Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Article
  • Published:

Artificial intelligence-aided wind plant optimization for nationwide evaluation of land use and economic benefits of wake steering


If clean energy pathways are to harness massive increases in wind power, innovations with broad geographic viability will be needed to support buildout in diverse locations. However, geodiversity in impact potential is seldom captured in technology assessment. Here we propose a scalable approach to plant-level optimization using artificial intelligence to evaluate land sparing and economic benefits of wake steering at more than 6,800 plausible onshore wind locations in the USA. This emerging controls strategy optimizes plant energy production by directing turbine wakes. On the basis of estimates from our artificial intelligence model trained on engineering wind flow simulations, co-optimizing plant layouts with wake steering can reduce land requirements by an average of 18% per plant (site-specific benefits range from 2% to 34%), subject to errors and uncertainties in the flow model, wind resource estimates, buildout scenario and geographic factors. According to model estimates, wake steering is predicted to increase power production during high-value (relatively low wind) periods, boosting the annual revenue of individual plants by up to US$3.7 million (equivalent to US$13,000 MW−1 yr−1) but producing negligible gains in some settings. Consideration of wake steering’s geographic potential reveals divergent nationwide prospects for improved economics and siting flexibility.

This is a preview of subscription content, access via your institution

Access options

Buy this article

Prices may be subject to local taxes which are calculated during checkout

Fig. 1: Nationwide modelling of wind power innovation impact potential.
Fig. 2: Nationwide benefits of wake steering.
Fig. 3: National land-use reductions and energy density benefits of wake steering.
Fig. 4: Drivers behind reduced cost of energy.
Fig. 5: Dynamic energy pricing and increased value capture through wake steering.

Similar content being viewed by others

Data availability

Wind plant simulation data from FLORIS, which were used for training and testing the WPGNN model, have been made publicly available through the Open Energy Data Intiative at (ref. 105). Data regarding the projected buildouts can found at Site-specific atmospheric data are obtained from NREL’s WIND Toolkit, which is available at Lastly, pricing data is from NREL’s Cambium database, available at Additional data supporting the results are available upon request. Source data are provided with this paper.

Code availability

The code for the trained WPGNN model and the wind plant layout generator is available on GitHub at (ref. 54).


  1. Shukla, P. et al. IPCC Climate Change and Land: An IPCC Special Report on Climate Change, Desertification, Land Degradation, Sustainable Land Management, Food Security and Greenhouse Gas Fluxes in Terrestrial Ecosystems (IPCC, 2021).

  2. Weinand, J. M. et al. Exploring the trilemma of cost-efficiency, landscape impact and regional equality in onshore wind expansion planning. Adv. Appl. Energy 7, 100102 (2022).

    Article  Google Scholar 

  3. Kitzing, L., Jensen, M. K., Telsnig, T. & Lantz, E. Multifaceted drivers for onshore wind energy repowering and their implications for energy transition. Nat. Energy 5, 1012–1021 (2020).

    Article  Google Scholar 

  4. Rinne, E., Holttinen, H., Kiviluoma, J. & Rissanen, S. Effects of turbine technology and land use on wind power resource potential. Nat. Energy 3, 494–500 (2018).

    Article  Google Scholar 

  5. Mai, T., Lopez, A., Mowers, M. & Lantz, E. Interactions of wind energy project siting, wind resource potential, and the evolution of the US power system. Energy 223, 119998 (2021).

    Article  Google Scholar 

  6. Harrison-Atlas, D., King, R. N. & Glaws, A. Machine learning enables national assessment of wind plant controls with implications for land use. Wind Energy 25, 618–638 (2022).

    Article  Google Scholar 

  7. Lopez, A. et al. Land use and turbine technology influences on wind potential in the United States. Energy 223, 120044 (2021).

    Article  Google Scholar 

  8. Beiter, P. et al. Wind power costs driven by innovation and experience with further reductions on the horizon. Wiley Interdiscip. Rev.: Energy Environ. 10, e398 (2021).

    Google Scholar 

  9. US EIA. Monthly Energy Review–March 2021 DOE/EIA 0035(2021/3) (2021).

  10. Wiser, R. et al. Wind vision: a new era for wind power in the United States. Electr. J. 28, 120–132 (2015).

    Article  Google Scholar 

  11. Larson, E. et al. Net-zero America: Potential Pathways, Infrastructure, and Impacts interim report (Princeton University, 2020).

