After the fault is corrected, tube-end leakage should stop without migrating to adjacent passes, treated-water deposits should not reappear on the fireside, and the boiler should complete monitored heat-up and cooldown cycles without the metallic banging returning. Leakage after only 833 operating hours is an early-life failure condition, not a routine first-maintenance event.
What is the operator actually seeing?
A York-Shipley four-pass dryback boiler developed a metallic banging sound near the burner front. Opening the boiler exposed phosphate deposits across fourth-pass tubes at the front and on the opposite side at the rear. Those deposits turn an ambiguous noise complaint into a pressure-boundary investigation: treated boiler water escaped to the fireside and evaporated, leaving treatment chemicals behind.
The pattern is more significant than one lightly weeping joint. The boiler contains 264 tubes, including 80 fourth-pass tubes. Hand re-rolling did not seal every leak. Mechanical re-rolling then expanded the work into almost 200 tube ends across the front and rear of the fourth pass, after which leakage shifted into the third pass. That migration indicates a system-level joint, geometry, or movement problem rather than one isolated loose tube.
| Operator observation | What it indicates | Next check |
|---|---|---|
| Metallic banging near the burner front | Possible movement of a pass division plate, tubes, tube sheet, casing, or restrained boiler structure | Inspect cold and during controlled warm-up; correlate each sound with burner state and temperature change |
| Phosphate on fireside tubes | Treated boiler water has crossed the pressure boundary and evaporated | Locate active tube-end or tube-wall leakage and document the deposit pattern before cleaning |
| Deposits at both ends of the fourth pass | The affected area is distributed rather than confined to one accessible joint | Map every wet joint, stained joint, and repaired joint on both tube sheets |
| Leaks move to the third pass after fourth-pass re-rolling | Repair work is redistributing stress or exposing marginal adjacent joints | Stop production-style re-rolling and investigate joint dimensions, tube-sheet condition, and thermal movement |
| Similar banging on two companion boilers | A shared design, installation, fabrication, or operating factor may exist | Inspect the other boilers using the same map and test sequence before leakage becomes extensive |
Which failure paths fit the pattern?
Four paths deserve comparison: marginal original tube expansion, cyclic differential expansion, restricted structural growth, and low-temperature fireside condensation. Water chemistry remains part of the review, but it does not explain every tube-to-tube-sheet failure mechanism.
| Failure path | Mechanism | Evidence to collect | Fit for this case |
|---|---|---|---|
| Original joint or fabrication problem | Insufficient, excessive, or inconsistent tube expansion leaves joints unable to tolerate normal pressure and temperature movement. Hole finish, tube-end condition, alignment, or tube-sheet distortion can affect many joints. | Fabrication and test records, joint dimensions, tube-end wall condition, hole condition, pass-by-pass leak map, and comparison with companion boilers | Strong candidate because leakage is extensive, early, and migrates during re-rolling |
| Short cycling or thermal shock | Repeated expansion and contraction can relax marginal rolled joints. Thermal shock adds a rapid temperature gradient rather than merely a high cycle count. | Burner starts, run duration, water temperatures, pressure trends, feedwater events, and timing of banging | Possible contributor; operating several times daily while warm does not quantify cycle severity or temperature gradients |
| Restricted thermal growth | A skid, footing, or attachment that should permit movement can transmit growth forces into the shell and tube sheets. | Manufacturer mounting drawing, fixed and sliding support locations, rear attachment condition, alignment, and cold-to-hot movement observations | Must be checked because tightened rear skid bolts were identified as a possible restraint mechanism |
| Fireside condensation | Gas cooled below its dew point forms condensate, normally most visible near the outlet of the last pass. Persistent condensate can attack fireside metal but is not by itself proof of a rolled-joint defect. | Measured outlet gas temperature, signs of wetness or corrosion, firing state, load, and return-water conditions | Relevant to the fourth or last pass, especially in cool conditions, but it does not directly explain why re-rolling moved leaks to another pass |
| Water-side chemistry problem | Scale, corrosion, or deposition can overheat or thin tubes and damage joints. | Independent water analyses, trends, blowdown records, deposit analysis, and internal inspection | Less persuasive here because staff and two independent treatment companies reported matching acceptable results, but internal inspection still closes the diagnostic loop |
Why is repeated re-rolling the wrong first response?
Re-rolling is appropriate when inspection identifies an under-expanded joint and the tube and tube-sheet hole remain suitable for additional expansion. It is not a root-cause test. Rolling every leaking joint until the pressure test passes can mask the distribution that identifies the cause.
Each additional rolling operation changes the joint. Excessive expansion can thin and work-harden the tube end, disturb the ligament between holes, alter local tube-sheet geometry, or transfer load into neighboring joints. A leak that shifts from the fourth pass to the third after nearly 200 ends have been worked is a stop condition. Continuing the same action risks turning marginal original joints into damaged repair joints.
| Approach | When it works | Main limitation |
|---|---|---|
| Re-roll only confirmed loose joints | A small number of joints are under-expanded and dimensional inspection shows adequate material for correction | Does not resolve a distributed fabrication, alignment, or restraint problem |
| Re-roll an entire pass | Only after engineering review identifies a consistently under-expanded population and defines a controlled repair method | Can redistribute stress and erase evidence of the original defect |
| Pause repairs and perform a mapped investigation | Leakage is widespread, early, bilateral, or migrates after repair | Requires outage time and coordination with the manufacturer and qualified boiler repair personnel |
The mapped investigation is the recommended approach. Preserve the leak pattern, determine why the joints lost tightness, and then select re-expansion, tube replacement, tube-sheet correction, mounting correction, or an operating change from measured findings.
