What an inspection record set tells you about an asset you are about to buy

What you are actually buying when you buy inspection records

In an asset transaction the inspection record set is a proxy for two things you cannot observe directly: the condition of the asset, and the discipline of the organisation that ran it. Condition you can partly verify by inspecting. Discipline you can only read from the records, and discipline determines whether the condition you were shown is the condition that exists across the parts you did not look at.

The seller supplies a programme — procedures, intervals, a written practice, a findings summary, an integrity management document. All of it can be genuine and none of it tells you what was performed. The programme is the intention. The records are the execution. The distance between them is the single most useful output of a due diligence review, and it never appears in the data room summary because the summary is built from the programme.

This matters because deferred inspection is invisible on a balance sheet and expensive after completion. A tank that has not had an internal inspection in nineteen years is not a problem until it is, and then it is a cleaning contract, a floor replacement, a turnaround and a shutdown that was never in the model. An independent record validation tells you which of those you are inheriting, before you own them.

Intervals on paper versus intervals performed

Start with dates, because dates are unambiguous and nobody can argue with them. Build one table of every statutory or programme inspection due in the review period, against the date it was performed and the report number that evidences it. This is mechanical work that a competent analyst can do quickly, and it produces the first real finding faster than any other activity in the review.

The interval framework gives you the yardstick. API 653 sets external inspection of aboveground storage tanks at intervals not exceeding five years or one quarter of the remaining corrosion life, whichever is less, with the first internal inspection within ten years of service and extension beyond that requiring a documented basis. API 510 places internal or on-stream inspection of pressure vessels at the lesser of half the remaining life or ten years, with external at the lesser of five years or the internal interval. API 570 requires Class 1 piping thickness surveys at intervals no greater than five years.

What you are looking for is not one missed date. It is a pattern of extension without a documented basis. An interval extended under a risk-based inspection analysis, with the analysis in the file and an engineer's name against it, is a managed decision that a buyer can evaluate. The same extension recorded only as a schedule change in a maintenance system, with no engineering document behind it, is deferred maintenance with a date on it.

The findings register is the fastest read in the set

Every finding raised in the period, its status, and the document that closed it. Three columns, and they carry more information per hour of review than anything else you will open. A healthy register shows findings raised, dispositioned, repaired and verified, with a closing document that exists, can be opened, and says what the register claims it says.

The failure mode is administrative closure. A finding marked closed with no repair package, no post-repair examination, no engineering disposition and no re-inspection has not been closed — it has been removed from a list. When you find one, count the rest. A register with a handful is a records problem. A register where a quarter of the closures have nothing behind them is a programme problem, and it changes what you are entitled to believe about everything else in the set.

Carried findings are the other side of this, and they are read the wrong way round by most buyers. Findings that are genuinely open, tracked, engineered and scheduled are a normal feature of a running asset and a good sign about the operator. What you are pricing is the work behind them, and that is a schedule you can build and cost. The register that should worry you is not the one with open items. It is the one with none.

Thickness data: read the readings, not the corrosion rate

Corrosion rates are calculated, and calculations carry assumptions. Ask for the raw condition monitoring location readings across the whole period rather than the summary rate, and plot them yourself. The question you are answering is whether the rate driving the seller's integrity model was derived from measurements, or fitted to produce a remaining life that suited the inspection interval.

Specific things surface in raw data. Readings that never decrease across ten years in corroding service. A CML that produces one low reading and is then quietly dropped from the survey. A location renumbered between campaigns so the history no longer joins up. A change of contractor or instrument coinciding with a step change in measured thickness. None of these proves anything on its own. Together they tell you the model is not measurement-driven, and the remaining life in the seller's model is not a number you can rely on.

Resolution matters too. A thickness table reporting three decimal places from an instrument and technique that does not support that precision is reporting certainty it does not have, and every remaining-life calculation built on it inherits the error silently. Read the procedure alongside the data, at the revision in force during each campaign, because a technique change mid-programme makes the two halves of the trend line incomparable.

Repair history and the post-repair examination that is missing

Repairs are where the inspection history and the maintenance history have to meet, and frequently they never do. For each recorded repair you want three documents: the engineering disposition that authorised it, the welding documentation that governed it, and the examination that verified it afterwards. The third is the one that goes missing, and it is the one that matters most to a buyer.

A weld repair recorded in a maintenance system with no post-repair examination report is an unverified repair that you now own. This is common in assets where inspection was contracted out and maintenance was performed in-house: the two record sets live in different systems, under different owners, and nobody ever joined them. Joining them is a specific, tractable due diligence task and it produces findings on almost every record set of any size.

The reverse case matters as much. Examinations that describe a repaired area with no repair record anywhere are either a records gap or an undocumented repair, and both limit what you can later claim about the asset to a regulator, an insurer or a future buyer. Operators who run inspection and maintenance against a single asset record have far fewer of these, and the absence of such a system in the seller's estate is itself a signal about how much joining work you should budget for.

