BRENNEN

Industries / Construction and property / Heavy civil / Marine construction

Offshore foundations

Offshore foundations are the structures that stand on the seabed and hold oil and gas platforms above the waves: steel lattice towers called jackets, and concrete bases that stay in place by their own weight.

6

Milestones achieved, from 1956 to 2007.

1

To watch for: announced, not yet confirmed outside.

2

Open deliverables. Nobody has shipped them.

The road

AchievedWatch forOpen

Achieved

  1. 1956

    Jackets slide off barges

    A jacket is the welded lattice of steel tubes that stands on the seabed and carries a platform's deck, pinned down by steel piles driven through its legs. As jackets grew too big for the cranes of the day, J. Ray McDermott in 1956 performed the first launch: the jacket was skidded off a barge into the sea, where it floated, and a floating crane then upended it and set it on the seabed. The method is still used for the largest jackets.

    texasce.org

    Credit

    J. Ray McDermott

  2. 1969

    A design code after Hilda

    Early platform designers used the highest wave recorded in the previous 25 years, and hurricanes in the 1960s toppled many Gulf of Mexico platforms. After experts met in November 1964 to study the damage from Hurricane Hilda, the American Petroleum Institute's Subcommittee on Offshore Structures published the first edition of API RP 2A, its recommended practice for fixed offshore platforms, in 1969; the industry later settled on designing for the wave of a 100-year storm, with an "air gap" between the wave crest and the deck.

    LessonTwenty-five years of wave records were too short to show the storms a platform would meet.

    offshore-mag.comtexasce.org

    Credit

    American Petroleum Institute, Subcommittee on Offshore Structures

  3. 1973

    A concrete tank at Ekofisk

    Oil from Norway's Ekofisk field was loaded into tankers at buoys, and production stopped whenever weather forced the tankers to disconnect. The field's owners ordered a concrete storage tank designed by the French firm C G Doris, and on 1 July 1973 the 215,000-tonne structure, towed out by six tugs, was flooded until it stood on the seabed under its own weight, with a steel skirt beneath it so currents could not wash the seabed out from under it. The Norwegian Petroleum Museum calls it the world's first concrete oil storage tank.

    ekofisk.industriminne.noekofisk.industriminne.no

    Credit

    C G Doris; Ing F Selmer; Høyer-Ellefsen

  4. 1988

    A jacket for deep water

    In the summer of 1988 Shell set its Bullwinkle platform in Green Canyon Block 65 of the Gulf of Mexico, about 150 miles south-southwest of New Orleans, drawing on its Cognac platform, set in the previous record depth of 1,025 ft (312 m). The Texas Section of the American Society of Civil Engineers gives Bullwinkle's depth as 1,356 ft (413 m) and its jacket as 49,375 tons, the heaviest structure moved to sea up to then, and it remains the deepest fixed platform installed in the Gulf.

    osti.govtexasce.org

    Credit

    Shell; Gulf Marine Fabricators

  5. 1991

    Sleipner's base sinks

    On 23 August 1991 the concrete gravity base for Statoil's Sleipner A gas platform sank in the Gands Fjord near Stavanger during its final tests: a wall in a tricell, the three-sided space where three round storage cells meet, cracked, and the base went down in about 18 minutes, with all 22 people on board taken off unhurt. Inquiries by Statoil and the builder, Norwegian Contractors, traced the failure to errors in the finite element analysis, a computer method that splits a structure into small pieces to calculate its stresses, which left the tricell walls too weak; the same error had been made on the Draugen base being built for Shell, which halted for about three months while outside analyses checked it.

    LessonCritical computer results need an independent check by another analysis or a physical test.

    equinor.industriminne.nodraugen.industriminne.no

    Credit

    Statoil and Norwegian Contractors (inquiries); Norwegian Petroleum Museum (account)

  6. 2007

    Storms beyond the design

    Hurricanes Ivan in 2004 and Katrina and Rita in 2005 brought winds and waves beyond design criteria at many Gulf of Mexico sites; of more than 3,900 fixed platforms, 123 were destroyed and more than 180 badly damaged, most of them older platforms designed before API RP 2A or to its early editions, by the account of API's Hurricane Evaluation and Assessment Team. The team found that the oldest storms in the industry's hindcast database, its reconstruction of past storms, were probably biased low, and by mid-2007 API had issued Bulletin 2INT-MET, interim guidance with markedly more severe conditions off eastern Louisiana.

