An exchanger bundle spends most of its life invisible. The turnaround is the one window where you can see it, measure it, and decide its future, and that window is measured in shifts, not weeks. Whatever you fail to capture before the bundle goes back in the shell is gone until the next outage.
That makes bundle inspection less a technical problem than a logistics and data problem. The techniques are well understood. What fails is the record.
The clock starts at the pull
A typical bundle sequence runs: pull, transport to the wash pad, clean, visual, eddy current or IRIS on the tubes, decision, plug or repair, reassemble, hydro, reinstall. Every step has a crew waiting on the previous one, and the tube examination sits squarely on the critical path.
The data problem shows up in three places:
- Identification. A wash pad with six bundles on it is a place where serial numbers get confused. A finding recorded against the wrong exchanger is worse than no finding at all.
- Tube maps. Eddy current results arrive as a vendor file keyed to a tube sheet map. If that map does not match the one your plugging crew uses, the wrong tubes get plugged. It happens more often than anyone admits in the closeout meeting.
- The decision record. Someone decides to plug forty tubes and run to the next turnaround. Five years later, the engineer planning that next turnaround needs to know what was found, what was plugged, and why the run decision was defensible. Too often what survives is a PDF and a rumor.
Set the record up before the pull
The teams that get clean data out of a bundle window do the record work before the crane moves:
- Pre-build the equipment records. Each exchanger, shell and bundle, exists in the system before the outage, with its tube count, layout, and prior plugging history loaded. Findings land against a known asset, not a free-text name.
- Agree the tube sheet convention with the NDE vendor in writing. Row and column origin, orientation, numbering direction. One page, signed, attached to the record. This single step prevents the wrong-tube plug.
- Define acceptance before results arrive. Wall loss thresholds, plug limits, and who makes the run-repair-retube call. When the results land at 02:00, the decision path is already drawn.
- Capture at the pad, not at the trailer. Visual findings, photos of the tube sheet, and plug counts recorded on the bundle's record while standing next to it. Transcription later is where identity errors breed.
Receiving the vendor data
Eddy current and IRIS results should come back as structured data, tube by tube, not only as a report PDF. When the results import against the same tube map the record already holds, the plugging list is generated, not retyped, and the percentage-plugged calculation that feeds the retube decision is arithmetic on the record rather than a spreadsheet built at midnight.
The bundle's history is the asset
A bundle's plugging history is one of the few genuinely cumulative datasets in a plant. Each outage adds a layer: tubes plugged, wall loss trend, cleaning method used. Read together across three turnarounds, that history tells you when the retube is coming and lets you budget for it two years early instead of discovering it mid-outage. Read separately, as three vendor PDFs in three folders, it tells you nothing.
That is the real argument for putting bundle data in a structured system: not this outage's efficiency, but the next one's foresight.
intelliNDT receives tube examination data against a consistent tube map, and intelliPROJECTS keeps the pull-to-reinstall sequence and its sign-offs on the exchanger's own record, so the history accumulates where the next planner will look for it. If your bundle records currently live in vendor PDFs, that continuity is the case for changing how the next window is run.
