NDT Weld Inspection Records

Definition

Non-destructive testing (NDT) weld inspection records document the examination of welds without damaging them — visual inspection, dye penetrant, magnetic particle, ultrasonic and radiographic testing — and tie each result back to a specific weld, welder, procedure and date. A complete record set links the weld map, the welding procedure specification, the welder qualification, the NDT technician's certification, the test report and the disposition of any defect found.

Why It Matters

Structural steel and pressure pipework are accepted on evidence, not appearance. A weld that looks perfect can contain lack of fusion or porosity that halves its capacity, and the failure mode is usually sudden. Regulators, insurers and clients all require the traceable chain, and on plant projects the authority having jurisdiction will not sign a system into service without it. The record set is also what makes rectification tractable: when one welder's technique drifts, the records tell you exactly which other welds to re-examine instead of re-testing the whole structure.

How the Record Set Is Built

1. Weld map. Every joint receives a unique identifier tied to a drawing location before fabrication starts.
2. Procedure and qualification. The welding procedure specification, its supporting qualification record, and each welder's current certification for that process and position.
3. Inspection and test plan. Defines the NDT method, extent (for example 10% UT on fillet welds, 100% RT on butt welds in critical lines) and acceptance standard.
4. Execution. Certified technicians perform testing to a written procedure; reports reference weld ID, technique, sensitivity and calibration block.
5. Disposition. Defects are recorded, repaired to an approved repair procedure and re-tested, with the repair traced to the original weld ID.
6. Compilation. Records are indexed into the quality dossier that accompanies handover.

Real-World Example

On a substation steelwork package, routine 10% ultrasonic sampling found lack of side-wall fusion in two connections. Because the weld map recorded welder IDs, the team could isolate every joint completed by that welder in the preceding three weeks — sixty-eight welds — rather than re-testing the eleven hundred welds in the package. Expanded sampling on those sixty-eight found four more defects; all were cut out and re-welded within a week. Had the records been kept by area instead of by welder, the only defensible response would have been full re-inspection, which the fabricator estimated at six weeks and a six-figure sum. The paperwork discipline, not the testing, is what contained the cost.

Practical Lessons Learned

• Traceability to the welder is the difference between surgery and demolition. Record it at the time; it cannot be reconstructed later.
• Expired welder certifications invalidate the work, not just the file. Track renewal dates like plant certificates.
• Repairs need their own trail. A repaired weld that carries only the original report is an audit finding waiting to happen.
• NDT rates should respond to results. Fixed 10% forever ignores what the data is telling you; increase sampling when defects appear, and justify reductions with evidence.
• Radiography has a programme cost. Exclusion zones stop adjacent work; schedule it into night shifts rather than pretending it is free.

Expert Tips

• Issue the weld map before fabrication, not after. Retrospective numbering is where traceability quietly dies.
• Audit one complete record chain per month end to end — drawing, procedure, welder ticket, NDT report, repair — and fix the gaps while the people involved are still on site.
• Keep NDT technician certifications and equipment calibration certificates in the same index as the reports; auditors always ask for them second.
• Photograph the stamped weld identification in place. It resolves arguments about which joint the report describes.
• Agree acceptance criteria and the repair procedure with the client's inspector before production welding begins, in writing.

Common Mistakes

• Sampling welds by convenience — accessible joints get tested and awkward ones do not, which inverts the risk profile.
• Allowing welders to work outside the position or process their qualification covers.
• Storing NDT reports as loose PDFs with no index, then spending three weeks assembling the dossier at handover.
• Repairing a defect without an approved repair procedure and re-test.
• Leaving the inspection and test plan unagreed until testing has already started.

Key Takeaways

• Acceptance rests on the record chain: weld ID, procedure, welder, technician, result, disposition.
• Welder-level traceability contains the cost of any defect discovery.
• Sampling rates should adapt to results rather than remain nominal.
• Repairs require their own procedure, record and re-test.
• Build the dossier continuously; it is a handover deliverable, not an afterthought.

Related Concepts

Pairs with Non-Conformance Report, Commissioning Witness Testing, Handover Pack, and Snagging Register.