Avoid turbulence in your compressed air system

Values for Di as a function of pipe schedule are tabulated in various engineering reference books. Although the formula assumes isobaric flow, it’s usually adequate for rough evaluations. Other more complex relationships accommodate flow with pressure differentials.

So what’s wrong with high velocity?

Some argue it’s simply better for air to get there faster. This is true. But what happens when that high-speed air goes through a block valve, crossing tee or other fitting? It produces turbulence, which, in turn, produces backpressure — two factors that degrade system performance. Here are some examples.

Textile mill

Short cycling in a 350-hp, oil-free, two-stage rotary screw compressor with two-step capacity control troubled this plant. The twin 1,500-cfm compressors were configured as shown in Figure 1. Compressor No. 2, the trim unit, was short cycling. The plant installed a blow off valve so the trim unit ran at part load with continuous blow off. This was obviously expensive because both units were at part load in demand, but at full electrical consumption. It operated like this for more than five years.

The air velocity for 1,500 cfm in a four-inch Sch 40 steel pipe at 90 psig was about 40 fps, twice the recommended value. The pressure loss in 200 ft. of four-inch pipe and one 90-degree short radius elbow was 0.89 psig — not very significant. Calculated values suggested this was an excellent piping arrangement.

The actual measured pressure drop from the compressor discharge to the dryer inlet was eight to nine psi — quite excessive. The air running through a 90-degree turn and dead heading into the block valve located only nine in. away caused this loss (see Figure 2). The block valve allowed the compressor and dryer combination to be alternated.

The turbulence sent a pressure signal to the capacity control pressure switch sensor to unload the compressor prematurely, which caused short cycling (load four seconds, unload four seconds). It also caused backpressure, which eliminated any effective storage for the 10-psig operating band.

With six-inch pipe, the velocity would fall to about 17.8 fps and an eight-inch line would have a velocity of about 10.3 fps. Either larger pipe would eliminate most of the pressure drop.

What is the annual cost of electrical energy to produce this eight-psi drop? At a rate of $0.05 kW-h, it’s more than $9,000. During the five years, the plant spent more $45,000 to generate pressure that was never used. The control problem that caused the plant to run in blow off mode, rather than short cycle and harm the equipment, could well have wasted another $25,000 to $30,000 annually.

Aluminum mill

This plant had a short cycling problem in one of two compressor rooms, both of which have three 150-hp 750-cfm compressors running at 90 psig (see Figure 3).

The three lubricant-cooled rotary screw compressors ran on two-step control and were short cycling (six seconds on, six seconds off). The lack of effective blow-down and idle consumed their full kW energy draw, which also led to high maintenance and repair costs for coolers, motors and airends. On the other hand, the three reciprocating units were loading and unloading in an acceptable manner (one minute on, 30 seconds off), saving power and causing no mechanical problems.

Both systems used tees to the header (see Figure 4). The only difference was the pipe size and resultant velocities. Room 2 with the rotary units had extremely high velocities, which cause high turbulence and pressure loss at the crossing tees, bends and valves.

Plant maintenance personnel explained that the rotary units used two-inch discharge pipe because that was the size of the port. The instruction manual also advised using pipe the same size or larger.

Short cycling induced by high velocity and piping configuration forced the plant to run three 150-hp compressors when the original specifications called for only two units. At $0.05 per kW-h, this plant wasted more than $60,000 per year in energy costs for more than 20 years — $1,2000,000.

Milk and dairy processing

In this case, the plant was running four 100-hp lubricant-cooled rotary screw compressors under modulating control. It was losing productive capacity because a 20-psi pressure drop made it impossible to maintain the required minimum 90 psig in the header.

The piping schematic in Figure 5a shows the original piping and Figure 5b shows its modification. Four 100-hp 490-cfm oil-cooled rotary screw compressors delivered air to a six-inch main header. The velocity in the four-inch interconnecting piping was as follows:

• 13.2 fps @ 490 cfm.
• 26.4 fps @ 980 cfm.
• 39.6 fps @ 1,470 cfm.
• 47.4 fps @ 1,760 cfm.

Four crossing tees added turbulence at these velocities. The total pressure loss with all machines at full load was 20 psig. When demand increased, the pressure in the main fell below 90 psig, shutting down production.

Two changes solved the problem. First, the four-inch crossing tees were changed to directional angle entry. The pressure drop fell to six psi and the main system now receives 104 psig that is easily regulated to a steady 90 psig. The connections were prefabricated and installed during a one-day maintenance shutdown at a cost of $4,200. This eliminated the production interruptions that had occurred for 20 years. Second, the compressor discharge pressure was reduced to 98 psig, which represents a power saving of six percent, equivalent to about $9,585 annually.

Short cycle and surge

Figure 6 shows three centrifugal compressors, each of which delivered about 3,900 cfm at 90 psig at full load through a six-inch line (46.3 fps), feeding a 12-inch pipe (11,700 cfm, 35 fps). Unit No. 1 fed the end of the 12-inch header coaxially. The line from unit No. 2 entered at a 90-degree crossing tee and that from unit No. 3 entered just upstream of the No. 2 connection but through a 45-degree directional entry.

In this case, three air streams converged within a nine-inch axial length of pipe. The turbulence at this connection was reflected in the pressure oscillating between 89 to 90 psi. That’s not bad, but centrifugal compressors are sensitive to backpressure. The measured flow from units No. 1 and No. 2 was a constant 3,900 scfm. No. 3 short cycled from 3,900 to 2,900 scfm every five seconds. Full turndown on No. 3 was 3,300 scfm. Thus, we concluded that the effect of the sensed pressure spike at the compressor controller was almost equal to the 10-psig operating band. Observation showed the blow off valve also trying to move. The annual loss of about 750 cfm cost the company about $42,000, in addition to the surging and productivity problems.

Lengthening the 12-inch header and bringing each compressor in with a directional entry oriented 30 degrees and 30 inches apart solved the problem. The pressure stabilized at 90 psig with no short cycling evident. Streamlined connections handle high velocity without significant problems.

These examples clearly show a key item often overlooked is compressed air velocity. During many years of troubleshooting compressed air systems, we find these problems in the newer systems installed during the past 20 years. Rotary screw compressors seem to be supplied with smaller and smaller NPT connections, which are subsequently construed as valid guidelines.

When installing compressors, always size the pipe according to the anticipated flow. Don’t be misled by the discharge flange or pipe size. Good piping practice dictates the use of directional angle entry instead of crossing tees and long radius ells rather than short radius fittings. Also eliminate dead heads and other sources of turbulence. Using air receivers and large headers effectively can offset many potential pressure drop problems.

Hank van Ormer is president of Air Power USA Inc. Don van Ormer is the company’s vice-president. They can be reached at [email protected] or (740) 862-4112.

Figures: Air Power USA, Inc.