Five leaks, one flaw: the ExxonMobil Fife CUI prosecution and the inspection that couldn't see

On 26 August 2026 the Health and Safety Executive published a prosecution release on ExxonMobil Chemical Limited’s Fife Ethylene Plant at Mossmorran, near Kirkcaldy. The company was fined £267,000 at Kirkcaldy Sheriff Court on 25 August after inspectors traced five separate leaks of extremely flammable hydrocarbons, all to a single degradation mechanism: corrosion under insulation. No one was hurt, and none of the leaks ignited. The prosecution is worth reading closely anyway, because it turns on the part of a process hazard analysis that is easiest to over-credit — the assumption that the inspection programme behind a passive safeguard can actually detect the thing it is meant to catch.

What the release describes

The plant is an Upper Tier COMAH establishment, subject to the strictest controls because of the volume of hazardous substances it holds. According to the HSE, five known leaks occurred between February 2018 and September 2019, each caused by corrosion under insulation (CUI). A pinhole leak of dimethyl sulphide (DMS) and ethane opened on a high-pressure 2-inch steel line running at about 115 °C in February 2018 and went unrepaired for roughly four months. In December 2018 a leak on a separate overheads line was spotted only by chance, when a worker noticed a haze coming through the insulation during unrelated work; corroded cladding had let four small holes form, and an estimated 82 tonnes of hydrocarbons escaped over about six weeks.

The investigation began after a routine inspection on 21 May 2019, when HSE inspectors smelled DMS and an operator confirmed a leak the company had known about for around four months while continuing to operate. HSE issued a Prohibition Notice banning access to the area and an Improvement Notice ordering repairs. Two further leaks came to light as the investigation and the ordered inspections proceeded — one on a small-bore connection to the acetylene converter, found to have corroded below its safe minimum thickness, and a further pinhole leak on the overheads line in September 2019.

Why the inspection regime failed

The regulator’s central finding was not that CUI occurred — the industry accepts that CUI cannot be entirely prevented where insulation is required — but that the arrangements for detecting it were, in the HSE’s words, not fit for purpose. Inspectors carried out visual “scoping” checks on each pipe; the results fed a computer tool that produced a risk score from 1 (high) to 4 (low), and the company followed up only on pipes scoring 1 or 2. In practice, the HSE found, overhead pipes were often checked from ground level, sometimes several metres away, with limited visibility and without removing insulation to look underneath. Despite repeated CUI leaks, the company did not recognise its inspection process was flawed until enforcement action forced the issue.

The prosecuting inspector’s framing is the sentence worth keeping: that none of the leaks resulted in a catastrophic event was “down to good fortune rather than good management.” ExxonMobil pleaded guilty to breaching Regulation 6(2) of the Provision and Use of Work Equipment Regulations 1998 and section 33(1)(c) of the Health and Safety at Work etc. Act 1974.

The lesson for study and revalidation teams

CUI is the textbook example of a hazard that lives in the mechanical-integrity programme rather than in any HAZOP node. A study credits the pressure boundary as a passive safeguard and, explicitly or not, assumes a written scheme of examination keeps the pipework above its minimum thickness. That credit is only as good as the examination method behind it. Here the examinations were nominally happening, but a ground-level visual check through intact insulation cannot see the wall loss that CUI produces — so the safeguard the study relied on was, in effect, unmonitored.

Two questions travel from this case to any revalidation. First, for each safeguard a study credits against a corrosion or wall-thinning failure, does the assigned inspection method genuinely detect that mechanism? CUI in particular needs insulation removal, stripping at suspect locations, or profile and real-time radiography — not a walk-down. Second, is there a bright-line rule that a confirmed live release of flammables takes the line out of service, rather than a “monitor and keep running” default while a repair is scheduled? The Fife sequence — a known leak run for months — is a reminder that the gap between a good procedure on paper and the decision made on the day is where the LOPA credit is actually won or lost.

The primary source is the HSE prosecution release of 26 August 2026.