Why a Steam System Loses Efficiency Without Anything Breaking

Why this matters

A steam plant that has not been walked in a decade is usually still holding pressure, still holding setpoint, and still passing its annual combustion check, while burning a materially larger share of its fuel for the same delivered heat than it did when it was commissioned. Nobody is hiding anything and nothing has failed in the sense that anyone in the building would recognise. The losses are real, they are large, and they are invisible for a structural reason worth naming precisely.

Every instrument a steam plant actually watches reads pressure, and pressure is the one variable that none of the quiet losses change. That is not bad luck. It follows directly from what steam is.

The hazards belong to the walk, not to any one fitting

This article sends you around a plant with a temperature probe, and the exposures come from the walking.

• Do not brush, cut, tear or lift pipe insulation to reach bare metal for a reading. Thermal system insulation on steam and condensate piping is presumed asbestos-containing material under 29 CFR 1926.1101 for construction work, with general industry duties at 29 CFR 1910.1001, until it has been sampled. Take readings on fittings, valve bodies and trap bodies that are already bare, or have removal done by a trained crew under the applicable standard. This is an inhalation route, so gloves and glasses are not the control, and a respirator is only a control inside a written program meeting 29 CFR 1910.134.
• Never trace a leak by hand, by feel, or by reaching toward a sound. A jet at header pressure is invisible for the first stretch out of the seat and cuts before you feel heat. Work from beside any discharge and keep a body length off it.
• Where a contact probe has to be seated on an already-bare fitting, wear gloves rated for that surface temperature and eye and face protection under your PPE hazard assessment per 29 CFR 1910.132, and brush the spot with a long-handled brush across the pipe rather than toward yourself.
• Do not enter a below-grade condensate pit, vault or tunnel until that space has been evaluated as a permit-required confined space under 29 CFR 1910.146, and treat sustained exposure to a blowing trap as noise under your hearing conservation program per 29 CFR 1910.95.

Why the gauge is blind by construction

Steam is the common working fluid in a building whose delivered temperature is read straight off a pressure gauge, because it is distributed at saturation. Any saturated fluid has its temperature set by its pressure, which is what a refrigerant P/T chart is; the difference is that on a steam system the saturation temperature is the process rather than a charging diagnostic. Hold 100 psig in a header and the steam in it is at roughly 338 F, everywhere in it, whether that header is delivering full load or nothing at all. That relationship is the whole design of a steam system, and it is also the reason the primary instrument tells you almost nothing about where the energy went.

A boiler gauge holding 100 psig says one thing: the burner is keeping up with whatever is leaving. It does not distinguish steam that condensed in a coil and heated a building from steam that blew through a worn trap seat into a return line and condensed there. Both consume the same latent heat. Both are satisfied by the same burner. The gauge reads identically.

Now put that next to what actually gets watched daily on a small plant: boiler pressure, header pressure, water level, and a burner that either fires or does not. Every one of those is a pressure or a level. None of them is a mass rate, a temperature difference, or a return rate, and the quiet losses only exist in those three.

What is deliberately not on this list

The list in the next section is short, and it is short because the more dramatic failures have been taken off it on purpose. Naming what came off is the useful half.

• A relief valve lifting is an event. It is loud, it is visible, and somebody calls that afternoon.
• A low-water cutoff trip shuts the boiler down and produces a service call.
• A visible atmospheric leak gets reported by the first person who walks past it.
• A cold zone or a cold tank produces an occupant complaint, which is the most reliable detector in any building.
• A fouled burner or a dirty fireside moves stack temperature and shows in the combustion analysis somebody already runs annually.

Every one of those changes something a human or an instrument in that building is already watching. Take all of them off, and what remains is not a leftover list. It is precisely the set of losses that have no detector anywhere in the plant, which is why they are the ones that get large.

The four losses that leave every gauge correct

Each of these is invisible on pressure and visible on exactly one other measurement. The measurement is the point of the entry.

Failed-open traps. A trap that no longer distinguishes condensate from steam passes live steam continuously into the return. The header does not care, the equipment still heats, and the only thing that changes is fuel. Detector: a trap test on the population. Secondary detector: a return main that reads above zero on a nominally vented system, which usually means somebody's failed-open trap on a higher-pressure branch is pressurising the common return, and that raises the back pressure against every other trap on it.

Air occupying steam space. Air is a gas that does not condense, it does not carry latent heat, and it does not leave a closed pipe on its own. A deleted or plugged vent leaves the far end of a main or the remote end of a coil filled with something that is not steam. Detector: metal temperature at the remote end compared against the saturation temperature for the gauge pressure at that equipment. A remote-end reading that plateaus more than about ten degrees below saturation means air or water is still occupying that space. The mechanism is owned by the sibling article on why air in steam is worse than air in water.

Missing, wet or crushed insulation. Detector: condensate rate, which is to say trap load and the wetness of the steam arriving downstream. What insulation does and does not do on a steam line is owned by its own article and is not re-derived here; the point that belongs in this list is that its loss shows up as a mass rate, and there is no mass instrument on most of these plants.