  12. Cherp, A., Vinichenko, V., Tosun, J., Gordon, J. A. & Jewell, J. National growth dynamics of wind and solar power compared to the growth required for global climate targets. Nat. Energy 6, 742–754 (2021).

    Article  Google Scholar 

  13. Denholm, P. et al. Examining Supply-side Options to Achieve 100% Clean Electricity by 2035 (NREL, 2022).

  14. Trainor, A. M., McDonald, R. I. & Fargione, J. Energy sprawl is the largest driver of land use change in United States. PloS ONE 11, e0162269 (2016).

    Article  Google Scholar 

  15. Turkovska, O. et al. Land-use impacts of Brazilian wind power expansion. Environ. Res. Lett. 16, 024010 (2021).

    Article  Google Scholar 

  16. Cole, W. J. et al. Quantifying the challenge of reaching a 100% renewable energy power system for the United States. Joule 5, 1732–1748 (2021).

    Article  Google Scholar 

  17. Wiser, R. et al. Expert elicitation survey predicts 37% to 49% declines in wind energy costs by 2050. Nat. Energy 6, 555–565 (2021).

    Article  Google Scholar 

  18. Veers, P. et al. Grand challenges in the science of wind energy. Science 366, eaau2027 (2019).

    Article  Google Scholar 

  19. Dykes, K. et al. Results of IEA Wind TCP Workshop on a Grand Vision for Wind Energy Technology (NREL, 2019).

  20. Howland, M. F. et al. Collective wind farm operation based on a predictive model increases utility-scale energy production. Nat. Energy 7, 818–827 (2022).

    Article  Google Scholar 

  21. Gebraad, P. M. et al. Wind plant power optimization through yaw control using a parametric model for wake effects-a CFD simulation study. Wind Energy 19, 95–114 (2016).

    Article  Google Scholar 

  22. Fleming, P. A. et al. Evaluating techniques for redirecting turbine wakes using SOWFA. Renewable Energy 70, 211–218 (2014).

    Article  Google Scholar 

  23. Adaramola, M. & Krogstad, P.-Å Experimental investigation of wake effects on wind turbine performance. Renewable Energy 36, 2078–2086 (2011).

    Article  Google Scholar 

  24. Bastankhah, M. & Porté-Agel, F. Experimental and theoretical study of wind turbine wakes in yawed conditions. J. Fluid Mech. 806, 506–541 (2016).

    Article  Google Scholar 

  25. Bartl, J. et al. Wind tunnel experiments on wind turbine wakes in yaw: effects of inflow turbulence and shear. Wind Energy Sci. 3, 329–343 (2018).

    Article  Google Scholar 

  26. Fleming, P. et al. Field test of wake steering at an offshore wind farm. Wind Energy Sci. 2, 229–239 (2017).

    Article  Google Scholar 

  27. Fleming, P. et al. Initial results from a field campaign of wake steering applied at a commercial wind farm—part 1. Wind Energy Sci. 4, 273–285 (2019).

    Article  Google Scholar 

  28. Fleming, P. et al. Continued results from a field campaign of wake steering applied at a commercial wind farm—part 2. Wind Energy Sci. 5, 945–958 (2020).

    Article  Google Scholar 

  29. Howland, M. F., Lele, S. K. & Dabiri, J. O. Wind farm power optimization through wake steering. Proc. Natl Acad. Sci. 116, 14495–14500 (2019).

    Article  Google Scholar 

  30. Gebraad, P., Thomas, J. J., Ning, A., Fleming, P. & Dykes, K. Maximization of the annual energy production of wind power plants by optimization of layout and yaw-based wake control. Wind Energy 20, 97–107 (2017).

    Article  Google Scholar 

  31. Jiménez, Á., Crespo, A. & Migoya, E. Application of a LES technique to characterize the wake deflection of a wind turbine in yaw. Wind Energy 13, 559–572 (2010).

    Article  Google Scholar 

  32. Sanz Rodrigo, J. et al. Mesoscale to microscale wind farm flow modeling and evaluation. Wiley Interdiscip. Rev.: Energy Environ. 6, e214 (2017).

