Thermosyphon systems have a solid reputation in the industrial refrigeration industry, and are a popular choice for many . . . as long as they work. But when a thermosyphon system has problems, those problems are sometimes not easy to identify or correct. Gary Gneiting, director of engineering at Mericle Mechanical, Inc., said a vortex could be one reason at the bottom of a poorly performing system.

“I wonder how many thermosyphon systems had design errors due to a vortex that were never resolved,” Gneiting said. “Now there is one more item to review in all thermosyphon systems to prevent vortices and costly field repairs.”

A vortex is a conical, rotating liquid void that forms in a fluid body as a result of a low pressure area. When a vortex occurs, the opening allows gas from the surface to be drawn down through the center of the vortex into the pipes. A vortex can occur in the refrigeration industry in five ways:

1. When oil is drained from oil pots;
2. When liquid enters the high pressure receiver dip tube;
3. When liquid leaves the thermosyphon receiver through the thermosyphon supply line;
4. With pump suctions for ammonia recirculators;
5. In water pumps for evaporative condensers.

In the case of a typical pump system, a vortex provides an opening for vapor to enter the liquid stream. Air is then drawn into the system, which inhibits water flow. “The pump is not designed to draw air; it’s made to pump a solid column of water,” Gneiting said. “So instead of 150 gallons per minute of flow, you might only get 10 because you’re moving nothing but air through the system.”

Gneiting said he uncovered the issue with vortices when he reverse-engineered a thermosyphon system at a large ammonia cold storage facility to determine the cause of high discharge temperatures for five ammonia screw compressors. He observed vapor in the thermosyphon supply pipes that increased significantly when the last screw compressor was operating.

“When four of the screw compressors in the system were running, you could see with sight glasses that it was just liquid, but when five compressors were operating it became overwhelmed with gas. You want to feed liquid, not gas, into your oil cooler. Liquid has a lot of heat transfer capability. When vapor is sent to the oil cooler, the needed heat exchange can’t be accomplished,” he said.

“There was no other explanation for where the gas was coming from than through a vortex,” Gneiting concluded. “It was happening because the liquid level in the thermosyphon vessel was too low.”

To illustrate the issue, Gneiting compared it to watching water flow down the drain in a bathtub. “When the water is high there isn’t a vortex,” he said. “It shows up as the last few inches are going down the drain.”

In researching the issue, Gneiting found a calculation for submergence issued by the Hydraulic Institute that predicts when vortices will occur. The calculation showed that 11.3 inches liquid height or submergence is necessary to prevent vortices; the thermosyphon system he was studying had only 9.5 inches. Using that discovery, Gneiting calculated that by reducing the thermosyphon ammonia flow by 33 percent, he could reduce the required submergence to 9.4 inches, thus preventing a vortex.

“The more flow you have the more height you need,” he said. “I reduced the flow by de-energizing one of the screw compressors. That was enough to prevent the vortex.”

At that point, the thermosyphon vessel could be replaced, repaired or the 1,000-horsepower screw could be converted to liquid injection. The de-energized screw compressor was operated only three months a year during blast freezing season, so it was decided to add liquid injection instead of replacing the vessel.

“I turned off the oil cooler and injected liquid directly into the screw compressor to perform the oil cooling,” Gneiting said. “The liquid injection came from a source other than the thermosyphon receiver, so it turned off part of the flow.”

Facility owners can now determine height levels in their thermosyphon system by plugging the appropriate numbers into the formula S=D + 0.574 * Q * 1/D1.5, with S equaling submergence inches, D equaling bell diameter inches and Q representing flow rate gallons per minute.

“I’ve been in this business for 35 years and nobody had ever asked me to check for a vortex in a thermosyphon system,” Gneiting said. “This calculation will provide guidance on whether you have a vortex and how best to deal with it. This will become a new item to consider when designing a system. If you have an old system you can check to see if a vortex is the problem.

“The end result is, the industry will have more reliable thermosyphon systems because people can utilize this information to prevent a vortex from forming,” Gneiting said.