The Trap That Tested Good and Was Passing Steam

Why this matters

Three consecutive annual surveys logged every trap in one bay of a plant as good. The atmospheric vent on that bay's condensate receiver plumed continuously the whole time and got heavier each year. Both records were accurate. The trap tests were correctly performed with correct instruments by three different technicians, and every one of them was measuring something other than the trap, because a trap test reads the outlet stub, and whatever else lands in that stub gets credited to the trap.

Approaching a receiver that is venting

The first temptation on a case like this is to walk up to the vent and see what is coming out of it. Do not.

A receiver vent discharging live steam is a burn hazard whose most dangerous zone is invisible, because a steam jet stays clear for the first stretch out of the opening and only becomes a visible plume once it has entrained enough room air to condense. Observe from a body length off and to the side, never from underneath or downwind, and never put a hand into a plume to judge its temperature. Where the vent discharges near a walkway or a work platform, the continuous noise falls under your hearing conservation program per 29 CFR 1910.95.

Everything in this case was diagnosed from records, gauges and valve handwheels. Nothing was opened, and nothing needed to be.

What the file actually contained

The reconstruction started from paper because the fault had been stable for years and there was no event to catch in the act.

• Three years of survey sheets for the bay, each trap with an ultrasonic level, an upstream temperature, a downstream temperature and a state. Every trap in the bay logged good all three years.
• Operations run-time records showing the bay's production hours essentially flat across those three years.
• A gauge log for the bay's steam header at 100 psig and the condensate return at 2 psig, both steady across the period.
• A maintenance work order from four years back, station N-114: "repack leaking valve, station bypass."
• A piping isometric of the trap stations in the bay.

Two things stood out immediately, and both were in the survey sheets rather than in the equipment. Station N-114's sheets carried a note in all three years reading "continuous background, header noise." And N-114's downstream temperature was recorded as warm every year, with the surveyor's note "consistent with condensate discharge."

Working backwards from the plume

The vent plume is the only hard physical fact in the file, so start there and ask what could produce it.

Could it be flash? Condensate arriving at a receiver at 2 psig from a 100 psig service re-boils a fraction of itself on the pressure drop, and that flash vapor is a large volume of visible cloud out of a small mass of liquid. On a receiver serving many traps the discharges overlap, so the plume is continuous even in a perfectly healthy system. The article on what flash steam is owns that mechanism and the fraction; it is not re-derived here. The relevant point is that a continuous plume at a multi-trap receiver proves nothing at all, which is why nobody had investigated it.

But it grew. Flash fraction is fixed by the two pressures, and both pressures were steady. Condensate load is set by the equipment, and production hours were flat. A flash plume at unchanged load and unchanged pressures cannot grow. A cracked valve can, and that is the mechanism: a valve held slightly open is being wire-drawn, the seat erodes, and the opening enlarges year over year. The growth is not evidence against a bypass, it is the signature of one. Something else was adding mass to that receiver.

Could it be a trap the surveys missed? Every trap in the bay was retested with an ultrasonic listener on the body and a contact probe on brushed bare fittings, on the same day. All read good again. So the surveys were not wrong about the traps.

Then the path is not a trap. If live steam is reaching the receiver continuously and no trap is passing it, there is a route around a trap. The isometric shows every station has a bypass, and the work order names the one bypass in the bay that had been touched.

The station that could not be tested the way it was being tested

At station N-114, the bypass line rejoins the trap outlet about 8 inches downstream of the trap body. That geometry is the whole fault.

A probe on the outlet reads the mixture of everything arriving in that stub. An ultrasonic sensor on the trap body hears the trap's own cycling clearly, because it is bolted to it, plus the bypass flow arriving as a steady, structure-borne background. Three surveyors heard a healthy trap snapping over a continuous hiss, and three surveyors wrote down what the trap was doing and characterized the hiss as noise from the header. The trap really was healthy. The station was not.

The temperature evidence was on the sheet the entire time and was read the other way. A warm downstream side is what a working trap produces, because it discharges hot condensate. There is no way to separate "warm because condensate is discharging" from "warm because steam is bypassing" without knowing what the number should be, and that requires both saturation temperatures, which the sheet did not carry.

Confirming it at the handwheel

The bypass valve at N-114 was found off its seat. The packing repair four years earlier had left it cracked, and no one had closed it because nothing on the work order asked anyone to.

Closing it is a hot valve operation, so it was done in heat-resistant gloves rated for the surface, standing to the side of the stem rather than in front of it, with the packing confirmed not blowing before anyone put a hand near the gland. A valve whose packing is actively leaking steam does not get operated at all until the line is isolated.

The confirming numbers came off the outlet stub with a contact probe on the same brushed spot before and after.

Saturation at the 100 psig inlet is roughly 338 F, and saturation at the 2 psig return is roughly 219 F. Those two define a 119 F span, and every temperature call in this case sits inside it.

• Bypass open: stub at 322 F, which is 103 F above return saturation, or about 87 percent of the span.
• Twenty minutes after closing it: stub at 224 F, which is 5 F above return saturation, or about 4 percent of the span.

The vent plume at the receiver went from continuous and heavy to intermittent bursts within minutes, matching the discharges of the traps feeding it.

Read the direction against what the mechanism predicts, because that is the check that the story holds. If the stub had been carrying condensate only, it would sit near the low end of the span both times and closing a bypass would change nothing. If it had been carrying live steam, it would sit near the high end and closing the path would drop it to the low end. It did the second thing, and it did it in the direction and by the magnitude the mechanism requires.

What this changes about how a trap gets tested

The generalizable finding is narrow and worth stating exactly: a trap test measures the outlet, and every path into that outlet is attributed to the trap. A cracked bypass, a leaking isolation valve on a parallel station, a tie-in from a drip leg upstream of the sensor, all of them read as trap performance.

That makes the first action in any trap test a valve position check, confirmed by hand on the stem, and it is not a measurement. It is the thing that makes the measurements mean anything. The testing article carries it as step one for exactly this reason.

There is a second, smaller finding in the paperwork. The survey sheet had no field for bypass valve position, and the work order that repacked the valve had no field for returning it to its normal position. Those are the same gap seen from two ends, and the cheap fix is a line on the survey sheet and a restore-to-normal line on any work order that operates a valve. Neither of those would have found the fault by itself, but either one would have made the three surveys able to see it.

Verifying you do not have the same fault

• Walk your bypasses on one bay and check stem position by hand, before you trust any survey of that bay. This is a fast pass and it does not require instruments.
• Compare each station's downstream reading against the two saturation temperatures, not against a general expectation of "warm." A downstream side sitting high in the span between inlet and return saturation is the signature, and it is invisible without both numbers on the sheet.
• Treat "continuous background" in a survey note as a finding, not as ambient. Three sheets carried that phrase and none of them generated a work order.
• Where a receiver vent's plume changes at unchanged load and unchanged pressures, stop attributing it to flash. The flash fraction cannot change while those two are fixed, so a change in the plume is a change in mass reaching that receiver.
• Close out any valve-operating work order with a stem-position line. The valve that gets repacked and never reseated is a specific, repeatable failure and it is not confined to steam.

References

• 29 CFR 1910.95, occupational noise exposure, for sustained work near a venting receiver or a blowing discharge
• 29 CFR 1910.132, personal protective equipment hazard assessment, covering hot surface contact during valve operation and probe placement
• Steam tables for saturation temperature at the measured inlet and return pressures, which are what make a downstream reading interpretable
• See related: How to Test a Steam Trap Without Taking It Apart; What Flash Steam Is and Why It Looks Like a Failure