Sampling strategy: what to pull when you cannot read everything

You will not read everything and you should not try. A forty-tank farm across twenty years, or a fleet of a dozen vessels across two special survey cycles, is tens of thousands of documents. Reading a random ten per cent of it produces a shallow impression of everything and a defensible conclusion about nothing, at full cost.

Sample deep and narrow instead. Pick the highest-consequence items — the equipment whose failure stops the plant, the tanks in the worst service, the vessels carrying the largest deferred scope — and read their complete history end to end, campaign by campaign, including the raw data. Three consecutive campaigns on one item tells you how the programme actually behaved. One campaign across thirty items tells you how the summary spreadsheet was assembled.

Then run three thin horizontal sweeps across the entire population: interval compliance dates, findings closure status, and raw-data retention. All three are cheap to check at scale and each one points you at the items that deserve a deep read. Layering a narrow-deep sample under a wide-thin sweep is how a records review of any size gets completed inside the window a transaction actually allows.

Red flags that change a price

Some findings are informational and some move numbers. The ones that move numbers share a single feature: they mean the asset's real condition is unknown rather than known-and-acceptable. An internal inspection deferred repeatedly with no engineering basis. A period with no records at all. Raw data destroyed, leaving only summaries that nobody can now verify against anything.

A change of inspection contractor immediately followed by a materially better report deserves attention, and so does the reverse — a new contractor whose first campaign finds substantially more than the previous one ever reported. Neither is evidence of anything improper. Both tell you the earlier data set and the later one are not comparable, which means the trend line the seller is showing you spans two different measurement regimes and is not a trend at all.

Personnel and procedure findings tie into this directly. Examinations performed by people whose certification did not cover the scope, or under a procedure revision issued after the examination date, produce records that cannot be relied on even where the underlying work was sound. Confirming this is straightforward and it belongs in every review; the same ground is covered systematically by an NDT programme audit for an asset you already own.

When records simply do not exist for a period

Gaps are normal and they are still findings. They cluster around predictable events: a change of operator, an outsourcing transition, a document management migration, an insolvency, a cancelled turnaround. The first question is whether the records were never made or were made and then lost, and the transition history in the corporate documents usually answers it in an afternoon.

A gap cannot be filled retrospectively and should not be. What replaces it is a bounded technical judgement: given the documented condition at the start of the gap, the service, and the condition at the end, what range of degradation is credible across the missing period, and what inspection would close it. That produces a defined scope with a defined cost that can go into the transaction, rather than an unquantified risk that neither party can price.

The commercial handling follows from that. A quantified gap becomes a price adjustment, a retention, a warranty, or a condition precedent requiring a defined inspection scope before or after completion. An unquantified gap is a discount argument that neither side can support and that both sides will simply assert at each other. Turning the second into the first is most of what a records review is worth in a live deal.

Marine and fleet assets: a different records regime

Ships and mobile offshore units carry a records regime driven by class rather than by an owner's own programme, which changes what you review and who you check. Thickness measurement firms working for class are approved service suppliers, and IACS UR Z17 sets the procedural requirements for that approval — including that the responsible supervisor is qualified to a recognised national or international NDT standard such as ISO 9712 Level II, and that operators hold Level I certification together with adequate knowledge of ship structures.

For a fleet purchase that gives you a checkable chain rather than an opinion. Was the firm that produced each thickness measurement report an approved service supplier on the date, and was the approval current? Were the supervisor and the operators qualified under the scheme the rule requires? Class survey status, outstanding conditions of class, and the thickness measurement reports behind the most recent special survey are the core of the read. The requirement itself is covered in more detail on IACS UR Z17.

The gap analysis is identical in principle to a fixed plant and different in its documents. Substructure records, ballast tank coating condition reports, and the closure evidence for conditions of class carry the same three questions: was it done, is it recorded, and does the record show a controlled process. Where a fleet's records are thin, the exposure sits in the next special survey rather than in the next turnaround, which changes the timing of the cost but not its size.

What the review produces, and how it feeds the deal

The deliverable is short and specific. A compliance table of intervals due against intervals performed, by item. A findings-closure analysis with every unsupported closure identified by finding number. A list of records gaps with their dates and a bounded inspection scope costed against each. A plain statement of what the record set does and does not establish about the asset, mapped to the equipment register so it carries into the buyer's own asset model after completion.

Written that way it goes straight into the transaction. Each gap becomes a scope with a cost. Each unsupported closure becomes a diligence question with a named document behind it, which the seller can answer specifically rather than deflecting a general assertion that the records are poor. A finding that the record set is sound is equally usable, because it removes an argument the other side would otherwise have had to price into the deal.

Timing decides how much any of it is worth. A records review run early generates questions while there is still time and incentive for the seller to answer them. The same review run in the final fortnight produces a list of items nobody can now resolve. If you are looking at a plant, a tank farm, a vessel or a fleet and the inspection history belongs to somebody else, start the review before the data room closes.

Request a consultation