NDT Procedure Development Consulting: Writing Procedures That Pass Audit

Every NDT procedure audit finding traces back to the same root cause: a procedure written once, filed, and never checked against the codes, essential variables, or acceptance criteria it was supposed to control. The technicians using it may be perfectly competent, the equipment may be properly calibrated, and the reports may look professional — and the procedure can still fail an audit, because the auditor is not evaluating whether the inspection worked. They are evaluating whether the procedure, as written, is legally and technically defensible against the code it claims to satisfy. That distinction is where most NDT procedure development goes wrong, and it is the difference between a procedure that passes and one that generates a finding.

Technique Sheets, Specific Procedures, and General Procedures Are Not the Same Document

ASME Boiler and Pressure Vessel Code, Section V governs the methods most NDT procedures reference — Article 4 for ultrasonic examination of welds, Article 5 for straight-beam UT, Article 6 for penetrant testing, Article 7 for magnetic particle testing, Article 9 for visual examination — and Section V draws a sharp line between a general procedure (covers a method across a range of applications), a specific procedure (tailored to a particular component, thickness range, and material), and a technique sheet or examination record (the field document showing the exact settings used on a specific job). A common finding is a company operating off a single general procedure for all UT work, with no specific procedure addendum for the actual thickness range, material, and configuration being examined, and no technique sheet on file proving the general procedure's parameters were actually demonstrated on that geometry. Auditors reviewing against ASME Section V, Article 1 essential variables (T-150 for UT) will ask for demonstration that a change in any essential variable — angle, frequency, couplant, scanning technique, reference reflector — triggers procedure requalification, and a company that cannot produce that demonstration record has, in the auditor's eyes, an unqualified procedure regardless of how good the actual welds are.

Where NDT Procedure Audits Find Problems

• No essential variable table. The procedure describes a method in prose but never isolates which parameters are essential (require requalification if changed) versus nonessential (can vary without requalification) per the governing Section V article.
• Qualification demonstration missing. For techniques like phased array or TOFD under Article 4, or automated UT under Article 5, the code requires demonstrated performance on a qualification block representative of the actual production geometry — and this demonstration record is frequently absent or generic.
• Acceptance criteria not tied to a code edition. A procedure references “ASME Section VIII acceptance criteria” without specifying the edition and addenda, which matters because flaw acceptance limits have changed across editions, particularly for Appendix 4 fracture-mechanics-based acceptance versus workmanship-based limits.
• Calibration block traceability gaps. The IIW block, ASME basic calibration block, or DSC block used to set sensitivity isn't traceable by serial number to a certificate, so the auditor cannot confirm the DAC or TCG curve behind a given weld inspection was built on a verified reflector.
• No revision control. Multiple versions of the same procedure circulate on different technicians' tablets or in different site folders, with no single controlled master and no record of which version was in effect on a given inspection date.
• Technician qualification mismatch. The procedure requires a Level II to perform and a Level III to interpret, but the personnel record shows the assigned technician's certification lapsed or was never issued at that level for that method.

A Worked Example: Shear-Wave UT on Girth Welds

Take a pipeline contractor performing shear-wave UT on girth welds per ASME Section V, Article 4, with acceptance to API 1104 (Welding of Pipelines and Related Facilities) workmanship standards rather than a fracture-mechanics-based alternative. A defensible specific procedure states the essential variables explicitly: nominal pipe wall thickness range (say, 0.250–0.500 inch), refracted angle(s) used — typically 45°, 60°, and 70° to cover the weld volume from multiple skew angles — nominal frequency (commonly 2.25 or 5 MHz depending on wall thickness and grain structure), couplant type, scanning technique (manual raster with defined overlap), and DAC curve construction referencing a specific calibration block with documented notch or side-drilled-hole dimensions. The technique sheet for a given weld then records the actual instrument settings, calibration verification readings taken at shift start and end, and any deviations noted. If API 1104's acceptance criteria are applied, the procedure must also state which edition's Appendix A (or the alternative fitness-for-service acceptance path) governs flaw sizing and disposition — because API 1104's workmanship limits for a given flaw length differ from what a fracture-mechanics-based evaluation would allow, and mixing the two without documenting which one applies is a defect discovered in almost every third-party procedure audit of pipeline UT programs.

PT and MT Procedures Have Their Own Parameter Traps

Penetrant and magnetic particle procedures generate fewer headline findings than UT phased array work, but they fail audits just as often, usually on measurable parameters that are easy to specify and easy to skip verifying in the field. For fluorescent penetrant inspection per ASME Section V Article 6, the procedure needs to specify black-light intensity at the examination surface — a minimum of 1,000 µW/cm² is the commonly cited threshold, verified with a calibrated UV radiometer at a stated frequency, not assumed from the bulb being new. Dwell time (penetrant and developer), ambient white-light level during fluorescent exam (must be below a specified lux/foot-candle threshold, again measured, not estimated), and developer type (dry powder, aqueous, or nonaqueous wet) are all essential variables that change sensitivity and therefore require documentation of what was actually used. For magnetic particle per Article 7, essential variables include the magnetization technique (yoke, prod, or coil), current type (AC versus half-wave DC, which behave very differently for surface versus subsurface indications), and field strength verification using a pie gauge or Hall-effect meter. A procedure that says “use a yoke per manufacturer instructions” without a stated lift-test verification (typically 10 lb for AC yokes at maximum pole spacing) will not survive a careful audit, even though the underlying inspection quality may have been fine.

