Records, Quality Assurance and Level 3 Decision Cases

Learning outcomes

By the end of this chapter, the learner should be able to define records needed to reconstruct an examination; distinguish raw records, reports and technical dispositions; plan audits, data retention and equipment traceability; resolve practical cases using bounded evidence; and communicate limitations clearly.

16.1 Record versus report

The examination record contains the detailed evidence created during the work: component identity, drawing and area, governing documents, personnel, date and location, equipment serial numbers, probe and wedge identifiers, calibration results, surface and temperature conditions, scan coverage, indications, calculations, raw data, limitations and any departures.

The report communicates the authorised result to the client or decision-maker. It may summarise the record, but it must retain enough traceability to locate the supporting evidence. A report that says only "UT acceptable" is a conclusion without a visible foundation.

16.2 Data integrity and retention

Electronic data should be protected by controlled naming, access, backup, checksum or other integrity controls where appropriate. The organisation should define native file retention, exported-view status, software dependencies, migration and disposal. Screenshots alone are rarely sufficient for encoded or advanced data.

Paper and electronic corrections must preserve the original entry, author, date and reason. Audit trails should show who changed an analysis and why.

16.3 Quality assurance system

A Level 3 quality program may include equipment registers, calibration status, probe performance trends, procedure qualification files, personnel authorisation matrices, training and eyesight records, controlled reference standards, internal audits, technical surveillance, nonconformance management and periodic review of false calls or missed indications.

Quality indicators should support learning rather than encourage concealment. Tracking only "number of rejects" says little about inspection quality. More useful measures include calibration failures, rescans, data gaps, report corrections, blind-trial performance and repeated interpretation disagreements.

16.4 Decision case A - failed end check

An examiner completes five welds and finds the reference response 3 dB below the permitted tolerance at the end check. The correct Level 3 response is to identify the last valid check, quarantine subsequent data, investigate equipment and coupling, restore the setup, document the nonconformance and rescan the affected scope unless an authorised evidence-based disposition proves validity. Averaging the beginning and ending values does not retrospectively calibrate the afternoon.

16.5 Decision case B - incomplete casting penetration

The far side of a coarse-grained casting cannot be resolved and the structural noise is close to the recording level. The Level 3 should not issue a full-volume clearance. Options include lower frequency, alternate mode, improved surface, another access, representative demonstration or a complementary method. Any residual unexamined volume must be reported.

16.6 Decision case C - conflicting PAUT and TOFD results

PAUT shows a strong planar response; TOFD shows no confirmed tip pair. Possible explanations include orientation, tip shadowing, dead-zone position, insufficient TOFD sensitivity, geometry or a non-crack planar reflector. Review raw data and setups, verify coordinates, use supplementary scans or another method, and apply the predetermined reconciliation rule. Absence in one technique does not automatically cancel presence in another.

16.7 Decision case D - outdated reference

A procedure cites an edition that is no longer the contractually nominated revision. Stop controlled issue, compare technical changes, assess qualification impact, revise references and obtain approval. Silently changing the cover-page date is not standards management; it is cosmetic archaeology.

Final Level 3 checklist

Before approving or directing an ultrasonic application, confirm:

• authority, scope and acceptance source are identified;
• the discontinuity mechanisms and inspection volume are defined;
• material and geometry assumptions are supported;
• equipment, calibration and essential variables are controlled;
• coverage and capability are demonstrated;
• sizing and interpretation limits are known;
• personnel are competent and authorised;
• records can reconstruct the work;
• unresolved limitations are visible to the decision-maker;
• the document has received formal technical and document-control approval.

Failure modes and decision control

Weak practice commonly includes retaining only the client report and deleting raw evidence; allowing editable exports to replace native data; closing nonconformances without affected-scope review; using quality metrics that reward under-reporting; reconciling conflicting methods by majority vote; and updating reference dates without technical review.