidity
for in the CFRP system design include: working pressure, transient controlled to meet specifications of the CFRP manufacturer. For
(surge) pressure, vacuum pressure, temperature differential, soil Hope Creek, and all projects requiring an expedited schedule, the
cover, height of the water table soil material properties, and any pipe interior is heated to aid the curing process.
vehicular traffic running on the ground above the pipe. Surface preparation was performed using sponge blasting as shown
CFRP systems are typically designed using a Load and Resistance in Figure 2. Sponge blasting is a unique abrasive blasting media
Factor Design (LFRD) approach, where factors are applied to loads consisting of abrasive embedded within a polymer. The sponge
combinations and material properties to account for uncertainties component helps to minimize airborne particulate and is recyclable,
within the design assumptions. As part of this design approach, which reduces the quantity of overall blasting media used on the
design limit states are analyzed separately and the CFRP lining project. It is much more environmentally friendly than typical
design is governed by the limit state that has the lowest demand sand blasting operations.
to capacity ratio for the particular design scenario. From this PCCP has is an inner concrete substrate and surface preparation
process it is determined how many layers of unidirectional CFRP is a critical component to the successful long term performance
will be applied both longitudinally and circumferentially within of the CFRP liner. Proper bond to the sound concrete substrate is
the designated pipe segment. Typically longitudinal layers provide achieved by preparing the surface to ICRI CSP3 finish, removing
resistance to thrust, thermal changes, and Poisson’s ratio, while laitance from the surface of the concrete, and exposing the aggre-
the circumferential layers provide strength in the hoop direction gate. This allows for a mechanical bond to the pipe surface. Joint
for internal pressure and external loads. detailing for the CFRP system requires carefully chipping down
Preplanning to the steel cylinder at the end joints and preparing the exposed
steel to SSPC-SP10 near white metal finish. Following surface
The 144-inch pipeline upgrade was a high profile project at Hope
preparation, bond tests were completed to insure proper adhesion
Creek during the fall 2013 outage, with an aggressive schedule
of the CFRP system. These tests are discussed in the Quality
and challenging working conditions. In order to ensure successful
Control section of the paper.
implementation, an extensive preplanning process was put in place
and soft mobilization to the site was scheduled well ahead of the Task Stage 2: Material preparation (saturation).
outage. To drive the preplanning process, the contractor and Hope CFRP systems are comprised of fabrics, glass and carbon, along
Creek initiated planning conference calls twice a week for the 6 with a 2-part 100% solids epoxy system. The carbon and glass fiber
months leading up to the outage. are dry rolls of fabric which are saturated in the field. The first step
These planning sessions included a wide variety of topics includ- in material preparation is mixing of the 2-part epoxy, Part A and
ing coordination of roles and responsibilities and creation of Part B. The Part A and Part B arrive at the jobsite in premeasured
extensive contingency plans for all aspects of the project. The 5-gallon buckets and are poured together per the manufacturer
contingency plans helped to insure all team members were aware specifications and mixed thoroughly to activate the epoxy.
of the processes for dealing with potential unforeseen develop- The epoxy system is used for multiple purposes throughout the
ments or conditions. installation process including saturation of the material and prim-
Cfrp Installation Process ing of the pipe substrate. A thickened version of the epoxy system
is also created using fumed silica. This thickened epoxy is used
The CFRP liner installation process consists of several typical
to fill imperfections in the pipe surface and as a coating between
steps which are broken down into five (5) task stages below. The
layers of CFRP.
task stages are as follows:
Prior to installation, the fabric is wet out in a mechanical saturation
1. Mobilization to surface preparation
machine used which has two (2) baths of epoxy and a calibrated
2. Material preparation (saturation)
rolling system. The fabric is fed through the saturation machine,
3. Application of the CFRP system and top coat and through the roller system the proper amount of epoxy is com-
4. CFRP repair system termination details pressed into the fabric.
5. Final curing of the system.
See figure 3 for example of mechanical saturation machine.
The Hope Creek project consisted of some unique challenges
which are discussed in a separate section of the paper. Following The materials preparation process takes place in a tent to eliminate
is a description of the task stages for Hope Creek: the chance for contamination and to reduce airborne particulate
during the use of fumed silica. To insure the long term durability
Task Stage 1: Mobilization to surface preparation. of the CFRP system, the materials preparation process has quality
Following mobilization of all personnel, materials and equipment control procedures which are discussed later in the paper.
to the Hope Creek site, an extensive amount of safety protocols
Task Stage 3: Application of the CFRP system and
were reviewed and put into place prior to entry into the pipe. All
confined space entry was performed in compliance with the Con-
top coat.
tractor, Hope Creek and OSHA regulations. Given the schedule As the material and surface preparation were completed at the
constraints at the Hope Creek project, environmental controls Hope Creek project site, installing the CFRP system began within
the pipe. The first step within the pipe interior, following surface

CTI Journal, Vol. 37, No. 2 9


preparation, was application of the primer coat of epoxy. Following 1. Removal of cooling tower muck
the primer coat, a layer of thickened epoxy is placed on the concrete Before access to the repair site could be established, the cooling
substrate, which is meant to infill imperfections and create a layer water tower basin needed to be cleared of the muck seen in Figure
of epoxy for proper bonding of the first layer of CFRP. 8. This situation was challenging in that the quantity of muck was
The saturated carbon fiber was then applied to the inside surface an unknown at the outset of the project. The CFRP contractor
of the host pipe in what is referred to as a wet lay-up process. The was required to set up a system to remove the muck as rapidly as
wet-out fabric was pressed to the inside surface of the host pipe to possible to gain access to the pipe (s) to be repaired.
achieve intimate contact. Any entrapped air between layers was
released or rolled out without wrinkling of carbon fibers. 2. Two-tiered scaffolding system with multiple vertical
slopes
When the CFRP was not properly aligned, the affected layer of
On typical CFRP projects, following the surface preparation and
the CFRP system was removed and replaced prior to curing. If the
prior to CFRP installation, portable scaffolding is erected, as
CFRP layer cannot be removed without affecting the integrity of
the surrounding carbon fiber, an additional layer was overlaid onto needed, spanning the pipe section to be lined. This is to give the
the off-axis fibers to restore the laminate structure to its intended installation crew the ability to apply the CFRP liner to all areas of
axial-strength requirements. the pipe section without the need to walk on the pipe. Best industry
practice to ensure a successful installation is that the application
For Hope Creek, unidirectional carbon fiber fabric was manually
takes place continuously in a manner that avoids contact with the
installed in both longitudinal (see figure 4) and circumferential
CFRP after installation and until cured. The pipe sections being
layers (see figure 5) to meet design requirements as presented in
repaired required scaffolding for the 144-inch diameter pipe seg-
the project drawings.
ments to be built on a vertical slope and span over an area which
Following installation of the prescribed number of layers on the had a vertical bend within the scaffold span. This required a
Hope Creek repair segments, a final top coat of epoxy was installed. specialized access system to be engineered and built in-place, as
The top coat installed over CFRP repairs adds to the long term shown in figure 10.
durability of the CFRP system, by providing protection from wear.
In cases where water constituents require chemical resistance, the 3. Single point of entry for repairs
top coat can be engineered to meet project requirements. On typical CFRP pipeline repair projects, two (2) points of entry
are established within the repair area for safety and ventilation
Task Stage 4: CFRP repair system termination needs. As shown in figure 11, at Hope Creek the project had to
details. be completed through a single entry point, with a butterfly valve
One of the most important individual components of the design as an obstruction. This presented several challenges for project
and implementation process for CFRP pipeline repairs, including execution; these challenges were addressed in the pre-planning
the Hope Creek project, is the water tightness of the system. Water process and steps to overcome were put in place and successfully
infiltration behind the CFRP repair system is a mode of failure,
implemented.
and this places a special importance on the points of termination,
most often at the joint regions on each end of the repair. The best 4. Removal of existing epoxy system during surface
practice for design of these “end details”, as they are referred to, preparation
includes removal of the concrete at the joints down to the steel At the pipeline segments to be repaired at Hope Creek there was
cylinder. A glass fiber is then installed directly onto the steel cyl- a build-up of epoxy on the pipe substrate, from previous repairs
inder and tied into the CFRP repair system. Glass fiber is used as performed by others, which had to be removed during surface
a dielectric barrier between the steel and the CFRP. After the glass preparation. This is shown in figure 12. There were several com-
fiber is installed and tied in, the joint is filled with an epoxy mortar ponents to this unknown condition including the amount of pipe
to create a natural, sloped transition (see figure 6). The end detail which had adhered epoxy, the thickness of the epoxy (up to ¼”) and
is finished with layers of CFRP, per the design, and this process the condition the pipe substrate would be in following removal of
as completed at Hope Creek is shown in figure 7. the epoxy. As with other unknowns, contingency plans were put
Step 5: Final Cure in place to address a variety of scenarios which might be present.
After top coat was installed at Hope Creek, the final cure of CFRP 5. Special details and substrate pitting
system took place under monitored conditions. See figure 8. In Once the substrate was sufficiently prepared at Hope Creek, and
general, CFRP systems are to have a minimum 80% cure before epoxy removed, extensive pitting, concrete patches, and other de-
the pipe is returned to service. Durability of CFRP is affected by tails such as penetrations by the temperature probe shown in figure
the degree of cure at the time the pipe is returned to full service. 13 were evident. The surfaces with pitting needed to be repaired
As discussed previously, the schedule at Hope Creek was expe- and for areas of penetration, such as the temperature probe, special
dited so the air temperature inside the pipe was elevated to over details such as shown in figure 14, were put in place.
100 degrees F to help accelerate cure. Testing has shown that the
CFRP system cures faster at higher temperatures. Qa/Qc Program
The pipeline repair project at Hope Creek included a compre-
Unique Project Challenges hensive Quality Assurance / Quality Control (QA/QC) program
From the onset of the project at Hope Creek there were several that involved multiple owner representatives, a fulltime quality
unique project challenges which were evaluated and addressed. assurance manager and a third party quality control inspector in
These issues are discussed below: a rigorous system of tracking. Up to nine (9) data points were

10 CTI Journal, Vol. 37, No. 2


CTI Journal, Vol. 37, No. 2 11
captured during each phase of the project. From environmental The extensive pre-planning process contributed significantly to
conditions where the materials were stored through to curing of the the project’s success and as well. The reliability of the cooling
installed system, each phase and step was documented and signed water system at Hope Creek was upgraded to meet long term
off throughout construction. requirements.
QA/QC documentation recorded each stage of repair implemen- Figures
tation, beginning with receipt of materials. Material lot numbers
were verified and recorded for all products shipped to the site.
QA/QC - surface preparation.
For surface preparation, all prepared concrete substrate achieved a
minimum surface profile of ICRI CSP-3 (figure 15). All prepared
steel surfaces where verified to have a near white metal blast of
SSPC-SP10, with a minimum roughness of 2 mils. The representa-
tive steel preparation is shown in figure 16. Once surface prepara-
tion is completed, the substrate was verified to be all cleaned and
dry prior to CFRP installation process begins.
QA/QC – mixing and saturation.
For mixing and saturation, lot numbers of fabrics and epoxies
are recorded as a starting point. A sample of the fabric is cut and
weighed prior to saturation in the mechanical saturation machine.
(Weight test shown in figure 17) Following this the sample is satu-
rated and weighed again. This weight test verifies ratio of fabric
Figure 1. Components of an ECP-type Prestressed
to epoxy is within tolerance (1:1 for carbon fiber fabric, 0.8:1 for Concrete Cylinder Pipe Section.
glass fiber fabric ±10%). In addition, the gap between saturator
rollers is measured and calibrated using weigh test.
QA/QC – CFRP liner installation.
During the installation process at Hope Creek there were several
check points for the pipe interior throughout the installation and
curing process. The environmental conditions monitored and re-
corded periodically included air temperature, surface temperature,
and humidity. In addition the physical installation is observed and
misalignment of over 5 degrees is addressed. One additional instal-
lation observation is development length of the installed fiber. For
Hope Creek, the design called for a minimum development length
of 12 inches in the fiber direction.
QA/QC – Terminations And Special Detailing.
Along with other data points collected throughout the pipe (envi-
ronmental), the termination points, joint details, have the epoxy Figure 2. Pipe segment during surface preparation.
mortar verified to insure a 2:1 slope.
QA/QC – cure and final walk-through.
During the final cure process at Hope Creek the environmental
conditions were monitored closely including air temperature,
surface temperature and humidity. In addition, Shore D hardness
values were recorded to assist in monitoring the progression of cure
of the CFRP system. A final check on cure time is verification of
testing in place to verify the degree of cure as determined by time
and temperature. Following cure of the CFRP system at Hope
Creek a final QA/QC walk-through took place, as seen in figure 18.

Conclusion
The use of CFRP at Hope Creek Nuclear Station provided a trench-
less, fully structural repair of the distressed 144-inch diameter
pipeline segments. Through the use of well-tested materials, a Figure 3. Mechanical saturation machine.
conservative design, experienced installers and a rigorous QA/QC
program, the project was safely and successfully implemented.

12 CTI Journal, Vol. 37, No. 2


CTI Journal, Vol. 37, No. 2 13
Figure 4. Installation of longitudinal layers of
CFRP at Hope Creek.
Figure 8. Final curing preocss being monitored

Figure 5. Installation of circumferential layers of


CFRP at Hope Creek.

Figure 9. Muck in the cooling tower basin that served as primary


access into the pipeline. Muck removed prior to installation.

Figure 6. Joint region following surface preparation and installation


of epoxymortar.

Figure 10. Two-tiered scaffolding system.

Figure 7. Joint region showing installation of CFRP layers.

14 CTI Journal, Vol. 37, No. 2


CTI Journal, Vol. 37, No. 2 15
Figure 14. Detailing of CFRP repair at
location of protruding temperature probe.

Figure 11. Single point-of-entry for pipeline repair


project at Hope Creek.

Figure 15. Verification of concrete surface


preparation utilizing ICRI CSP profile sample.

Figure 12. Epoxy build-up on pipeline interior.

Figure 16. Verification of steel surface preparation t


o near white metal SP-10.

Figure 13. Pitting of concrete and temperature


probe penetration requiring special detailing.

Figure 17. Weighing of fabric prior to saturation process.

16 CTI Journal, Vol. 37, No. 2


Figure 18. QA/QC final walk-through.

References
ASTM D3039, Standard test method for tensile properties of
polymer matrix
composite materials. American Standard for Testing and Materi-
als (ASTM)
ASTM D4541, Standard test method for pull-off strength of coat-
ings using portable adhesion: American Standard for Testing and
Materials (ASTM)
ICRI Guideline No. 310.2, Selecting and specifying concrete
surface preparation for sealers, coatings and polymer overlays.
International Concrete Repair Institute (ICRI)
SSPC-SP No.10 / NACE 2 Near-White Blast Cleaning. Society
for Protective Coatings (SSPC) and the National Association of
Corrosion Engineers International (NACE)

CTI Journal, Vol. 37, No. 2 17


Proposed Methodology For CTI ATC-
128 Sound Certification Of Factory
Assembled Towers
John Dalton and Larry Burdick • The outdoor location is a compromise, so ef-
Spx Cooling Technologies forts such as testing at night are made to minimize
The Cooling Technology Institute has a very background sound levels and to monitor the levels
successful, long standing program for thermal during testing. ATC-128 provides limits.
capability certification of factory assembled • The test pad is a reflective plane most suitable
towers, but this type of confidence or 3rd party for this type of equipment. No adjustments are
validation for published sound levels does not made to the measured sound levels due to the
exist within the industry. This paper discusses an reflective plane.
approach taken, with its successes and challenges, • Personnel are trained in equipment use, setup,
to establish a sound data set for an entire pack- and testing.
aged product model line that accurately reflects • The products being tested are part of a CTI ATC-
sound emission of all models within the line, and 201 certified product line.
proposes basic ATC-128 certification requirements • The sound equipment meets requirements for
and guidelines. accuracy per CTI ATC-128, e.g. Type 1 equipment.
John Dalton • Sound equipment calibrations are current per
Methodology for model selection, testing, and CTI ATC-128 requirements.
data analysis is provided to facilitate the breadth of data and the
• An adequate sampling of models in the product line must
considerations needed to establish information suitable for cooling be initially tested in order to apply the trends with small
tower site design and certification. adjustments to the remaining untested variations.
Sound Data Quality • The manufacturer and a 3rd party CTI licensed Test
Agency witness and take part in the sound measurements.
Ideally, sound measurements are made for each model in a qualified
test chamber and meet very high standards for testing, but this type • The manufacturer uses the measured trends and applies
small adjustments to closely estimate sound levels for the
of testing is typically reserved for equipment in sound-sensitive
product line.
areas such as indoor office spaces. Even then, it is often impracti-
• The data is reviewed and must be approved by a 3rd party
cal to test every product line variation, so good estimates based on
CTI licensed Test Agency.
empirical data are applied. Table 1 describes some of the different
ways sound data may be determined with progressively more er- Adjustments:
ror from left to right. Ambiguous data can often be recognized by A key factor in judging how many units and variations must be
vague descriptions and lack of units. tested is the determination of how far test data can be reliably ex-
Table 1: Data Acquisition & Quality of Results trapolated. Howden [2] developed equation 1 to calculate fan sound
power and presented it as part of CTI Technical Paper TP93-03
on fan noise reduction. Equation 1 states that sound power can be
calculated by combining empirical data, C, with adjustments for
tip speed, net power delivered to the air by the fan (not including
efficiency losses), fan diameter, and adjustments for variances
such as inlet and outlet obstructions, tip clearance, fan guards,
and beams located near the fan. The equations below are used to
adjust data from test conditions to published design conditions
with good success. The adjustments are applied uniformly across
all octave bands.

