The complaint never says "fouling". It says the batch that took forty minutes now takes an hour (lesson D1's coil, running out of breath), or that the hot-water outlet cannot hold temperature any more, or — the quietest tell — that a control valve which used to sit at 60% open now lives wide open (lesson D3's authority, spent). Inside the shell, scale and films have been growing for months. This Clinic lesson is the ten-minute measurement that catches them growing: one number, U, computed from four temperatures and a flow, trended until it names the day the exchanger must be cleaned. Lesson D13 built the hardware; this lesson checks its pulse.
The one number an exchanger owns
Every heat exchanger obeys Q = U × A × LMTD (lesson A11). Area is on the nameplate and cannot change. LMTD comes from temperatures you can read. So if you measure the duty Q and the temperatures, the equation hands you U — the overall heat-transfer coefficient, the exchanger's fitness certificate. Clean metal has a high U; every layer of scale, mud or oil film adds its resistance in series and drags U down. You cannot see inside a working shell, but U sees for you: fouling is invisible everywhere except in the arithmetic.
The duty side is one line: for a water heater, Q = water flow × 1 kcal/kg·°C × temperature rise. The LMTD side carries the one trap in this lesson — the steam temperature must be the saturation temperature at the shell's own gauge, never the header's. In normal running the control valve throttles, and the shell floats at whatever lower pressure condenses the load (D1's floating-shell idea) — take 148 °C from the 3.5 kg/cm²g header while the shell actually sits at 1.4 kg/cm²g (126 °C, lesson A5) and your "measured" U is fiction. The protocol below removes the trap entirely.
The protocol — same conditions, every time
A trend is only honest if every point is taken the same way. Once a month, ten minutes: steam valve wide open (so the shell sits at line pressure and the gauge-trap disappears — and never during the turndown conditions where lesson G3's stall lives), water flow at its normal full rate, then read five things: water in, water out, water flow, shell pressure, and the date. Saturation temperature from the shell gauge (A5), LMTD, one division — U goes on the chart. The absolute value carries honest slop (nameplate area, a plant flowmeter); the slope is what convicts, which is why the conditions must repeat.
The unit: 8 m³/h of water to be heated 60 → 85 °C on 3.5 kg/cm²g steam (Ts 148 °C), A = 3 m², designed at U = 1,000 with fouling allowed for (lesson D13's sizing).
Note what the fouled reading looks like on the plant floor: even with the valve wide the water tops out at 78 °C against a required 85 — the exchanger physically cannot deliver its duty any more. The month-12 reading was the appointment card; today's is the penalty for missing it.
Reading the slide — which side, what cure
U falls; the next question is which side grew the layer, because the cure differs. Check the air vent before blaming any deposit: air blanketing on the steam side (lesson C8) mimics fouling perfectly — resistance where condensing film should be — and a seized vent or post-shutdown airlock can knock U down overnight. Fouling grows over months; air arrives in a day. The recovery speaks too: a U that collapses after every shutdown and partially recovers as the system vents is air; a U that slides smoothly down the calendar is deposit. On the water side, hardness scale is the default suspect above ~60 °C surfaces (lesson B7's chemistry applied downstream) — it follows the water quality, so a softener upset shows up on the U-chart weeks later. On the steam side, oil films and dirt are rarer but real. D13's dirty-side-first architecture decides how cleanable each side is — tube-side deposits meet a brush or a chemical circulation; shell-side ones usually mean a bundle pull.
After any cleaning, take a protocol reading the same week: that post-clean U is the new baseline, and it is also the receipt — a "cleaned" exchanger whose U did not recover was not cleaned where it mattered.
- Pick your three duty-critical exchangers and give each a card: nameplate A, design U, and a line per month — commissioning-week reading first if you still can.
- Take every reading valve-wide at normal full water flow; log the shell gauge, not the header. If the plant cannot spare a valve-wide window, trend at a fixed repeatable valve position instead — repeatability beats correctness of the absolute.
- U dropped suddenly? Vent first, diagnose second — C8's fix is free and takes an afternoon.
- Watch the two proxies between readings: the control valve's resting position creeping open, and batch heat-up times stretching (D1). Either one is the chart asking to be read.
- After cleaning, take the receipt reading the same week — and if U did not come back, the wrong side was cleaned.
- U = Q ÷ (A × LMTD): four temperatures and a flow. Fouling is invisible everywhere except in this arithmetic.
- Tₛ comes from the shell's own gauge — a throttled shell runs far below header pressure, and the header number fabricates loss.
- The calorifier's story: 1,400 clean → 980 at the design line (book the clean) → 610 today (duty lost, valve wide).
- The cleaning trigger is the design (dirty) U — not "when production complains".
- Sudden drop = air, vent it (C8). Slow slide = deposit, clean it. Post-clean reading = the receipt.
Fouling is the oldest heat-transfer problem in domestic history — every kettle in a hard-water town grows its "fur", and for centuries the answer everywhere was the calendar: brewers scoured their coppers between brews, laundries chipped their pans each season, engine-room crews scaled condenser tubes on a rota whether they needed it or not. Cleaning by calendar survives on many plants today, and it fails in both directions at once — units opened clean (money spent, production lost, gaskets disturbed for nothing) while the one actually choking waits its turn in the schedule. The U-chart is the modern correction: it lets each exchanger book its own appointment, and cancel the ones it does not need.
FAQ
How do I measure fouling in a heat exchanger without opening it?
Compute U from operating readings: duty (water flow × temperature rise) divided by area × LMTD, with the steam temperature taken from the shell's own pressure gauge. Repeat monthly under the same conditions — valve wide, normal flow — and plot it. The falling trend is the fouling, quantified; the design U is the line that books the cleaning. Ten minutes a month, no tools beyond thermometers and the gauges the unit already carries.
Why must the steam valve be wide open for the reading?
Two reasons. Repeatability — a trend only convicts if every point is taken the same way. And the gauge trap: with the valve throttling, the shell floats at whatever low pressure condenses the load, so the saturation temperature in your LMTD must come from the shell gauge; valve-wide, the shell sits at line pressure and the ambiguity disappears. If production cannot spare a valve-wide window, a fixed repeatable valve position is the honest second best.
U dropped 30% overnight — can fouling really grow that fast?
Almost never. Deposits grow over weeks and months; an overnight collapse is air on the steam side — a blanket of non-condensables doing a perfect impression of fouling — or a flooded shell (the trap or a stall condition, G3). Check the air vent (C8) and the condensate drainage first; both cures are free compared with a bundle pull.
What U should a steam-to-water exchanger show?
Teaching bands: clean steam-to-water service commonly measures in the 1,200–1,500 kcal/h·m²·°C region, and designers buy it at 800–1,000 with the fouling allowance built in — D13's calorifier was sized at 1,000 exactly so it could foul and still deliver. Your absolute number will differ with velocity and geometry, and it does not matter much: the baseline you measure at commissioning, and the slope after it, carry the diagnosis.
When exactly should I clean — and how do I know the cleaning worked?
Clean when the trended U reaches the design (dirty) value — that is the whole meaning of the design margin: the day it is spent, the duty has no reserve left. Waiting for production to complain, as the calorifier's story shows, means running months below duty with the valve pinned wide. Afterwards, take a protocol reading the same week: U back near the clean baseline is the receipt; U unmoved means the wrong side was brushed.
Check yourself
Five quick questions on this lesson. No marks, no records — the score is for you.