    LessonDesign storms drawn from sparse old records can understate the sea.

    offshore-mag.com

    Credit

    API Hurricane Evaluation and Assessment Team

Watch for

  1. Digital twins for old jackets

    Keyang Liu, Yuanchang Liu, Baoping Cai and Jeom Kee Paik propose a method that computes wave loads and the remaining ultimate strength of an aging jacket, the load at which it would collapse, from measured sea conditions and measured defects such as corrosion, fast enough to monitor its safety as conditions change. They validated it on one case-study platform.

    Achieved whenan operator or regulator uses the method on a working platform and publishes its predictions against inspection findings or an independent analysis.

    discovery.ucl.ac.uk

    Claimed by

    Liu, Liu, Cai and Paik, in Ocean Engineering, volume 343, 2026

Open

  1. 01

    One table of storm losses

    One public, machine-readable table of every fixed platform destroyed or badly damaged by Gulf of Mexico hurricanes, with its install year, water depth, deck height and design vintage, and a short analysis of how loss rates fell with newer design rules.

    The brief
    Credit · unclaimed
  2. 02

    Blind test for jacket models

    A scale model of a steel jacket frame, a few meters tall and welded from steel tubes, tested to collapse under sideways load in a university structures lab, some with chosen braces dented or cut to mimic damage. The geometry, material tests and damage are published first, teams using digital twins and finite element codes post predictions, and then the measured collapse loads and failure sequences are released.

    The brief
    Credit · unclaimed

Open 01 · the brief

One table of storm losses

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

One public, machine-readable table of every fixed platform destroyed or badly damaged by Gulf of Mexico hurricanes, with its install year, water depth, deck height and design vintage, and a short analysis of how loss rates fell with newer design rules.

Why it is still open

Energo Engineering's reports for the US Minerals Management Service list destroyed platforms storm by storm in PDF appendices: 116 fixed platforms for Katrina and Rita in the 2007 report, and 60 of 3,245 exposed for Gustav and Ike in the 2010 report. Mark Kaiser of Louisiana State University totaled the five storms of 2004 to 2008 at 181 structures destroyed (Oil & Gas Journal, July 2014), but as counts and maps, not a public per-platform table, and nothing opened covers storms after 2008.

bsee.govbsee.govogj.com

Built with what exists

pdftotext or a table extractor such as Camelot; a vision-language model to read any scanned tables, with every row checked against its page; the Bureau of Safety and Environmental Enforcement's (BSEE) public platform structure records for install year and location; Python with pandas.

  1. First step

    Turn Appendix A of the Katrina and Rita report, 116 platforms, into a CSV and match ten rows by hand to BSEE records.

  2. Done looks like

    A CSV covering the five storms from Ivan to Ike, each row traced to a report page, at least 95% of rows matched to a BSEE structure record, and a chart of platforms destroyed per hundred exposed, by decade of installation.

Open 02 · the brief

Blind test for jacket models

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

A scale model of a steel jacket frame, a few meters tall and welded from steel tubes, tested to collapse under sideways load in a university structures lab, some with chosen braces dented or cut to mimic damage. The geometry, material tests and damage are published first, teams using digital twins and finite element codes post predictions, and then the measured collapse loads and failure sequences are released.

Why it is still open

The last large frame test program listed by the US regulator, a joint industry project led by Billington Osbourne-Moss Engineering from 1999 on the reserve and residual strength of jacket frames, shared its final report only with its participants. New digital-twin methods for aging jackets are each validated on a single case-study platform chosen by their authors.

bsee.govdiscovery.ucl.ac.uk

Built with what exists

Standard steel tube; a lab reaction frame and hydraulic actuator; strain gauges; digital image correlation, which tracks a painted speckle pattern with cameras to measure strain, using open software such as Ncorr; open finite element codes such as OpenSees or Code_Aster for the organizers' own prediction.

  1. First step

    Ask a university structures lab for a test slot, and post the intact frame's drawings and material test plan publicly.

  2. Done looks like

    Three frames tested (intact, one damaged brace, two damaged braces), with predictions from at least five outside teams posted before the results, and each team's error in collapse load published.