Industries / Energy, water and waste / Power generation / Wind turbines
Offshore foundations
Offshore foundations hold a turbine up at sea, from the bottom of its tower down into the seabed, either standing on the seabed or floating and held in place by mooring lines.
5
Milestones achieved, from 2003 to 2023.
1
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
2003
Monopiles reach the open sea
Horns Rev 1, built by the Danish utility Elsam 14 to 20 km off Denmark's west coast in the North Sea, stands 80 turbines with a total capacity of 160 MW on monopiles: each turbine rests on one steel tube driven into the seabed. Vattenfall, which now owns 60% of the farm, calls it "the world's first large-scale offshore windfarm" and "one of the first" to use monopiles. It was inaugurated in 2003.
Credit
Elsam; owned today by Vattenfall and Ørsted
2008
Limits on piling noise
Hammering a pile into the seabed sends out underwater sound that experiments showed can cause temporary hearing loss in harbor porpoises. From 2008 Germany's maritime agency BSH wrote two limits into every approval, to be met 750 m from the pile: 160 dB sound exposure level (the sound energy of one hammer blow) and 190 dB peak. The acoustics firm itap, in a 2020 report for the German environment ministry, says BSH set this dual limit "for the first time worldwide". Noise-reducing systems, such as a bubble curtain (a ring of air bubbles rising from a perforated hose laid on the seabed around the pile), became mandatory in 2011, but the limits could not be met reliably until offshore-ready systems arrived in 2014. By the end of 2019, 1,447 foundations with more than 2,400 piles had been driven in German waters, by the report's count.
Lessona limit set before the tools to meet it existed pushed industry to build them, but meeting it reliably took six years.
Credit
BSH, on scientific advice from the Federal Environment Agency (UBA); the industry that developed the systems; report by Michael Bellmann and colleagues at itap
2009
Grouted joints slip
Most early monopiles carried a transition piece, a short steel sleeve slid over the top of the pile to hold the tower, fixed in place by grout (a cement mortar) poured into the gap. In the fall of 2009, after reports that transition pieces at finished European wind farms had slid down their piles, putting the turbine's weight on temporary brackets not designed to carry it, the certifier DNV began reviewing its design rules. A team of experts from 11 utilities, engineering firms, suppliers and contractors found that steel tubes distort at large diameters, that tolerances were poorly controlled, and that bending from wind and waves wore away the grout's bond. DNV's new guide called for slightly conical joints, which press the grout tighter if the sleeve settles, and for added supports on existing straight joints. ENR, reporting in February 2011, said monopiles held two thirds of Europe's 1,136 grid-connected offshore turbines at the end of 2010, and that no industry-wide count of affected turbines existed.
Lessonthe UK's Sheringham Shoal project changed its joint design while the steel was still being made, at a cost its project director put at about $16 million, which ENR contrasts with costly repairs at sea.
Credit
DNV and its study team
2018
Buckets instead of hammers
A suction bucket is a steel cup turned upside down on the seabed. When water is pumped out of it, the pressure of the sea pushes it into the ground, so no hammering is needed. On 30 July 2018 the last of 20 suction bucket jackets, steel lattice frames standing on such buckets, was set at Ørsted's Borkum Riffgrund 2 off Lower Saxony, Germany, to carry 8 MW turbines, after a prototype at Borkum Riffgrund 1. The Norwegian Geotechnical Institute (NGI), which designed the buckets and ran the suction with Framo, calls the method smoother and "less noisy" than hammering, but gives no noise figures.
Credit
Ørsted; NGI and Framo; GeoSea (part of DEME) with the vessel Innovation
2023
Concrete floats, shared anchors
Hywind Tampen, about 140 km off Norway, floats 11 turbines on concrete spars, tall cylinders ballasted so they float upright, in water 260 to 300 m deep, held by a common anchoring system rather than separate anchors for each turbine. It began producing power in 2022 and was opened in August 2023 to supply the Snorre and Gullfaks oil and gas platforms; after an upgrade to 8.6 MW per turbine its capacity is 94.6 MW. Equinor reports an investment cost per megawatt about 35% lower than for its Hywind Scotland farm, adjusted for prices, while the total cost rose from an estimated NOK 5 billion to about NOK 7.4 billion, which it blames partly on late deliveries and quality problems from suppliers.
Lessona floating design proven at sea still met supply-chain delays and quality problems when built at a larger scale.