Radiographic Procedures Carry Their Own Essential Variable Set

Radiographic testing under ASME Section V, Article 2 gets less attention in procedure audits than UT phased array, largely because RT feels more standardized — and that assumption is exactly what lets gaps slip through. Film radiography, computed radiography (CR), and digital detector arrays (DR) each carry a distinct set of essential variables under Article 2 and its mandatory appendices, and a procedure written for film does not automatically cover a shop that has since switched to DR, even though the weld being examined hasn't changed. For film, the procedure needs to specify source type (Ir-192, Se-75, or X-ray) and source size, source-to-film distance and the resulting geometric unsharpness calculation, film system class, and the image quality indicator (IQI) type and required sensitivity — commonly 2% for single-wall exposures on the thickness range involved, verified by the smallest IQI wire or hole visible on the actual film density range of 1.8 to 4.0 for single-film viewing. Digital systems replace film density with signal-to-noise ratio (SNR) and basic spatial resolution requirements per the applicable digital imaging appendix, and a procedure that simply substitutes “DR” into a film-based document without addressing SNR, normalized SNR at the correct exposure, and detector calibration frequency has not actually qualified the digital technique — it has just relabeled an old document. This is one of the more common findings on shops that upgraded equipment faster than they updated the paperwork behind it.

Client Specifications Add Requirements the Code Doesn't

ASME Section V and the referencing construction code set the floor, not the ceiling. Major owner-operators layer their own engineering specifications on top — a refiner's internal inspection specification might require a tighter IQI sensitivity than Article 2's minimum, a lower acceptable indication size than the construction code's workmanship limits, or a specific brand and model of digital detector array with a documented calibration history. A procedure that satisfies ASME Section V but ignores the client's overlay specification will still generate a finding, because the contract, not just the code, defines the acceptance bar for that job. This is compounded by third-party vendor prequalification platforms — ISNetworld, Avetta, and similar systems used by major industrial owners to screen contractors before award — which increasingly ask contractors to upload procedure documentation directly for review before a bid is even considered, meaning a procedure gap can cost a contract before an inspector ever reaches the jobsite. Building a procedure library that tracks which client specifications layer onto which base code, and keeping that mapping current as clients revise their own specs, is tedious work that pays for itself the first time a bid deadline lands the same week a client specification changes.

Auditors Look at the Whole System, Not Just the Procedure Text

A procedure that reads perfectly on paper still fails if the records behind it don't match — calibration due dates that lapsed mid-project, technician certifications that expired between the procedure's effective date and the actual inspection, or report templates that omit a field the procedure requires be recorded. This is where procedure governance intersects directly with the systems a company runs day to day. Tracking calibration intervals, certification expirations, and procedure revision status in a spreadsheet works until the company scales past a handful of technicians, at which point Atlantis NDT ERP earns its keep by flagging a calibration due date or an expiring certification before it becomes an audit finding rather than after. On the output side, NDT reporting software that enforces the procedure's required fields — essential variable values, calibration verification readings, acceptance criteria edition — at the point of report generation catches the gap where a technician skips a required field, rather than an auditor catching it months later. For asset owners layering inspection history onto a digital twin platform, the procedure and technique-sheet data become part of the permanent asset record, queryable against the specific weld or CML location years after the original inspection date.

Where Consulting Fits: Writing, Qualifying, and Mock-Auditing Procedures

An ASNT Level III consulting engagement on procedure development typically works through three phases. First, a gap review of existing procedures against the specific codes and client specifications they claim to satisfy — checking essential variable tables, acceptance criteria edition references, and qualification demonstration records against ASME Section V, API 1104, AWS D1.1, or whichever code governs the client's work. Second, drafting or revising the procedures and technique sheets themselves, including building the qualification demonstration package for any technique requiring it. Third, and often the most valuable phase, running a mock audit — pulling a sample of completed inspection records and technique sheets and checking them against the procedure exactly the way a third-party auditor or client quality team would, before the real audit happens. Companies that go through this exercise consistently find issues in their own records they didn't know existed; finding them internally, with time to correct the underlying process, is a materially different position than finding them during a client's vendor-qualification audit with a contract on the line.

Building Long-Term Procedure Governance

A procedure library that passes one audit and then drifts out of compliance over the following two years has not solved the underlying problem. Durable governance requires a revision cadence tied to code edition updates — ASME Section V and the referencing construction codes update on a multi-year cycle, and a procedure that still cites a superseded edition after the client's contract requires the current one is a finding waiting to happen — plus a cross-reference matrix mapping every active procedure to the codes and client specifications it satisfies, so a code update triggers a known, bounded review rather than a company-wide scramble to figure out which procedures are affected. Assigning a named Level III as the technical owner of the procedure library, with a defined review interval independent of any specific audit or client request, is the single change that most reliably keeps a procedure set audit-ready year-round instead of only in the weeks before a scheduled review.

Record retention is the part of governance most often left undefined until it becomes a problem. Contracts and referencing codes frequently require radiographs, UT technique sheets, and final reports be retained for the life of the component, not just a fixed number of years — which for pressure vessel or pipeline work can mean decades. A procedure library should state the retention requirement explicitly for each record type and specify where and how records are archived, including a plan for what happens to physical film archives as a company transitions to digital radiography, and how digital records are protected against format obsolescence over a multi-decade retention window. Companies that discover a retention gap usually discover it the hard way — during a failure investigation or a fitness-for-service evaluation years after the original inspection, when the original technique sheet or radiograph that would resolve the question cannot be located.