Data-Driven Machine Learning for Wind Plant Flow Modeling

King, Ruth; Adcock, Christiane; Annoni, Jennifer; Dykes, Katherine

doi:10.1088/1742-6596/1037/7/072004

Cited by 11 publications

(9 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Bringing together physics-based models and data-driven techniques through data assimilation may improve the combined modeling accuracy for a variety of applications and may also reduce computational costs. An example of such an approach is the use of data to improve the fidelity of less-expensive physics-based models of plant flow for design and control applications (King et al 2018;Adcock et al 2018).…”

Section: Data-driven Modeling and Simulationmentioning

confidence: 99%

IEA Wind TCP: Results of IEA Wind TCP Workshop on a Grand Vision for Wind Energy Technology

Dykes

Veers

Lantz

et al. 2019

View full text Add to dashboard Cite

Research advanced erosion-resistant composite materials for use on the leading edge of wind turbine blades Research, development, and scale-up of thermoplastic resin systems to reduce wind turbine blade manufacturing cycle times, reduce blade costs, enable thermal welding of bond lines, increase recyclability, and enable advanced in-field blade repair techniques Research advanced thermoset materials that can be recovered and reused Research metals and other noncomposite materials for development of lightweight, high-strength, and high-stiffness turbine components Grand Challenge 2: Automation Develop comprehensive techno-economic models for automation and a virtual factory (or digital twin of a factory) to understand advantages and disadvantages of targeted automation for both wind turbine blade molding and finishing operations Identify and quantify all cost inputs for wind turbine component manufacturing (e.g., blades, towers) to identify production steps that would benefit from automation Develop automated production methods for the continuous manufacturing of one-piece wind turbine towers for on-site manufacturing Develop advanced design tools to optimize blade and tower design with respect to manufacturing with automation Research and develop automation technology able to locate complex wind turbine blade geometry in space to allow for effective automated finishing operations, such as sanding, drilling, and cutting Develop methods to automate the delivery, placement, and inspection of core material into blade skin molds during the composite laminate skin lay-up Integrate real-time inspection and quality control into automated production steps for wind turbine blades, towers, and other components Develop large-scale, low-cost, high-throughput, high-performance automated additive manufacturing technologies (e.g., 3D printing, automated fiber placement and tape layup, filament winding) for the production of towers, blade skins, blade spars, and other turbine components Develop embedded metrology and out-of-mold indexing technologies Grand Challenge 3: On-Site Manufacturing Develop robust composite materials for use in broad on-site environmental conditions Research the chemistry and processing of in-situ thermoplastic resin systems to enable thermally welded joints in an on-site manufacturing environment Develop targeted wind turbine blade finishing automation techniques designed to reduce the overall footprint of on-site finishing operations and minimize production facility floor space Research additive manufacturing to print wind turbine blade molds and tooling at on-site manufacturing locations IEA Wind TCP Task 11

show abstract

Section: Data-driven Modeling and Simulationmentioning

confidence: 99%

IEA Wind TCP: Results of IEA Wind TCP Workshop on a Grand Vision for Wind Energy Technology

Dykes

Veers

Lantz

et al. 2019

View full text Add to dashboard Cite

show abstract

“…In terms of novelty, the authors know of only two examples of data-driven turbulence modeling applied to wind farms, namely those from Adcock et al [31] and King et al [32]. The papers employ quite a different approach to us, and do not go beyond a two-dimensional model.…”

Section: Introductionmentioning

confidence: 99%

Data-driven RANS closures for wind turbine wakes under neutral conditions

2022

View full text Add to dashboard Cite

“…Surrogate models provide streamlined approximations of more costly models and offer reduced computational time 38 . These data‐driven models can be used to augment or correct existing models, 39 as a complete stand‐alone model, 40 or as part of a multifidelity hierarchy of several models 41 . The computational cost of data‐driven models is generally front‐loaded into the model training process, while evaluating a trained model is comparatively inexpensive.…”

Section: Introductionmentioning

confidence: 99%

Machine learning enables national assessment of wind plant controls with implications for land use

2021

View full text Add to dashboard Cite

Summary As deployment of wind energy continues to expand, computationally efficient tools for predicting wind plant performance over a wide range of layout designs, technology innovations, and spatial locations are increasingly important for policy and investment decisions. We demonstrate two approaches to training a surrogate model to predict annual energy production (AEP) of parameterized wind plant layouts: one using a Gaussian process (GP) and the other using a fully convolutional neural network (FCNN). We leverage the powerful FCNN architecture by encoding wind plant design parameters and output response surface as an image. The FCNN produces more accurate results than the GP with mean absolute errors equivalent to 1% and 1.9% of plant rated power, respectively, although the GP performs well under limited training data and provides useful uncertainty information. We also evaluate a surrogate model for wake steering, enabling a nationwide assessment of the impact of plant control strategies and plant layout decisions. Across two million locations, we find that wake steering strategies boost AEP with relative gains upwards of 3%. Gains are most pronounced at sites without a dominant wind direction and where layout optimization is less fruitful. Additionally, we perform a nationwide sensitivity analysis showing that wake steering can mitigate wake losses from higher density plant layouts. Our results suggest that regions which have not been previously viable for wind deployment due to moderate wind resources are especially well enhanced by wake steering strategies that could help overcome land constraints and inflexible layout options, potentially identifying new deployment opportunities.

show abstract

Data-Driven Machine Learning for Wind Plant Flow Modeling

Cited by 11 publications

References 10 publications

IEA Wind TCP: Results of IEA Wind TCP Workshop on a Grand Vision for Wind Energy Technology

IEA Wind TCP: Results of IEA Wind TCP Workshop on a Grand Vision for Wind Energy Technology

Data-driven RANS closures for wind turbine wakes under neutral conditions

Machine learning enables national assessment of wind plant controls with implications for land use

Contact Info

Product

Resources

About