On the afternoon of Tuesday, 5 May 2015, in the southern Bay of Campeche, a 7,000 ton jackup was moving into position beside one of PEMEX's producing wellhead platforms. The rig was the Troll Solution, contracted through Typhoon Offshore Services, a Mexican operator owned by conglomerate Grupo Salinas, for well maintenance in the Abkatún-Pol-Chuc shallow water complex, adjacent to the Caan-Alfa (CAAN-A) platform. The operation was routine. The crew was experienced. And then a leg went.
One of the three spud cans punched through the seabed. Seven thousand tons of steel redistributed onto the two remaining legs in a fraction of a second. The hull tilted, hard. Within hours the rig had listed further, taken on water, and begun the slow collapse that would end with it on the seabed in roughly 30 metres of water, two metres from the active CAAN-A platform it had come to service.
Two Typhoon Offshore workers were killed. PEMEX's initial communications reported two minor injuries. Typhoon Offshore said 10 injured. Local Mexican media put the count at 28. Roughly 100 remaining personnel were evacuated. For an incident of that scale, in that basin, in that year, the record is remarkably thin. No public investigation report was ever released by PEMEX or by Mexico's offshore safety regulator, ASEA, the same regulator that six months earlier had quietly closed the file on PEMEX's other 2015 fatal offshore incident with no sanctions.
What actually happened
The Troll Solution was a self elevating unit built by COSCO Nantong in 2010, Vanuatu flagged out of Port Vila, IMO 8770338, owned by Trollrig1 Ltd. and managed by Harren and Partner. Industry sources describe it as a modified Friede and Goldman Super M2, a three legged, self propelled design repurposed from drilling to well service, fitted with a 400 ton crane in place of a drilling package. That asset conversion matters, and we will return to it.
During positioning, according to accounts from parties involved in the subsequent wreck removal, preloading was performed with a significant air gap: hull elevated well clear of the sea surface before the deliberate weight transfer that proves out each spud can's foundation. Preload with too much air gap magnifies the consequences of any asymmetric leg penetration. The moment one can breaks through, the other two are levered by the tallest cantilever the rig will ever have.
The failure mode itself is well known. A punch through is a sudden, uncontrolled penetration of a spud can through a competent surface layer into a softer stratum beneath it. It has been the primary geotechnical risk in jackup positioning since the rig type came into wide use, and the mechanics are documented across decades of SNAME and ISO guidance. On paper it is a solved problem. Site specific assessment before spud in, sub bottom profiling, soil sampling, layered bearing analysis, a defined preload procedure with go and no go criteria, all of it is supposed to prevent this.
In practice it keeps happening. It keeps happening because seabed conditions vary at metre scale across a single platform footprint, because last year's survey is not this year's seabed, and because positioning is a live operation with wind, current, DP thrust, and a producing platform two metres away all narrowing the window in which stop and reassess is a real option.
The punch through did not kill the crew
The tilt was survivable. The evacuation was not. Freefall and davit launched lifeboats are engineered for release from a defined attitude. Neither system tolerates a 7,000 ton rig moving unpredictably through unexpected angles while a boat is being boarded or released. Routine abandonment drills, the muster exercises every offshore worker sits through, do not simulate this. They cannot, safely. The two workers who died on the Troll Solution died during evacuation, not from the initial list.
That distinction is the single most transferable operational lesson from the incident. Evacuation from a listing jackup is a distinct training scenario, not a subset of the standard muster drill.
The pattern PEMEX did not want to explain
The Troll Solution was PEMEX's second fatal offshore incident in 35 days. On 1 April 2015 a fire on the Abkatún Permanente processing platform, inside the same Abkatún-Pol-Chuc complex, killed four workers and injured sixteen. Three others were never found. ASEA's investigation concluded the fire was caused by a rarely used gas fuel line that had suffered accelerated corrosion driven by sulfate reducing microorganisms and sulfuric acid, ruptured, and ignited. No sanctions were issued. The regulator ruled there had been no failure to comply with applicable regulations.
That regulatory posture is the backdrop for the Troll Solution's missing report. Two fatal incidents in five weeks, in the same operator's fleet, in the same basin, is the kind of pattern regulators elsewhere treat as a systemic signal. Without a public investigation, findings, contributing factors, corrective actions, the industry does not get to learn from the Troll Solution. It only gets to remember it.