    Google Scholar 

  33. Binz, C. & Truffer, B. Global innovation systems—a conceptual framework for innovation dynamics in transnational contexts. Res. Policy 46, 1284–1298 (2017).

    Article  Google Scholar 

  34. Fleming, P. A., Ning, A., Gebraad, P. M. & Dykes, K. Wind plant system engineering through optimization of layout and yaw control. Wind Energy 19, 329–344 (2016).

    Article  Google Scholar 

  35. Song, D. et al. Multi-site and multi-objective optimization for wind turbines based on the design of virtual representative wind farm. Energy 252, 123995 (2022).

    Article  Google Scholar 

  36. Kondziella, H. & Bruckner, T. Flexibility requirements of renewable energy based electricity systems—a review of research results and methodologies. Renewable Sustain. Energy Rev. 53, 10–22 (2016).

    Article  Google Scholar 

  37. Sprague, M. A., Ananthan, S., Vijayakumar, G. & Robinson, M. ExaWind: a multifidelity modeling and simulation environment for wind energy. J. Phys. Conf. Ser. 1452, 012071 (2020).

    Article  Google Scholar 

  38. Wu, G. C. et al. Low-impact land use pathways to deep decarbonization of electricity. Environ. Res. Lett. 15, 074044 (2020).

    Article  Google Scholar 

  39. Rand, J. & Hoen, B. Thirty years of North American wind energy acceptance research: what have we learned? Energy Res. Social Sci. 29, 135–148 (2017).

    Article  Google Scholar 

  40. Wang, S. & Wang, S. Impacts of wind energy on environment: a review. Renewable Sustain. Energy Rev. 49, 437–443 (2015).

    Article  Google Scholar 

  41. Rehbein, J. A. et al. Renewable energy development threatens many globally important biodiversity areas. Glob. Change Biol. 26, 3040–3051 (2020).

    Article  Google Scholar 

  42. Enevoldsen, P. & Jacobson, M. Z. Data investigation of installed and output power densities of onshore and offshore wind turbines worldwide. Energy Sustain. Dev. 60, 40–51 (2021).

    Article  Google Scholar 

  43. Stehly, T., Beiter, P. & Duffy, P. 2019 Cost of Wind Energy Review (NREL, 2020).

  44. Jorgenson, J., Awara, S., Stephen, G. & Mai, T. A systematic evaluation of wind’s capacity credit in the western United States. Wind Energy 24, 1107–1121 (2021).

    Article  Google Scholar 

  45. Hirth, L. The market value of variable renewables: the effect of solar wind power variability on their relative price. Energy Econ. 38, 218–236 (2013).

    Article  Google Scholar 

  46. Dalla Riva, A., Hethey, J. & Vītiņa, A. Impacts of wind turbine technology on the system value of wind in Europe (IEA, 2017).

  47. Wiser, R. H. et al. Land-based Wind Market Report: 2021 Edition (LBNL, 2021).

  48. Short, W., Blair, N., Sullivan, P. & Mai, T. ReEDS Model Documentation: Base Case Data and Model Description (NREL, 2009).

  49. SOWFA: Simulator For Wind Farm Applications; (NREL, 2022).

  50. NREL. FLORIS version 2.3.0. Github (2021).

  51. Simley, E., Fleming, P., King, J. & Sinner, M. Wake Steering Wind Farm Control with Preview Wind Direction Information (NREL, 2021).

  52. Bensason, D. et al. Evaluation of the potential for wake steering for US land-based wind power plants. J. Renewable Sustain. Energy 13, 033303 (2021).

    Article  Google Scholar 

  53. Stanley, A. P., Bay, C., Mudafort, R. & Fleming, P. Fast yaw optimization for wind plant wake steering using boolean yaw angles. Wind Energy Sci. 7, 741–757 (2022).

    Article  Google Scholar 

  54. Harrison-Atlas, D., Glaws, A., King, R. & Lantz, E. Wind Plant Graph Neural Network (WPGNN) and Plant Layout Generator (PLayGen) version 1.0.0. Github (2022).

  55. Jakeman, J. D., Gorodetsky, A., Eldred, M. S., Geraci, G. & Smith, T. M. MFNETS: Multi-fidelity Data-driven Networks for Data Analysis (Sandia National Lab, 2021).

