Even though thermal-fluid heating systems have been widely used for over 80 years, they still provoke a certain amount of fear and trepidation in users. This anxiety is reinforced by any number of horror stories that usually involve systems that suddenly develop a "problem" after years of trouble-free operation. Maybe it's a fire around a pump or another piece of equipment. Or maybe the heater room fills up with smoke. What makes these incidents into oft-repeated horror stories is the lack of a readily apparent explanation for the event. Accident investigators might want to know why the pump seal was leaking, when the flash/fire/autoignition of the fluid was last tested, or whether a higher-temperature thermal fluid should have been specified. Then, whoever has the most experience with these systems will recite some myths and old wives' tales about thermal fluids and systems, and blame the incident on one of them.
Because thermal fluid systems do involve a combustible liquid under moderate pressure at an elevated temperature, they should inspire some degree of caution. However, few plants have people who've had time to develop a real understanding of how such systems work; thus, the myths and old wives' tales get accepted. So here, to help you reduce the potential for safety incidents, let's dispel five of the more common myths.
Myth 1 — New thermal fluid always contains water. It's easy to see how this myth keeps going. After new fluid is added to the system, suddenly hot fluid "geysers" out of the expansion tank vent into the catch tank and possibly onto the heater room floor. Or the pump begins to cavitate at every startup. Typically these incidents result in calls to the supplier, berating it for sloppy quality control and requesting an immediate shipment of "good" fluid. Ignoring the plant's own role in this problem (e.g., storing drums of fluid unprotected outside), the truth is all new fluid does contain trace amounts of water, typically less than 150 ppm. This is why new system startup instructions include a boil-out ("seasoning" or "conditioning the fluid") step. A proper boil-out not only removes any water in the new fluid but also eliminates water remaining in components from hydrotesting and in piping left exposed to weather prior to installation.
Unfortunately, often the boil-out is terminated prematurely, as soon as the pump cavitation stops or the scheduled time is up. Any water left in the system will collect in the bottom of the expansion tank and in other cold (less than 212°F) low points such the thermal buffer tanks. The fluid samples shown in Figure 1 come from the main loop and thermal buffer tank of a system that had been operating steadily at 550°F. Covered by the thermal fluid (because it's heavier and insoluble in water), water can't evaporate and, as a result, can remain hidden for years. Documented incidents testify to water remaining in place long enough to corrode through the metal, causing a thermal fluid leak (which creates its own excitement because these tanks are usually installed above the heater) even though the system has been run continuously. Normal operating-temperature swings might not generate enough flow past the hidden water to pull it into the main loop. However, a large drop in temperature at shutdown or adding new "good" fluid through the expansion tank may cause enough flow to dislodge the water. Problems might begin at the expected temperatures during startup (when the heater outlet temperature reaches 220–250°F) or might not occur until the outlet temperature exceeds 350°F in a large multiple-loop system. The most effective way to prevent geysers is to take the time to do a thorough boil-out after every cold shutdown. Circulate fluid through all branches and all heat exchangers. Leave the boil-out valve open until the expansion tank temperature is above 230°F and no more steam is visible from the vent. Justification for this extra time is simple — water will keep a system from operating.
Myth 2 — The thermal fluid must be bad because my pump seal is leaking. Leaking pump seals should be a concern because they are involved in a majority of thermal fluid fires. Seals are the only components of a thermal fluid system where there isn't a solid piece of metal between the hot combustible fluid and air. So while the ignition source isn't always identifiable, the fuel source is always the leak from the pump seal. What perpetuates this myth is that leaking seals often will have carbon buildups on the outer surface. While this carbon can force the seal faces apart and cause a leak, it doesn't necessarily indicate the fluid needs replacing.
Mechanical seals are lubricated by a thin film of fluid that seeps through the gap between the faces. Once this fluid hits the air on the outside surface it acts just like the fluid around a leaking flange, producing some smoke and lots of carbon. Worn or poorly installed replacement seals will allow more fluid to seep out between the faces, producing more carbon. While it's true that carbon is formed in the fluid as it degrades, these particles typically are soft, soot-like and small enough to pass between the faces (think graphite). Even larger pieces of "coke" that flake off heater tubes can't cause seal leaks because they're still soft enough to be ground up between the tungsten carbide and silicon carbide faces typically specified for thermal fluids. Only metal particles are hard enough to score a seal face; this is why a 60+ mesh screen should be installed in the pump suction Y-strainer on a new system or on any system that has had the piping opened up or new components added.