UPWARDS · Project

Faster, Cheaper Wind Turbine Design Using High-Performance Simulation and Machine Learning

energyTestedTRL 5

Designing a wind turbine today is a bit like building a car without ever test-driving it — engineers rely on simplified models that miss how the wind, the blades, the tower, and nearby turbines all interact at once. UPWARDS built a supercomputer-powered simulation environment that models the entire turbine system together, including the messy real-world effects like turbulence from neighboring turbines and ocean waves. Think of it as a virtual wind tunnel for complete wind farms, powered by machine learning to run fast enough to be practical. The result is that turbine designers can test new blade shapes, predict noise problems, and spot material fatigue issues before bending a single piece of steel.

By the numbers

consortium partners across the full wind energy value chain

countries represented in the development consortium

industry partners validating the simulation tools

54%

industry participation ratio in the consortium

total project deliverables produced

5000

minimum website visits targeted for dissemination

The business problem

What needed solving

Wind turbine designers today rely on simplified models that simulate components in isolation — the blades separate from the tower, the wind field separate from the terrain. This leads to unexpected failures, inaccurate power predictions, and noise problems that only surface after expensive physical prototyping or worse, after installation. The industry needs a way to test complete turbine system behavior virtually, fast enough to fit into real design cycles.

The solution

What was built

The project delivered an open-source integrated simulation platform coupling aerodynamics, structural dynamics, acoustics, and material behavior into one system. Key concrete outputs include: an HPC-integrated simulation environment tested and ready to use, a machine learning pipeline for real-time turbine data analysis, and a Model Order Reduction Toolbox that speeds up detailed physics simulations. In total, 28 deliverables were produced across the 4.5-year project.

Audience

Who needs this

Wind turbine OEMs designing next-generation rotors and bladesOffshore wind farm developers needing accurate power yield and load predictionsBlade composite manufacturers dealing with fatigue and failure issuesWind farm acoustic consultants managing noise compliance for permittingHPC and simulation software vendors expanding into renewable energy

Business applications

Who can put this to work

Wind turbine manufacturing

enterprise

Target: Wind turbine OEMs and blade manufacturers

If you are a turbine manufacturer spending months on physical prototype testing and still getting unexpected blade failures in the field — this project developed an open-source integrated simulation platform that couples aerodynamics, structural loads, and acoustics in a single run. With 13 partners including 7 industry players validating it, this tool lets you test new rotor designs virtually and catch material failure risks before production.

Wind farm development and operations

mid-size

Target: Wind farm developers and owner-operators

If you are a wind farm developer struggling with energy yield predictions that miss real-world wake effects and terrain turbulence — this project built an HPC simulation environment that calculates the complex wind field including interactions from nearby turbines, waves, and terrain. More accurate force and power predictions mean better site planning and fewer underperforming assets after commissioning.

Engineering simulation software

any

Target: CFD and engineering software vendors

If you are a simulation software company looking to expand into wind energy applications — this project created a Model Order Reduction Toolbox that speeds up detailed physics simulations, plus machine learning algorithms integrated into a stream processing pipeline for real-time analysis. The modular, open-source platform from 9 countries of development could be the foundation for a commercial product or add-on module.

Frequently asked

Quick answers

What would it cost to adopt this simulation platform?

The serial integrated simulation platform was delivered as open-source software, meaning the code itself is free to use. However, running the full HPC-integrated version requires access to high-performance computing clusters, which carries infrastructure or cloud compute costs. Based on available project data, no licensing fees for the core platform were indicated.

Can this handle industrial-scale wind farm simulations?

Yes. The HPC-Integrated Simulation Environment was specifically built and tested on the Beehive computing cluster, designed to handle the computational load of full turbine system simulations. The Model Order Reduction Toolbox was developed specifically to increase simulation speed for industrial-scale use cases.

What is the IP situation — can we use this commercially?

The serial simulation platform was delivered as open-source to project partners. However, with 13 consortium partners including 7 industry organizations across 9 countries, the IP landscape for specific modules (particularly the ML algorithms and HPC optimizations) may involve shared ownership. Contact the coordinator SINTEF AS in Norway for licensing specifics.

How does this compare to existing wind simulation tools?

Unlike conventional tools that simulate components in isolation, this platform couples aerodynamics, acoustics, structural dynamics, and material behavior into one integrated run. The machine learning layer adds predictive capability that traditional CFD-only tools lack. The project had 28 deliverables covering the full physics chain from wind field to blade material response.

Is this ready to use today or still experimental?

The key deliverables were described as 'implemented, tested, and ready to use,' including the HPC simulation environment and the ML pipeline. The project closed in September 2022, so the tools have been available for over three years. However, as a research output, expect integration effort rather than a plug-and-play commercial product.

What specific problems can this predict that current tools cannot?

Based on the project objectives, the platform better predicts acoustic phenomena (noise complaints are a major permitting barrier), material fatigue issues in turbine blades, and wake effects from nearby turbines. It also models wave-structure interaction for offshore installations, which most current design tools handle poorly.

Consortium

Who built it

The UPWARDS consortium is unusually industry-heavy for a research project, with 7 out of 13 partners (54%) coming from industry — a strong signal that this work was designed with real-world application in mind. Led by SINTEF, one of Europe's largest independent research organizations based in Norway (a major offshore wind market), the project spans 9 countries including key wind energy markets like Denmark, Germany, the Netherlands, and Spain. The mix of 3 universities providing fundamental physics, 3 research institutes handling HPC and integration, and 7 industry partners ensuring practical relevance creates a credible path from simulation code to commercial tooling. The presence of partners from Argentina suggests offshore wind ambitions beyond Europe.

SINTEF ASCoordinator · NO
WAVESTONE LUXEMBOURG SAparticipant · LU
SIEMENS GAMESA RENEWABLE ENERGY ASparticipant · DK
SIEMENS INDUSTRY SOFTWARE SASthirdparty · FR
UNIVERSIDAD NACIONAL DEL LITORALparticipant · AR
UL SERVICES SPAIN SLparticipant · ES
SIEMENS INDUSTRY SOFTWARE NVparticipant · BE
IN EXTENSO INNOVATION CROISSANCEparticipant · FR
VON KARMAN INSTITUTE FOR FLUID DYNAMICSparticipant · BE
WAGENINGEN UNIVERSITYparticipant · NL
AALBORG UNIVERSITETparticipant · DK

How to reach the team

SINTEF AS is a major Norwegian research institute — their wind energy division is publicly listed. The project coordinator can be reached through SINTEF's website or via Google search for 'UPWARDS wind project coordinator SINTEF'.

Order a report on this consortium See SINTEF AS profile →

Next steps

Talk to the team behind this work.

Want an introduction to the UPWARDS team at SINTEF? SciTransfer can connect you with the right researcher for your specific simulation needs — whether blade design, acoustic compliance, or wind farm layout optimization.

Order a report on this topic More in Energy & Climate