  56. Moriarty, P. et al. American WAKe ExperimeNt (AWAKEN) (LLNL, 2020).

  57. Hamilton, N., Doubrawa, P., Naughton, J. & Kelley, C. Rotor aerodynamics, aeroelastics, and wake (RAAW) campaign overview. Bull. Am. Phys. Soc. Vol. 67, No. 19, Abstract L14.00003 (2022).

  58. Harrison-Atlas, D., Lopez, A. & Lantz, E. Dynamic land use implications of rapidly expanding and evolving wind power deployment. Environ. Res. Lett. 17, 044064 (2022).

    Article  Google Scholar 

  59. Dai, K., Bergot, A., Liang, C., Xiang, W.-N. & Huang, Z. Environmental issues associated with wind energy—a review. Renewable Energy 75, 911–921 (2015).

    Article  Google Scholar 

  60. Bruck, M., Sandborn, P. & Goudarzi, N. A Levelized Cost of Energy (LCOE) model for wind farms that include Power Purchase Agreements (PPAs). Renewable Energy 122, 131–139 (2018).

    Article  Google Scholar 

  61. Gagnon, P., Frazier, W., Cole, W. & Hale, E. Cambium Documentation: Version 2021 (NREL, 2021).

  62. Beiter, P. et al. Toward global comparability in renewable energy procurement. Joule 5, 1485–1500 (2021).

    Article  Google Scholar 

  63. Kanev, S., Bot, E. & Giles, J. Wind farm loads under wake redirection control. Energies 13, 4088 (2020).

    Article  Google Scholar 

  64. Schmidt, J., Requate, N. & Vollmer, L. Wind farm yield and lifetime optimization by smart steering of wakes. J. Phys. Conf. Ser. 1934, 012020 (2021).

    Article  Google Scholar 

  65. Dana, S., Ivanov, H. & Doubrawa, P. Lifetime fatigue response due to wake steering on a pair of utility-scale wind turbines. J. Phys. Conf. Ser. 2265, 022106 (2022).

    Article  Google Scholar 

  66. Bastankhah, M. & Porté-Agel, F. A new analytical model for wind-turbine wakes. Renewable Energy 70, 116–123 (2014).

    Article  Google Scholar 

  67. Annoni, J. et al. Analysis of control-oriented wake modeling tools using lidar field results. Wind Energy Sci. 3, 819–831 (2018).

    Article  Google Scholar 

  68. Brugger, P. et al. Lidar measurements of yawed-wind-turbine wakes: characterization and validation of analytical models. Wind Energy Sci. 5, 1253–1272 (2020).

    Article  Google Scholar 

  69. Doekemeijer, B. M., Simley, E. & Fleming, P. Comparison of the Gaussian wind farm model with historical data of three offshore wind farms. Energies 15, 1964 (2022).

    Article  Google Scholar 

  70. King, J., Fleming, P., King, R., Martínez-Tossas, L. A., Bay, C. J., Mudafort, R. & Simley, E. Controls-oriented model for secondary effects of wake steering. Wind Energy Sci. 6, 701–714 (2021).

    Article  Google Scholar 

  71. Bay, C. J. et al. Addressing deep array effects and impacts to wake steering with the cumulative-curl wake model. Wind Energy Sci. 8, 401–419 (2023).

    Article  Google Scholar 

  72. Simley, E., Fleming, P. & King, J. Field validation of wake steering control with wind direction variability. J. Phys. Conf. Ser. 1452, 012012 (2020).

    Article  Google Scholar 

  73. Simley, E., Fleming, P. & King, J. Design and analysis of a wake steering controller with wind direction variability. Wind Energy Sci. 5, 451–468 (2020).

    Article  Google Scholar 

  74. Farrell, A. et al. Design and analysis of a wake model for spatially heterogeneous flow. Wind Energy Sci. 6, 737–758 (2021).

    Article  Google Scholar 

  75. Lundquist, J., DuVivier, K., Kaffine, D. & Tomaszewski, J. Costs and consequences of wind turbine wake effects arising from uncoordinated wind energy development. Nat. Energy 4, 26–34 (2019).

