Non-Destructive Testing and Inspection Insurance

Non-destructive testing produces data, and then a person decides what it means. The second part is where the claims live.

An ultrasonic trace, a radiograph, a magnetic particle indication or a penetrant bleed shows something. Whether that something is porosity within tolerance, a slag inclusion, a lack of fusion or a crack, and whether it is acceptable against the applicable standard, is an interpretation made by a certified technician under time pressure, frequently on site, sometimes in poor conditions. Pass something that should have been rejected and the weld goes into service in a vessel, a structure, a pipeline or a lifting component, and it fails later under load.

So the records are the defence and they need to be substantive. The procedure used and its sensitivity, the equipment and its calibration status, the technician and their certification level for that method, the areas examined and the areas not examined with the reason, the indications found with their location and dimensions, and the acceptance standard applied. A report that records a pass with none of that is impossible to defend because nothing in it shows what was looked at or how. Retaining radiographs and traces rather than discarding them after a job is part of the same point.

In inspection work, silence is read as coverage, and that is how an inspection business ends up answering for something it never looked at.

Access on site is frequently incomplete: a weld behind insulation, a joint inside a vessel, a connection obscured by steelwork, a section under water or buried, or a run that could not be reached in the time the possession allowed. If a report lists the joints examined and says nothing about the rest, the recipient reasonably assumes the scope was complete. When a failure occurs at an unexamined joint, the inspection business is in the frame for a weld it never saw.

So the scope has to be explicit in both directions. State what was examined and, as a separate statement, what was not and why: not accessible, insulation not removed, outside the agreed scope, or time not available. Quantify it where possible, so a report says which percentage of welds were examined rather than implying all of them. And where a client restricts the scope for cost or programme reasons, record that it was their decision. That single paragraph is the difference between an incomplete inspection and a negligent one.

Most professional work carries no physical hazard to third parties. Site radiography is an exception and it needs managing as one.

Industrial radiography uses sealed sources or x-ray sets that produce radiation at levels requiring a controlled area, barriers, warning signals and exclusion of everybody who is not involved. On a live industrial site, at night, with other contractors around, the realistic failures are somebody walking into a controlled area, a barrier that was moved, a source not returned fully to its container, a source left unattended, and in the worst case a lost source. Each of those is a regulatory event with its own reporting obligations, and some are potential injuries with consequences that appear much later.

So the controls are the regulated ones and an insurer will expect them described. A radiation protection adviser and supervisor arrangement, controlled area set up with physical barriers and signals rather than cones, dose monitoring for technicians, source accounting with checks before and after every exposure, secure storage and transport of sources, and a written procedure for a source that does not return. Working at night to reduce the number of people on site is standard in this work for exactly this reason, which brings fatigue in as a separate factor.

NDT is unusual among inspection disciplines because two properly certified people can look at the same data and reach different conclusions, and the business carries that.

A laboratory measurement is a number: another competent technician running the same test gets the same answer. An ultrasonic examination is not. Reading a trace, distinguishing a geometric echo from a genuine reflector, judging the significance of a small indication in a thick section, and deciding whether a radiograph shows acceptable porosity or unacceptable lack of fusion depends on the individual's method knowledge, their experience of that product form, and the conditions they were working in. Certification establishes a minimum rather than a uniform output.

So the controls are about people rather than about equipment. Certification tracked by method and level with renewals managed before expiry, technicians working only within the methods and levels they hold, a second opinion arrangement for marginal or unusual indications rather than a lone judgement under programme pressure, periodic review of reports by a senior technician, and a culture where a technician can say an indication needs another look without that being a problem. A business that can describe how it handles a marginal call is describing the thing that actually produces claims.

Interpretation and radiography are the two distinct questions. We are an FCA regulated broker and will not tell you we are the best choice. These are the questions that decide it.