Furnace Tube Burst Eight Months After a Burner Retrofit
A 40,000 lb/hr watertube boiler (150 psig design) tripped on low water and was found to have a burst generating tube in the furnace wall directly opposite the burner. The removed tube shows a fish-mouth opening about 5 in. long; the wall thins from 0.165 in. at the original thickness to roughly 0.045 in. at the ruptured lip, and the fracture edges are thin and ductile with bright, clean metal and almost no steam-side scale. Neighboring tubes in the same panel show no bulging and no measurable scale. Eight months ago the burner was replaced; since then the flame has been visibly longer and brighter, and full-load stack temperature has risen about 60 F. Water treatment logs show two brief hardness excursions that were corrected, and the low-water cutoff carries a current test tag.
1. Classify the failure mechanism and explain why the water treatment history should not be your first suspect.
Read the worked answer
Classification: short-term, high-temperature overheating.
The physical evidence points one direction:
- A wide, open fish-mouth rupture with a thin, ductile lip is the classic short-term signature - the metal was heated past its strength limit quickly, then tore open while still ductile.
- Long-term (gradual) overheating looks different: a thick-lipped, narrow, often brittle longitudinal crack, heavy steam-side oxide, and obvious internal scaling. None of that is present.
- Almost no steam-side scale rules out a deposition-driven mechanism as the primary cause. Scale insulates and drives long-term overheating; this tube was clean.
The hardness excursions are real but they were brief and corrected, and they left no deposit behind to testify against them. The timeline is a much stronger clue: the failure appeared within eight months of a burner retrofit, and the observable flame got longer and brighter at the same time. A longer, more intense flame in the furnace panel opposite the burner raises local heat flux, and once the local heat flux exceeds what the tube can shed by nucleate boiling, a stable vapor film forms (departure from nucleate boiling) and that patch of wall climbs in temperature very fast. Flame impingement on the wall does the same thing locally.
Conclusion: localized overheating driven by the new flame pattern / heat flux, not by water chemistry. Confirm by checking whether the tube rows involved sit in the burner's direct radiation path and whether circulation in that circuit is restricted.
2. The tube is replaced and the boiler is ready for a hydrostatic test. What do you inspect and verify before it goes back on line, and what do you ask the burner vendor?
Read the worked answer
Replacing the tube fixes the symptom, not the cause. Work through the cause systematically:
Fireside / burner
- Compare the actual flame envelope to the vendor's commissioning drawings: length, width, impingement on walls, quarl/tile condition, register and swirler settings.
- Ask the vendor for the commissioned combustion data at high, mid, and low fire, and whether the retrofit changed the maximum firing rate or the required furnace volume.
- Confirm the retrofit was commissioned with a flue gas analyzer, not by eye, and obtain the curve.
Waterside / circulation
- Inspect the entire affected panel plus adjacent tubes for bulging, discoloration, and oxide; check tube IDs for deposits or obstructions that could have starved that circuit.
- Review the boiler water level records and the low-water cutoff history for the days before the trip - an unnoticed low-water episode can produce exactly this appearance too.
Records / limits
- Verify the pressure gauge accuracy and the high-pressure limit setting.
- Confirm the repair (tube replacement, welding, NDE) was done by qualified personnel and the hydrostatic test is witnessed and logged per the jurisdiction's rules before return to service.
3. The plant manager says: "Cut out the burst tube, plug the ends, and fire up - we need steam by 6 a.m." How do you respond?
Read the worked answer
A plugged tube is a legitimate emergency measure in many codes, but it is not a fix, and it is not free of limits:
- Code limits apply. Jurisdictions cap how many tubes may be plugged (commonly a percentage of the tubes in a given section or of total heating surface) and require the plugs to be properly installed and recorded. Exceeding those limits changes the pressure-part design and needs engineering and jurisdictional approval. Verify the actual limit in your state's rules before you sign anything.
- Capacity drops. Each plugged tube takes away heating surface. If you are already capacity-limited, the plug makes the process problem worse, not better.
- The root cause survives. The flame is still long and still hitting that wall. If you restart without re-commissioning the burner, expect the neighboring tube to fail next - and the second failure may not be caught by a low-water trip in time.
Practical position: I would not refuse a properly executed, code-compliant plug as a short-term measure, but I would insist on (1) written record of the plug and the number of plugged tubes, (2) reduced firing rate, (3) combustion re-commissioning and flame verification before light-off, and (4) a scheduled permanent repair. If the limits are exceeded or the firing rate cannot be reduced, the boiler stays down. Put the recommendation and the manager's direction in the log - that record protects both of you.