What a Desiccant Dryer Costs You in Purge Air

Why this matters

A desiccant dryer takes a slice of the air it just dried and throws it out of a muffler on the floor, continuously, for as long as the plant is pressurised. That slice is real flow, made at full compressor power, and it is claimed before any tool in the building sees a molecule of it. Almost nobody writes it on the demand list. Shops end up buying a second compressor to cover a demand line that was already installed and running, and they buy it without ever seeing the number.

This article is the accounting. The measurement methods for dryness itself and for system pressure drop live in sibling articles and are not repeated here.

Before you go near a tower

A desiccant dryer holds two pressure vessels, and the off-line tower is at or near atmospheric while the on-line tower is at full line pressure, with a switching valve between them. Never open, vent, top up or sample a tower without isolating the dryer at both air connections, venting both towers to atmosphere through their own vents, and confirming zero on each tower's gauge with the vents still open, then locking and tagging the isolation under 29 CFR 1910.147, which is the standard covering stored mechanical and pressure energy. A tower that has just switched can be at full pressure with a controller that thinks it is not.

Desiccant itself carries two hazards the safety block on a generic pressure job will not name. Refilling or vacuuming a bed releases respirable dust from activated alumina or molecular sieve, so that work needs a respirator selected under a written program meeting 29 CFR 1910.134, because the route is inhalation and gloves do not address it. And the material is a desiccant, so it pulls moisture from eyes and skin on contact; wear sealed eye protection and gloves for handling in addition to the respirator, not instead of it.

The purge muffler is the third one. A heatless dryer blows down loudly, on cycle, all shift. Where measured exposure puts workers in scope, hearing protection and the rest of the program obligations under 29 CFR 1910.95 apply, and a muffler element that has plugged raises both the noise and the back pressure on regeneration.

What the bed is actually doing

Desiccant dries by adsorption: water molecules stick to the enormous internal surface area of the bed material rather than condensing as liquid. That is why a desiccant dryer reaches pressure dew points a refrigerated dryer physically cannot, which are down around the minus 40 F class and below rather than the above-freezing range a refrigerated unit is limited to by its own condensate.

The catch is that adsorption is reversible and the bed fills. Every gram of water that goes in has to come back out before that tower goes back on line, and the only way out is to lower the water's partial pressure over the bed, raise the bed's temperature, or both. That is regeneration, and regeneration is what purge air buys.

Purge is not overhead, it is a demand line

Write it on the demand list, at the top, above the tools. Three properties make it different from every other line on that list.

It is continuous. A fixed-cycle heatless dryer purges on a timer whether the plant is running one grinder or forty.

It is first in line. Purge is taken from dried product air downstream of the dryer's own inlet, so the compressor makes it before anything else is served, and it is charged at the compressor's full specific power.

It is invisible. It leaves through a muffler at floor level in the compressor room, which is the one place nobody stands.

What each regeneration type costs

Published purge fractions, all quoted at a reference inlet pressure, commonly 100 psig, and at the manufacturer's stated inlet conditions. Read the actual figure off the data plate rather than off this table; the point of the table is the spread between the options.

The heatless unit is the one most small shops own, because it is the one that quotes cheapest, and it is the one whose running cost never appears on a quote at all.

What moves the number

Line pressure. Regeneration works by expanding dried air to near atmospheric pressure over the off-line bed. At lower line pressure there is less expansion available per unit of air, so a larger fraction of product has to be spent to strip the same water. Purge fraction therefore rises as line pressure falls. That is a genuine offsetting term against the general rule that lower system pressure saves energy, and it is a small one: the net effect of system pressure on total plant energy is owned by the distribution article, not resolved here.

Control type. A fixed timer sizes every cycle for design conditions, which means it regenerates a bed that is nowhere near full most of the time. A dew point demand controller switches towers on measured bed loading instead, so purge falls in proportion to how far below design the real moisture load is running. The saving is real and it is not a constant, so it is read from the controller's own logged cycle times after a week, never assumed from a brochure.

Inlet condition. Every gram of water arriving at the bed has to be purged back out. Air that reaches the dryer at a lower temperature carries less water, which is why the aftercooler and the separator ahead of the dryer are part of the purge cost, not separate from it.

The artifact: a filled-in flow budget

One sheet, four lines, every line with the method that produced it. This is the sheet for a shop with one rotary screw and a heatless desiccant dryer whose rated flow matches the compressor's rated flow, so every figure below is expressed as a percentage of that one rated flow.

Read what that sheet says. Purge is 15 of the 95 percentage points the compressor actually has to produce, so about 16 percent of everything the machine makes goes straight out of a muffler. Purge and leaks together are 33 of 95, about 35 percent, which is air no tool in the building ever touches. And the compressor is running at 95 percent of rated capacity with no headroom, which is why this shop was quoting a second machine.

The leak figure sits just under the lower edge of the 20 to 30 percent range the Department of Energy's compressed air guidance associates with plants that run no leak program, which is consistent with a plant that fixes the loud ones and ignores the rest. The same guidance puts under 10 percent within reach where a program is actually run.

Now the two corrections, applied to the same sheet.

Required production falls by 16 percentage points, from 95 to 79, which is about 17 percent less air made for exactly the same work done. The second compressor is now a machine this plant does not need to buy, and the existing one has gone from 95 percent of rated capacity to 79 percent, which is the headroom the shop thought it was shopping for.

Two honest caveats on that sheet, both of which belong in the report rather than in a footnote. The 9 percent purge figure is a one-week average under one set of ambient conditions, and it will rise in a humid month because the bed genuinely fills faster then; the controller is not a fixed discount. And the 8 percent leak figure is a measured value on the day of the retest, which drifts back up between surveys, so it is a starting point for a program rather than a permanent state.

References

• 29 CFR 1910.147 for isolating and verifying depressurisation of both towers before any work on a desiccant dryer
• 29 CFR 1910.134 for the written respiratory protection program covering desiccant dust during a bed change, and 29 CFR 1910.95 for occupational noise exposure where purge muffler discharge puts workers in scope
• Department of Energy compressed air system guidance for the leak-loss ranges cited, and dryer manufacturer data plates for purge fraction, reference inlet conditions and pressure dew point rating
• See related: How to Tell Whether a Dryer Is Doing Its Job; What Pressure Drop Through Distribution Actually Costs