Pragmatic cooling tower and large industrial product testing best


practice usually blends “Ideal” and “Good Estimates” categories
to produce sound data. For example:

18 CTI Journal, Vol. 37, No. 2


PWL = the sound power level, W
C = fan characteristic value, dB(A)
utip = fan tip speed, m/s
q = fan air flow, m3/s
Ptot = total pressure = static pressure + velocity pressure, Pa
Dfan = fan diameter, m
∆dB = correction terms, dB(A)
Based on sound radiating spherically from the source, sound power
is directly related to sound pressure by equation 2, [2], so it fol-
lows in equation 3 that changes in tip speed and / or a change in
fan power directly correlate to sound pressure level.
Eq. 2: PWL = SPL + 10 LOG (S / SREF) dB
PWL = sound power level, dB (re 10-12 W)
Chart 1: Measurement locations at 1.5m (5’) and 15.2m (50’) from unit.
SPL = sound pressure level, dB (re 20x10-6 Pa)
S = surface area of a sphere, 4 π r2, m2
SREF = reference area, 1m2
C = test data, sound pressure at a specific operating condition, distance,

and location, dB (re 20x10-6 Pa)


TSdesign = tip speed of the fan whose sound level is being determined
TStest = tip speed of the test fan
BkWdesign = fan motor power of the fan whose sound level is being
determined
BkWtest = fan motor power of the test fan
Note that the correction for environment disappears. This is because
the basis data, C, is test data of the fan in the tower environment, not
a set of additions and subtractions based on theory. It includes how
Photo 1: Microphone stationed at 1.5m (5’) from a tower Louver Face (LF).
the fan and unit actually respond acoustically at the tested operat-
ing conditions. Likewise, the fan diameter term disappears as it is
absorbed in the adjustment for the ratio of fan speeds. Finally, the
adjustment for power includes the fan and drive efficiency losses,
all of which contribute to the sound characteristics of the system.

Comparison Of Overall Sound Pressure Level


Adjustments To Test Data
Tables and charts 2a-2d compare overall A-weighted SPL test
averages to A-weighted SPL predictions using equation 3: Table
2a shows base case measurements used for equation 3 predictions;
tables 2c-2d compare predictions to actual measurements. The dif-
ference between tested sound pressure levels at various conditions
and predicted sound pressure levels is a simple subtraction, rounded
to 0.1 dB. The example test data is from a large double-flow modu-
lar crossflow package cooling tower nominally 14’long x 20’high Tables 2a-2d show examples of equation 3 predicting overall sound
x 22.4’wide and with a 3658mm (12’) fan. Four fan speeds and levels to within 0.9 dB when the fan rpm is within +/- 20% of the
respective flow rates associated with the towers’ capacity at 35°C tested fan rpm, but when fan speed is different from the test con-
hot water – 29.4°C cold water – 25.6°C wet bulb conditions, (95- dition beyond 25%, error increases rapidly to as much as 1.8 dB.
85-78°F) are considered. Sound pressure levels were measured at The increasing error occurs because equation 3 models changes in
locations 5’ and 50’ from the louver face (LF), the cased face (CF), magnitude and does not capture changes in frequency.
and above the fan (AF), as depicted in Chart 1.

CTI Journal, Vol. 37, No. 2 19


Table 2e, below, is another example illustrating the closeness between linear and A-weighted SPL at low frequencies shown
of predictive results. The tower is a modular doubleflow with in Table 6 minimize the impact of error at these frequencies on
a 3658mm (12’) 8-blade fan tested in 1998 at various speeds. overall SPL. However, octave bands are important in predictive
Predictions yield accuracy within 2 dB of actual at fan speeds as sound modeling, and low frequencies such as 63Hz are also the
high as 120% of the base 214 rpm conditionTable 2e: Example of most difficult to attenuate. One of the reasons ultra-quiet fans have
predicting sound at faster than tested fan speeds. low overall sound pressure levels is A-weighting; another is that
blade pass frequencies are often shifted to the 31.5 Hz band which
we currently do not report.
Chart Group 4: Comparison of sound test data at 242 rpm

Predicted Individualoctave Band Sound Pres-


sure Levels Versus Actual Test Data
Mechanical equipment operating frequencies create sympathetic
vibrations in the tower structure and accessories. Consequently,
resonances show up as octave bands variations that occur at dif-
ferent fan speeds. Tables 3 - 5 compare the predictive model using
equation 3 to the octave bands that make up the overall sound
pressure levels shown in Appendix B.
When limited to +/-20% of tested fan speed, it is evident from
tables 3-5 that equation 3 often predicts the middle frequencies
reasonably well, but does not account for frequency variations at
different conditions. The lower frequencies are also close most of
the time, yet the most pronounced error is at the low end in the
63Hz band due to fan speed. Since water noise dominates the upper
frequencies, separating water and fan noise contributions, applying
equation 3 to the fan portion to make predictions and recombining
with water noise may yield the least error. In fact the data shows
that the sound pressure level remains essentially constant at 8000
Hz across all four fan speeds at the 1.5m (5’) louver face location.
Therefore the predicted sound level becomes more accurate if no
adjustment is made to the base data at 8000 Hz for the 1.5m (5’)
louver face location; the absolute error is reduced from 4.5 dB to
0.9 dB at 1.5m (5’) louver face in table 4.

A-weighting (dBA) reflects human sensitivity to sound at differ-


ent octave band frequencies, so the overall sound pressure level is
usually listed in dBA. Octave bands are typically communicated as
“linear,” or unweighted (dB) sound levels. The large adjustments

20 CTI Journal, Vol. 37, No. 2


CTI Journal, Vol. 37, No. 2 21
Predictions When The Fan Profile Is Constant
And The Fan Size Changes
Table 7 contains overall tested SPL’s for a small and a medium
sized crossflow package tower having different fan diameters. The
smaller tower is nominally 8.5’ long x 18’wide x 11’high. The larger
tower is nominally 12’long x 22.4’wide x 11’high.

Using the data in table 7 and equation 3, adjustments are made


to the small box (S) test data to predict medium sized box (M)
sound levels.

22 CTI Journal, Vol. 37, No. 2


CTI Journal, Vol. 37, No. 2 23
Chart 6 and table 9 show that: extrapolations of fan size by 40%
yielded 4 dB errors. (Values rounded to 0.1)
A second case of predicting sound levels of one tower size using
test data from a tower of another size considers five fan speeds. The
mid-size box (M) was nominally 12’long x 11’high x 22.4’wide
and utilized a 3353mm (11’) diameter fan. The larger box (L) was
nominally 14’long x 11’high x 22.4’wide and utilized a 3658mm
(12’) diameter fan. The data includes five fan speed and cor-
responding fan power increments. The comparison of 50’ above
fan is not included as these two boxes were tested at 24’ and 33’
above fan, respectively and adjustments could introduce error into
this comparison.

Predictions Across Both Cabinet Size And Fan


Size
Charts 7a-e and table 11 depict how much even incremental varia-
tions in box and fan size can affect predictions.

Application For Quiet Fans At Low Fan


Power And Rpm
Use of the coefficients in equation 3 will under-predict the unit
sound level when applied to very quiet fans, especially at low
power and rpm. As fan induced sound fades off with lower and
lower power, falling water noise stays relatively constant and
mechanical noise from the motor and gear reducer or belt drive
becomes audible. Changes in fan induced noise become so quiet
that they are inconsequential, so continuing to predict quieter op-
eration using equation 3 (generic black curve in chart 8) results in

24 CTI Journal, Vol. 37, No. 2


CTI Journal, Vol. 37, No. 2 25
unrealistic sound levels. Water noise is a good sound level floor Setup And Testing
for an operating cooling tower. Set up in a free-field area with a reflective plane per CTI ATC-128.
Measure fan rpm with a calibrated strobe light or other means to
match actual fan speeds with VFD settings. Measure fan power with
a calibrated power meter for each condition, then calculate motor
output power based on the motor nameplate efficiency. Measure
water flow initially using a pitot tube to validate flow meter read-
ings, then set the flow to match tower capacity at 95-85-78°F for
each model being tested. Measure sound pressure level at 1.5m
(5’) and 15.2m (50’) distances from the louver face, cased face,
and above the fan. Above fan measurements can be facilitated
with a pneumatic mast at the 1.5m elevation, or by mounting the
microphone to an extension mounted to a man-lift basket. Ensure
that background levels are adequately low and record two measure-
Chart 8: Tower sound levels become asymptotic as air noise lessons. Limits are relatively constant mechanical noise and water
ments of each point to eliminate occasional anomalies. Test with a
noise. Large crossflow package tower, 3658mm (12’) diameter very low noise fan.
3rd Party CTI licensed Test Agency witness. Compare and affirm
Choosing Units To Sound Test collected data with the Test Agency witness.
Extrapolating test data far from the test points can quickly lead to
differences between the predicted and actual sound levels, there-
fore testing multiple configurations in many box sizes is necessary
to provide high quality data. Minimizing both the number of boxes
that need to be tested initially while at the same time minimizing
how far data is extrapolated quickly reduces the number of test
candidates a manufacturer may consider. Additionally, test data
may already exist for a unit and this can further narrow the selec-
tion process.
Exceptions arise for large, low blade count fans operated at fan
blade tip speeds near 61m/s (12,000 fpm) under difficult condi-
tions. In this case, pulsing and unit vibration occur yielding sound
levels that do not trend with higher blade count variations. Addi-
tionally, if beams or obstructions are near the fan, a pulse frequency
is more likely to be generated with a small blade count or a blade Photo 2: Sound test setup.
count coincidental with the number of obstructions.
• Once towers are selected, test as many models as pos- Analysis
sible within that unit size. Recommend testing additional Apply equation 3 or other adjustments that fit test data trends, and
points that aid trending and closely align with points of test data nearest the design conditions for published literature.
adjacent box sizes that may not be tested at this time. A Trend data for very low sound models as mechanical equipment
variable frequency drive (VFD) allows many fan speed and water noise become predominant. Gain approval through the
changes in a short time. Model selection within the box 3rd Party CTI licensed Test Agency.
size can be determined by mapping in a spreadsheet the
motor speed, motor power, fan speed, drive ratio, and Conclusion
blade count. . It is clear that making adjustments to test data with the goal of
• Based on utilizing a VFD, use affinity laws to predict predicting sound levels of different, untested configurations is
power draw as a function of fan rpm for each drive ratio, imperfect, and therefore a variety of models in many cabinet sizes
motor rpm divided by fan rpm, at a given fan pitch. need to be tested to arrive at published overall sound levels with
• For each rpm associated with a different model, check an accuracy of +/- 3 dB at all locations. Testing of an initial group
that the motor power is close to the design power, per of towers coupled with use of existing sound test data yields some
CTI-ATC-128.
sound predictions that are quite accurate while others are “ball-
• The goal is to be able to mimic several models without park” estimates. ATC-201 thermal certification provides customer
major mechanical changes.
confidence that the towers they purchase perform at the published
• Change the fan pitch between test points as needed during levels claimed by manufactures. This paper provides justification
the test period to extend the number of test models. for the concept of a parallel ATC-128 sound certification program,
Counterflow towers present a few other challenges including summarized in table 11, and targeted at achieving the same goal.
determination of adjustments for falling water height variations, The process begins with testing an initial group of towers and
air inlet height variations, and fill height variations. transforms “ballpark” sound data into “good estimates” gradually,
by using adjustments based on annual recertification testing.

26 CTI Journal, Vol. 37, No. 2


CTI Journal, Vol. 37, No. 2 27
Appendix B: Octave band data, Test,
Predictions, Errors.

Appendix A: References
[1] Table adopted from TRANE training video, Evaluating Sound
Data, video @ t =14:25/51:00 minutes. http://www.trane.com/com-
mercial/north-america/us/en/products-systems/education-training/
educational-resources-by-type/continuing-education/Evaluating-
Sound-Data-HVAC-Acoustics.html
[2] Reduction of Noise Generation by Cooling Fans, CTI 1993
Annual Meeting, TP93-03, Ir. Henk F. van der Spek, Ventilatoren
Sirocco Howden B.V., Hengelo, The Netherlands
[3] CTI ATC-128 (14), February 2014, Equation 4.

28 CTI Journal, Vol. 37, No. 2


CTI Journal, Vol. 37, No. 2 29
Safety In Cooling Tower Maintenance
Magose Abraham Eju
Energy Business Total Solutions, Ltd

Abstract
Maintenance of Cooling Towers
usually poses quite a number of
occupational / personal safety
challenges. For example, the
process of removing and replac-
ing packing (fill) in a cooling
tower involves working at height
in most cases. If not well man-
aged, this exercise can result to
accident of falling; leading to Magose Abraham Eju
Figure 1.1: Picture of the 9-Cell Cooling Tower
injury and/or fatality. In order to
avert such safety incidences during Cooling Tower Maintenance, a
robust safety management system needs to be developed for every
maintenance work.
This paper uses a case study to show the various safety hazards
that can be associated with maintenance of Cooling Towers, as
well as, suggest ways that these hazards / risks could be mitigated.

Introduction
A major maintenance activity was carried out at one of the 9-cell
cooling tower of the Nigeria Liquefied Natural Gas Company. The
work scope include the production of new cooling water packing
(fill), removal and replacement of the old fill, the revamp of the
cooling water tower demister, cooling fan refurbishment, water
basin draining / cleaning / refilling and the cleaning of associated Figure 1.2: Line Diagram of NLNG Cooling Tower System
heat exchanger tubes. (Courtesy of NLNG Operating Manual)

Gas liquefaction in this plant is performed with the aid of three Cooling is achieved by pumping treated fresh water to process
cooling circuits: the cooling-water circuit, the propane refrigerant exchangers and other equipment (“users” in Figure 1.2) and then
circuit and the mixed refrigerant circuit. The mixed refrigerant back to the induced draught film type counter-flow cooling towers.
(which is a mixture of Nitrogen, Methane, Ethane and Propane The water to be cooled is fed into the top of the cells and the air
gases) liquefies the natural gas by refrigerating it down to a tem- and water streams come into intimate contact as the water falls by
perature of -160oC. The propane refrigerant circuit pre-cools the gravity into the three cooling water basins that feed the cooling
natural gas and partially condenses the mixed refrigerant. The water pump basin. Within the cells, the water is distributed to the
cooling water circuit condenses and sub-cools the propane. The spray nozzles which spray the water evenly unto the PVC pack-
entire liquefaction circuit is a complete refrigeration cycle consist- ing (see Figure 1.3 below for cooling tower cell schematic). This
ing of condensers, compressors, expanders and Joule-Thompson packing ensures that there is proper heat exchange between the air
valves, in addition to a cryogenic heat exchanger, which acts as and water by breaking up the water into fine particles.
the evaporator where the actual liquefaction of the natural gas oc- In March 2012, a comprehensive revamp maintenance of the cool-
curs. The cooling tower provides cooling water to both process and ing tower system was done. The scope of work included:
equipment heat exchangers using water for cooling or condensing • Changing defective/damaged nozzles.
within the Liquefied Natural Gas (LNG) Plant. A picture of the
• Replacing fill packs (PVC packing): This was the major
9-cell cooling tower is shown in Figure 1.1 while, Figure 1.2 is a part of the job requiring a large number of personnel
line diagram of the cooling tower system at NLNG plant. Each of working at height. All safety precautions were taken as
three LNG Plant Production Train is equipped with an independent directed in the Job Hazard Analysis. 3,000 cubic meters
open re-circulation water-cooling system. of fill packs were removed and replaced with new ones
in a specified arrangement.
• Flushing of the drift eliminators: The drift eliminators
were removed and washed with high pressure water to
clear the algae on it.