    Article  Google Scholar 

  76. Maas, O. & Raasch, S. Wake properties and power output of very large wind farms for different meteorological conditions and turbine spacings: a large-eddy simulation case study for the German Bight. Wind Energy Sci. 7, 715–739 (2022).

    Article  Google Scholar 

  77. Lopez, A. et al. Impact of siting ordinances on land availability for wind and solar development. Nat. Energy 8, 1034–1043 (2023).

    Article  Google Scholar 

  78. Unnewehr, J. F., Jalbout, E., Jung, C., Schindler, D. & Weidlich, A. Getting more with less? Why repowering onshore wind farms does not always lead to more wind power generation—a German case study. Renewable Energy 180, 245–257 (2021).

    Article  Google Scholar 

  79. Koziel, S. & Leifsson, L. Surrogate-based Modeling and Optimization (Springer, 2013).

  80. Jiang, P., Zhou, Q. & Shao, X. Surrogate Model-based Engineering Design and Optimization (Springer, 2020).

  81. LeCun, Y., Bengio, Y. & Hinton, G. Deep learning. Nature 521, 436–444 (2015).

    Article  Google Scholar 

  82. Goodfellow, I., Bengio, Y. & Courville, A. Deep Learning (MIT Press, 2016).

  83. Rolnick, D. et al. Tackling climate change with machine learning. ACM Comput. Surv. 55, 1–96 (2022).

    Article  Google Scholar 

  84. Hulsman, P., Andersen, S. J. & Göçmen, T. Optimizing wind farm control through wake steering using surrogate models based on high-fidelity simulations. Wind Energy Sci. 5, 309–329 (2020).

    Article  Google Scholar 

  85. Battaglia, P. W. et al. Relational inductive biases, deep learning, and graph networks. Preprint at (2018).

  86. Park, J. & Park, J. Physics-induced graph neural network: an application to wind-farm power estimation. Energy 187, 115883 (2019).

    Article  Google Scholar 

  87. DeepMind Graph Nets version 1.1.0. Github (2020).

  88. Denholm, P., Hand, M., Jackson, M. & Ong, S. Land Use Requirements of Modern Wind Power Plants in the United States (NREL, 2009).

  89. Damiani, R. et al. Assessment of wind turbine component loads under yaw-offset conditions. Wind Energy Sci. 3, 173–189 (2018).

    Article  Google Scholar 

  90. Fleming, P. et al. Simulation comparison of wake mitigation control strategies for a two-turbine case. Wind Energy 18, 2135–2143 (2015).

    Article  Google Scholar 

  91. Fleming, P. et al. A simulation study demonstrating the importance of large-scale trailing vortices in wake steering. Wind Energy Sci. 3, 243–255 (2018).

    Article  Google Scholar 

  92. Campagnolo, F. & Bottasso, C. L. On the effectiveness of one-sided wake steering-a wind tunnel study with dynamic direction changes. In 2021 American Control Conference (ACC) 3070–3075 (IEEE, 2021).

  93. Bortolotti, P. et al. IEA Wind TCP Task 37: Systems Engineering in Wind Energy-WP2. 1 Reference Wind Turbines (NREL, 2019).

  94. Kruskal, J. B. On the shortest spanning subtree of a graph and the traveling salesman problem. Proc. Am. Math. Soc. 7, 48–50 (1956).

    Article  MathSciNet  Google Scholar 

  95. Hagberg, A. A., Schult, D. A. & Swart, P. J. Exploring network structure, dynamics, and function using NetworkX. In Proc. 7th Python in Science Conference 11–15 (2008).

  96. Kingma, D. P. & Ba, J. Adam: a method for stochastic optimization. Preprint at (2014).

  97. Draxl, C., Clifton, A., Hodge, B.-M. & McCaa, J. The Wind Integration National Dataset (WIND) toolkit. Appl. Energy 151, 355–366 (2015).

    Article  Google Scholar 

  98. Porté-Agel, F., Bastankhah, M. & Shamsoddin, S. Wind-turbine and wind-farm flows: a review. Boundary Layer Meteorol. 174, 1–59 (2020).

