Building a thickness record a class surveyor can sign without argument

The audit is a test of the record, not of the steel

By the time a class surveyor, a vetting inspector or a port state control officer is standing on deck, the condition of the steel is whatever it is. What is still in play is whether the record explains it. A special survey does not go wrong because a plate is 3.2 mm below as-built. It goes wrong because nobody can show what the as-built was, who gauged it, under which qualification, with which instrument, verified when.

This is why audit preparation that consists of collecting documents is the wrong shape of work. The documents were always there. What is missing is the linkage — report to reading to member to certificate to calibration record, each pointing at the others. Assembling that linkage in the fortnight before an attendance is precisely how errors enter: dates get approximated, a certificate number is typed from memory, a reading is attributed to whoever appears on the crew list for that week.

A registry moves the assembly work to the moment of capture, where it costs almost nothing. The technician's certificate, the instrument, the verification block and the structural member are all selected at the reading rather than reconstructed months later. Preparing for an attendance then becomes an extraction from a system rather than an excavation through folders.

Diminution allowances are per member, and that changes the data model

Fixed-equipment inspection uses a single retirement thickness per location, computed from design conditions and allowances. Marine does not work that way. Every structural member has its own permissible diminution, set by the classification society's rules and varying with member type, position in the ship, and whether the vessel is a tanker, a bulk carrier or another type. Deck plating, side shell, longitudinals, web frames and brackets in the same tank carry different allowances.

The registry therefore needs three values per point rather than one: as-built thickness taken from the ship's scantling drawings or midship section, the applicable corrosion addition or permissible diminution for that member, and the current gauged value. Wastage then becomes expressible both in millimetres and as a percentage of allowance, which is the form a surveyor actually reasons in when deciding what happens next.

Substantial corrosion — wastage that has consumed 75 to 100 percent of the allowance — should be a computed status rather than a judgement made at the gangway. It drives extended gauging, close-up survey scope and, on older ships at intermediate survey, whether the survey expands significantly. A record that presents raw thickness and leaves the arithmetic to the attending surveyor invites disagreement about the scope of the survey itself.

Reading through coating: the 0.7 mm that is not there

Ships are painted. In a refinery a technician removes insulation to reach the steel; on a hull, blasting every gauging point is often impractical, and readings are routinely taken through an intact coating system. A conventional single-echo thickness gauge calibrated to a steel velocity of about 5,920 m/s will include that coating in its answer, and coating sound velocity is roughly 2,300 m/s — less than half.

The arithmetic is unkind. A 0.3 mm paint film adds about 0.26 microseconds of round-trip transit time, which the gauge converts into roughly 0.77 mm of apparent steel. Set against a corrosion addition of two or three millimetres on a shell plate, that single error consumes a quarter to a third of the entire allowance, and it always reads in the reassuring direction. A structure that appears comfortably inside limits can already be at substantial corrosion.

Echo-to-echo measurement removes the problem by timing between successive back-wall echoes and discarding the coating. Class-approved procedures are written accordingly. The registry's contribution is to make the measurement mode a required field on the reading rather than a line in a procedure nobody opens — so that a reviewer three years later can tell whether a suspiciously stable location was genuinely stable or was measured through a fresh coat of paint.

Renewal resets the baseline, and most spreadsheets never notice

Nothing in fixed-equipment integrity behaves like a steel renewal. When a wasted shell plate or a set of side shell frames is cropped out and replaced in the yard, the location physically becomes new material at the drawing scantling. Every reading taken before that event describes metal that no longer exists on the ship.

A spreadsheet has no way of knowing this. It carries the column forward, computes the difference between last survey's 9.1 mm and this survey's 14.0 mm, and produces a negative corrosion rate. More often the negative gets noticed, the row is quietly deleted to make the sheet look sensible, and the location loses its entire pre-renewal history along with the evidence that a repair ever happened. Both outcomes damage the record a later survey depends on.

