Inspection software built around the PHMSA records a pipeline contractor is audited on

Operator qualification is a per-person, per-task, per-date record

PHMSA does not ask whether your technicians are certified. It asks whether the individual who performed a covered task was qualified for that task on that date. 49 CFR 192.803 defines a covered task by four tests: it is performed on a pipeline facility, it is an operations or maintenance task, it is required by Part 192, and it affects the operation or integrity of the pipeline. ASME B31Q supplies the industry task list. ASNT or API certificates are separate evidence and do not substitute.

That distinction changes the data model. A certification tracker stores a certificate with an expiry. An operator qualification record stores a person, a covered task, an evaluation method (written exam, oral exam, on-the-job observation, simulation), an evaluator, a date, and a re-evaluation interval that the operator sets, not you. One technician working for three operators can hold three different qualification sets for the same physical task, and the software has to represent all three at once without collapsing them.

The retention rule runs two clocks at once. Records supporting a current qualification stay live while the individual performs the task. Records of prior qualification, and records for people who have stopped performing covered tasks, are held five years. Liquid pipeline contractors carry the identical duty under 49 CFR 195.507. Software that deletes a leaver's file on termination breaks the second clock, and the gap surfaces the moment an operator audits a job that person worked.

Girth weld traceability starts with the weld number

A pipeline weld record is not a report line. It is an object carrying the weld number, the spread and segment, the milepost or engineering station, the joint numbers either side, the welder or welding operator identity, the welding procedure specification used, the NDT method, the technician who interpreted it, the interpretation procedure and revision, the acceptance decision, and the repair or cut-out that followed a reject. Every one of those fields is a question an operator will eventually ask you to filter on.

49 CFR 192.243(f) makes the aggregate mandatory: by milepost, engineering station, or geographic feature, the number of girth welds made, the number nondestructively tested, the number rejected, and the disposition of the rejects, for the life of the pipeline. That aggregate is trustworthy only if it is computed from the per-weld records rather than typed. Typed totals drift the moment a repair weld is added late or a reject is re-shot and passed.

The traceability chain also runs forwards. When a weld fails in service years later, the operator works backwards from the weld number to the technician, the procedure revision, the instrument, and the qualification held that day. If any of those four links was recorded on a paper field sheet later scanned as a flat image, the chain is broken at that link. Offline field capture that writes structured fields on the right-of-way keeps it intact.

API 1104 acceptance is recorded per weld, not per report

49 CFR 192.241(c) sends weld acceptability to Section 9 or Appendix A of API Std 1104. Section 9 is workmanship-based and organised by imperfection type: inadequate penetration, incomplete fusion, internal concavity, burn-through, slag inclusions, porosity, cracks, undercutting. Each type carries a limit on individual indication length, a limit on aggregate length in any continuous 12-inch length of weld, and for several types a further limit at 8 percent of the weld length.

That structure decides the data model. A record storing only accept or reject cannot be re-adjudicated, because the decision depends on values nobody kept: indication type, its length, its position around the circumference, and what else was found in the same 12 inches. Cracks show why type matters. Under Section 9.3.10 a crack of any size or location is a defect unless it is a shallow crater or star crack no longer than 5/32 inch.

Section 9.2 adds a right of rejection: the company may reject a weld that appears to meet the acceptance standards if it judges the depth of an imperfection detrimental. That is a judgement, and judgements need an author and a rationale attached to the weld. Appendix A is the separate alternative acceptance criteria route requiring fracture mechanics analysis, and it is not interchangeable with Section 9. Record which route governed each weld, on the weld.

The NDT personnel record API 1104 Section 8.4 requires

API 1104 Section 8.4 imposes a personnel record separate from operator qualification. NDT personnel are certified to Level I, II or III per ASNT SNT-TC-1A, ACCP, or another recognised national certification program acceptable to the company, and only Level II or III personnel may interpret test results. Section 8.4.2 requires a record of certified NDT personnel holding the certification test results, the agency and person granting certification, and the date of certification.

The recertification clock is explicit: Level I and Level II at least every three years, Level III at least every five. That is a third set of dates alongside operator qualification re-evaluation and any client-specific requirement, running on a different cycle for the same person. Section 8.3 adds documentation of education, experience, training and qualification examination results for welding inspection personnel, retained by the company and acceptable to it.

Three clocks per technician is precisely what a recertification spreadsheet handles badly. Columns work until the interpretation restriction has to be enforced, because a Level I cannot interpret and scheduling must therefore know the level, not just that a certificate exists. A certification tracking system carrying level, method, issuing body, issue date, expiry and a next-action flag at 90, 60 and 30 days makes the audit answer and the dispatch decision fall out of one record.

Proving the NDT percentage by class location

49 CFR 192.243(d) scales the required percentage of field butt welds tested to the class location: at least 10 percent in Class 1 outside offshore areas, at least 15 percent in Class 2, and 100 percent in Class 3 and Class 4 locations, inside railroad or highway rights-of-way, at crossings of major waterways, and offshore, dropping to at least 90 percent only where 100 percent is impracticable. Tie-ins, including replacement section tie-ins, are 100 percent.