A falling condensate return rate. Detector: the make-up meter, or the softener regeneration log, or the chemical drum change frequency, all of which scale with make-up volume and are already written down somewhere. The compounding cost of a lost pound of condensate is owned by the sibling on why condensate is the expensive half of a steam system.

Why it drifts instead of stepping

If any of these arrived as a step change, somebody would date it and fix it. They do not.

Traps accumulate failures at a few percent of the population per year where nobody tests them, and published trap-survey results on untested populations vary widely with age, pressure and trap family, and where you quote a band, cite the survey it came from and say whether it counted failed-open only or every failure mode. That is not one trap failing; it is one trap every few months for years, and no single one is detectable against the noise of a fuel bill that also moves with weather, occupancy and fuel price.

Insulation comes off in patches. A valve is repacked and the section around it never goes back on. A pipe is re-routed and eight feet of it stays bare. Each patch is a defensible decision on the day.

A vent gets plugged during one repair because it was passing steam, which it was, because it had failed. The plug fixes the visible symptom permanently and removes the function permanently, and the person who did it has left.

None of those is negligence and none of them announces itself. What is missing is not attention. It is a cumulative indicator, and a steam plant does not have one.

Worked example: a morning where every reading was correct

A plant with one boiler on a 100 psig header, last trap survey nine years ago, no alarms in the previous nine months. The walk starts with the readings the site already takes.

Everything the plant watches, checked first. Boiler gauge 100 psig. Header gauge 98 psig at the far end, a normal drop for that run. Water level normal, no alarm history, combustion analysis from the annual service within the burner manufacturer's range, stack temperature consistent with three years of records, building holding setpoint. On the plant's own instrumentation there is nothing to report, and that is the finding rather than the absence of one.

Then the four measurements nobody takes.

A trap test on a 30-trap sample drawn from a 118-trap population returned 6 failed open and 2 failed closed. Six of thirty is 20 percent of the sample failed open, which lands inside the fifteen to thirty percent band that an untested population of that age is expected to sit in, so the sample is behaving as the general case predicts rather than being an outlier that needs explaining.

Remote-end metal on the five steam coils in the plant, compared against saturation for each coil's own gauge pressure: two of the five were within the ten degree working limit, and three sat between 40 F and 70 F below saturation. Those three are air-bound, water-bound, or both, and every one of them was delivering heat and holding its space, just less of it.

The return main gauge, on a system whose drawings label the return as vented, read 6 psig. A vented return should sit near zero. Six psig on it means live steam is entering that return continuously, which both confirms the trap finding and raises the back pressure on every trap discharging into it.

The softener regeneration log, which the site keeps because the water treatment vendor asks for it, showed 3 regenerations per week in the earliest year on file and 8 per week now. That is a factor of about 2.7 in make-up volume, on a plant whose steam demand has not grown.

What the four readings have in common. Not one is a pressure at the boiler. One is a trap test, one is a set of temperatures compared against a table, one is a pressure at a place nobody has a reason to stand, and one is a tally in a binder kept for a different purpose. All four were available on the day the plant was commissioned, and none of them is on any log the site maintains.

What this example deliberately does not do. It does not add the four findings into one percentage of generation. That conversion has its own denominator problem and its own independent cross-check, and it belongs to the sibling article on putting a number on what a trap population is wasting.

The failure mode if this walk is skipped and a proposal is written anyway. The usual outcome is a burner tune and a boiler replacement quote, because the boiler is the only component anybody has instruments on. That work is often defensible on its own terms and it touches none of the four losses above, so the plant gets a new boiler feeding the same three air-bound coils through the same worn trap population into the same pressurised return. The measured improvement comes in well under what was promised, and the reason is not the boiler.

Getting these four readings right

Take metal temperature with a contact probe or on high-emissivity tape, never by pointing an infrared thermometer at bare shiny steel, which reads low, sometimes by tens of degrees, and fakes exactly the fault you are looking for.

Match the windows. A trap sample walked in February against a make-up figure totalised over a year compares a peak to an annual average, and on a heating plant those differ by more than the finding does.

Note that three of the four are a single visit. Only the trap survey scales with plant size, and the other three will usually tell you in advance whether that survey is going to find anything, which is the practical argument for taking them before quoting any work.

References

• ASME Steam Tables, or an equivalent saturated-steam table, for the saturation temperature at each gauge pressure you read
• 29 CFR 1926.1101 and 29 CFR 1910.1001 for presumed asbestos-containing thermal system insulation, 29 CFR 1910.134 for respiratory protection programs, 29 CFR 1910.132 for PPE hazard assessment, 29 CFR 1910.146 for below-grade condensate spaces, 29 CFR 1910.95 for noise exposure, and 29 CFR 1910.147 for isolation before a gauge connection is opened
• See related: How to Put a Number on What a Trap Population Is Wasting; Why Condensate Is the Expensive Half of a Steam System; What Insulation Does and Does Not Do on a Steam Line; Why Air in Steam Is Worse Than Air in Water