    Article  Google Scholar 

  99. Cole, W., Carag, V. & Gagnon, P. 2021 Standard Scenarios Report: A US Electric Sector Outlook (NREL, 2021).

  100. Nocedal, J. & Wright, S. J. Numerical Optimization (Springer, 1999).

  101. Pryor, S. C., Shepherd, T. J. & Barthelmie, R. J. Interannual variability of wind climates and wind turbine annual energy production. Wind Energy Sci. 3, 651–665 (2018).

    Article  Google Scholar 

  102. Eberle, A., Roberts, J. O., Key, A., Bhaskar, P. & Dykes, K. L. NREL’s Balance-of-system Cost Model for Land-based Wind (NREL, 2019).

  103. Virtanen, P. et al. SciPy 1.0: fundamental algorithms for scientific computing in Python. Nat. Methods 17, 261–272 (2020).

    Article  Google Scholar 

  104. TIGER/Line Shapefiles (US Census Bureau, 2023);

  105. Dakota, R., Glaws, A., King, R., & Harrison-Atlas, D. Flow Redirection and Induction in Steady State (FLORIS) wind plant power production data sets. Open Energy Data Initiative (2021).

Download references


We are grateful to P. Brown, T. Mai, P. Veers, O. Roberts, A. Lopez, P. Fleming and M. Sprague for feedback on earlier drafts of this work. This work was authored by the National Renewable Energy Laboratory, operated by Alliance for Sustainable Energy, LLC, for the US Department of Energy (DOE) under contract number DE-AC36-08GO28308. Funding provided by the US Department of Energy Office of Energy Efficiency and Renewable Energy Wind Energy Technologies Office. The research was performed using computational resources sponsored by the Department of Energy’s Office of Energy Efficiency and Renewable Energy and located at the National Renewable Energy Laboratory. The views expressed in the article do not necessarily represent the views of the DOE or the US Government. The US Government retains and the publisher, by accepting the article for publication, acknowledges that the US Government retains a non-exclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this work, or allow others to do so, for US Government purposes.

Author information

Authors and Affiliations



D.H.-A. contributed to conceptualization, methodology, software, formal analysis, investigation, visualization, project administration, supervision, writing (original draft, review and editing) and funding acquisition. A.G. contributed to methodology, software, data curation, formal analysis, validation, investigation, visualization and writing (original draft, review and editing). R.N.K. contributed to methodology, investigation, supervision and writing (original draft, review and editing). E.L. contributed to conceptualization, writing (original draft, review and editing) and funding acquisition.

Corresponding author

Correspondence to Dylan Harrison-Atlas.

Ethics declarations

Competing interests

This work was performed by the authors while employed by the National Renewable Energy Laboratory.

Peer review

Peer review information

Nature Energy thanks the anonymous reviewers for their contribution to the peer review of this work.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data

Extended Data Fig. 1 Potential expansion of wind energy under a decarbonized power sector and associated buildout characteristics.

Spatial distribution of onshore wind plant buildouts (n=6,862, capacity=720 GW) projected under our focal scenario5, which captures plausible changes to the bulk power system with a goal of reducing 95% of carbon emissions from the energy sector by 2050 (a). Under energy scenarios that rely heavily on wind power, like the one considered here, technologies must be viable in diverse settings. As captured by the breakdown of buildouts for this scenario, important characteristics of future buildouts include plant size (b), quality of wind resource (c), and land constraint (d). Using the projected plant sites as meaningful growth targets that reflect the scope and magnitude of possible future wind energy deployment, we perform a series of plant layout optimization studies to evaluate impacts of wake steering on land use and economic objectives. State boundaries from US Census Bureau104. Locations of water bodies from NaturalEarth (

Extended Data Fig. 2 Randomly selected examples from test set A showing WPGNN predictive error for turbine and plant power output.

Each example depicts a unique plant layout and inflow condition (arrows indicate the inflow direction). Differences between WPGNN predictions and FLORIS predictions for plant power production are quantified using the relative root mean squared error (RRMSE) metric. Examples contained in test set A are used exclusively for model validation. Note that all of the turbines shown in these specific cases have randomly sampled yaw angles. The high accuracy of WPGNN across the test set, as illustrated here for a few examples, demonstrates its ability to generalize well to new plant layouts and different inflow conditions.

Extended Data Fig. 3 Predictive error of the WPGNN for plant power against test set A visualized parametrically across key dimensions including wind speed, turbulence intensity, layout type, turbine spacing and plant capacity.