The proof is arithmetic across two dimensions the field crew does not think in. Welds made by each welder each day, against welds tested, sliced by the class location of the segment where each weld sits. If class location is not a field on the weld record, the percentage cannot be computed at all and has to be reconstructed from alignment sheets afterwards, under audit pressure, by someone who was never on the spread.

Building it in is straightforward: attach class location to the segment, attach the weld to the segment, and let the count fall out. The contractor then sees a shortfall on day three of a spread rather than on day ninety, when the pipe is buried. That is the difference between a compliance dashboard and a compliance report, one actionable and one a post-mortem. See audit preparation software for how that pattern generalises.

Integrity dig records and the anomaly-to-closure chain

Pipeline audit preparation is one of the most searched phrases in this segment, and integrity digs are where most audit findings live. The dig record starts before the excavation: the in-line inspection call, its dimensions, its location by odometer and girth weld reference, and the urgency the operator assigned, immediate, scheduled, or monitored, under their integrity management plan. All of that arrives from the operator and has to be stored, not re-typed.

In the ditch, the record captures what was actually found: coating condition, corrosion mapping or crack sizing by the NDE method used, the technician and equipment, remaining wall thickness, and the comparison between the ILI call and the field measurement. That comparison is not incidental. Operators use it to validate tool performance, and PHMSA looks at whether the operator did. A dig record with no ILI-versus-field field forces someone to build the comparison manually later.

Then closure: disposition against the operator's repair criteria under 49 CFR 195.452(h) or 192.933, the repair method applied, the post-repair verification, and the date the condition was corrected. The operator owns the decision; the contractor owns the evidence behind it. Software treating a dig as a document rather than a set of measured fields makes the operator re-key your findings into their integrity database, and re-keying breeds the discrepancies an auditor finds.

The audit pack an operator requests

Operators do not usually ask for everything. They sample. A typical request names a project, a date window, and a technician, then asks for the qualification record covering that person for the covered tasks performed, the NDT procedures at the revision in force, the calibration status of the instruments used, the per-weld interpretation records, and the summary counts required by 192.243(f). Five record types, one query, one deadline that is rarely generous.

Assembled by hand from folders, that request takes a coordinator days and surfaces inconsistencies: a procedure revision superseded mid-project, a calibration certificate that expired between two jobs, a technician whose re-evaluation date passed unnoticed. None of those are new failures. They are old failures that stayed invisible until someone assembled the pack, which is the worst moment to find them and the most expensive moment to explain them.

Assembled from a system where instrument, procedure, person and weld are linked records with effective dates, the same request is a filtered export. More importantly, the same query can be run on yourself, monthly, before the operator runs it. Compliance tracking that surfaces expiries and revision mismatches continuously turns audit preparation from an event into a background process nobody has to schedule.

Offline capture on a right-of-way with no signal

Pipeline construction happens where connectivity does not. Any software requiring a live connection to record a weld interpretation gets bypassed within a week, and the bypass is a paper sheet transcribed by someone who was not there. Transcription is the single largest source of the traceability breaks described above, because the transcriber fills gaps with assumption rather than observation, and the assumption inherits the authority of the record it lands in.

The requirement is that the device holds the full weld register for the assigned segment, accepts interpretations, photographs, radiographic reader sheets and signatures offline, and reconciles on reconnection without overwriting a concurrent edit. Atlantis supports offline field capture on this pattern, cloud or on-premise. Reconciliation behaviour is the part to test in a demo, because that is exactly where field data quietly disappears in most systems.

Test it with the ugly case: two technicians edit records for the same segment offline, one loses their device, and a third person amends a weld from the office in between. Ask to see what the system does. A vendor answering with conflict-resolution behaviour rather than a feature list is describing something built for the field. See the RFP requirements checklist for the rest of the questions worth asking.

Choosing inspection software for US pipeline work

Weight the evaluation towards the records regime rather than the report layout. A pipeline contractor's reports get read once; the underlying records get audited for decades. Ask whether the system can express a covered task separately from a certification, whether retention rules can differ by record type, whether class location is a first-class field, and whether an interpretation can carry indication dimensions rather than a bare verdict.

Then ask about the exit. Records you must hold for the life of the pipeline outlive most software contracts. A full structured export, not a PDF dump, of welds, people, qualifications, procedures and digs, on demand and without vendor assistance, is a hard requirement for this segment specifically. Vendors treating export as a professional-services engagement are asking you to carry a compliance risk on their behalf.

Atlantis builds this on Odoo, deployed cloud or on-premise, with a configurable record model rather than a fixed one. Configurability decides this segment, because no two operators define their covered task lists identically and a contractor working four operators carries four definitions at once. Related reading: the pipeline inspection services we deliver. Affordable, accessible, fully customizable. Request a demo and bring a real audit request.

Request a consultation