Isolation valves, centrifugal solids separators and clean-in-place connections improve thermal performance.
by Chuck Sherman

An equipment provider received a phone call from a disgruntled user. The company thought their process cooling system was not working properly. They requested that someone come look at it immediately.

The equipment provider made a trip to the site for troubleshooting. The system, sized for 320 gallons per minute (gpm), consisted of a hot side closed loop system serving the process and a cold side open loop system with a cooling tower to dissipate heat. The two loops shared a plate and frame heat exchanger, and the pumps were fitted with typical trim—wye strainers and isolation valves on the pump suction, isolation and check valves on the discharge.

Image 1. Debris in the mesh prevented the system from functioning normally. (Courtesy of Fluid Cooling Systems)Image 1. Debris in the mesh prevented the system from functioning normally. (Courtesy of Fluid Cooling Systems)

A visual inspection of the system revealed no problems. After checking flow rates with a sonic flowmeter, it was discovered that the cold side of the system was operating at a fraction of rated capacity. Discussions with the maintenance staff revealed a maintenance schedule including weekly blowdown of the wye strainer and inspection of the cooling tower basin.

The system was shut down, and the strainer basket was removed. This revealed the source of the problem—debris was wedged in the mesh, almost completely blinding the strainer (see Image 1). The debris was so tightly embedded that a torch was required for cleaning. Upon startup, the system functioned as designed.

Heat Exchanger History

What happened in this system, and how can the system be designed to optimize service intervals?

To understand this, the history of heat exchanger design and selection must be examined.

For many years, the standard practice of the process cooling industry was to use a single pass shell and tube heat exchanger with marine waterboxes to dissipate heat (see Image 2). The design of the heat exchanger is simple and straightforward: A series of tubes are rolled into a flat tubesheet, and the tube bundle, complete with baffles, is inserted into the shell.

Image 2. The standard design for heat exchangers (Courtesy of Xylem Bell and Gossett)Image 2. The standard design for heat exchangers (Courtesy of Xylem Bell and Gossett)

The less aggressive of the two fluids is typically circulated through the shell, and the more aggressive fluid is circulated through the tubes. The choice of fluid circulation is also a function of viscosity and materials of construction, since the shell is commonly fabricated from carbon steel. Marine waterboxes were common in process cooling applications because head plates may be removed to facilitate cleaning in place (via rodding) of the tube bundle.

To ensure optimal service intervals, the heat exchanger was selected with a fouling factor to account for deposition of debris on heat-transfer surfaces. As late as the mid-1980s, a common fouling factor was 0.001 for industrial applications. The fouling factor could increase the heat exchanger surface area by 50 percent or more depending on heat transfer efficiency, expressed as the U value.

The U value defines thermal transmittance, or the heat exchanger's ability to transfer heat. The lower the number, the less effective the heat exchanger is in its ability to transfer heat. Heat exchangers with low U values require more surface area to accomplish their task. The combination of large surface areas, open passageways (shell and tube heat exchangers often use ½- or ¾-inch tubes, for example) and generous fouling factors made the shell and tube heat exchanger tolerant of debris in the system.

In 1923, Dr. Richard Seligman invented the plate and frame heat exchanger (see Image 3, page 111). This design relies on a series of thin gasketed corrugated plates, typically stainless steel, to transfer heat. It came into widespread use approximately 35 years ago when computerized selection software facilitated sizing. Characteristically, these heat exchangers are much more effective per square foot of heat transfer surface area, with U values commonly exceeding 1,000—meaning the unit is four times as effective as its counterpart.

Image 3. A plate and frame heat exchanger (Courtesy of Sondex, Inc.)Image 3. A plate and frame heat exchanger (Courtesy of Sondex, Inc.)

Increased efficiency comes at a cost. Plate and frame heat exchangers rely on very small passages between plates to maximize velocity and turbulence. Introduction of a fouling factor rapidly increases the size of the plate pack, which increases cost and reduces efficiency. To combat this problem, application engineers frequently select units with little, if any, fouling factor. The resultant selection, while effective in transferring heat, can be extremely sensitive to fouling. Without upstream filtration, a plate-type heat exchanger on an open-loop system can quickly foul, requiring disassembly and cleaning to restore the unit to design parameters.