Preventing Legionnaires’ Disease in Cooling Towers

It has been more than 30 years since Legionnaires' disease (also known as Legionellosis) reared its head. Since then, our knowledge of the disease and its causes has grown dramatically, and state-of-the-art prevention not only is possible, but practical and affordable.

This article will address prevention of Legionnaires' disease from a specific point source—cooling towers—but does not imply that cooling towers are to be singled out as the only point source. Any water source can serve as a source of Legionella bacteria. Also, keep in mind that although this article presents a "state-of-the-art" prevention regimen, if contamination and illness lead to a court case, one could pose the following question to a jury: "What more could we have done?"

Let us look at what is known about this illness. The causative agent of the most serious form of Legionnaires' disease in the United States is Legionella pneumophila SG1 (L. pneumophila). Some would point to Legionella anisa, the agent for Pontiac fever, a mild form of the illness with no reported deaths, as another serious source. Be that as it may, if we address L. pneumophila in an efficient manner, we would, in all probability, also be addressing all aquatic forms of Legionella.1

There are many different configurations of open HVAC systems. Figure 1 illustrates a system with a cooling tower, a chiller, a plate-and-frame heat exchanger, and a load-following, water-meter-actuated chemical feed/bleed system. The plate-and-frame heat exchanger is included because such devices increasingly are being used to achieve energy savings in green-focused buildings.

Because Legionella bacteria are aquatic organisms, one can assume contamination of a system is most probable via the makeup-water system, regardless of the source. Although there are papers that point to dissemination via air for long distances, one must assume that aquatic organisms normally are associated with aquatic systems. That being the case, let us address the issue by seeing what can be done to minimize makeup water as a source of entry.

The use of an iodine-feeding device or a bromine feeder may appear to be an acceptable approach, but this only addresses planktonic or "free-swimming" organisms. It misses bacteria within sloughed-off slime films or protozoan cysts, failing to address the combined issues of contact, contact time, and concentration.2

Figure 2 shows a dual-bag filtration system that prevents dirt and other material from entering a tower pan. Protozoan cysts, which can contain thousands of Legionella bacteria, are quite large and can be filtered out.

Installing a sand filter, rather than a dual-bag system, is not advised. Consider the surface area of a grain of the course sand used in a sand filter. Now, multiply that by the number of grains of sand in the filter. Do you see the point? Achieving low-micron filtration requires fine sand (or garnet) and a large surface area. In time, such a system may experience high biofilm growth, which can serve as an additional inoculant.

Two bag filters are shown because one does not want to bypass unfiltered water into a cooling tower during bag changes, and the filtration system must provide sufficient capacity to address greater flow during cleaning and flushing in the spring and fall.

Now that we are feeding clean water to the tower pan, on what should we focus next? It stands to reason that a cooling tower is nothing more than an efficient air washer in its role as a heat exchanger. So, we look toward sidestream filtration of system water at 10-percent total system flow.

Now, you can see why the plate-and-frame heat exchanger was included in Figure 1. This system is stagnant, valved off most of the time, and an excellent incubator for undesirable microbes. Consider that when its internal surface area is valved off, it will not get any chemical feed. The first time the valve is opened to the unit, whatever is inside will be dumped into the operating system.

Figure 3 shows a dual-filter-bag system piped so that one filter can handle full flow. The other filter is valved on during bag changes. This allows complete draining and flushing of a dirty unit without contaminating the system. A sand filter can be used, but it must have the means to have the sand "fluffed" no less than monthly.

Figure 3 also shows a circulating pump off the down-comer pipe of the cooling tower. Water is pumped through the filter in service and piped back to the system through the line feeding the plate-and-frame heat exchanger. Do not "dead-head" this line. This will ensure that only clean, filtered, and treated water is flowing into the plate-and-frame heat exchanger at all times, minimizing biofilm growth and the likelihood of bacterial amplification.

If a system does not have a plate-and-frame heat exchanger, feed the clean water into the intake pipe of the condenser. It is assumed that the main condenser-water pump has a check valve to prevent backflow when the pumps are off. Again, do not dead-head this line.

Use of a centrifugal separator or mechanical filtration system also is not advised, as it cannot be defended in court. Remember, we are talking a range of 5 microns and less. A bag-filter system can be tailored to fit any system at any micron rating. Most system designers will settle for 50-micron bags; however, one can go as far as submicron, if desired. In an open cooling tower, we are not as concerned with filtering bacteria as we are with filtering dirt.

To achieve a state-of-the-art system, one additional piece of equipment is recommended: an air screen. Air screens minimize the impact of bugs, birds, leaves, grass clippings, and airborne debris on biocide feed systems. Also, they prevent plastic fill from becoming packed with debris and collapsing. An air screen loaded with debris can be "broomed off" during the season or washed out at season's end.

Now that we are at the current state-of-the-art, we must return to the question that could be asked in court: "What more could have been done?" The answer is that towers can be tested in-house for Legionella with a 30-min test kit. Such a test allows an in-house staffer to ascertain if conditions meet facility standards for risk. One always can send samples to a certified laboratory for a complete evaluation and paper trail, if required by health authorities.

There are two tests from which one can choose: a 30-min field test with a sensitivity of 100,000 colony-forming units (cfu) per liter or an industrial test with a sensitivity of 100 cfu per liter. The benefits of testing are obvious: It can be done quickly in-house, and the results are immediate. There also are some minor caveats: The kit only tests for L. pneumophila (although that is the organism responsible for more than 85 percent of outbreaks), and its sensitivity leaves a lot to be desired. But even buffered charcoal yeast agar, with its dilution factor, is prone to variables that can be challenged in court. Microbiological sampling, plating, and testing is an art, not a science. The science comes in the identification of the colonies.

Because the focus is achieving a reading of 0 cfu, any other reading is a cause for action, especially in a health-care setting, motel/hotel environment, and areas where close proximity to cooling towers is an issue.

There you have a look at a state-of-the-art Legionella-prevention regimen for an open-condenser circulating system. If you already are there, congratulations. If not, your facility is at risk.