A steam trap has no alarm. Failed open, it leaks into a closed pipe where nobody sees steam; failed shut, it announces itself only through someone else's process complaint. So trap populations decay silently — and audit after audit across Indian industry finds the same picture: in a plant with no testing programme, typically 15–25% of traps have failed, most of them open, most of them years ago. This lesson is the working method for finding them, pricing them, and — the part most plants skip — keeping them found.
What a failed population costs
Give the practice plant a realistic register: about 80 traps — drip stations every 30–50 m along the mains (lesson C4), tracers, and the process drains on the dryer, jacket and hot-water duties. Let it decay to the unmanaged average: say 15 failed, ten of them open. A failed-open trap is an internal leak; through a typical DN15 orifice at main pressure even a partially failed trap passes 10–20 kg/h of live steam (lesson C9's table puts a fully open 3 mm path at 40 kg/h). Take a defensible 15 kg/h average: ten traps × 15 kg/h × 7,300 h ≈ 1,100 tonnes of steam a year. At fuel-only value that is ≈ ₹6.6 lakh; at lesson F1's marginal cost (₹675/t, the honest rate for a saving) it is ≈ ₹7.4 lakh — every year, from one mid-size plant, through valves whose job was to pass nothing. Against that, a full survey costs a few days and the failed traps a few thousand rupees each. Trap management is not maintenance hygiene; it is one of the highest-return line items in lesson F2's waste list.
The five tests, honestly ranked
Sight — the test tee or sight glass lesson D5's station provided. Crack the tee: flash steam (lesson A8 — condensate re-evaporating, normal) is a lazy, white, drooping cloud; live steam (a failed trap) is a hard, almost invisible cone at the orifice that turns white only downstream. The distinction takes an hour to learn and settles arguments no instrument can. Where there is no test point, there is no cheap truth — which is why D5 specified one.
Temperature — the most used and most misused test. An inlet reading proves the trap is in service (hot = steam is arriving) and a cold trap is failed shut or air-bound (lesson G1) — that much is solid. But inlet-vs-outlet ΔT says almost nothing about leakage: downstream of a healthy trap sits flash at saturation temperature, nearly as hot as the inlet. A pyrometer alone cannot convict a leaking trap. It can only acquit the frozen ones.
Sound and ultrasound — the workhorse. A ₹30–80k ultrasonic gun (the same one as lesson C9's leak survey) hears the trap's rhythm through the noise: a cycling trap (disc, bucket) should have a rhythm — open, discharge, shut, silence; continuous rushing through a trap that ought to cycle is the signature of blow-through. Modulating float traps are subtler — continuous light flow is their normal voice — which is where the tester's skill, and a comparison with a known-good twin on the same duty, earns its keep.
Thermal imaging — the pattern-finder. A camera walk along a return line finds the anomalies worth testing properly: one trap discharge glowing far hotter than its neighbours, a "closed" bypass valve with a hot outlet (the survey's classic bonus find), a flooded stretch running cool. The camera does not test traps; it tells you which ten of the eighty to test first.
Built-in sensors — conductivity or acoustic chambers fitted at the trap, wired or wireless, turning the annual survey into a live dashboard. They shine on critical drains (turbine protection, the dryer's syphon traps) where a failure costs more than the instrumentation; fitted plant-wide they are a budget conversation between instrument cost and steam price that lesson F1's numbers let you have honestly.
From survey to management
A survey is a photograph; management is the film. The difference is four disciplines, none of them expensive:
Tag everything. Every trap gets a numbered tag and a register line: location, duty, family and model, size, set pressure, differential it was sized on (lesson D5), test point yes/no. Eighty lines of a spreadsheet — one afternoon, done once, and suddenly "the trap near the old compressor" has a name, a history and a spare.
Route and rhythm. Test the population on a route, half-yearly for process-critical and high-pressure traps, yearly for the rest. A route with the register on a clipboard (or one of the survey apps the instrument makers now ship) takes two days for eighty traps and turns testing from an event into a habit.
Close every finding with a why. A replaced trap without a diagnosis is a return appointment. Failed shut on a syphon duty → steam locking, fit the release (lesson G1). Failed open at a drip leg → dirt, where is the strainer? Three float traps dead in a year on one duty → water hammer or a differential problem, not bad luck (lesson C5, G1). The register's "cause" column is where the plant actually learns.
Rationalise the spares. Eighty traps of thirty models is a stores nightmare; the same eighty in five or six standard families (lesson D5's matrix chooses them) means the right spare is always on the shelf and every fitter knows every trap. Standardisation is the quiet half of trap reliability.
- No register? Start with the count: walk the plant and number the traps. Most plants discover 20% more traps than anyone guessed — each one somebody's forgotten decision.
- First-ever survey: expect 15–25% failed, and budget the repairs before the survey so findings become work orders, not a report on a shelf.
- Learn the flash-vs-live-steam sight test at a known-good trap this week — it is the one skill in this lesson that costs nothing and transfers everywhere.
- Put the ten most critical drains on the half-yearly rhythm (turbine and main drips first — lesson C4's water-hammer stakes), whatever happens to the rest.
- Track one number for management: % of population failed at last test. Under 5% is a run plant; over 15% is lesson F2's waste #4 wearing a tag.
- Unmanaged trap populations run 15–25% failed; the practice plant's realistic bill is ≈ ₹7 lakh/year — silent.
- Temperature acquits (cold = shut); it cannot convict. Leak verdicts need ultrasound rhythm or the sight test.
- Flash steam droops lazy and white; live steam drills a hard clear cone. Learn the difference at a test tee.
- Survey → register → route → cause → standard spares. The register line matters more than the instrument brand.
- A replaced trap without a "why" is a return appointment — lesson G1 has the usual whys.
The plants that discovered trap losses first were not the ones with the best engineers — they were the ones with steam meters and night shifts. Metered plants kept noticing the same puzzle: the Sunday-night baseline, with every process idle, refused to fall to zero. Boilers ticking over, nothing running — and hundreds of kilograms an hour still leaving the header. Chasing that "ghost load" down the mains led, again and again, to the drip traps: dozens of small valves, failed open one by one over the years, each too small to notice and together the size of a running process. The ghost load remains the fastest business case for a survey: read your meter at 3 a.m. on the quietest night, and ask the traps where it went (lesson F8 turns this into method).