The removal that told the story instead
The wreck itself became the largest salvage of 2016. Salvor Ardent, the newly merged Titan and Svitzer entity, spent months on the operation. Because the hull had come to rest within two metres of a live production platform, conventional cutting methods were closed off. Pull cuts require tunnelling under the wreck. Push cuts require an external framework. Neither was compatible with an active wellhead complex at arm's length.
Ardent deployed the Conquest MB-1 crane barge, with 1,400 ton lifting capacity, previously used in the Costa Concordia removal off Italy, together with a 1,000 ton hydraulic wreck grab from the Netherlands, and its newly designed Guided Guillotine, a chain cutting system engineered to work underwater with the precision required to section the rig without an external framework. The team cut the deckhouse into six pieces with Ardent chain pullers and the hull into 31 pieces with the Guided Guillotine, then lifted every fragment from the seabed.
The salvage record is better documented than the accident record. That, more than anything, says what the state of offshore transparency in the basin actually is.
What site specific has to mean now
Every jackup positioning decision is a bet on what the seabed looks like right now, not what it looked like the last time somebody surveyed it. That bet gets safer when three things are true. The seabed picture is current, with bathymetry, sediment character, and known hazards updated from real survey data rather than five year old paper charts. The seabed picture is integrated, with pipelines, cables, wellheads, existing platform footprints, and anchor patterns from prior operations, all of which both constrain positioning and change the local geotechnics, sitting on the same map as the bathymetry rather than in three separate GIS files on three laptops. And the seabed picture is connected to what is happening above it, so vessel positions, weather, and the DP approach live in the same operational view as the seabed data, and when a current shifts or an approach vector changes, the reconciliation with the survey happens in the moment, not in a report six weeks later.
This is what Atlas is built for. Atlas is Blackwell's seabed intelligence product, a live geospatial view of everything beneath the surface: bathymetry, seabed conditions, pipelines, cables, wellheads, other subsea infrastructure, inside the same platform that carries vessel tracking and weather. For rig move and jackup positioning specifically, the value is not a new algorithm. It is the elimination of the seam between survey data, chart data, asset data, and the live operational picture. One shared map. Fewer surprises during positioning. A faster path from we have a plan to we have a plan grounded in what is actually there.
None of this replaces site specific assessment under ISO 19905-1 and 19905-4 or SNAME T and R Bulletin 5-07. It makes the assessment usable at the moment it matters, during the approach, not filed in a report from six weeks ago.
The lesson that never got formalised
The Troll Solution is a compact case study in every failure mode the jackup standards exist to prevent. Geotechnical assumption where site specific assessment was required. A preload procedure that amplified rather than absorbed the failure. An asset repurposed from drilling to well service without a matching update to its positioning and evacuation regime. An evacuation scenario the crew had never trained for. And a regulator that let the record close without a public finding.
Rig positioning is not a passive operation. Every move carries geotechnical risk that has to be actively managed on the day, with current data, integrated tools, and rehearsed procedures for the moment a leg does not sit where you expected it to.
Two men from Typhoon Offshore should be alive.
References
- Reuters via gCaptain, Accident Aboard Pemex Jack-Up Kills Two, Injures 10, 5 May 2015. https://gcaptain.com/accident-aboard-pemex-jack-up-kills-two/
- OE Digital, Ardent completes Troll Solution removal, 7 December 2016. https://www.oedigital.com/news/447742-ardent-completes-troll-solution-removal
- OE Digital (Melissa Sustaita), Cause of deadly PEMEX Abkatún fire determined, 19 October 2015. http://www.oedigital.com/component/k2/item/10636-cause-of-deadly-pemex-abkatun-fire-found
- Maritime Directory MarineLink, Troll Solution vessel record, IMO 8770338. https://directory.marinelink.com/ships/ship/troll-solution-7542
- gCaptain (Mike Schuler), PHOTOS: Largest Wreck Removal of 2016 Completed in Gulf of Mexico, 7 December 2016. https://gcaptain.com/photos-troll-solution-largest-wreck-removal-of-2016-completed-in-gulf-of-mexico/
- Blackwell, Atlas seabed intelligence. https://blackwellgeo.com/products/atlas
- ISO 19905-1 and 19905-4, Petroleum and natural gas industries, Site specific assessment of mobile offshore units, Jack-ups; SNAME Technical and Research Bulletin 5-07, Guidelines for Site Specific Assessment of Mobile Jack-Up Units.