What should be checked before another tube is rolled?
- Freeze the current condition. Stop further rolling, clean only enough area to identify leakage, and photograph both tube sheets. Number or map every active leak, dried phosphate trail, previously rolled end, and unaffected comparison joint.
- Confirm the leak path. Use the pressure or leak test specified by the manufacturer, repair organization, and governing jurisdiction. Distinguish leakage at the rolled interface from a tube-wall crack, tube-sheet crack, stay, gasket, or another pressure-boundary path.
- Inspect representative joints. Examine leaking, repaired, and apparently sound joints from the third and fourth passes. Measure the features required by the approved repair procedure rather than judging tightness solely from roller feel.
- Review fabrication records. Compare original rolling and factory test documentation with the physical joint measurements. Look for pass-to-pass or front-to-rear patterns rather than treating each leak independently.
- Check thermal-growth provisions. Compare every skid and footing attachment with the manufacturer drawing. Verify which points are intended to remain fixed and which must accommodate growth. Do not loosen rear bolts merely because they appear tight; obtain the manufacturer’s mounting requirement first.
- Review operating trends. Extract burner starts, run times, pressure, water-temperature changes, feedwater events, firing position, and the time of each banging event. Short cycling is a count-and-duration problem; thermal shock is a rate-of-temperature-change problem.
- Inspect the gas outlet region. Record wetness, corrosion, deposits, and measured gas temperature at relevant firing states. Compare measured temperature with the applicable fuel and flue-gas dew-point basis supplied by the boiler or burner documentation.
- Inspect the companion boilers. The other two units also produced banging. Open and map them using the same method so common pass, end, mounting, and fabrication patterns remain comparable.
Can firing changes solve both cycling and condensation?
Longer burner runs can reduce the number of expansion cycles, but a minimum firing position is not automatically the correct answer. Lower firing can extend run time while also lowering outlet gas temperature. The controlling measurements are cycle duration and stack temperature, not the firing command alone.
| Setting or condition | Location to verify | Effect to evaluate |
|---|---|---|
| Minimum firing position | Burner control and combustion setup | May lengthen runs; may also lower gas temperature |
| Burner start frequency | Operating log or controller history | High frequency increases thermal-cycle count |
| Water-temperature change | Boiler and feedwater trends | Rapid change identifies thermal shock more directly than start count |
| Outlet gas temperature | Last-pass or stack measurement point | Shows whether the operating state approaches the applicable dew point |
| Standby temperature | Boiler water indication and logs | Shows whether a nominally warm boiler actually experiences large temperature swings |
Choose a firing strategy only after measuring both sides of the tradeoff. If a longer low-fire run keeps the measured outlet gas temperature above the applicable dew point and reduces starts, it addresses both concerns. If outlet temperature falls into the condensing range, correct load matching, sequencing, or another documented operating constraint instead of holding low fire indefinitely.
How should the final repair be selected?
Select the repair from the failure boundary. A few dimensionally acceptable, under-expanded joints may justify controlled re-rolling. Tube ends that have been repeatedly expanded, thinned, cracked, or work-hardened may require tube replacement rather than another roller pass. Damaged holes, distorted tube sheets, or abnormal pass alignment require an engineered correction before joint work continues.
If the same location or pattern appears in all three boilers, treat it as a common-cause investigation. Compare manufacturing dates, construction details, tube and tube-sheet materials, pass geometry, rolling records, mounting arrangements, and commissioning operation. Similar banging alone does not prove identical leakage, so inspect before assigning one remedy to all units.
Keep water treatment records in the repair package. Matching readings from plant staff and two independent water-treatment companies provide useful separation between water-side control and mechanical joint behavior. Retain samples of unusual deposits when their identity or origin affects the repair decision.
How is the repair verified before normal service?
- Repeat the approved pressure or leak test and record every joint examined, repaired, or replaced.
- Perform a controlled cold-to-hot startup while observing the front and rear tube-sheet areas, skid attachments, and the location associated with the banging.
- Record burner state, run duration, water-temperature change, pressure, and outlet gas temperature so the mechanical observations can be aligned with operating conditions.
- Complete monitored firing and cooldown cycles representative of service. Check for renewed moisture, phosphate deposits, joint weeping, abnormal movement, and metallic impact sounds after each cycle.
- Reinspect the third and fourth passes after the monitored cycles. Acceptance requires dry joints, no new treated-water deposits, no migration of leakage into adjacent passes, and no recurrence of the unexplained banging.
Frequently Asked Questions
Can I expect tube leaks during the first two years?
No routine timeframe makes early leakage acceptable. Extensive leakage after 833 operating hours, especially across many tube ends and more than one pass, calls for a fabrication, joint, mounting, and operating investigation.
Does correct boiler-water chemistry rule out every water-side cause?
No. Matching results from staff and two independent treatment companies reduce the likelihood of a chemistry-control problem, but internal inspection and trend review must still check for localized scale, corrosion, or overheating.
Can I stop the leaks by re-rolling every tube end?
Only re-roll joints that inspection identifies as suitable for controlled additional expansion. When almost 200 ends have been worked and leakage shifts from the fourth pass to the third, stop and determine whether joint dimensions, tube-sheet condition, alignment, or restrained growth is driving the failure.
Does a passed pressure test mean the boiler is ready for service?
Not by itself. Complete representative monitored heat-up and cooldown cycles, then perform the final verification step: reopen or reinspect the third and fourth passes for dry joints, no fresh phosphate deposits, no migrated leakage, and no returning metallic bang.