30 CTI Journal, Vol. 37, No. 2


• Cleaning the cooling water cells and pump basin (water
jets were used to flush the cooling water cell walls and
fan blades. Shovels and water jets were then used to clean
the basin into which all the dirt from the top had fallen).
• Cleaning the trash screens at the pump inlet.
• Back-flushing of piping
Maintenance work such as this usually poses a number of occu-
pational / personal safety challenges and hence requires a robust
safety program in order to avert the hazards resulting in accidents /
incidence. This discourse shall highlight some of these safety issues
that were highlighted / considered during this particular cooling
tower maintenance by reviewing the activity of removing the old
(fouled / damaged) packing (fill) of the NLNG cooling tower and
replacing them with new ones.

Figure 2.1: Work Hazard Analysis Sheet for Removal of Old Fill

Production Of Pvc Packing / Fill


Each cooling water cell contains 2,058 (1800 X 300 X 300mm)
packs and 126 (600 X 300 X 300mm). Since there are 9 cells, the
target production of the packs was as follows:
• 20,000 packs (1800 X 600 X 600mm)
• 2,000 packs (600 X 300 X 300mm)
The raw materials come in sheets and 15 sheets make up a pack.
The sheets are held together by an adhesive produced from a
Figure 1.3: Cooling Tower Cell Schematic
mixture of chemicals.
(Courtesy of NLNG Operating Manual) The fill packs were produced on site with 3 production lines each
with 4 men working 30 minute shifts. The work steps were as
Removal Of Old Packing / Fill From The follows:
Cooling Tower
1. A valid permit to work is obtained from Operations Depart-
The removal of the old fill was a labour-intensive job because of
ment
the volume of material involved. The workers / personnel were
arranged in chain across the cell and out of it so that the old packs 2. The Job Hazard Analysis (see Figure 3.1 below) is discussed
could be moved in an organized way. Two 40-tonne cranes with with the men at the beginning of every 12 hour shift.
man-baskets were used to move the removed packs to the ground 3. Mixing of the chemicals was done in a mixer in the following
floor from the top of the tower (33 feet / 10 meters height). proportion:
Safety And Job Hazard Analysis • Cyclohexanone 100% weight
The job hazard analysis was conducted with the lead men on • Tetrahydrofuran 1-3% weight
the job, safety professionals and the occupational hygienist. The • PVC powder 4-5% weight
prepared document is shown below and was discussed with the 4. Stir the mixture for about 5 minutes
workers at the beginning of each 12 hour shift on the job. The 5. Charge the rolling machines with the solution
analysis was done with the Materials Safety Data Sheet (MSDS)
6. Run the PVC sheets through the machine ensuring that the
and all hazards were examined with recommended mitigation
joining points adequately make contact with the solution.
measures proposed.
7. Stack 15 sheets together to make one pack.

CTI Journal, Vol. 37, No. 2 31


Safety And Job Hazard Analysis The pictures below show some of the work activities, the hazards
The job hazard analysis was conducted with the lead men on and the safety precautions that were applied:
the job, safety professionals and the occupational hygienist. The
prepared document is shown below and was discussed with the
workers at the beginning of each 12 hour shift on the job. The
analysis was done with the Materials Safety Data sheet (MSDS)
and all hazards were examined with recommended mitigating
measures proposed.

Picture 1: Damaged Fill of the Cooling Tower

Picture 2: Fouled Fill of the Cooling Tower

Figure 3.1: Work Hazard Analysis Sheet for On-Site Production of New Fill

Picture 3: Cooling Tower after Removal of Old Fill

32 CTI Journal, Vol. 37, No. 2


CTI Journal, Vol. 37, No. 2 33
Picture 4: Cooling Tower Removal of Old Fill
Picture 7: Height of the Cooling Tower – 33 feet

Picture 5: Old Fill Removed from Cooling Tower

Picture 8: Workers Installing Fan Blade

Picture 6: Lighting inside the Cooling Tower

Picture 9: Flushing of Pump Suction Pipe – 72-inch

34 CTI Journal, Vol. 37, No. 2


Picture 10: New Fill Installed in Cooling Tower

Conclusion
Cooling Tower maintenance work usually involves personnel
from different fields of engineering, and sometimes, from different
companies / organizations. In order to avoid downtime / loss time
injuries due to accidents, it is essential that a Job Hazard Analysis
is carried out prior to commencement of work to identify all the
possible hazards and recommend / develop mitigating measures
that will have to be put in place. Also, proper safety orientation
of the various personnel on the work site has to be ensured. For
example, it is necessary to have a safety brief / tool box session
between supervisors and workers before the start of work. At this
safety brief, all the work for that day, the possible hazards and the
necessary precautions will be discussed in details.
Finally, it is recommended that CTI develop a “Cooling Tower
Maintenance Safety Guide” that will outline minimum safety
requirements, based on best practices, for different types of main-
tenance work on Cooling Towers.

References
Atuchukwu, C. (2013), Cooling Tower Revamp. In: Nigeria
Society of Engineers Projects Report, Nigeria.
British Standards Institute (1988), Specification for Water
Cooling Towers, BS 4485, Part 2 and Part 3.
Nigeria Liquefied Natural Gas Plant Operating Manual, (2002).

CTI Journal, Vol. 37, No. 2 35


Advancements in Cleaning and
Passivation of Cooling Water Systems
Raymond M. Post, P.E., Prasad Kalakodimi, Ph.D.
Jeffrey O’Brien, and Richard H. Tribble, Chemtreat all the areas of the system where the deposits are
located. Chemical cleaning is more versatile in
Abstract being able to reach all areas of the system and
Newly installed piping and heat exchangers should more effective in terms of removing all traces of
be pre-cleaned to remove surface rust and pas- rust and deposition.
sivated to protect against initial corrosion prior Traditional methods employed for chemically
to entering service. Traditionally, high levels of removing corrosion products relied upon strong
polyphosphate [1], organic phosphate [2], ortho acids and required special precautionary measures
phosphate, and combinations have been used to for handling and application. Additionally, the
both preclean and passivate in one step. Both poly- extreme pH risked excessive loss of the base metal,
phosphate and organic phosphate have good iron se- particularly in systems with mixed metallurgy.
questering and rust removal properties. Chromate
and zinc are effective at forming a protective film, More recently, considerably safer neutral pH clean-
but have fallen out of favor due to environmental Raymond M. Post ing programs for rust removal were introduced.
restrictions. Nitrite is an effective passivator in These formulations essentially consist of (1) a
closed cooling loops, but promotes microbiological activity in open strong reducing agent to reduce ferric (Fe3+) iron oxide deposits
cooling tower systems and does not become incorporated into a into a soluble ferrous (Fe ) form, and (2) an organic phosphate
2+

durable, protective film. High levels of molybdate are also used chelant to complex the dissolved iron. The system is thoroughly
for pretreatment, but molybdate is expensive for large systems and flushed after this process to remove the dissolved iron. Though
increasingly subject to environmental restrictions. safer to handle, neutral pH cleaners still can still be very aggressive
to carbon steel base metal. Provided that the final pH of the solu-
Periodically it may also be necessary to clean equipment to remove tion is greater than approximately 6.0, these neutral pH cleaners
accumulated corrosion products and deposits in order to restore can provide some degree of passivation.
thermal performance. Traditional cleaning methods include
mechanical brushes and scrapers, acid cleaning, or neutral pH Regardless of whether the cleaning is performed mechanically or
cleaning using high levels of sequestrants or chelants. Regardless chemically, freshly cleaned steel surfaces are extremely vulner-
of whether a mechanical or chemical cleaning method is used, able to flash rust. A separate passivation step is required after the
the freshly cleaned steel surfaces should also be passivated im- cleaning process to protect the bare metal from flash rust and to
mediately following cleaning to protect against flash rust until the protect against the high initial corrosion rate that occurs when the
normal treatment program is established. Proper initial passivation equipment is initially placed into service.
has been reported to approximately double the life expectancy of To achieve a one-step cleaning and passivation, there exists a need
mild steel heat exchangers [3]. for a strong corrosion inhibitor chemistry that protects the steel
This paper describes the development and application of a new surfaces from the aggressive cleaning chemistry and provides a
non-phosphate pre-treatment chemistry that forms a truly passive strong passive film to protect the surfaces as they are placed into
film on steel surfaces. This passivation treatment can be used in service. The starting point for corrosion inhibitor development
combination with either mildly acidic or neutral pH solutions to was an inhibitor chemistry originally developed to comply with
provide a one-step cleaning and passivation. The use of electro- emerging restrictions on phosphorus discharge. The development
chemical techniques in the development and evaluation of the and application of this corrosion inhibitor for use in cooling towers
chemistry, including cyclic polarization and electrochemical im- under normal operation has been described in earlier publications
pedance spectroscopy, are discussed. A case history is provided [4] [5]. This corrosion inhibitor chemistry was selected because
to illustrate the effectiveness of this technology for cleaning and it interacts directly with metal surfaces to form a reactive polyhy-
passivation as part of a comprehensive program for protecting droxy complex (RPC) that is independent of calcium, pH, or other
industrial cooling systems from corrosion. water chemistry consitutents. It also possesses many of the other
desired attributes including an excellent aquatic effects profile. The
Introduction ability of the RPC chemistry to protect and passivate the base metal
Industrial cooling systems rapidly form iron oxide deposits which during preoperational cleaning is the focus of this study. Efficacy
reduce their heat transfer efficiency. It is common to mechanically of the RPC chemistry was compared to traditional polyphosphate
clean these systems when the iron oxide deposits become excessive. and organic phosphate cleaning and passivation chemistries.
Mechanical means such as scraping, brushing, sand blasting, and
high pressure water washing are few of the well-known mechanical
Laboratory Studies
cleaning methods. Mechanical cleaning, while effective in many Most metallic corrosion occurs via electrochemical reactions at the
cases, is time-consuming, laborious and expensive. Also, it is interface between the metal and an electrolyte solution. Since cor-
not always possible for the physical removal methods to access rosion phenomena are electrochemical in nature, they are governed
by the measurement of equilibrium corrosion potential (Ecorr) of

36 CTI Journal, Vol. 37, No. 2


the metal surface. During the late ‘70s and early ‘80s, corrosion Open Circuit Potential (OCP) Evaluation – Open circuit poten-
specialists began to discover that electrochemical instruments and tial measurements were made in blank water with bare mild steel
techniques could be valuable tools for mechanistic understand- (MS) coupons and also on the coupons passivated as described
ing and problem solving. Electrochemical methods can be very above using the organic phosphate, polyphosphate, and RPC
helpful in rapidly evaluating the persistence of a passive layer chemistries. OCP measurements were taken at 5 minute intervals
formed on a metallic specimen in the presence of inhibitor when for an extended period of 36 hours. Comparing the OCP profiles
that specimen is transferred into a second solution not containing obtained for these electrodes indicate that after the electrode surface
a corrosion inhibitor. was passivated using RPC, the OCP is >200 mV more positive than
that the MS coupons passivated with polyphosphate and organic
Electrochemical Experimental
phosphate chemistries. These results indicate that the RPC pas-
The passivation chemistries selected for laboratory evaluation were sive film is thermodynamically more stable than phosphate and is
two well-established phosphate chemistries and the RPC chemistry, very effective in reducing corrosion. Moreover, the passive film
which were compared to untreated blank water samples for their retained its effectiveness in the untreated blank water, with the OCP
ability to form a persistent protective film. The organic phosphate decreasing <100 mV during the 36-hour exposure. In the cases of
(HEDPA) was evaluated at a concentration of 1,250 mg/L as PO4=, the polyphosphate and organic phosphate passivations, the OCP
the polyphosphate (TKPP) was evaluated at a concentration of 320 values quickly become more negative, indicating rapid dissolution
mg/L as PO4=, and the RPC was evaluated at a concentration of of the passive film by diffusion of aggressive ions and subsequent
12 mg/L. The blank water contained 150 mg/L Ca as CaCO3, 100 increase in corrosion rates.
mg/L Mg as CaCO3, 100 mg/L M-Alkalinity as CaCO3, 50 mg/L
chloride as Cl-, 10 mg/L silica as SiO2, at pH 8.0.
Working electrodes were made from cylindrical shaped carbon
steel electrochemical coupons. Corrosion resistance performance
of different passivation chemistries was evaluated using electro-
chemical techniques including open circuit potential (OCP), cyclic
polarization (CP), and electrochemical impedance spectroscopy
(EIS) measurements. Carbon steel coupons were passivated for 8
hours in baths containing the passivation chemistries. After pas-
sivation, the electrochemical coupons were rinsed with RO water
and placed in the blank water. The pH of the blank solution was
maintained at 8.0-8.2 using caustic soda. Electrochemical mea-
surements were performed in a three electrode cell consisting of a
graphite counter electrode, an Ag/AgCl reference electrode, and
Figure 2. Open circuit potential after passivated coupon
the carbon steel coupon. Electrochemical impedance spectroscopy is placed in untreated water
(EIS) measurements were performed using the excitation signal
of 10 mV sinusoidal potential through a frequency domain from
100 kHz down to 10 mHz. The impedance diagrams were recorded
at the equilibrium OCP, while polarization scans were traced at a
rate of 0.5 mV s-1. Unless otherwise stated, each experiment was
conducted in at least triplicate using freshly prepared solution
in each case and with the average of the reliable data reported.
Figure 1 shows the electrochemical testing apparatus used for the
experimental work.

Figure 3. Cyclic polarization after passivated


coupon is placed in untreated blank water
From the cyclic polarization curves, some electrochemical cor-
rosion kinetic parameters can be obtained, such as the corrosion
potential (Ecorr), Tafel slopes, corrosion current density (Icorr), and
corrosion rates in Table 1.

