BRENNEN

Industries / Energy, water and waste / Power generation / Wind turbines

Blades

Blade making designs and molds the long composite blades that catch the wind, from glass or carbon fiber, resin and a light core of wood or foam, and tests them before they fly.

5

Milestones achieved, from 1957 to 2020.

2

To watch for: announced, not yet confirmed outside.

2

Open deliverables. Nobody has shipped them.

The road

AchievedWatch forOpen

Achieved

  1. 1957

    Glass fiber blades

    Ulrich Hütter's StGW-34, a two-bladed turbine 34 m across, was connected to the grid at a test field in the Swabian Alb in December 1957 and began testing in 1958. Its blades were made of glass-fiber-reinforced plastic, and Hütter anchored them with what became known as the Hütter flange: glass fibers running down the blade were wrapped around the bolt holes of a wide flange at the root. The machine had long outages, threw a blade in 1961 and was scrapped about a decade after it started. The University of Stuttgart calls it "the original model of modern wind turbines," and the wind historian Paul Gipe writes that "Hütter gave us the modern wind turbine blade."

    Lessona test machine that failed still passed on its best part, the light glass-fiber blade and its root joint.

    wind-works.orgifb.uni-stuttgart.de

    Credit

    Ulrich Hütter

  2. 2001

    Blades cast in one piece

    Many blades are built from shells glued together along their edges, and Sandia National Laboratories' 2013 survey lists poor bonds at the leading and trailing edges among the common defects. A patent with a priority date of 13 November 2001, whose inventors include Henrik Stiesdal, describes making a blade "in a closed process and mainly in one piece without any glue joints," with resin infused around a mold core inside a closed mold. Siemens Gamesa still makes blades this way: in 2021 it cast its first 108 m B108 blade in one piece.

    patents.google.comoffshorewind.bizosti.gov

    Credit

    Siemens (patent inventors including Henrik Stiesdal and Peder Bay Enevoldsen); Siemens Gamesa

  3. 2013

    Defects traced to the factory

    Sandia National Laboratories and the US Department of Energy started the Blade Reliability Collaborative in 2010 to find out why blades fail. Its 2013 report says blade repair contractors in the US "report that as much as 80% of the blades they repair have never been operated," and that some wind farms had to replace every blade after a defect was found across the fleet. It cites wrinkles in the laminate (the stacked layers of fiber), bond lines and out-of-plane bending of the layers as the leading causes of blade failures in one study. Montana State University then measured wrinkles and porosity in failed commercial blades and tested coupons made with the same flaws.

    Lessonmuch blade trouble starts in the mold, before the blade has turned once.

    osti.govosti.gov

    Credit

    Sandia National Laboratories (Thomas Ashwill, Alistair Ogilvie, Joshua Paquette); Montana State University

  4. 2020

    Balsa demand outruns plantations

    Many blades use balsa wood as a light core between their fiber skins. InSight Crime reported in 2021, from Ecuador's environment ministry data, that seizures of illegal balsa in Ecuador rose from 700 cubic meters in 2019 to 1,973 in 2020, as Chinese and US buyers rushed to finish wind farms before subsidies ended; a local consultant said a balsa tree sold for 5 dollars in 2019 and up to 45 in 2020. In 2024 the Environmental Investigation Agency (EIA), a campaign group, reported that Ecuador grows over 90% of the world's balsa, that exporters say blade cores need wood of 120 to 150 kg per cubic meter, and that they reach this by mixing lighter plantation wood with wild wood from the Amazon, in ratios they put at between 50-50 and 70-30. GE Vernova, TPI Composites, 3A Composites Core Materials and Ecuabalpro denied wrongdoing in replies to EIA.

    Lessona density specification says nothing about where the wood was cut.

    insightcrime.orgeia.org

    Credit

    reporting by InSight Crime and Primicias; investigation by the Environmental Investigation Agency

  5. 2020

    Longer than the test hall

    A blade must pass full-scale tests before a certifier signs it off: a static test bends it once to its highest design load, and a fatigue test bends it back and forth many times over to stand in for years of service. LM Wind Power's 107 m blade for GE's Haliade-X turbine was tested twice, at the Wind Technology Testing Center in Boston and at ORE Catapult in Blyth, England, and received its component certificate from TÜV Nord in November 2020. ORE Catapult said it applied loads "50 per cent greater than ever previously applied in a blade test," and the blade's tip had to be cut off so the test hall door would close.

    Lessontest halls have to grow with the blades they prove.

    offshorewind.biz

    Credit

    LM Wind Power; GE Renewable Energy; ORE Catapult; Wind Technology Testing Center; TÜV Nord

Watch for

  1. A 153-meter blade on test

    Dongfang Electric says three blades 153 m long and weighing 83.5 tonnes each, for a 26 MW offshore turbine it is developing, finished all static tests in May 2025 and were shipped from Yantai on 1 August 2025 for fatigue testing.