The correct model is a renewal event that closes the previous series and opens a new one at the same location, carrying the date, the extent, the new as-built and the repair documentation. The point identity persists so the surveyor can follow the whole story, while the trend restarts so the rate is honest. How a system handles this single case is one of the quickest ways to tell whether it was designed for marine work or adapted to it afterwards.

Service supplier approval: the qualification evidence belongs to the reading

Thickness measurement for classification purposes is not something an owner can simply arrange internally. Under the IACS unified requirement for service suppliers, the firm carrying out the gauging must be approved by the society, must work to an approved written procedure, and must employ operators qualified under a recognised scheme — commonly ISO 9712 Level 2 in ultrasonic testing — with the surveyor attending as the survey type requires.

The audit consequence is sharper here than in most industries. A reading taken by a technician whose certification had lapsed, or by an operator outside the approved supplier's listed scope, is not merely a weak reading. It is inadmissible, and its removal can leave an incomplete gauging pattern for a transverse section, which in turn puts the completeness of the survey itself in question rather than just the quality of one number.

So certificate state has to be bound to the reading at the moment of capture: scheme, method, level, certificate number and expiry as they stood on that date, together with the approval scope of the firm. A folder of PDFs on a shared drive is not the same thing. The test is whether you can answer, for any single reading three years old, whether that person was qualified for that work on that day, without opening anything else.

Enhanced survey programme points: where the arguments actually happen

The enhanced survey programme prescribes what gets gauged and where: transverse sections at defined frame positions, close-up survey of specified structural members, and scope that increases with the ship's age at second, third and subsequent special surveys. The gauging pattern is not the technician's choice. It is a diagram, and the report is expected to correspond to it member by member.

On single-side-skin bulk carriers, side shell frames and their end brackets are the classic contested area — a small number of members carrying a disproportionate share of the survey's outcome, where a handful of readings can decide whether the ship trades or goes into the yard. On tankers, ballast tank internals and deck longitudinals occupy the same position. At these locations a missing reading is not a data gap; it is an argument waiting to happen.

A registry that carries the gauging diagram as structure — sections, members and required points — rather than as an attached PDF can tell the superintendent what is still outstanding before the surveyor arrives. A spreadsheet can hold exactly the same readings and never answer the only question that matters in the final week of preparation, which is what has not been done yet.

Vetting and port state: the same data, a different audience

A class survey is one demand on the record, not the only one. A tanker owner will face vetting inspections where a hull condition question is answered from the same thickness history, and where an inspector's observation about record quality is scored and visible to charterers. Dry bulk operators face equivalent scrutiny from ship-vetting services with their own inspection regimes and their own tolerance for untidy paperwork.

Port state control adds a third audience with the least patience of the three. An officer working from a targeting profile who asks for the last thickness measurement report is not going to learn your filing convention. Producing the report, the substantial-corrosion summary and the repair close-out evidence within a few minutes is a materially different position from offering to email everything after departure.

The evaluation implication is that a marine registry has to produce several output shapes from one dataset: the class report format, a condition summary by tank and member for vetting, and a repair-history extract. If those exist as three separately maintained documents, they will eventually disagree, and the disagreement will be discovered by whichever audience you were least prepared for.

Evaluating a registry six weeks before the surveyor lands

Start with completeness rather than features. Load one transverse section from the last special survey and ask the system to tell you what is missing against the required pattern. A tool that can only display readings you have already entered is a database. A tool that knows what should exist, and can name what does not, is a survey preparation instrument.

Then test the ugly cases in front of the vendor. Enter a renewal and see whether the trend restarts or continues. Enter a single-echo reading on a coated surface and see whether it is flagged. Enter a reading against a certificate that expired the day before. Change a stored as-built value and ask for the change history. Anything that passes silently in a demonstration will pass silently during the survey too.

Finally, look hard at who is expected to use it. Gauging happens in tanks and in dock, with cold hands, poor light and no signal. If capture needs a laptop and a connection, readings will be written on a pad and typed up afterwards — and everything the system claims about provenance becomes fiction, because the person, time and instrument it records are those of whoever did the typing.

Request a consultation