Figure 1. Potentiostat arrangement used for electrochemical studies Table 1. Cyclic polarization data of various passivation treatments

CTI Journal, Vol. 37, No. 2 37


An immediate observation from comparing the Ecorr (Figure 3 It is clear that the semicircle diameter increases in the order:
& Table 1) and OCP (Figure 2) is that both the electrochemical Polyphosphate ≤ Unpassivated < Organic phosphate << RPC
techniques are in good agreement. Small differences could be
Since the diameter of the capacitive semicircle represents the resis-
attributed to the effect of polarization. It is clearly seen from Fig.
tance of the coating, an increase in diameter represents increased
3 that the coupon passivated with RPC chemistry has lower pas-
resistance. Figure 4 illustrates that the corrosion resistance of RPC
sivation current and lower currents at all applied potential values.
passive film is much greater than other traditional organic phos-
The corrosion resistance of different passivation treatments was
phate and polyphosphate passivation treatments, consistent with
calculated using Stern-Geary Equation:
OCP and CP measurements in Figures 2 and 3. The EIS results
Icorr = babc/ ( ba+bc)2.303 Rp (1) also confirmed that RPC passivation forms a highly protective
From the graph and the table, it is clear that the corrosion current film on mild steel in the test water, significantly outperforming
for different passivation treatments increases in the order: the polyphosphate and organic phosphate passivation chemistries.
(Icorr)RPC < (Icorr)Organic phosphate < (Icorr)Polyphosphate (2)
Comparison with Nitrite and Molybdate Passiv-
Inhibition efficiency (% Inh) of various passivation treatment
ation
programs has been calculated using the formula:
Carbon steel coupons were passivated for 6 hours in baths contain-
% Inh = (Icorr)blank – (Icorr)treatment / (Icorr)blank * 100 (3)
ing the treatment solutions of interest in blank water (described
It is clear from Table 1 that higher inhibition efficiency was ob- above). Baths 1 and 2 contained the RPC corrosion inhibitor at
tained by the RPC passivation treatment than the polyphosphate 12 mg/L and 6 mg/L respectively, bath 3 contained 100 mg/L of
and organic phosphate chemistries, indicating the stability and molybdate (as MoO4=), and bath 4 contained 1,200 mg/L of nitrite
persistent nature of the RPC non-phosphate passive film. (as NO2-). Following the 6-hour passivation period, the coupons
Electrochemical Impedance Spectroscopy Measurements – The were rinsed in DI water and placed into untreated blank water for
long-term corrosion behavior of the three passivation treatments a period of 3 days to evaluate the persistence of the passive film.
on the surface of carbon steel electrodes was probed by electro- After exposure, the coupons were removed and photographed.
chemical impedance spectroscopy (EIS). Electrical resistance is a Average corrosion rate was determined by weight loss. The coupon
well-known parameter for characterization of the ability of a circuit appearance and corrosion rate are shown in Table 2.
element to resist the flow of electrical current. Similar to resistance,
impedance is also a measure of the ability of a circuit to resist the
flow of electrical current. However, unlike resistance, impedance
is not limited by the simplifying properties of a single resistor.
Electrochemical impedance is usually measured by applying an
AC potential to an electrochemical cell and then measuring the AC
current response through the cell. EIS is one of the most useful
techniques to discern the corrosion mechanism, evaluate protective
films on metals in aggressive solutions, and analyze film formation
at the metal-electrolyte interface. When a metal surface is covered
(passivated) with a protective layer, the corrosion is controlled by
the transport of the species in the protective film through diffusion
channels resulting from passive film breakdown or micro cracks.
The response can be represented as the relationship between the Table 2. Corrosion rate and coupon appearance after 6-hour
passivation and 3-day exposure to untreated blank water.
imaginary and real components of impedance, referred to as a
Nyquist plot. Figure 4 presents the Nyquist impedance diagram
recorded at OCP for various passivation treatments. Preoperational Cleaning Passivation OCP
Evaluation
One of the major needs for passivation chemistry is during pre-
operational cleaning when the freshly cleaned surfaces are most
vulnerable to flash rust just prior to the equipment entering service.
To simulate the condition of new steel heat exchangers subject to
environmental rusting during the construction period, a new steel
heat exchanger tube was first allowed to rust outdoors for several
days, then cut into sections approximately 3 in (7.5 cm) long. One
rusted tube section was placed into a beaker containing a neutral
pH, precleaning and passivation solution containing organic
phosphate, reducing agent, and surfactant. A second section from
the same tube was placed into another beaker containing the same
chemistry with the addition of the RPC passivating chemistry.
The solutions were stirred for 6 hours until the both specimens
Figure 4. Nyquist plots for different passivated coupons in untreated water were substantially free of rust. The pH of the cleaning solutions
remained in the 6.5-6.7 range during the 6-hour exposure. Figure 5
and 6 show the specimens in the beakers before and after cleaning.

38 CTI Journal, Vol. 37, No. 2


CTI Journal, Vol. 37, No. 2 39
metal. Following the cleaning, both specimens were placed in baths
of untreated blank water for 24 hours to evaluate their ability to re-
sist corrosion using open circuit potential as illustrated in Figure 7.
In this test, the precleaning solution containing the RPC chemistry
exhibited an open circuit potential more than 200 mV higher than
the solution without RPC (Figure 8).

Figure 5. Precleaning and passivation study at start of 6-hour exposure.


Specimen on right contains the RPC passivation chemistry.
Figure 8. OCP of cleaned specimens over 24 hours in untreated water, with and
without RPC passivation chemistry
Figures 9 and 10 show the two specimens after the 24-hour expo-
sure to the untreated blank water. The steel tube section that was
exposed to the cleaning solution containing the RPC passivation
chemistry clearly resisted rusting, while the specimen not exposed
to RPC readily rusted after cleaning.

Figure 9. Tube section cleaned and passivated with RPC, organic phosphate,
and reducing agent after 24-hour exposure to untreated blank water.

Figure 6. Precleaning and passivation study at completion of 6-hour exposure.


Specimen on right contains the RPC passivation chemistry.

Figure 10. Tube section cleaned with organic phosphate and reducing agent
after 24-hour exposure to untreated blank water.

Copper Sulfate Verification Test


A simple, visual test for passivation is to expose steel or stain-
less steel specimens to a copper sulfate solution. Immersion in a
copper sulfate solution is one of the verification tests described in
the ASTM A967 standard specification for chemical passivation
treatments [6]. Copper ions in solution act as electron acceptors
Figure 7. OCP measurement of cleaned tube specimens at the cathode of the corrosion cell and readily electroplate onto
unpassivated or weakly passivated surfaces, producing a very
As shown in Figure 6, after 6 hours, the cleaning solution without
pronounced copper discoloration to the steel surface (Figure
the RPC passivation chemistry is noticeably darker than the solu-
11). The test can be performed in the field as an acceptance test
tion with RPC chemistry due to additional corrosion of the base

40 CTI Journal, Vol. 37, No. 2


CTI Journal, Vol. 37, No. 2 41
42 CTI Journal, Vol. 37, No. 2
CTI Journal, Vol. 37, No. 2 43
criteria for equipment passivation procedures. A well passivated
steel coupon will exhibit considerable resistance to copper plating
even when exposed to a solution containing a high concentration
of copper ions.

Figure 13. Passivated steel coupon after 2 day exposure to pH 3


cleaning solution containing RPC

Figure 14. Coupon surface magnified 40x


Carbon steel coupons exposed during the cleaning process were
Figure 11. Mechanism of copper electroplating onto steel surfaces
further evaluated to determine the corrosion resistance of the
Exposure to a copper sulfate solution was used to evaluate the ef- passive RPC film. The copper sulfate verification test described
fectiveness of RPC passivation chemistry compared to established earlier was performed on the exposed coupon shown in Figure 13
organic phosphate and polyphosphate passivation chemistries. and compared to a new, carbon steel coupon. The lower half of
Mild steel coupons were treated for 6 hours in passivation solutions each coupon was exposed to a 15% copper sulfate solution for 20
and then exposed to a 15% copper sulfate solution for 20 seconds. seconds. As shown in Figure 15, the new coupon is copper plated
Following exposure, the coupons were rinsed with deionized water, on its lower half, while the coupon passivated during the acidic
dried, and visually inspected for copper plating as an indication of chemical cleaning effectively resisted copper plating.
passivation effectiveness. A relatively high concentration of cop-
per sulfate was used in order to produce severe test conditions that
would discriminate among the passivation chemistries. As shown
in Figure 12, the RPC passivation chemistry clearly demonstrated
effectiveness superior to organic phosphate and polyphosphate
in the copper sulfate exposure test,
and at significantly lower treatment
concentrations.
Organic phosphate passivation
640 mg/L as PO4=, for 6 hours
Surface is copper plated.
Passive film failed.
Polyphosphate passivation Figure 15. New coupon (top) and coupon passivated with RPC during the acidic
320 mg/L as PO4=, 6 hours chemical cleaning (bottom) after half-exposure to a 15% copper sulfate solution
Surface is copper plated. for 20 seconds. Note the copper plating on the new coupon as compared to the
Passive film failed. absence of copper plating on the RPC-passivated coupon.

RPC passivation A second carbon steel coupon that was exposed during the chemi-
3.8 mg/L as RPC, 6 hours cal cleaning process was evaluated for passivation by placing it
Steel resisted copper plating, indicating a strong passive film. in untreated blank water and comparing its OCP to a new coupon
Figure 12. Passivated coupons after 20 sec. exposure to 15% CuSO4 solution over time as described in the Experimental section. The RPC pas-
sivated coupon maintains its potential for several days in untreated
Preoperational Cleaning Application water (Figure 16).
A Midwest manufacturing plant resumed operation after being idle
for 18 months. Prior to resuming operation, the cooling system
was cleaned using an acidic cleaning solution with surfactant and
dispersants recirculated at a target pH of 3.0. The cleaning solution
was inhibited using a triazole together with the RPC passivation
chemistry. Mild steel coupons were installed to evaluate base
metal loss and passivation effectiveness. The steel coupons were
removed for after 2 days showed a dull grey passivated surface
(Figures 13 and 14). The presence of RPC on the coupon was
confirmed by surface analytical methods.

44 CTI Journal, Vol. 37, No. 2


CTI Journal, Vol. 37, No. 2 45
Figure 18 shows the coupon appearance before and after cleaning
in the neutral pH formulations with and without RPC.

Figure 16. OCP measurement over time on carbon steel coupon passivated with
RPC during chemical cleaning compared to a new carbon steel coupon.
As further visual evidence of passivation, the untreated OCP test Figure 18. Before and after cleaning photos of the corroded coupons in standard
baths were photographed after 3 days. Figure 17 shows the rust- neutral pH cleaners with and without RPC. Cleaning duration was 6 hours for all.
colored solution containing the new coupon on the left and the Formulations #1, #2, and #3 with RPC cleaned more effectively
clear solution containing the RPC-passivated coupon on the right. and left the coupons with a dull grey finish consistent with pas-
sivation. A fresh corrosion coupon was inserted into each of the
cleaning bath to study the effect of the cleaner product on the base
metal. Corrosion rates were measured by weight loss. The clean-
ing solutions containing RPC also exhibited a ~60-70% reduction
in corrosion rate as compared to the standard neutral pH cleaner.
Conclusions
Effective cleaning and passivation of heat exchangers and pip-
ing is required during initial start-up and periodically during the
operational life of the system. Using advanced electrochemical
techniques, a powerful new passivation chemistry has been devel-
oped that protects the base metal during the cleaning process and
provides a persistent passive film that resists corrosion for several
days as the system is placed back in service. The new Reactive
Polyhydroxy Complex (RPC) based chemistry was found to be
superior to traditional polyphosphate and organic phosphate based
Figure 17. OCP test baths after 3 days showing the rust-colored
solution containing the new coupon on the left and the clear
treatments in terms of forming a stable and persistent passive film.
solution containing the RPC-passivated solution on the right. The RPC chemistry can be used in conjunction with either acidic
or neutral pH cleaners to achieve simultaneous cleaning and pas-
Neutral pH Cleaning Formulations With RPC sivation, eliminating the need for a separate passivation step.
Neutral pH formulations containing various levels of RPC were
studied for their effectiveness in cleaning rusted coupons and were Acknowledgements
also evaluated for corrosion rate on fresh coupons. The authors gratefully acknowledge DeAnn Wills-Guy of Chem-
Treat for conducting the electrochemical studies.
Cleaning solutions were prepared according to Table 3 and applied
at a dosage of 10%. Nomenclature
ba Anodic Tafel slope, Eq. (1)
bc Cathodic Tafel slope, Eq. (1)
Beta Anodic Tafel slope, mV/decade, Table 2
cfu Colony forming units
CP Cyclic polarization
Ecorr Corrosion potential at equilibrium, mV or V
EIS Electrochemical impedance spectroscopy
HEDPA 1-Hydroxyethylidene-1,1-diphosphonic acid

Table 3. Table showing neutral pH cleaning solution


Icorr Corrosion current at equilibrium, mA or A
chemistries used to clean coupons in Figure 18.

46 CTI Journal, Vol. 37, No. 2


mHz Megahertz References
mpy Mils per year, 0.001 inch/yr., equivalent to 0.0254 mm/yr. 1. Initial Conditioning of Cooling Water Equipment, NACE
OCP Open circuit potential mV or V RP0182-95.

Rp Polarization resistance, Eq. (1) 2. Rust Removal and Composition Thereof, Waller, J.E., Gray,
J.A., and Aston, D.A., US Patent 4,810,405, 1989.
RPC Reactive polyhydroxy complex
3. Proper Initial Passivation of Cooling Water Heat Exchangers
TKPP Tetrapotassium pyrophosphate
- Is This a Lost Art? P.R. Puckorius, 63rd International Wa-
Z Impedance (of surface), Ohm-cm2 ter Conference, IWC-02-62. Engineers' Society of Western
Zimag Imaginary component of surface impedance, Ohm-cm2 Pennsylvania, Pittsburgh, 2002.
Zreal Real component of surface Impedance, Ohm-cm2 4. Development of Next Generation Phosphorus-Free Cooling
Water Technology. R.M. Post, R.H. Tribble, & J.R. Richard-
son, International Water Conference, IWC-10-23. Engineers
Society of Western Pennsylvania, San Antonio, 2010.
5. Development and Application of Phosphorus Free Cool-
ing Water Treatment. R.M. Post, R.P. Kalakodimi, & R.H.
Tribble, CTI Journal Vol. 35, No. 2, pp. 52-67, 2014.
6. Standard Specification for Chemical Passivation Treatments
for Stainless Steel Parts, ASTM A967/A967M-13, 2013.

CTI Journal, Vol. 37, No. 2 47


Mechanical Behavior of Polymer Fills
Nina Woicke, PH.D, Daniel Dierenfeld,
ENEXIO Water Technologies GmbH, 2H PP normally performs well under Gardner condi-
Components and Solutions tions, because it is more ductile then PVC.
To visualize this effect, in Figure 2 the results of
Abstract instrumented indention test are shown. The di-
The paper outlines the mechanical properties of ameter of the test “dart” in this case was 1.65 mm
polymer fills and discusses the influence of dif- (0.065’’). Two foils with the same weight, one PP
ferent parameters (e.g. PP vs. PVC, foil thickness, and one PVC, have been tested. The results can
design) with particular regard to cooling tower be seen in Figure 2. The PVC breaks, when the
aspects. maximum force is reached. PP on the other hand
Introduction will stretch locally more than 100% extra distortion
before the final failure.
This paper will give an overview on the mechani-
cal behavior of cooling tower fills. The main focus Therefore normally impact tests with notched
Nina Woicke test bars like the notched Izod test are used for
is thereby the short term behavior. Temperature
dependent and time dependent properties play their role in the the characterization of PP (blue specimen in Figure 3). The notch
cooling tower fill life as well, but here these subjects will only increases the local stress, so that as well PP will break at the hit of
touched briefly. the hammer and a distinctive value can be given. Unfortunately
the values of Gardner tests and notched Izod tests are not directly
The paper is structured in three sections:
comparable, since the overall test set-up is quite different.
1. Material properties
Heat deflection temperature
2. Mechanical theory on cooling tower fills
3. Results of specific tests in comparison The heat deflection temperature (ASTM D648) is a technological
test. A small specimen is put in an oil bath and preloaded with a
Material properties certain load (either 66 or 264 PSI) under bending. The bath then
The two main materials used for cooling tower fills are polyvinyl- is slowly heated. The heat deflection temperature is reached, when
chloride (PVC) and polypropylene (PP). In the following Table 1 the specimen has a certain deflection (0.2%). Even though this
the mechanical material properties (minimum values) defined by test is used in the STD 136 as a reference, neither the load case
CTI STD 136 are summarized. These properties are for the non- nor the deflection limit is in linked to the ones of drift eliminators
engineered material only and are mainly used for general quality or fills. Therefore this temperature cannot directly be related to
control. temperatures in use in a cooling tower.
Strength The mechanical behavior of PP and PVC is different as the tempera-
The strength is the maximum load a tensile bar can take under ture increases. PVC is very rigid and stiff at temperatures between
tensile load (ASTM D638 or EN ISO 527-1/-2) or a bending bar room temperature and about 55 °C (131 °F). The change in modulus
(ASTM D790 or EN ISO 178) under flexural load. In both cases is very low for this regime. Above 55 °C the material reaches its
the usual standard test bars are much thicker (about 10 times) than glass transition and changes its behavior drastically, so there is not
the final foils. That means if you want to test 100% according to a lot structural integrity left above 65 °C (149 °F). To reduce this
the standard it is not possible to measure this property directly effect, PVC can be partly mixed with other polymers like ABS or
from the foils. SAN. This composition is is normally referred as “HPVC”. Another
possibility is to use chlorinated polyvinyl chloride (CPVC), which
Modulus as well has higher glass transition temperature.
The modulus defines the stiffness of the material. That means the
PP on the other hand has its main melting onset at about 140 °C
stiffer the material the less deformation at certain loads. Therefore it
(284 °F). Therefore upto 110 °C (212 °F) the mechanics if sized
is not recommended to use the strain as an absolute failure criteria
properly can be still strong enough and applicable to be used [5].
over both materials discussed.
Nevertheless the softening slope of the material is there as well
Both PP and PVC modulus can be measured under tensile or and at the same time the thermal degradation through oxidation
flexural conditions since they are not directional reinforced (like has to be carefully addressed through engineering and product
e.g. FRP). design characteristics. Therefore at higher temperatures the use of
Impact resistance specially heat stabilized and probably thicker material is advisable.
There are several different standards available to measure the Density
impact resistance. One standard is the Gardner impact strength, Additionally to the properties mentioned in the CTI STD 136
which is also called “falling dart” test, which already describes the material density has to be recognized for the full understanding
method of impact (test set-up in Figure 1). of the mechanical behavior of cooling tower fill.

