High-fidelity wind farm simulation methodology with experimental validation

Hsieh, Alan; Brown, Kenneth; deVelder, Nathaniel; Herges, Thomas; Knaus, Robert; Sakievich, Philip; Cheung, Lawrence; Houchens, Brent C.; Blaylock, Myra L.; Maniaci, David Charles

doi:10.1016/j.jweia.2021.104754

Cited by 17 publications

(16 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…After another thin transition of hexahedra cells, the outermost mesh uses tetrahedra cells of 1.25 m to make approximately 39 million cells for the total domain. We did not perform a resolution study for this effort, but these cell sizes are in line with previous experience and validation studies that included resolution tests in Nalu-Wind [27]. The turbines are placed on the transverse centerline at 3.4D and 8.4D from the inlet.…”

Section: Nalu-windmentioning

confidence: 96%

“…As an LES, Nalu-Wind is inherently unsteady and requires significantly more computational time than CACTUS. In particular, Nalu-Wind uses a node-centered finite volume discretization and solves the acoustically incompressible Navier-Stokes equations with an approximate pressure projection technique [27]. The ALM models the wind turbine blades, tower, and hub as series of discrete points and couples Nalu-Wind to OpenFAST in the Exawind software such that fluid properties can be sampled at the ALM points to be used by OpenFAST to compute quantities of interest (QoI) on the turbine such as blade and tower motions, generator power, and structural loads.…”

Section: Nalu-windmentioning

confidence: 99%

See 1 more Smart Citation

Comparison of a mid-fidelity free vortex wake method to a high-fidelity actuator line model large eddy simulation for wind turbine wake simulations

Houck

deVelder

Kelley

2022

J. Phys.: Conf. Ser.

View full text Add to dashboard Cite

The need for mid-fidelity wind turbine simulations is increasing due to their balance of accuracy and computational efficiency. A common form of mid-fidelity code is the free wake vortex method (FVWM), but these are known to inaccurately represent the far-wake depending on their implementation. Herein, we compare CACTUS, a FVWM code, to Nalu-Wind, an actuator line model large eddy simulation (ALM-LES) code with particular focus on the far wake. Specifically, we compare one CACTUS simulation, which necessarily has steady and uniform inflow, to three Nalu-Wind simulations using a steady and uniform inflow, a low turbulence inflow, and a high turbulence inflow. The comparison reveals interesting parallels between CACTUS and all Nalu-Wind cases, but ultimately demonstrates that CACTUS is not an appropriate tool for the simulation of the far-wake of a wind turbine.

show abstract

Section: Nalu-windmentioning

confidence: 96%

Section: Nalu-windmentioning

confidence: 99%

Comparison of a mid-fidelity free vortex wake method to a high-fidelity actuator line model large eddy simulation for wind turbine wake simulations

Houck

deVelder

Kelley

2022

J. Phys.: Conf. Ser.

View full text Add to dashboard Cite

show abstract

“…A validation case for the lidar processing techniques is derived from data at the Scaled Wind Farm Technology (SWiFT) facility in Lubbock, Texas, USA as illustrated in Figure 7. The site features level terrain with minimal surface roughness, and characterization of the atmospheric conditions is given in Kelley and Ennis (2016) with recent benchmarking and validation activities given in Doubrawa et al (2020) and Hsieh (2021). Occasional spurious spikes in the signal are removed in pre-processing using a median absolute deviation filter with a length of 10,000 data points, or 100 seconds.…”

Section: Facilitymentioning

confidence: 99%

“…interactions represent two areas needing further advances and areas for which accurate wind-field sensing around the turbine is imperative. Such sensing is enabled through Doppler lidar instruments, and nacelle-mounted lidar, in particular, have made recent inroads with applications in monitoring and control (Harris et al, 2006;Mikkelsen et al, 2013;Simley et al, 2014;Simley et al, 2018) and model validation (Doubrawa et al, 2020;Brown et al, 2020;Hsieh, 2021). Continuing investment in such lidar technology includes efforts to reduce the uncertainty of wind field measurements over the whole field of view, which is critical for both forward-mounted lidar used in feedforward control applications and rear-mounted lidar used in wake measurements for model validation.…”

mentioning

confidence: 99%

High-Fidelity Processing of Instantaneous Line-of-Sight Returns from Nacelle-Mounted Lidar including Supervised Machine Learning