    Achieved whenthe fatigue and post-fatigue tests are finished and a certifier issues a component certificate that is made public.

    offshorewind.biz

    Claimed by

    Dongfang Electric Corporation, through the Xinhua News Agency, August 2025

  2. Sensors that steer resin infusion

    The TURBO project, led by Siemens Gamesa with 12 partners and funded by Horizon Europe and Innovate UK, built a system that watches resin flow while a blade is cast and adjusts the process to prevent dry spots, plus a camera system using thermal imaging and optical coherence tomography (a light-based scan beneath the surface) to find defects early. The UK's National Composites Centre says nearly every blade needs some rework today and that the approach "is projected to reduce composite blade defects by up to 80%."

    Achieved whena blade factory publishes measured defect or rework rates from production before and after using the system, checked by an outside party.

    nccuk.com

    Claimed by

    National Composites Centre for the TURBO consortium, June 2026

Open

  1. 01

    Photos of wrinkled layups

    An open set of photos and 3D scans of dry glass fabric laid in a curved, blade-shaped mold, some smooth and some with wrinkles whose height and length are measured by hand, and a published test of how well current vision-language models and a small trained detector find the wrinkles before resin is pulled in.

    The brief
    Credit · unclaimed
  2. 02

    Tell wild balsa from plantation

    A lab test that tells balsa grown on plantations from balsa cut in natural forest, using density, near-infrared spectra and wood structure under a microscope (with stable isotope ratios from a commercial lab if needed), checked on samples of known origin and then applied to sticks cut from the cores of scrapped blades.

    The brief
    Credit · unclaimed

Open 01 · the brief

Photos of wrinkled layups

Credit · unclaimed

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

What to build

An open set of photos and 3D scans of dry glass fabric laid in a curved, blade-shaped mold, some smooth and some with wrinkles whose height and length are measured by hand, and a published test of how well current vision-language models and a small trained detector find the wrinkles before resin is pulled in.

Why it is still open

Wrinkles are among the leading causes of blade failure in Sandia's survey, and the University of Southern Denmark says the layup is still placed by hand and takes up to 50% of the time a blade mold is in use. Industry is working on it in-house: the Siemens Gamesa-led TURBO project inspects with thermal imaging and optical coherence tomography, a University of Southern Denmark doctoral project (part of the ALMA project with DTU, Airborne and Siemens Gamesa) builds defect models from simulations and lab layups, and an LM Wind Power patent granted in 2023 describes a device that measures wrinkles in the fiber layup of a blade shell. No image set from any of these could be found. The nearest open work is from Manyar and colleagues at the University of Southern California (2022): 10,000 synthetic and 1,000 real images of carbon fiber prepreg sheets (fabric already soaked in resin) with wrinkle-like defects, and a detector that scored 0.98 mean average precision on their production images. It covers small carbon sheets, not dry glass fabric in a curved blade mold, and it finds wrinkles without measuring their height. Five searches of Zenodo found no open image set of blade layups.

osti.govsdu.dknccuk.compatents.google.comdoi.orggithub.com

Built with what exists

A 2 to 3 m curved mold made of foam or plywood; offcuts of the non-crimp glass fabric blades use; a phone camera and a low-cost structured-light or laser line scanner; a depth gauge; the USC group's open synthetic-image pipeline as a starting point; current vision-language models and an open object detector such as Ultralytics YOLO.

  1. First step

    Lay 20 stacks of fabric in a curved mold, put wrinkles of known size into half of them, and photograph them under shop lighting.

  2. Done looks like

    At least 2,000 labeled images from 100 or more layups, each wrinkle's height measured, published under an open license, with each model's detection rate by wrinkle height at a stated false-alarm rate, set beside the rate of two experienced laminators checking the same layups by eye.

Open 02 · the brief

Tell wild balsa from plantation

Credit · unclaimed

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

What to build

A lab test that tells balsa grown on plantations from balsa cut in natural forest, using density, near-infrared spectra and wood structure under a microscope (with stable isotope ratios from a commercial lab if needed), checked on samples of known origin and then applied to sticks cut from the cores of scrapped blades.

Why it is still open

EIA reports that blade core blocks are glued up from sticks of plantation and wild balsa, mixed until the block reaches the density buyers want, that blade makers rarely asked where the wood came from, and it calls for a traceability system for balsa. A test that reads origin from the wood itself would let a buyer check the mix. A Europe PMC search on Ochroma pyramidale (balsa's botanical name) found studies of plantation wood density but none on telling origin, and a web search in November 2026 found origin tests for other tropical timbers, by isotopes, genetics and trace elements, but none for balsa.

eia.org

Built with what exists

Reference samples of documented origin from plantations and from wood collections; a handheld near-infrared spectrometer; a balance and calipers for density; a microscope camera; a small classifier; an isotope lab for a subset.

  1. First step

    Gather 30 documented plantation samples and 30 from a wood collection with recorded forest origin, and measure their density and spectra.

  2. Done looks like

    A published accuracy for plantation against natural forest on at least 200 samples of known origin held out from training, and the method applied to sticks from the cores of 10 scrapped blades.