48 CTI Journal, Vol. 37, No. 2


The densities of the two materials differ quite a lot due to the fact Asp: Specific surface of the fill (ft²/ft³). This is a fill type specific
that the chloride in the PVC is a rather “heavy” atom. Additionally number and is normally provided by the fill supplier. Nevertheless
both materials can be added with inorganic fillers. The density in Table 2 there are indication numbers for some usual fill types.
ranges about 0.92-1 kg/l (0.033-0.036 lb/in³) for PP and 1.35- The area has to be divided by two because of the two sided of a foil.
1.55 kg/l (0.049-0.056 lb/in³) for PVC. Therefore, if you compare Weighing a fill is a simple method to cross check, if the foil thick-
two products (PP and PVC) with the same material gauge, the PP ness is about what the supplier stated.
product will be lighter than the PVC.
Buckling under compression
Mechanical behavior of cooling tower fill If the foil bonding is done properly cooling tower fills under pres-
The overall mechanical fill properties are a combination of the sure normally fail by buckling of the foil. In Figure 5 you can see
material properties, the fill design, the thickness respectively the a completely vertical fluted fill, which is buckling between the
overall weight of the product and the quality of the bonding. The connection points.
second aspect is not the main focus of this paper, but of course has The theory of buckling is rather complicated; nevertheless it is
to be taken into account as well. A fill that has been badly bonded necessary to understand the main idea to understand the fill be-
can have totally different strength than one that is bonded properly. havior. To keep it as simple as possible we model it as a column
In this case the number of bonding points and the bonding quality between two supports, which are represented by the bonding points
of a single point is important. (see Figure 6).
Foil thickness Euler established a formula for the maximum force a slender
In former times nearly all film fills were produced by thermoform- column can carry (Euler case 2 [3]). Of course, cooling tower fills
ing. At that time it was established, that a fill was defined by its are far from a perfect slender column, but still the equation can be
foil (sheet/film) thickness (normally measured in mil), especially used for the general understanding:
before forming. In that case you have a flat foil, where it is pretty
easy to measure the thickness at any point. 2
After the corrugation process, the measurements already become
a lot more complicated depending on the complexity of the
F: The maximum Force
corrugation or forming. After bonding only the outer edges are
available for any check on site. Even with a single foil you will E: The modulus, influenced by the material as discussed in 3.2
have a differential of the thickness throughout the sheet and it is LS: The unsupported length. This is influenced by the design as
then difficult to measure to a fair level of accuracy, the actual foil well as the quality of bonding. For example, a 1 ft high fill, which
thickness for the entire sheet. is only bonded at the edges, is four times more likely to Eulerian
Nowadays there are several more methods to produce a fill; e.g. buckling than a fill, which has an additional bonding in the middle
injection molding for the so called “hybrid” fills or the direct of the fill.
extrusion processes without any defined semi-finished product in
IF: is the so called momentum of inertia of the foil, which is in-
between. These methods even allow having an optimized, uneven
fluenced by the design (overall design as well as the corrugation
foil thickness to enhance higher loaded parts. Especially the edges
design). This is hard to calculate for a so complex geometry like
of the fill are more likely to be battered by the water impact or
inspection traffic. Erosion is therefore a typical cooling tower fill a corrugated cooling tower foil. Nowadays the only meaningful
problem. method to calculate the buckling is finite elements analysis (FEA)
(see Figure 7 as an example). Only with this method you can get
In this respect an uneven thickness distribution to strengthen the the height and distribution of the load.
edges (Figure 4) of the fill can help to minimize this problem with
PP as well as with PVC fills. If you have a look at one single, rectangular element, the momen-
tum of inertia is then easy to calculate:
Therefore it is recommendable to use the fill weight as an extra
measure. It is, of course an overall property and doesn’t give you
any indication on the specific distribution, but a fill from the same 3
material and same design, bonded the same way, but has e.g. 30%
higher weight will behave stronger under mechanical load. IE: Momentum of inertia of element
To get an idea how the weight and the mean foil thickness are WE: Width of the element
related, use the following equation:
TE: Thickness of the element, which normally is directly related
to the thickness of the foil. That means the thickness of the foils is
1 influencing the local momentum of inertia by the power of three
and is therefore an important figure.
TF: Thickness of the foil (mil)
M: Mass of the fill (lb) Bending
L: Length of the fill (ft) The fill in the cooling tower is not only loaded under direct pres-
W: Width of the fill (ft) sure, but the bottom layer is normally on a support structure, which
H: Height of the fill (ft) implies local compression point loading as well as bending into
ρ: Density of the fill material (lb/in³)
the fill. In most of the cases that makes the bottom layer to the

CTI Journal, Vol. 37, No. 2 49


most critical layer. This is not only influenced by the fill, but also One approach of comparison is to use the same specific fill weight.
by the distance between the supports, the width of the support and Due to the differences in densities this results in differences in the
the arrangement of the fills. As an example, according to elastic foil thickness. Under these circumstances, the pure compression
bending theory [4] two 3 ft long blocks will bend more than double strength of PVC fills is lower.
to one 6 ft long block on the same support structure. For the same border conditions the deflection under load looks dif-
Similar it is true that the same fill one or two feet high will bend ferent: Both, the bending and the local deformation for the PP fill,
differently. Even though there may be a small drawback in per- are slightly higher due to the lower modulus of the material. For
formance, normally it is recommendable to use a 2 ft high block a direct visual comparison it is useful not to take the direct strain
as a bottom layer to improve the overall stability. but its inverse (which is called bending resistance and compression
Bending as well depends on the stiffness and the momentum of resistance in this paper).This is done, so higher values represent
inertia, but this time not only of the local foil, but of the whole fill “better” and not the opposite.
structure. This is again not really computable by easy equations, The overall comparison summarized in Figure 11.
so the best possibility to get an impression of the flexural behavior Fill weight/foil thickness
is a direct test of the fill.
The foil thickness (fill weight) has obviously a major influence on
Results of specific cooling tower fills the fill mechanics (for one material the thicker the better). If you
To illustrate the theoretical background of the previous chapters, look at the same fill weight (see also 5.3) the compression strength
in this chapter specific data is given, mainly in a qualitative manor of the PP is higher. In contrast to that, if you look on the same
to emphasis the main differences. They all refer to data measured mean foil thickness of the formed foil, the PVC fill has the higher
with 2H® fill. The measured fill were the KFP 619/KVC 619 for compression strength. This variation is visualized in Figure 12.
the 19 mm crossfluted fill, the KPP 612 for the 12mm crossfluted Concerning the bending the influence of the specific thickness is
fill, the KVP 623 for the vertical fluted fill and KGP 620/KBP 620 even harder to determine. The general behavior is of course that
for the offset fill types. All results are plotted qualitatively to show the deflection is lower the higher the fill weight respectively the
the general influence of the parameters. foil thickness is. As an example 22 kg/m³ (1.37 lb/ft³) and 27 kg/
Compression test set-up m³ (1.68 lb/ft³) PP 19 mm crossfluted fill is compared in the local
compression as well as in bending in Figure 13. You can see that
The pure compression tests have been conducted with a 305 mm
the lower weight has significantly higher deformations.
(1 ft) x 305 mm (1 ft) x 610 mm (2 ft) set-up at room temperature
and a test speed of about 2 ft/min (see Figure 9). The respective If you choose a too low weight (too thin foil) for a certain distrib-
strain rate in this test is in the magnitude than the strain rate to uted load, the fill will fail at the support by a local buckling (see
determine the tensile strength (ASTM D638). So these results can Figure 14). Therefore one must carefully choose the right fill for
be used for direct correlations between the material strength and the design load.
the fill strength. Of course this only tests the short time behavior of Influence of fill design
the fill. Normally the same test is done with 3 different specimens. Beside all aspects discussed before, the fill design has a tremendous
In Figure 8 one can see the normal output of the force deflection influence on the overall stability and performance. And again, the
curve. For the evaluation of the tests the mean value of the maxi- aspect of buckling (due to compression) has to be looked at as
mum force is taken. well as the bending.
Bending test set-up for the bottom layer Fills with cross flutes have normally a good combination between
The tests to simulate the bending situation of the bottom layer the necessary properties. They even have an extra value in their
were done with 610 mm (2 ft) high and 2440 mm (8 ft) long test design: They distribute local load along the flute inclination into
specimens. The support structure was spaced 900 mm (2.95 ft) and the fill. Due to the bonding of the foils, the cross-fluted fills behave
a cantilever of 320 mm (1.05 ft) on each side (see Figure 10). The similar to carpenters work in house. Fills with offset are normally
width of the support was 60 mm (2.4 ’’). There were seven measur- a lot more sensible to pure compression than the fills mentioned
ing points along the side of the fill. The different points were chosen before, while for the bending issue the offset again helps to sta-
to separate the local deformation at the supports from the bending bilize the fill.
between the supports, which adds up to the overall deflection. In Figure 15 you can see the comparison of different compression
Static loads were applied on top of the fill (see Figure 10). The stability of different fill types in PP. The blue columns are for a
distance of fill points to the flooring was detected before loading comparison at the same foil thickness of the final product and the
and after 24 h after applying the load. red columns at the same specific weight, which equals a calcu-
Material lated mean thickness of the 15 mil of 19 mm crossfluted fill. The
compression strength has been normalized to the value of the 19
Since all of the mechanical data is very different, it is not so easy
mm fill, so this diagram can be used in principle for other weights/
to directly compare PP with PVC fill. And as well it very much de-
thicknesses and PVC similarly.
pends on the specific design, how high a certain effect is. Therefore
in the following only relative measures are given to illustrate the What you can see is that you have to be careful when you compare
general tendencies. As a reference 27 kg/m³ (1.68 lb/ft³) 19 mm different fills, because due to the differences in corrugation height,
cross fluted fill, thermally bonded, has been taken as a reference the general tendencies can vary (see also 5.3).
of being 100%.

50 CTI Journal, Vol. 37, No. 2


CTI Journal, Vol. 37, No. 2 51
At the same time there is the bending behavior of the different fills. Figures
From the pure compression point of view the fully vertical flute
is the best, because the column-like structure supports strongly in
the vertical direction.
These columns on the other side have only a low stability into the
other two directions. Therefore bending stability of this design
is not as good. Therefore the bottom layer of such fill should be
stiffened with flat sheets as shown in Figure 16. This problem is
due to the design and has to be addressed for PP and PVC alike.
The results of the deformations of these variations are summarized
in Figure 17. Again here fill with 27 kg/m³ (1.68 lb/ft³), thermally
bonded, is compared. That reflects about 15mil for the 19mm
crossfluted and
Summary
This paper illustrates different aspects of the cooling tower fill
mechanics:
PP and PVC differ in all properties and therefore can’t be directly Figure 1: Gardner “falling dart” test set-up
compared by comparing pure material values
• The material thickness is an important measure, but hard
to check. Therefore the specific weight of the fill is the
better measure for a comparison for all product types
• The design has equal if not more influence as the material
on the overall mechanical strength of a fill
• To judge the overall mechanical behavior of a fill, it is
necessary to evaluate the bending properties as well as
the compression strength of the final product
Even though it has not been verified by this study, the mechanical
strength of the fill also depend on the type and frequency of the
bonding points.
For the final decision for a certain cooling tower fill it is therefore
necessary to evaluate carefully all aspects of engineering and
product design.
Figure 2: Instrumented indention test of PP and PVC foils of the same weight
Tables

Table 1: Overview of material properties of PP and PVC [1]

Figure 3: Notched Izod test set-up

Table 2: Overview of specific surface area of different fill types Figure 4: Sketch of a 2 ft wide foil/sheet with optimized thickness distribution

52 CTI Journal, Vol. 37, No. 2


CTI Journal, Vol. 37, No. 2 53
Figure 5: Foil buckling (thermally bonded)

Figure 7: FEA visualizing the stress distribution of two bonded fill sheets

Figure 6: Principle of buckling [2]

Figure 8: Force-deflection curve of a compression test

Figure 9: Picture of the test set-up

54 CTI Journal, Vol. 37, No. 2


CTI Journal, Vol. 37, No. 2 55
Figure 12: Relative compression strength of 19 mm cross fluted fill, thermally
bonded in dependence of different parameters (A: Spec. weight, B: Foil
thickness) (test method as described in 5.1)

Figure 10: bending test set-up

Figure 13: Relative deformation for two different fill weights (thermally bonded)

Figure 11: Comparison of PP and PVC for the same fill weight (27 kg/m³,
19 mm cross fluted fill, thermally bonded)

Figure 14: Fill under local buckling (set-up like in Figure 10)

56 CTI Journal, Vol. 37, No. 2


Figure 17: Comparison of deformation behavior of different fill
types in PP (27 kg/m³, thermally bonded)
Figure 15: Compression strength comparison for different
fill types (thermally bonded) Literature
[1] CTI STD 136-2010
[2] Wikimedia Commons
[3] DUBBEL - Handbook of Mechanical Engineering, Wolfgang
Beitz, Karl-Heinz Küttner, Springer, 1994, page B45
[4] DUBBEL - Handbook of Mechanical Engineering, Wolfgang
Beitz, Karl-Heinz Küttner, Springer, 1994, page B19-B22
Figure 16: Flat foil will improve the bending properties of fully vertical flutes [5] DUBBEL - Handbook of Mechanical Engineering, Wolfgang
Beitz, Karl-Heinz Küttner, Springer, 1994, page D57