Credit
Equinor
Watch for
Monopiles without the hammer
Ecowende and Van Oord say three monopiles at the Hollandse Kust West wind farm off the Netherlands were installed to their target depth with VibroJet, GBM Works' method of fluidizing the soil with water jets inside the pile, combined with CAPE Holland's vibration tool, so the piles sank under their own weight. They call it the first use of the method at commercial scale. Van Oord's vessel Boreas installed all 52 of the farm's monopiles, three of them this way. The release gives no noise figures and says the data will be used to check models of underwater sound.
Achieved whenunderwater noise measured by an independent party is published for these piles, and the method installs the foundations of a whole wind farm.
Claimed by
Ecowende and Van Oord, April 2026
Open
01
Check piling noise forecasts
A public table of every pile driven for US offshore wind farms whose underwater sound was measured, giving pile diameter, hammer energy, water depth, the noise-reducing systems used, the level forecast before piling and the level measured 750 m away, with a test of how well the published German rules of thumb would have predicted the US results.
The briefCredit · unclaimed02
Spot gaps in bubble curtains
Software that reads drone video of a bubble curtain at the sea surface and maps where its ring of bubbles is thin or broken, so a crew can find a clogged or torn hose before the next pile is driven.
The briefCredit · unclaimed
Open 01 · the brief
Check piling noise forecasts
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A public table of every pile driven for US offshore wind farms whose underwater sound was measured, giving pile diameter, hammer energy, water depth, the noise-reducing systems used, the level forecast before piling and the level measured 750 m away, with a test of how well the published German rules of thumb would have predicted the US results.
Why it is still open
The German experience report, built on measurements of almost 2,000 piles, says noise-reducing systems are proven only for monopiles up to 8 m across in water up to 40 m deep, and that larger piles in deeper water "may require further development". At Vineyard Wind 1 off Massachusetts, the measuring firm JASCO found that "the received levels of the first piles exceeded the predicted levels," so 12 monopiles and a jacket were measured instead of the three planned. Germany's data sit in BSH's MarinEARS system, which BSH says holds "the world's largest data set on impact pile driving noise events"; it offers a public noise registry, but its data portal opens on a login page. The US reports are posted one by one as PDFs on NOAA Fisheries' site, and no table combining them was found.
Built with what exists
NOAA Fisheries' published sound field verification and monitoring reports for US projects; BSH's public noise registry; the relationships between pile size, depth and noise in Bellmann and colleagues' 2020 report; a language model that pulls figures from report tables and cites the page each came from (see the Note "Findings that cite notes"); Python.
First step
Copy the per-pile results of the Vineyard Wind 1 final verification report into a spreadsheet, next to the forecasts the same report lists.
Done looks like
At least 50 piles from at least four US projects in the table, each row linked to its report page, with the forecast error at 750 m given for each project and for the German rules of thumb.
Open 02 · the brief
Spot gaps in bubble curtains
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
Software that reads drone video of a bubble curtain at the sea surface and maps where its ring of bubbles is thin or broken, so a crew can find a clogged or torn hose before the next pile is driven.
Why it is still open
German approvals require bubble curtain hoses to be checked with side-scan sonar and, "if necessary," operation to be documented "by means of recordings with drones," and the 2020 German experience report describes no automatic reading of that footage. At Vineyard Wind 1 the crew adopted a new routine partway through: re-drilling the hoses' holes each time they were recovered, to clear them of sediment. A search of journal records found acoustic measurements and computer models of bubble curtains, but no method that reads images of them. The nearest open data is underwater video of a 31 m test curtain in MARIN's offshore basin, filmed in 2021 by an underwater drone and a fixed camera and published under CC0 by Dominik Krug and Simon Beelen; it was made to study the curtain's flow and sound, not to find gaps. A November 2026 web search found crews checking curtains by air pressure, air flow and sound, not by reading images.
Built with what exists
A consumer camera drone; the MARIN basin footage for a first look; a perforated air hose and an air compressor laid in a test basin, or in a harbor with permission; an open image-segmentation model such as Meta's Segment Anything 2; OpenCV; a pair of hydrophones to check that the gaps found let more sound through.
First step
Lay a 20 m perforated hose in a basin, block one tenth of its holes, and film the surface from a drone at three heights.
Done looks like
30 basin runs with known sections of hose blocked, publishing the share of blocked sections found and the false alarms; then, with footage shared by a wind farm developer, a comparison across at least 10 piles between the gaps found and the noise measured 750 m away.