Brown

Herges

2022

Preprint

Self Cite

View full text Add to dashboard Cite

Abstract. Wind turbine applications that leverage nacelle-mounted Doppler lidar are hampered by several sources of uncertainty in the lidar measurement, affecting both bias and random error. Two problems encountered especially for nacelle-mounted lidar are solid interference due to intersection of the line of sight with solid objects behind, within, or in front of the measurement volume, as well as spectral noise due primarily to limited photon capture. These two uncertainties can be reduced with high-fidelity quality assurance/quality control (QA/QC) processing techniques. Our work compares three QA/QC techniques, including conventional thresholding, advanced filtering, and a novel application of supervised machine learning with ensemble neural networks, based on their ability to reduce uncertainty introduced by the two observed non-ideal spectral features. The approach leverages data from a field experiment involving a continuous-wave (CW) SpinnerLidar from the Technical University of Denmark (DTU) that provided scans of a wide range of flows both unwaked and waked by a field turbine. Independent measurements from an overlapped meteorological tower permit experimental validation of the instantaneous velocity uncertainty remaining after QA/QC processing that stems from solid interference and strong spectral noise, which is a validation that has not been performed previously. All three methods perform similarly for non-interfered returns, but the advanced filtering and machine learning techniques perform better when solid interference is present, which allows them to produce overall standard deviations of error between 0.2 and 0.3 m/s, or a 1–22 % improvement versus the conventional thresholding technique, over the rotor height for the unwaked cases. Between the two improved techniques, the advanced filtering produces 3.5 % higher overall data availability, while the machine learning offers a faster runtime (i.e, ~1 second to evaluate) that is therefore more commensurate with the requirements of real-time turbine control. The QA/QC techniques are described in terms of application to CW lidar, though they are also relevant to pulsed lidar. Previous work by the authors (Brown and Herges, 2020) explored a novel attempt to quantify uncertainty after a lidar QA/QC process using simulated lidar returns; this article provides true uncertainty quantification versus independent measurement and does so for three rather than one QA/QC techniques.

show abstract

“…Real-time control of turbines within the stochastic atmosphere and better understanding of turbineto-turbine wake interactions represent two areas needing further advances and areas for which accurate wind field sensing around the turbine is imperative. Such sensing is enabled through Doppler lidar instruments, and nacellemounted lidars, in particular, have made recent inroads with applications in monitoring and control (Harris et al, 2006;Mikkelsen et al, 2013;Sjöholm et al, 2013;Simley et al, 2014Simley et al, , 2018 as well as wake aerodynamics model validation (Doubrawa et al, 2020;Brown et al, 2020;Hsieh, 2021). Continuing investment in such lidar technology includes efforts to reduce the uncertainty of wind field measurements over the whole field of view, which is critical for both forward-mounted lidar used in feed-forward control applications and rear-mounted lidar used in wake measurements for model validation.…”

Section: Introductionmentioning

confidence: 99%

High-fidelity retrieval from instantaneous line-of-sight returns of nacelle-mounted lidar including supervised machine learning

Brown

Herges

2022

Atmos. Meas. Tech.

Self Cite

View full text Add to dashboard Cite

Abstract. Wind turbine applications that leverage nacelle-mounted Doppler lidar are hampered by several sources of uncertainty in the lidar measurement, affecting both bias and random errors. Two problems encountered especially for nacelle-mounted lidar are solid interference due to intersection of the line of sight with solid objects behind, within, or in front of the measurement volume and spectral noise due primarily to limited photon capture. These two uncertainties, especially that due to solid interference, can be reduced with high-fidelity retrieval techniques (i.e., including both quality assurance/quality control and subsequent parameter estimation). Our work compares three such techniques, including conventional thresholding, advanced filtering, and a novel application of supervised machine learning with ensemble neural networks, based on their ability to reduce uncertainty introduced by the two observed nonideal spectral features while keeping data availability high. The approach leverages data from a field experiment involving a continuous-wave (CW) SpinnerLidar from the Technical University of Denmark (DTU) that provided scans of a wide range of flows both unwaked and waked by a field turbine. Independent measurements from an adjacent meteorological tower within the sampling volume permit experimental validation of the instantaneous velocity uncertainty remaining after retrieval that stems from solid interference and strong spectral noise, which is a validation that has not been performed previously. All three methods perform similarly for non-interfered returns, but the advanced filtering and machine learning techniques perform better when solid interference is present, which allows them to produce overall standard deviations of error between 0.2 and 0.3 m s−1, or a 1 %–22 % improvement versus the conventional thresholding technique, over the rotor height for the unwaked cases. Between the two improved techniques, the advanced filtering produces 3.5 % higher overall data availability, while the machine learning offers a faster runtime (i.e., ∼ 1 s to evaluate) that is therefore more commensurate with the requirements of real-time turbine control. The retrieval techniques are described in terms of application to CW lidar, though they are also relevant to pulsed lidar. Previous work by the authors (Brown and Herges, 2020) explored a novel attempt to quantify uncertainty in the output of a high-fidelity lidar retrieval technique using simulated lidar returns; this article provides true uncertainty quantification versus independent measurement and does so for three techniques rather than one.

show abstract

High-fidelity wind farm simulation methodology with experimental validation

Cited by 17 publications

References 47 publications

Comparison of a mid-fidelity free vortex wake method to a high-fidelity actuator line model large eddy simulation for wind turbine wake simulations

Comparison of a mid-fidelity free vortex wake method to a high-fidelity actuator line model large eddy simulation for wind turbine wake simulations

High-Fidelity Processing of Instantaneous Line-of-Sight Returns from Nacelle-Mounted Lidar including Supervised Machine Learning

High-fidelity retrieval from instantaneous line-of-sight returns of nacelle-mounted lidar including supervised machine learning

Contact Info

Product

Resources

About