CTI Journal, Vol. 37, No. 2 57


ASHRAE Legionella Standard 188:
Evidence-Based Interpretation And
Application
Janet E. Stout, Ph.D.
Special Pathogens Laboratory & in the building water systems. The standard tells
University Of Pittsburgh
the team what to do, but provides little specific
information on how to successfully control Le-
Abstract gionella in building water systems. Reliance on
The first U.S. standard for the prevention of evidence-based decisions validated from peer-
Legionnaires' disease was published by the reviewed scientific investigations will strengthen
American Society of Heating, Refrigerating and the effectiveness of the Team and defensibility of
Air-conditioning Engineers (ASHRAE) on June the Program and close the “gaps.”
26, 2015. The normative sections of the standard The “Gaps”
include development of a Water Management Plan
Following are a few instances of potential gaps
for building water systems and devices including Janet E. Stout within the standard:
open and closed circuit cooling towers and evapo-
rative condensers. ASHRAE Standard 188 is a process standard and 1. Critical knowledge for effective disease prevention
critical decisions are left to the discretion of the Water Management • The “Water Management Program Team" must have
Team. Information will be shared to help inform these decisions "knowledge of the building water system design and
so that they are evidence-based and defensible. water management as it relates to Legionellosis."
• This critical knowledge of Legionella will not typi-
Introduction cally be available in-house as most building owners
ANSI/ASHRAE Standard 188-2015 Legionellosis: Risk Manage- don’t possess the necessary specialized knowledge
ment for Building Water Systems is a good step in preventing of Legionella in the built environment.
Legionnaires’ disease in building water systems. However, the
• For example, the presence and growth of Legionella
standard provides only minimum requirements and compliance
can be impacted by how water systems are operated
would not insure Legionella control. Building owners are presented
and maintained. The Water Management Team
with a framework for a risk management approach to the preven-
should understand the differential effectiveness of
tion of Legionellosis, but are left to rely on their own judgment
control strategies, including defining control limits
and knowledge in deciding the most appropriate course of action.
and monitoring of the water treatment/biocide
This gap between the understanding of the how, what and why of
program for cooling towers and secondary water
compliance, and the creation of an effective Water Management
treatment of the building water system if applicable.
Program could leave building owners vulnerable to liability claims
if cases occur. At the end of the day, building owners could develop • For the first time, many building owners may be
a plan without Legionella control. confronted with decisions such as whether or not
to install a disinfection system. Selecting the most
Implementing The Standard appropriate and cost-effective technology requires
Compliance is based on defining water systems in buildings, de- knowledge of Legionella and engineering, as well
termining if they are covered by the standard, and going through as water treatment expertise.
risk management steps in the standard, but not necessarily dem- 2. The microbiology of prevention
onstrating the effectiveness of the plan in controlling Legionella. • In the case of outbreaks involving cooling towers,
This represents a critical "gap" that could be closed with in-depth the causative agent has almost exclusively been
knowledge and decision-making. ASHRAE has placed the standard Legionella pneumophila, serogroup 1. The majority
on "continuous maintenance” providing a mechanism for changes of drinking water-associated outbreaks are caused
throughout this process that could address these “gaps” in future by Legionella pneumophila. Overall, most reported
iterations of the standard. cases of Legionnaires’ disease are caused by L.
The plan must include confirmation that the Program effectively pneumophila, serogroup 1, with other serogroups
controls the hazardous conditions throughout the building water (mostly 3, 4 and 6) and Legionella micdadei, and
systems. Practically speaking, the “hazard” is growth and exposure Legionella longbeachae accounting for a small
to Legionella bacteria in the building water systems and subsequent percentage of cases.
risk of illness. A Water Management Team is formed to develop and • Interpreting the relative risk of these disease-causing
implement the Water Management Program for Legionella control strains is important for determining and deploy-
ing resources for risk management. The Water

58 CTI Journal, Vol. 37, No. 2


Management Team is tasked with decisions for drinking water–associated outbreaks that resulted
testing (if, where, how many locations) and results in 431 cases of illness, 102 hospitalizations, and 14
interpretation. deaths. All 14 outbreak-associated deaths reported
3. Sources of exposure to Legionella were caused by Legionella, including 12 (86%)
• For sporadic community-acquired cases, the public cases associated with health care facilities. Legio-
health threat from improperly managed cooling nella in building plumbing systems was among the
towers is relatively unknown—with the exception two most commonly identified deficiencies [factors]
of well-defined outbreaks like the 2015 outbreak in leading to drinking water–associated outbreaks.
the Bronx in New York City that reportedly caused 4. ASHRAE 188 does not require Legionella testing to validate
138 cases and 16 deaths. In contrast, warm water that your risk management program is working.
distribution systems inside buildings are a well- • Studies show that the only reliable way to validate
defined source of exposure. efficacy of your risk management program and
• The risk of illness from building water systems residual disinfectant (if you are using one) and the
was established with our 1982 publication in The threat from Legionella is to test for Legionella.
New England Journal of Medicine. We showed • Culture of viable Legionella is the standard method
that hospitalized patients acquired Legionnaires’ for detection. The International Organization for
disease from the water in their hospital rooms, not Standardization: ISO 11731 -1998, “Water Qual-
from cooling towers. Since then, the link between ity - Detection and Enumeration of Legionella” is
illness and exposure to Legionella from building a method that is ‎accepted internationally as best
warm water distribution systems in hospitals, hotels, practice for Legionella culture.
senior high rise apartments, prisons, nursing homes, 5. ASHRAE 188 requires hazardous conditions be redressed to
and private homes has been well established. “acceptable” levels but doesn’t define “acceptable.” That’s
• In 2006, a National Research Council report cited left up to the Water Management Team.
Legionella as “the single most common etiologic The approval of ASHRAE standard 188 has the potential to in-
[disease-causing] agent associated with outbreaks troduce a new era Legionnaires’ disease prevention. The success
involving drinking water.” of this effort will depend upon well-informed Water Management
• Recently, the Centers for Disease Control and Pre- Team and evidence-based Water Management Programs. Such
vention (CDC) reported that for the period between efforts could prevent illness and death from a preventable water-
2011 and 2012, Legionella accounted for 66% of the borne disease.

CTI Journal, Vol. 37, No. 2 59


New York Legionella Regulations:
Are They Missing The Boat?
by Sarah Ferrari
Legionella concentrations’ discussed within this
Abstract paper pertain only to the virulent, or infectious,
form of the bacteria.
A large outbreak of Legionnaires’ disease in the
Bronx in 2015 prompted NYC to enact law and The vast majority of LD occurs as apparently
NYS to propose emergency regulations on the isolated cases. Of cases reported to the CDC, 96%
registration and maintenance of cooling towers. are classified as sporadic and are not typically
This paper describes the fundamental character- investigated4. A cluster of cases is classified as an
istics of point sourced vs. potable water sourced outbreak when two or more people are exposed to
outbreaks and discusses the Bronx outbreak Legionella and get sick in the same place at about
from those perspectives. Ultimately a case is the same time. Recognized outbreaks of LD are
made that these new regulations will not have rare; but when they occur, they provide opportuni-
a measurable impact on reducing the incidence ties to understand the epidemiology of the illness
of Legionellosis. Rather, more detailed and and improve prevention strategies. This opportu-
open-minded investigations of future outbreaks, nity is wasted if the extensive data that is generated
including investigation of potential potable wa- during an outbreak is not evaluated impartially.
ter sources, are called for to inform appropriate Sarah Ferrari The recent outbreak in the South Bronx has re-
regulations and disease prevention activities. sulted in New York City enacting a local law5 on the registration
and maintenance of cooling towers in the city. Also, the State of
Introduction New York has proposed emergency regulations for the registration
Legionnaires’ disease (LD) is a severe form of pneumonia which and maintenance of cooling towers state-wide6. As more facts have
is contracted by inhaling or aspirating water droplets containing emerged, it appears that the hastily prepared emergency regulations
Legionella deeply into the lungs. For many years it was believed have fallen victim to the “detection bias” referred to in the first
the disease could be transmitted only by large equipment which paragraph above and that authorities have not only squandered
emits aerosols or by equipment designed to aerosolize. Thus spas, an opportunity to expand our understanding of the disease but
decorative fountains, grocery misters, spray humidifiers, cooling imposed regulations and cost on cooling tower owners that have
towers and other aerosol sources were the only water systems little chance of reducing the incidence of disease.
investigated when an outbreak occurred. In the early 1980s inves-
tigations of potable water systems in hospital outbreaks indicated This paper will describe the outbreak in the Bronx that instigated
that the potable water is also a vector in disease transmission, either these regulations. While a specific cooling tower was identified
via aspiration of Legionella from the mouth into the lungs 1 or via as the source of the original outbreak, two subsequent outbreaks
inhalation of droplets emitted by sinks and showers. It now appears occurred. Potable water in the building where people lived was
that many LD outbreaks were initially blamed on cooling towers positively identified as the source for second outbreak. Two months
due to a “detection bias” that has not been widely recognized, and after the third outbreak ended a cooling tower was identified as the
that these outbreaks were actually caused by potable water issues. source, even though all of the cooling towers in the area had been
In more recent years it has been found that the primary source of recently cleaned in accordance with the newly enacted NYC laws.
hospital-acquired Legionnaires’ disease is potable water 2. The Bronx outbreak and regulatory response has many similarities
In the United States there have been requirements to address Le- to a French outbreak in the winter of 2003-2004. In Pas-de-Calais a
gionella in hospital potable water systems from the Joint Commis- large outbreak was attributed to a cooling tower and resulted in the
sion on the Accreditation of Healthcare Organizations (JCAHO), promulgation of regulations for the registration and maintenance
Allegheny County (Pittsburgh), Maryland, New York, and others; of cooling towers. The inconsequential result of those regulations
however, until recently there have been no similar mandates in the on the reduction of the incidence of disease will also be described
United States for cooling towers. Although not mandated, many in this paper.
industrial groups such as CTI, ASHRAE, and AWT have published Outbreak Characteristics
best practice guides that describe methods for maintaining equip-
ment to minimize the risk of Legionellosis. Aerosol Point-Source
More than fifty Legionella species have been identified, but not all There have been many Legionnaires’ disease outbreaks traced to an
have been linked to disease. Legionella pneumophila serogroup aerosolized point source of the bacteria. Point sources of aerosols
1 is the most virulent strain causing the majority of infections3. include decorative fountains, spas, grocery misters, and cooling
Virulence varies not only between strains and their subtypes but towers. Investigation of these outbreaks typically reveals a close
can also vary within a particular cell. There are two major phases relationship between time spent near the source and incidence
to the life cycle; a non-pathogenic vegetative phase and a virulent of infection. A dose-response relationship between exposure and
transmissive phase. The concepts of ‘infectious dose’ or of ‘relative illness has been demonstrated in numerous carefully investigated

60 CTI Journal, Vol. 37, No. 2


CTI Journal, Vol. 37, No. 2 61
studies. This is particularly true for indoor sources where proxim- the ubiquitous nature of this bacterium in the environment, that
ity to the aerosol source has been identified as an essential factor in order to cause disease, the number of Legionella in the water
for infection. For cooling tower sources the same relationship would need to be much higher than that found in most normal
exists, i.e., a close correlation between proximity to the source aquatic habitats11.”
and incidence of disease. However, for some outbreaks where the The virulence of a particular genetic type of Legionella is not a
purported source is a cooling tower there is little spatial correla- constant and may change during its lifecycle and also may change
tion between the incidence of disease and the alleged source. The by exposure to chemicals, heat, or interaction with amoebae. Nev-
unstated assumption is that there is a hidden variable that causes ertheless, the ubiquity of Legionella in nature implies that everyone
disease seemingly randomly a great distance from the source. This has been exposed to at least a low concentration of the bacteria.
lack of spatial correlation was evident in the South Bronx outbreak
where there was no clustering of cases near the purported tower In addition to dilution of the aerosol over distance, cooling towers
at the Opera House Hotel. What that hidden variable could be is have made dramatic improvements in drift eliminator technology.
not clear. The following paragraphs describe several well studied Drift eliminators are the component of cooling towers that separates
outbreaks where, as one would expect, proximity to the source recirculating water from exiting air. Over a period of about five
strongly affected the incidence of disease. years ending around 2000 all major manufacturers of factory-built
cooling towers developed low-drift eliminators. These modern
In 1999 a spa at a floral trade show in the Netherlands was identi- drift eliminators reduce aerosol emissions by an order of magnitude
fied as the source of a large outbreak of Legionnaires’ disease. from the previous generation of towers. If the water in a cool-
Among exhibitors a close correlation was found between elevated ing tower was contaminated with 1,000 CFU/ml of Legionella, a
antibodies to Legionella and proximity of their booth to the source. person standing on top of the tower and breathing in only tower
The most important visitor-related risk factor was pausing at the exhaust for 1 hour would breathe in droplets containing a total of
whirlpool spa display7. 10 bacteria12. As a comparison of transmission pathways between
Clive Brown et al. investigated a 29-case community outbreak of inhalation and aspiration, a person drinking 4 ounces of water
Legionnaires’ disease in 1994 linked to a hospital cooling tower8. from a potable water source contaminated with only 10 CFU/ml
This investigation showed that the risk of infection decreased by would consume over 1,000 bacteria in the few seconds it took to
20% for each 0.1 miles from the hospital and increased by 80% drink the water. These bacteria would be in the mouth and not in
for each visit to the hospital. The paper described an Aerosol the lungs, but the numbers of bacteria that can enter the body are
Exposure Unit defined as the ratio of time spent near the source significantly higher with an aspiration route from potable water
and the distance from the source and found a strong correlation than inhalation from a modern cooling tower.
between exposure and incidence of disease. The image in Figure 1 As the C. Brown hospital study, the Netherlands flower show
taken from this study illustrates the expected distribution of cases study, and the Michigan prison study imply, there should be a dose-
of disease from a single aerosol source. The cases are clustered response with exposure and incidence of disease. The higher the
near the source and taper off rapidly at increasing distance. dose that a particularly susceptible individual receives, the higher
Proximity was well demonstrated in a 1993 outbreak at a Michi- will be the likelihood of infection. Also, because of Legionella’s
gan prison which was traced to a hospital cooling tower. Fourteen ubiquitous presence in nature, there should be a threshold dosage
(0.6%) of 2253 prisoners who used exercise yards each day within below which there is no disease.
100 yards of the prison hospital were infected, compared with only The key characteristics of an aerosolized point source exposure
two (0.1%) of the 2270 inmates who used yards at least 400 yards to Legionella are:
from the prison hospital.9 This equates to roughly a 50% reduction
in infection risk in 0.1 miles. 1. The original source of the bacteria is colonization of the
potable water supply.
The influence of prevailing wind on aerosol dispersion was ad-
2. While many water-based devices could be contaminated due
dressed by P. Wilmot et al. using geographic information system
to an upset in the potable water system, only in one will the
(GIS) data10. They developed plume dispersion models to help
bacteria find hospitable growing conditions for amplification
locate a potential cooling tower source during outbreak investiga-
and susceptible individuals to infect.
tions. A plume refers to the moist exhaust air from a cooling tower.
The plume will mix with ambient air as it travels away from the 3. The bacteria reproduce in an individual aerosol producing
tower becoming more and more dilute. If a case occurs outside of device. This usually requires the water in the equipment to
the plume dispersion model “the chance that it originated from that reach temperatures that permit amplification.
cooling tower remains highly unlikely”. Figure 2 displays a case 4. The bacteria are emitted from a single point.
where the cooling tower would not be considered a likely source. There is a strong correlation between incidence of infection and
The plume is diluted with increasing distance from the tower re- proximity of the individual to the aerosol source in both time and
ducing the concentration of Legionella in the air. Legionella can space.
live only a relatively short time in the air and contamination at a While a correlation between proximity to the aerosol source and
large distance from the tower is unlikely. A person would need to disease incidence seems fundamental, there are many outbreaks
spend sufficient time near to the tower in order to inhale an infec- where investigations concluded that this was not the case. This is
tious dose of bacteria. particularly true of outbreaks attributed to cooling towers. There
Dr. Richard Miller at the University of Louisville has stated: have been numerous cases, including the July 2015 outbreak in
the Bronx, where a cooling tower in the general area was blamed
“Legionnaires' Disease, like all infectious diseases, requires a
for causing disease without any apparent association of patients
minimum infectious dose in order to cause disease. While the exact
with the aerosol source. This lack of spatial correlation between
number required for humans varies depending on the susceptibility
disease incidence and the purported source is an extremely strong
(i.e. immune status) of the individual, it is likely that the number
indication that the specific cooling tower is not the source of the
for most individuals is relatively large. It should be apparent from
outbreak.