For the continuous plots, the solid lines represent the median error, the colored regions show the 25-75% range of errors, and the dotted lines show the 1.5x interquartile range. Similarly, the box plots display the median error line in each box, the 25-75% box limits, and the 1.5x interquartile range as whiskers. The error rates are plotted separately for two cases, one where turbines are randomly yawed and another where yawing has been optimized. Random yaws are used in training and testing data to ensure that WPGNN learns to generalize about how wake effects propagate through a wind plant via complex turbine interactions. Overall, the plant power relative errors are small, except as the wind speed approaches the lower cut-in limit (4 m/s). Here, relative errors are driven up by vanishingly small normalization values. Rug plots along the x-axes show the relative frequency of training data along these dimensions.

Source data

Extended Data Fig. 4 Predictive error of the WPGNN for turbine-level power against test set A visualized parametrically across key dimensions including wind speed, turbulence intensity, layout type and turbine spacing.

For the continuous plots, the solid lines represent the median error, the colored regions show the 25-75% range of errors, and the dotted lines show the 1.5x interquartile range. Similarly, the box plots display the median error line in each box, the 25-75% box limits, and the 1.5x interquartile range as whiskers. The error rates are plotted separately for two cases, one where turbines are randomly yawed and another where yawing has been optimized. Random yaws are used in training and testing data to ensure that WPGNN learns to generalize about how wake effects propagate through a wind plant via complex turbine interactions. The majority of the turbine relative errors are within a couple of percent. Rug plots along the x-axes show the relative frequency of training data along these dimensions.

Source data

Extended Data Fig. 5 Randomly selected examples, taken from test set B, of wind plant layouts and corresponding accuracy of the WPGNN predictions for complete response surfaces.

For a given layout (a) and (d), the response surfaces (b) and (e) visualize predicted plant power output for all possible combinations of inflow direction and wind speed (shown here above turbine cut-in speed of 4 m/s). The cases shown here assume reference controls (that is, no wake steering via yaw). Predictive error of the WPGNN (evaluated relative to FLORIS) across the response surfaces is captured in (c) and (f). As predicted, maximal waking occurs for the plant in (a) when inflows have greatest alignment with the orientation of the turbine strings, at approximately 30 and 200 degrees in this case. For both visualized scenarios, relative errors are within an acceptable range of a couple percent with the largest errors occurring as the plant approaches the lower bound cut-in threshold.

Extended Data Fig. 6 Predictive error for the relative improvements obtained from the WPGNN-based optimization using test set C across the three objectives.

The box plots display the means as scatter points, the median lines in each box, the 25-75% box limits, and the 1.5x interquartile range as whiskers. Note that for the validation of the land area minimization objective, we report errors in predictions of energy density as the more meaningful metric for evaluating WPGNN outputs since the plant footprint is not dependent on FLORIS power outputs. Errors are obtained by computing FLORIS power outputs for 10% of the output designs from the three optimization studies and comparing to the WPGNN reported improvements. The 10% of cases were chosen to span all of the geographic regions and a wide range of plant capacities. The small discrepancy between the WPGNN-predicted improvements and the target FLORIS quantities provide confidence in the scalable surrogate modeling approach.

Source data

Extended Data Fig. 7 Stylized examples of the four canonical wind plant layout types generated by the Plant Layout Generator (PLayGen).

For a given layout type and number of turbines, PLayGen is able to generate randomized plant layouts that are consistent with a specified average turbine spacing parameterization. Hypothetical wind plants developed by PLayGen are used as inputs to train and test the Wind Plant Graph Neural Network (WPGNN).

Supplementary information

Supplementary Information

Supplementary Figs. 1–4 and Table 1.

Source data

Source Data Fig. 4

Statistical source data for Fig. 4b.

Source Data Extended Data Fig. 3

Statistical source data for Extended Data Fig. 3.

Source Data Extended Data Fig. 4

Statistical source data for Extended Data Fig. 4.

Source Data Extended Data Fig. 6

Statistical source data for Extended Data Fig. 6.

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Harrison-Atlas, D., Glaws, A., King, R.N. et al. Artificial intelligence-aided wind plant optimization for nationwide evaluation of land use and economic benefits of wake steering. Nat Energy (2024).

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI:


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing