Steam is distilled water. It should be the kindest fluid a pipe ever carries — and a steam system still eats itself, reliably, in the same half-dozen places. The attackers are never the steam: they are the two gases that ride with it, and the velocities it reaches when it flashes. This closing Clinic lesson is pipe forensics — what CO₂, oxygen and two-phase flow each do to metal, how to read a failed component like a case file, and which upstream lessons hold each cure. Because every one of these failures is a chemistry or sizing decision wearing through, years later, as a hole.
Carbonic acid: the condensate line's slow saw
Lesson B7 put alkalinity in the feedwater story; here is where it ends up. The bicarbonates in make-up water break down in the boiler and release carbon dioxide as a gas — which leaves with the steam. Rough arithmetic for the practice plant: at ~300 ppm (as CaCO₃) alkalinity, each tonne of make-up releases of the order of 0.2 kg of CO₂ (breakdown is 50–90% complete; the chemistry is approximate, the direction is not). The CO₂ travels harmlessly with the dry steam — until the steam condenses. CO₂ dissolves eagerly in fresh condensate and makes carbonic acid, dropping the purest water on site to pH 5–6. That mild acid then flows, for years, along the bottom of every condensate line.
Its signature is unmistakable: grooving — a smooth channel dissolved along the bottom of the pipe where the acid ran, and thread-first failures at screwed joints, because threads are the thinnest metal in the system (the fitter's rule: when a return line weeps at the threads, the whole run is on notice). The cures are all upstream: reduce what breaks down — dealkalisation or degassing in the water plant (lesson B7); neutralise what remains — amine dosing carried with the steam, with the food-plant caveat lesson D12 explains; or outlast it — stainless returns, the standard answer on new clean-duty systems. Downstream palliatives (thicker schedule, welded joints) buy years, not absolution.
Oxygen: the pitting sniper
Where carbonic acid saws broadly, oxygen drills. Air enters at every shutdown (lesson C8's refilled system), and survives poor deaeration (lesson B5's whole purpose). Dissolved O₂ in hot water attacks iron at flaws in the magnetite film, and the attack concentrates: isolated deep pits under crusty tubercles, each a tiny corrosion cell drilling toward a pinhole leak — in feed lines, economiser tubes, receiver shells and the top of flooded returns. A grooved line dies over a decade; a pitted one can pinhole in a couple of monsoons.
Cures, again upstream: a deaerator doing its measured job (lesson B5's 0.007 ppm class of performance, polished by the scavenger dosing of lesson B7); shutdown discipline — either drain-and-dry completely or lay up wet and treated, never the damp middle; and honest gasketing on systems that breathe (lesson D11's vacuum plants live or die by this).
Velocity: erosion and wiredrawing
Lesson C11 showed that condensate crossing a trap becomes a 99%-by-volume steam flow. Aim that two-phase jet at metal and it machines it. The first elbow after a trap is the classic victim — droplets carried in the flash cannot turn the corner and hammer the outer radius thin (survey crews call it the sacrificial elbow). Cures: give the discharge its C11-sized line so velocity stays civil, make the first bend long-radius (or a blanked target tee, which sacrifices a replaceable plug instead of pipe wall), and put schedule thickness where geometry cannot improve.
Wiredrawing is the same weapon against valve seats. A globe valve left 10% open to throttle flashing condensate — the stalled equipment's folk remedy (lesson G3) — passes a needle jet of two-phase flow across its seating faces, and within months the seat carries fine parallel score lines, as if drawn with wire. The valve will never shut tight again. The rule it teaches: on condensate, valves are for isolation — open or shut; only the trap is allowed to throttle. Wet-steam impingement after inadequate separation (lesson A9, C10) does the same to PRV internals and meter runs — the separator ahead of them is armour, not garnish.
FAC — the one you escalate
For completeness and humility: flow-accelerated corrosion is a distinct disease of hot feed and economiser circuits (roughly 130–180 °C, single phase, low-oxygen chemistry, turbulent geometry) in which the protective magnetite layer dissolves faster than it reforms, thinning pipe from the inside at rates that produce sudden large ruptures — it has killed people at power stations (the Mihama accident of 2004 is the industry's standing memorial). It is chemistry-dependent in ways that surprise non-specialists (ultra-low oxygen can make it worse). The Clinic's guidance is deliberately short: know it exists, keep wall-thickness surveys on hot feedwater and economiser circuits in the IBR inspection scope (lesson B10), and treat it as boiler-inspector territory, not a plant-fitter fix.
Reading the corpse
| What you find | Who did it | The cure lives in |
|---|---|---|
| Smooth groove along the pipe bottom; screwed joints weep first | carbonic acid (CO₂) | B7 water chemistry · D12 dosing limits · stainless returns |
| Deep isolated pits under rusty mounds | dissolved oxygen | B5 deaeration · shutdown lay-up discipline |
| First elbow after a trap thinned on its outer radius | two-phase erosion | C11 line sizing · long-radius / target tee |
| Fine parallel scores across a valve seat that won't shut | wiredrawing | G3 — pump instead of throttle; valves open/shut only |
| General inside thinning of hot feed / economiser pipe, orange-peel surface | FAC | specialist survey — escalate (B10) |
- Next shutdown, cut open one retired stretch of condensate return and read it against the table above — ten minutes of forensics that sets your whole water-side agenda.
- Ultrasonic thickness-gauge the first elbow after your five biggest trap discharges yearly; trend, don't spot-check.
- Walk the throttled valves on condensate duty (there are always some — G3 explains why); schedule each one's proper cure before its seat is scrap.
- Ask the lab for condensate pH at the far ends of the return system: 5-and-a-bit is the acid at work now, whatever the boiler-house logbook says.
- Confirm feed/economiser wall-thickness checks are in the IBR survey scope — FAC is found by programme, never by luck.
- Steam doesn't corrode pipes; its passengers do — CO₂ from alkalinity, O₂ from shutdowns and weak deaeration.
- Bottom groove = acid · pits under mounds = oxygen · thin elbow = erosion · scored seat = wiredrawing. Position is diagnosis.
- ~0.2 kg CO₂ per tonne of make-up at practice-plant alkalinity — every kilogram ends its journey as acid in a return line.
- On condensate, only the trap throttles. A valve asked to throttle flashing water is a consumable.
- FAC is real, deadly and specialist — survey hot feed circuits and escalate; don't improvise.
Condensate grooving was mapped by plumbers before chemists. The old trade name for a return line's bottom-channel wear was the "smile" — cut a failed nipple lengthwise and the groove grins up at you — and experienced fitters would specify extra-heavy pipe for returns while offering no better reason than "condensate is hungry". The chemistry — bicarbonate to CO₂ in the boiler, CO₂ to carbonic acid in the line — was assembled decades later by the water-treatment industry, which then sold the cure by the drum as neutralising amine. The fitters' surcharge and the chemists' dosing pump are the same insight, purchased in different centuries: the purest water on site comes back with an appetite.