62 CTI Journal, Vol. 37, No. 2


CTI Journal, Vol. 37, No. 2 63
Potable Water Supply Source may infect only a few patients per year and appear somewhat
Municipal potable water systems in the United States and Eu- sporadically.
ropean countries have been very effective at reducing, but not The key factors of a potable water outbreak of Legionnaires’
eliminating, waterborne diseases. Potable water is sanitary but disease are:
not sterile. Disinfection is designed to eliminate many pathogens 1. The original source of the bacteria is colonization of the
transmitted by the oral-fecal route, such as cholera and typhoid, potable water supply.
however many other bacteria that are natural inhabitants of aquatic
environments may survive. Most water pipes contain a layer of 2. While many buildings could be contaminated due to an upset
biofilm. This biofilm may harbor many non-pathogenic bacteria in the potable water system, only in some buildings will the
but can also harbor bacteria such as Legionella. When a known bacteria find hospitable growing conditions for amplification
upset occurs, such as a power outage or water main break which and susceptible individuals to infect.
causes loss of system pressure, warnings to boil water before using 3. The bacteria reproduce within several building potable wa-
are sent to the system users. There are an estimated 240,000 water ter piping systems to reach infectious levels. This usually
main breaks per year in the United States13. Minor upsets such as requires the water in the system to reach temperatures that
pressure surges that could disturb biofilms in the pipes may seem permit amplification.
unremarkable or go unnoticed yet could release bacteria into the 4. The bacteria are emitted at multiple points in multiple build-
water stream. A study conducted in Wales and northwest England ings throughout the affected portion of the municipal water
from 2001 to 2002 found a very strong association between self- distribution system.
reported diarrhea and reported low water pressure at the home tap14. 5. Since the bacteria are emitted from widespread sources there
The investigators hypothesized that most of the reported episodes is a seemingly random distribution of cases of disease all
of pressure loss were due to main breaks in which contamination within the same municipal water distribution system.
entered the distribution system. As the infrastructure ages, the Potable water has also been identified as the source for Legion-
frequency of upsets that can potentially cause contamination in naires’ disease in non-hospital settings, but again usually with very
the system has increased. low infection rates. High risk buildings tend to be tall with complex
The first incidence in which a municipal water system was found to piping and many residents. If Legionella from the municipal water
be the vector for disease transmission occurred during the cholera supply colonize sediment and biofilm in a building water system
epidemics in mid-19th century London. Sir John Snow, an English there may be occasional incidents of disease. When multiple cases
physician, prepared a map of where cholera deaths had occurred. occur in a single building, the health department will evaluate that
This map15 clearly showed that most of the deaths were in build- building for contamination but rarely investigate other buildings
ings that received their water from a particular contaminated well receiving water from the same municipal system. An incident in
that was the source of the disease. However, it took 20 years after the municipal potable water supply system that contaminated many
Snow generated his map and 8 years after Snow passed away buildings over a short period of time could result in an outbreak of
before the correlation between contaminated water and cholera disease occurring in the area downstream from the incident over a
was accepted16. The “detection bias” that cholera was caused by relatively short period of time.
airborne “miasma” was too firmly entrenched in the nineteenth
century zeitgeist to be easily dislodged. Figure 3 reproduces Sir Epidemiology – Not an Exact Science
John Snow’s map. Cholera fatalities are indicated by small red Determining the source of a particular outbreak requires gather-
circles; public drinking water wells are indicated by larger blue ing and sifting through large amounts of information. Patients
circles. The cases are spread uniformly over the area where the are interviewed, commonality between sites visited by patients is
contaminated water was used. evaluated, possible sources are examined, and an attempt is made
A more recent potable waterborne infection occurred in Denmark to match the DNA of infectious bacteria grown from patient isolates
in 200717. There an outbreak of gastroenteritis affected a high to that found in the environment. It is only when all of these fall
percentage of residents in one section of the city. Investigation in line that a source of the disease can be definitively imputed.
showed massive contamination of a part of the water distribution As of this writing, there are approximately 2,000 different Legio-
system, while other parts of the distribution system appeared to nella pneumophila genotypes known worldwide, but only 10%
be unaffected. The source was eventually identified as backflow of those are known to be associated with disease in the US19 20. A
of sewage into that portion of the drinking water system. Figure DNA match between environmental and patient isolates is not as
4 shows the case map for this outbreak with contamination in one determinative of the infection source as one might expect. In a
subsection of the water supply. specific geographical region there are usually fewer than several
Legionellosis is a waterborne disease. Since the early 1980’s, dozen genotypes endemic to the water systems21. Since potable
potable water has been known to be a vector for Legionnaires’ water is the ultimate source of the bacteria, many water features in
disease. The best studied cases with a potable water source are an area can be contaminated with the same genotype of bacteria.
hospital acquired infections. A hospital’s internal piping system In fact particular genetic types of Legionella can become endemic
can become contaminated with Legionella from the municipal to a given water system with the same genetic match appearing
potable water supply. In warm areas of piping, particularly if there in seemingly unrelated environmental sources. Thus the lack of
is a biofilm on surfaces or sediment in the system, the Legionella a DNA match can disprove that a particular feature is the source,
can multiply and occasionally release large numbers of bacteria but a DNA match on its own cannot prove that it is the source.
into the water18. The bacteria may be transmitted to many hospital The difficulty of determining a source was clearly seen in an out-
patients via aspiration of contaminated drinking water or ice chips break in South Dakota in 200722. There had been a dramatic in-
or, more commonly, via aerosols generated by sinks and showers. crease in the incidence of Legionnaires’ disease over a short period
Since Legionella are not as virulent as cholera this contamination of time. There was little in common with the patients except that
they spent time in Rapid City. To the investigators this at first ap-

64 CTI Journal, Vol. 37, No. 2


CTI Journal, Vol. 37, No. 2 65
peared to be a likely cooling tower issue, though there was no clear A particular strain is likely to be present throughout the potable
relationship with time spent near a tower and incidence of disease. water system, so commonality in activity among patients relative
Cooling towers throughout the city were located and sampled. to a particular potential source must be established in order to
Many had detectable levels of Legionella and were required to be deduce the specific source(s) of infections.
disinfected. However, none of the towers Legionella were a genetic The particular cooling tower located at the Opera House Hotel
match to the patients’ isolates. The investigators continued to look circulated water at 800 gpm; at full fan power it moves 49,700
for other aerosol sources without success as infections were still CFM; and it is equipped with drift eliminators that reduce the drift
occurring. Then the investigators noticed that many of the patients to less than 0.001%. The drift rate then is: 0.00001 x 800 gpm x
had eaten at the same restaurant during their likely infection period. 3785 ml/gal = 30 ml/min. At full fan power this drift results in a
When investigators revisited the restaurant they sampled a small concentration of: 30/49700CFM x 1CF/7.48 gal x 1 gal/3.785 liters
fountain and found both a high level of Legionella in the fountain = 0.00002 ml of drift /liter of exhaust air. If each ml contained
and an exact genetic match to the Legionella found in the patients. 1000 CFU of Legionella, then on average there would be 0.02
The fountain was removed and no additional infections occurred. Legionella per liter of undiluted tower exhaust air or in every 50
It was fortunate that none of the cooling towers testing positive liters of air there would be a single bacteria. If a typical person
for Legionella were a genetic match to the particular bacteria that has a 500 ml tidal volume and takes 15 breaths per minute, and
infected patients. Had there been a match, even with little data if they were breathing only undiluted tower exhaust for an hour,
showing that infected persons spent more time near the towers they would inhale 0.5 x 15 x 60 = 450 liters of air or only about
than the general population, that tower may likely have been de- 10 bacteria. As one moves away from the tower the exhaust air
clared the source and the fountain would have continued to infect containing the drift becomes more dilute and a person would re-
restaurant patrons. quire much longer time in the diluted exhaust air to inhale a similar
number of bacteria.12
2015 Bronx Outbreaks Recall that proximity has been well established in aerosol point-
sourced outbreaks, with reduced infection rates at increasing
Initial Outbreak in South Bronx distance from a source. The circle around the Opera House Hotel
During the summer of 2015 a large outbreak of Legionnaires’ in Figure 6 indicates a 0.1 mile radius. This is the area where one
disease occurred in the South Bronx. The onset dates, 2 to 10 would expect the highest cluster of cases, but there is no cluster of
days after the infection dates, were between July 8th and August cases at or immediately adjacent to the hotel. There is no gradient
3rd. The outbreak was officially declared over on August 20th. with radial or, presuming wind, with directional distance. Rather,
During this outbreak 133 individuals were diagnosed with the the cases appear randomly across a broad area, producing a case
disease with 16 deaths. The graph in Figure 5 is from the NYC map with more similarity to the waterborne cholera outbreak. Cool-
Department of Health23. This epidemic curve is characteristic of ing towers and fountains were sampled as part of the investigation,
a common source outbreak where the individuals were exposed to but potable water was not. City officials repeatedly claimed that
a source of the bacteria over a short period of time. Based on the “the drinking water is unaffected”. This claim was not substanti-
shape of the curve, this common origin could be an aerosolized ated with data by NYC because no testing of potable water was
source with an upset condition or a municipal water system with performed. In fact it was dramatically disproven in only a few
an upset condition. weeks when the cases associated with the Melrose Houses were
The source of this outbreak was identified by the NYC DOH as a more fully evaluated.
cooling tower on the Opera House Hotel. Note that the particular
cooling tower at the Opera House Hotel that was identified as the Outbreak at Melrose Houses
source of the outbreak had a small 7.5 HP fan and was equipped There was a second outbreak of four cases of Legionnaires disease
with modern low-drift eliminators that reduce by an order of at the Melrose Houses in the South Bronx identified after the initial
magnitude the quantity of drift that the tower emits. The cooling outbreak. An arrow in Figure 6 indicates the location of the Melrose
tower was disinfected on August 1st with the last reported case Houses, less than 0.4 miles from the Opera House Hotel. The first
onset on August 3rd. The infection date for all of these cases could case in the Melrose Houses outbreak occurred in March 2015 and
have been prior to the cleaning of the cooling tower, though the was not investigated at that time. Two more cases occurred during
declining case rate in late July indicates that the outbreak had likely the July outbreak and were originally included in that outbreak,
stopped before the tower was disinfected. It is likely that none of and the latest occurred in late August after the July outbreak was
the testing and cleaning of cooling towers demanded by the NYC declared over. No potable water sources were investigated by the
DOH had any effect on ending the initial outbreak. NYC DOH until four cases were identified at a single location at
The cases were distributed fairly uniformly over a large area of the Melrose Houses. When the potable water at Melrose Houses with
South Bronx. The NYC DOH has provided the map in Figure 6 over two thousand residents was investigated, Legionella were
showing approximately 85 of 133 case locations and the locations found in the potable water and the potable water was positively
of cooling towers.23 Circles represent cases and triangles cooling identified as the source of infection. Point-of-use water filters were
towers with the large red triangle at the location of the cooling installed on all faucets and showerheads and a copper-silver ioniza-
tower at the Opera House Hotel. Red circles are cases where ge- tion system was installed in the buildings’ potable water piping.
netic testing of the Legionella was performed. The patient samples In spite of 3 cases occurring in the same facility over a period of
and the bacteria found at the Opera House Hotel were of the same only a few months, potable water was not sampled. It was only
genetic type, but as discussed above, genetic typing alone can rule after the original outbreak was declared over yet an additional case
out a potential source but cannot, on its own, definitively establish occurred that the potable water was sampled and identified as the
a source. In fact the particular strain found in the patients and source of the infections.
in the Opera House Hotel cooling tower was also found in other
features in the area and had caused previous outbreaks in NYC24.

66 CTI Journal, Vol. 37, No. 2


CTI Journal, Vol. 37, No. 2 67
Outbreak at Morris Park contamination. As a result of this outbreak regulations were promulgated
A third outbreak of Legionnaires’ disease occurred in the Morris Park for the control of cooling towers. Cooling tower regulations for hospitals
neighborhood of the Bronx. Morris Park is about 4 miles from the Opera had been enacted in 2003 but with the Pas-de-Calais outbreak they were
House Hotel in the East Bronx. There were 13 cases in this outbreak and extended to all cooling towers32.
1 death. Figure 7 shows the epidemic curve for this outbreak. The onset The 2004 regulations require frequent testing of cooling towers for Legio-
dates for all but one of the cases included in the outbreak were between nella. The regulations mandate monthly testing for Legionella unless 12
September 14th and September 21st with infection occurring 2 to 10 days consecutive monthly tests are less than 1 CFU/ml then the testing can be
prior. On August 6 a city wide order had been issued which mandated that: reduced to quarterly. If the reading exceeds 100 CFU/ml the tower must
“Regardless of the outcome of the evaluation required by item (2) above, be immediately shut down and cleaned. More frequent testing is then
direct the environmental consultant to carry out a disinfection/treatment required until the system again meets the strict low limits.
sufficient to remove organic material, biofilm, algae and other contami- The graph in Figure 9 indicates that the incidence rate for LD infection
nants and disinfect in a manner sufficient to control for the presence of has been hovering around 2 per 100,000 of population for over a decade.
Legionella organisms within 14 days of receipt of this letter”25 After the regulations were issued at the end of 2004, there was an increase
Due to this city-wide order, all cooling towers in the area had been dis- in the reported incidence of LD. This could well be due to a heightened
infected by August 20th a few weeks prior to the earliest possible date awareness of the disease from both the issuing of the regulations and the
of infection. If indeed a cooling tower was responsible for the outbreak, widely publicized outbreak at Pas-de-Calais. There has been a gradual
then the remediation requirements in the NYC newly enacted laws were decline in the reported incidence since 2005, but a dramatic reduction has
ineffective at preventing an outbreak. If potable water or another feature not been observed since the regulations were implemented. The author be-
was the source or sources then the newly enacted laws which focus only lieves that these results indicate that the cooling tower regulations had little
on cooling towers actually hindered a proper investigation of the outbreak. to no effect on the incidence of sporadic cases of Legionnaires’ disease.
In late September, thirty-five cooling towers in the area were tested for Conclusion
Legionella and 15 had detectable levels of the bacteria. These 15 towers There are outbreaks of Legionnaires’ disease which have been clearly
were disinfected for a second time beginning September 29th, significantly linked to a sole aerosolized source of bacteria. In these outbreaks, acute
after the outbreak had ended. A press release issued by the NYC DOH on proximity and duration were shown to govern exposure and incidence of
November 20, fully two months after the outbreak had ended, identified Legionnaires’ disease. However, there are many outbreaks attributed to a
a cooling tower as the source26. With this outbreak it is unequivocally cooling tower source which do not adhere to these rules. These outbreaks
clear that none of the testing and cleaning of cooling towers demanded have case profiles which are more random and cover larger areas. These
by the NYC DOH had any effect on ending the outbreak. outbreaks mimic the documented profile of certain potable water outbreaks
In the large South Bronx outbreak and also in the Morris Park outbreak, where a single ‘upstream’ source contaminates multiple exposure sites.
infection occurred over a relatively large area in a short time-span. Al- The source(s) of future Legionnaires’ disease outbreaks must be inves-
though the pattern could be due to an upset in the municipal water system tigated more thoroughly, exploring the possibility of multiple exposure
simultaneously infecting a large number of buildings, the only sources sites including potable water, in order to advance our understanding of
that were investigated were aerosol point sources such as cooling towers, LD transmission. The traditional epidemiological models that assume cau-
spas and fountains. sality with a genetic match but with only a tenuous exposure mechanism
have resulted in the promulgation of regulations that have not resulted in
Impact of regulations on public health a significant reduction in the incidence of disease. Increased awareness
and the issuing of explicit orders for the care of cooling towers did not
Incidence of Disease in the United States and Europe prevent the outbreaks from occurring at Melrose Houses or in Morris
Legionnaires’ disease is a reportable illness in the United States and many Park. Strict cooling tower regulations imposed in France after the Pas de
European Countries. Records are made available by the CDC27 in the US Calais outbreak have not resulted in any significant reduction of disease.
and the ECDC28 in Europe. For 5 European countries these records go Is it time we say, perhaps, that it is not always the cooling tower? And
back at least to 2003. Figure 8 shows the reported incidence of LD for what of the other 96% of reported LD cases, those considered sporadic?
these countries and the US in illness per 100,000 of population.
A 2006 report issued by the National Academy of Sciences states “Communi-
There are many factors which can affect the shape of the curves besides ties should squarely address the problem of Legionella, both via changes to
actual incidence of disease. Legionnaires’ disease is believed to be un- the plumbing code and new technologies.”14 The report goes on to call for
derreported. The US reported incidence rate has steadily increased over research projects which specifically address potential problems arising from
the period shown. The CDC believes that this may be partially due to: premise plumbing. Regarding outbreaks the report states, “Environmental
“An increasing population of older persons contributed to the increase assessments of outbreaks should begin to incorporate new insights and al-
in reported legionellosis cases. Other factors that might have contributed low possible cause-and-effect relationships to be established.” This would
include an increasing population of persons at high risk for infection; include a much greater emphasis on dose reconciliation in outbreaks in order
improved diagnosis and reporting, possibly stimulated by the 2005 CSTE to develop basic practical data on dose-response relationships.
endorsement of more timely and sensitive legionellosis surveillance; and Clearly the scientific community recognizes the health risks associated
increased use of urine Legionella antigen testing29”. with premise plumbing, or building potable water systems. These potential
The graph in Figure 8 shows that there is little difference in the incidence risks were also well-recognized by the team of experts who contributed
rate of Legionnaires’ disease between Western Europe and the United to ASHRAE Standard 188 Legionellosis: Risk Management for Build-
States. This lack of difference exists in spite of burdensome regulation ing Water Systems. The Standard provides minimum Legionellosis risk
of cooling towers which has been in place in Europe for many years. management requirements for buildings and their associated potable and
non-potable water systems. Subsequent to the recent outbreaks, NYC
Legionella Regulations in France adopted a small subset of the Standard, the portion addressing cooling
France provides an interesting study of the effects of cooling tower regula- towers. Adoption of the full Standard, as intended by the expert authors,
tions on the incidence of disease. Guidelines to improve underreporting is necessary in order to have a significant positive effect on public health.
of the disease were written in 1997 along with introduction of the urinary
antigen detection test30. In the winter of 2003 to 2004 a large outbreak
with 86 confirmed cases and 18 fatalities occurred in Pas-de-Calais,
France31. An industrial cooling tower was implicated as the source for the

68 CTI Journal, Vol. 37, No. 2


CTI Journal, Vol. 37, No. 2 69
Figure 1 – Case Map from Aerosolized Point-Source

Figure 4 – Gastroenteritis Epidemiology Denmark 2007

Figure 2 – Case Outside Dispersion Model

Figure 5 – Epidemic Curve of Initial Outbreak in South Bronx 2015

Figure 3 – Cholera Epidemiology London 1854

Figure 6 – Map of Cases and Cooling Towers in


South Bronx July 2015 Outbreak

70 CTI Journal, Vol. 37, No. 2


CTI Journal, Vol. 37, No. 2 71
7. Den Boer, J. (2002) A Large Outbreak of Legionnaires’ disease at a
Flower Show, the Netherlands, 1999, Emerging Infectious Diseases
(Jan).
8. Brown, C. et al. (1999) A Community outbreak of Legionnaires’
disease linked to hospital cooling towers: an epidemiological method
to calculate dose of exposure, International Epidemiological Associa-
tion Vol 26:353-359.
9. CDC Morbidity and Mortality Weekly Report (1994) July 15.
10. Wilmot, P. et al., (2004) Modelling Cooling Towers Risk for Legion-
naires’ Disease using Bayesian Networks and Geographical Informa-
tion System, University of Adelaide.
11. Miller, R., Reducing the risk of Legionnaires’ Disease, Environmental
Safety Technologies
12. Bugler, T. et al. (2008) Cooling Towers, Drift, and Legionellosis,
IWC-08-21, International Water Conference.
13. 2013 Report Card for America’s Infrastructure. American Society of
Civil Engineers. Web. 2 Nov. 2015.
14. Committee on Public Water Supply Distribution Systems: Assessing
and Reducing Risks, Water Science and Technology Board, Division
on Earth and Life Studies, National Research Council, National
Figure 7 – Epidemic Curve for the Third Outbreak in East Bronx 2015 Academies Press, Dec 22, 2006.
15. http://www.udel.edu/johnmack/frec480/cholera/cholera2.html
16. Competing Theories of Cholera. UCLA Department of Epidemiology,
School of Public Health. Web. 2 Nov. 2015.
17. Vestergaard, LS. et al. (2007) Outbreak of Severe Gastroenteritis
With Multiple Aetiologies Caused by Contaminated Drinking Water
in Denmark, January 2007, Eurosurveillance, Vol 12 Issue 13.
18. Stout, J. et al. (1984) Ecology of Legionella pneumophila within
Water Distribution Systems, Applied and Environmental Microbiol-
ogy (Oct).
19. Kozak-Muiznieks N. et al. (2013) Prevalence of Sequence Types
among Clinical and Environmental Isolates of Legionella pneu-
mophila Serogroup 1 in the United States from 1982 to 2012, Journal
of Clinical Microbiology.
20. EWGLI Sequence-Based Typing (SBT) Database for Legionella
Figure 8 – Incidence of Legionnaires’ Disease in the US and Western European Countries pneumophila. The European Working Group for Legionella Infec-
tions. Web. 21 Dec 2015.
21. Drenning S. et al. (2001)Unexpected Similarity of Pulsed-Field Gel
Electrophoresis Patterns of Unrelated Clinical Isolates of Legionella
pneumophila, Serogroup 1, Journal of Infectious Disease.
22. O’Loughlin, R. E. (2007) Restaurant outbreak of Legionnaires’
disease associated with a decorative fountain: an environmental and
case-control study, BMC Infectious Diseases (Aug).
23. http://www.nyc.gov/html/doh/html/diseases/cdlegi.shtml
24. Private Communications with CDC.
25. NYC DOH Order of the Commisioner, August 6, 2015.
26. “Bronx Psychiatric Center Cooling Tower Blamed For Morris Park
Legionnaires’ Outbreak”, CBS New York, 20 November 2015, Web,
12 Dec. 2015.
27. Morbidity and Mortality Weekly Report (2015) October 23.
28. ECDC Annual Surveillance Report Legionnaires’ disease in Europe.
29. CDC Morbidity and Mortality Weekly Report (2011) August 19.
Figure 9 – Incidence of Legionnaires’ Disease in France 30. Campese, C. et al. (2011) Progress in the surveillance and control of
Legionella infection in France 1998-2008, International Journal of
References: Infectious Diseases.
1. Stout, J. et al. (1982) Ubiquitousness of Legionella Pneumophila in
the Water Supply of a Hospital with Endemic Legionnaires’ Disease, 31. Eurosurveillance (2004) Vol.9 Issue 1-3.
New England Journal of Medicine (Feb. 25). 32. Arrêté du 13 décembre 2004 relatif aux installations de refroidisse-
2. Sabria, M; Yu, V. (2002) Hospital-acquired Legionellosis: solutions ment par dispersion d’eaudans un flux d’air soumises à autorisation
for preventable infection, THE LANCET (Jun). au titre de la rubrique no 2921, Journal Officiel de la République
Français.
3. Yu et al. (2002) Distribution of Legionella species and serogroups
isolated by culture in patients with sporadic community-acquired Acknowledgements
legionellosis: an international collaborative survey. J Infect Dis, The author gratefully acknowledges Dr. Claressa Lucas of the Centers for
186:127–128
Disease Control and Prevention National Center for Infectious Respira-
4. CDC Morbidity and Mortality Weekly Report (2015) August 14. tory Diseases, Division of Bacterial Diseases for her careful review of the
5. Local Laws of the City of New York No. 77, Article 317 Cooling technical portions of this paper.
Towers
6. NYS Department of Health Emergency Rule Making, Protection
Against Legionella, HLT-35-15-00005-E, NYS Register/December
2, 2015.

72 CTI Journal, Vol. 37, No. 2


CTI Journal, Vol. 37, No. 2 73
74 CTI Journal, Vol. 37, No. 2
CTI Journal, Vol. 37, No. 2 75
76 CTI Journal, Vol. 37, No. 2
CTI Journal, Vol. 37, No. 2 77
78 CTI Journal, Vol. 37, No. 2
CTI Journal, Vol. 37, No. 2 79
Cooling Technology Institute
Licensed Testing Agencies
For nearly thirty years, the Cooling Technology Institute has provided To be licensed as a CTI Cooling Tower Performance Test Agency,
a truly independent, third party, thermal performance testing service the agency must pass a rigorous screening process and demonstrate
to the cooling tower industry. In 1995, the CTI also began providing a high level of technical expertise. Additionally, it must have a suf-
an independent, third party, drift performance testing service as well. ficient number of test instruments, all meeting rigid requirements for
Both these services are administered through the CTI Multi-Agency accuracy and calibration.
Tower Performance Test Program and provide comparisons of the Once licensed, the Test Agencies for both thermal and drift testing
actual operating performance of a specific tower installation to the must operate in full compliance with the provisions of the CTI License
design performance. By providing such information on a specific Agreements and Testing Manuals which were developed by a panel
tower installation, the CTI Multi-Agency Testing Program stands in of testing experts specifically for this program. Included in these
contrast to the CTI Cooling Tower Certification Program which certifies requirements are strict guidelines regarding conflict of interest to
all models of a specific manufacturer's line of cooling towers perform insure CTI Tests are conducted in a fair, unbiased manner.
in accordance with their published thermal ratings.
Cooling tower owners and manufacturers are strongly encouraged to
utilize the services of the licensed CTI Cooling Tower Performance
Test Agencies. The currently licensed agencies are listed below.

Licensed CTI Thermal


Testing Agencies
License Type A, B* Licensed CTI Drift
Clean Air Engineering Testing Agencies
7936 Conner Rd, Powell, TN 37849
800.208.6162 or 865.938.7555 Clean Air Engineering
Fax 865.938.7569 7936 Conner Rd, Powell, TN 37849
www.cleanair.com / khennon@cleanair.com 800.208.6162 or 865.938.7555
Contact: Kenneth (Ken) Hennon Fax 865.938.7569
www.cleanair.com / khennon@cleanair.com
Cooling Tower Technologies Pty Ltd Contact: Kenneth (Ken) Hennon
PO Box N157, Bexley North, NSW 2207
AUSTRALIA McHale & Associates, Inc.
+61.2.9789.5900 / (F) +61.2.9789.5922 4700 Coster Rd, Knoxville, TN 37912
coolingtwrtech@bigpond.com 856.588.2654 / (F) 865.934.4779
Contact: Ronald Rayner www.mchale.org / ctitesting@mchale.org
Contact: Jared Medlen
Cooling Tower Test Associates, Inc.
15325 Melrose Dr., Stanley, KS 66221
913.681.0027 / (F) 913.681.0039
www.cttai.com / cttakc@aol.com
Contact: Thomas E. (Tom) Weast
McHale & Associates, Inc.
4700 Coster Rd, Knoxville, TN 37912
856.588.2654 / (F) 865.934.4779
www.mchale.org / ctitesting@mchale.org
Contact: Jared Medlen

* Type A license is for the use of mercury in glass thermometers
typically used for smaller towers.
Type B license is for the use of remote data acquisition devices
which can accommodate multiple measurement locations
required by larger towers.

80 CTI Journal, Vol. 37, No. 2


Cooling Technology Institute
Sound Testing
Cooling towers are used extensively wherever water is used as a cooling medium or pro-
cess fluid, ranging from HVAC to a natural draft cooling tower on a power plant. Sound
emanating from a cooling tower is a factor in the surrounding environment and limits on
those sound levels, and quality, are frequently specified and dictated in project specifica-
tions. The project specifications are expected to conform to local building codes or safety
standards. Consequently, it may be in the interest of the cooling tower purchaser to contract
for field sound testing per CTI ATC-128 in order to insure compliance with specification
requirements associated with cooling tower sound.

Licensed CTI SoundTesting Agencies


Clean Air Engineering McHale Performance
7936 Conner Rd 4700 Coster Rd
Powell, TN 37849 Knoxville, TN 37912
800.208.6162 or 865.938.7555 865.588.2654
Fax 865.938.7569 Fax 865.934.4779
www.cleanair.com www.mchale.org
khennon@cleanair.com ctitesting@mchale.org
Contact: Kenneth (Ken) Hennon Contact: Jared Medlen

Cooling Technology Institute Certification Program


STD-201 for Thermal Performance
As stated in its opening paragraph, CTI Standard 201... "sets forth a program whereby the Cool-
ing Technology Institute will certify that all models of a line of water cooling towers offered for
sale by a specific Manufacturer will perform thermally in accordance with the Manufacturer's
published ratings..." By the purchase of a "certified" model, the User has assurance that the
tower will perform as specified, provided that its circulating water is no more than acceptably
contaminated-and that its air supply is ample and unobstructed. Either that model, or one of its
close design family members, will have been thoroughly tested by the single CTI-licensed testing
agency for Certification and found to perform as clained by the Manufacturer.
CTI Certification under STD-201 is limited to thermal operating conditions with entering wet
bulb temperatures between 12.8oC and 32.2oC (55oF to 90oF), a maximum process fluid temperature of 51.7oC (125oF), a cooling range
of 2.2oC (4oF) or greater, and a cooling approach of 2.8oC (5oF) or greater. The manufacturer may set more restrictive limits if desired
or publish less restrictive limits if the CTI limits are clearly defined and noted int he publication.
Those Manufacturers who have not yet chosen to certify their product lines are invited to do so at the earliest opportunity. You can
contact Virginia A. Manser, Cooling Technology Institute at 281.583.4087, or vmanser.cti.org or PO Box 681807, Houston, TX 77268
for further information
Licensed CTI Thermal Certification Agencies
Agency Name Contact Person Telephone/
Address Website / Email Fax
Clean Air Engineering Kenneth (Ken) Hennon 800.208.6162 or
7936 Conner Rd www.cleanair.com 865.938.7555
Powell, TN 37849 khennon@cleanair.com (F) 865.938.7569
Cooling Tower Test Associates, Inc. Thomas E. (Tom) Weast 913.681.0027
15325 Melrose Dr. www.cttai.com (F) 913.681.0039
Stanley, KS 66221 cttakc@aol.com
McHale Performance Jared Medlen 865.588.2654
4700 Coster Rd www.mchale.org (F) 865.934.4779
Knoxville, TN 37912 ctitesting@mchale.org

CTI Journal, Vol. 37, No. 2 81


82 CTI Journal, Vol. 37, No. 2
CTI Journal, Vol. 37, No. 2 83
Index of Advertisers

Aggreko Cooling Tower Services........ 42-43


AHR Expo.................................................... 69
Amarillo Gear Company......................... IBC
AMSA, Inc.................................................... 15
Bailsco Blades & Castings, Inc..................... 4
Bedford Reinforced Plastics....................... 25
Brentwood Industries................................... 3
ChemTreat, Inc............................................ 39
CTI Certified Towers............................ 74-79
CTI License Testing Agencies.................... 80
CTI Sound Testing/Thermal Performance... 81
CTI ToolKit............................................ 82-83
Cooling Tower Resources........................... 55
Denso............................................................ 21
Dynamic Fabricators................................... 45
EvapTech, Inc............................................... 17
Gaiennie Lumber Company......................... 6
Hewitech....................................................... 47
Hudson Products Corporation.................. 13
IMI Sensors.................................................. 41
Industrial Cooling Towers....................IFC, 2
Kipcon........................................................... 53
KIMCO Cooling Towers............................ 73
Midwest Cooling Towers............................ 29
North Street Cooling Towers..................... 23
Paharpur....................................................... 51
Paltech Cooling Towers & Equipments.... 61
PowerGen..................................................... 63
Research Cottrell Cooling.......................... 59
Rexnord Industries...................................... 27
C.E. Shepherd Company, LP........................ 7
Simpson Strong-Tie.................................... 67
Solid State Technologies LLC..................... 57
Spraying Services, Inc................................. 33
SPX Cooling Technologies.....................OBC
Strongwell..................................................... 11
Tower Performance, Inc....................... 35, 84
Turbo Machinery......................................... 71
Walchem (Iwaki America)......................... 65

84 CTI Journal, Vol. 37, No. 2