According to an article in Nature Energy by researchers at the U.S. Department of Energy’s National Renewable Energy Laboratory (NREL), the wind industry stands to gain significant advantages by leveraging artificial intelligence (AI) for the design and deployment of wind plants.

These researchers developed an AI-driven surrogate model named the Wind Plant Graph Neural Network (WPGNN), which underwent training using simulations encompassing over 250,000 wind plant layouts generated randomly under diverse atmospheric conditions, plant configurations, and turbine operations. The simulation data originated from another NREL-developed model known as the FLOw Redirection and Induction in Steady State (FLORIS) tool. Subsequently, the AI utilized this data to determine the most optimal wind plant design. By employing AI, the calculation of optimal plant layouts and operational strategies for achieving various objectives, such as land footprint reduction or revenue maximization, is facilitated.

Wake steering

The study centered on wake steering, a strategy aimed at optimizing a plant's energy output by directing the wake generated by an upstream turbine away from a downstream turbine. Employing AI enabled the researchers to assess the effects of wake steering on three distinct objectives: land utilization, costs, and revenue.

While the benefits of wake steering have been previously demonstrated at the level of individual plants, prior studies have been constrained in spatial scale and in the variety of optimization objectives examined. The Wind Plant Graph Neural Network (WPGNN) utilized by the NREL team effectively modeled wake interactions as a directed graph, enabling a comprehensive exploration of optimal configurations for both turbine placement and nacelle yaw across a national wind energy portfolio.

Decarbonize power sector

The utilization of wind energy as a renewable power source is poised to play an increasingly vital role in the nation's efforts to decarbonize its power sector. However, hurdles persist, as certain communities have imposed restrictions on the placement of wind turbines. In an AI-guided scenario, researchers explored the potential deployment of 6,862 wind plant expansions nationwide, generating a cumulative 721 gigawatts of power, with the aim of achieving a 95% reduction in carbon emissions from the energy sector by 2050.

Potential expansion of wind energy under a decarbonized power sector and associated buildout characteristics.

Advantages of wake steering

The adoption of wake steering tactics could lead to an average 18% reduction in land requirements for future wind plants and up to a 60% reduction in some cases. This translates to national land savings of approximately 13,000 square kilometers, equivalent to 28% of the wind energy footprint in the United States.

Wake steering is valuable because merely spacing out turbines often fails to mitigate wake losses, and some wind plants lack the necessary space for further expansion. Moreover, wind plants optimized for wake steering could accommodate a denser concentration of turbines, addressing the concerns of local communities regarding land usage restrictions. By installing more turbines within a smaller footprint, developers could gain greater flexibility in site planning, potentially unlocking economies of scale for larger projects.

Furthermore, the researchers discovered that the implementation of wake steering consistently reduces the cost of energy for wind deployments. Leveraging AI, they identified regional disparities where this strategy would yield the greatest benefits.