Meet the team: AIOLUS (12/09/24)

The University of Warwick’s ICSE researchers are aiming to tap into the power of AI to revolutionise wind farm management through AIOLUS.

The UK is powering ahead in wind energy, with ambitious plans to achieve 60 GW of offshore wind capacity and 30 GW of onshore wind energy by 2030. Setting up these targets is crucial as it will reduce our carbon emissions, allowing the UK to meet its energy and environmental goals. However, there is a problem on the horizon; current technologies for managing wind farms are not keeping up with their rapid growth. These current methods will struggle to support the growing complexity of modern wind farms and won’t be able to deal with the huge amounts of data being generated, undermining wind energy’s efficiency.

One major challenge for wind farms is something called the “wake effect”, where an upstream wind turbine slows down the wind and reduces the energy output for the turbines behind it. This causes between 5–20% energy production loss under conventional ‘greedy’ control strategies, where each turbine focuses solely on maximising its own power production. Additionally, existing methods of modelling wind farm dynamics often struggle to balance efficiency and accuracy, with high-quality simulations taking days to complete even on powerful computers. As wind farms expand, these issues become a bigger problem and risk hindering effective prediction, optimisation, and management. To overcome this, new control methods are needed to boost efficiency and reliability.

“Our goal with AIOLUS is to harness AI to revolutionise wind energy management, targeting a 10% reduction in operation and maintenance costs and a 3-5% increase in energy production—enough to power 1 million UK households based on the 2022 wind generation capacity”

– Professor Xiaowei Zhao, Control Engineering, The University of Warwick & Director of EPSRC Supergen Network+ in AI for Renewable Energy, AIOLUS Lead

New Solutions for Wind Farm Management

University of Warwick’s AIOLUS, named after the Greek god of wind ‘Aeolus’, introduces a solution designed to tackle the most pressing challenges in wind energy management. Through its innovative AI solution, AIOLUS aims to reshape the wind energy sector and deliver optimal wind farm performance with three key technologies:

• AI-Enhanced Wind Farm Simulations

Current wind farm simulation tools either fail to capture critical wind flow features or are too expensive for industrial applications. AIOLUS aims to address this with an AI-powered Wake Model that combines artificial intelligence and computational fluid dynamics. This approach is anticipated to slash simulation times by 1000x compared to current tools. These real-time simulations can run on personal laptops, removing the need for supercomputers and making wind farm simulations accessible to everyone.

• AI-Driven Control and Decision-Making

Existing solutions managing wake effects struggle with issues like adaptability, data inefficiencies, lack of generality and scalability. AIOLUS tackles these challenges with reinforcement learning algorithms, which are designed to handle complex tasks more effectively and ensure safety-critical operations. This will improve the feasibility and trustworthiness of the decisions being made.

• AI-based  Digital Twins for Wind Farms

The AI-powered digital twins create a virtual model of the wind farms, enabling real-time insights for predictive maintenance and smarter decision-making. These digital twins will revolutionise wind farm management, presenting advanced measurements and predictions to provide a complete picture of the wind flows. By merging wind flow measurements, physical knowledge, aerodynamics and deep learning, AIOLUS sets the stage for more reliable and efficient wind farm operations.

Advancing Wind Energy

By integrating AI controllers with digital twins, AIOLUS hopes to enhance wind farms’ ability to make smarter, real-time decisions, optimising management and significantly boosting power output. If this were to be rolled out across the UK, current projections estimate this technology could increase annual energy production in the UK by up to 5%, while simultaneously reducing operation and maintenance costs by 10%, lowering energy costs for UK households. AIOLUS could support the UK’s goals for a greener energy grid.

“The Manchester Prize is a crucial catalyst for AIOLUS. Its impact will expand as wind energy grows. We are proud to be a Manchester Prize finalist and be the first in Europe to develop intelligent wind farm control technology via deep reinforcement learning. Our team also introduced the world’s first digital twin specifically for wind farm flow systems, bridging a critical gap in digital twin technology.”

– Professor Xiaowei Zhao

Whether you’re from academia, industry, policymaking, or simply passionate about leveraging AI for the greater good, if you’d like to get in touch with the team, they’d welcome you to connect.

Connect with the team

“Together we can harness cutting-edge technology to create more efficient, sustainable, and cost-effective wind energy solutions.”

– Professor Xiaowei Zhao

Header image owned by ICSE research group at University of Warwick, created using MidJourney