National Wind Technology Center Research Articles

Comparison of power probability density functions for vertical sound propagation

When sound travels through the atmosphere, fluctuations in the wind velocity and temperature generate fluctuations in the sound amplitude and phase, which makes source detection, localization, and identification more difficult. While there are several parametric probability density functions for the signal power, the previous literature has not shown which model performs best in a wide variety of environments. This paper compares the noncentral Erlang, gamma, compound gamma, generalized gamma, and log exponentially modified Gaussian distributions' ability to approximate the signal power distributions collected at the National Wind Technology Center (Colorado, United States) in 2018. This weeklong experimental campaign measured the vertical and slanted sound propagation of a non-moving, sinusoidal source (600-3500 Hz) on the ground using microphones mounted to a meteorological tower up to 130 m high. The dataset included several different atmospheric conditions, which were recorded by the meteorological tower and resulted in a wide range of scattering regimes in the measured acoustic data. When evaluated using the Kullback-Leibler divergence, the gamma distribution was the best two-parameter model, and the log exponentially modified Gaussian distribution was the best three-parameter model.

Evaluation of a gamma power, wrapped normal phase model for vertical and slanted atmospheric sound propagation

As sound propagates through the atmosphere, turbulence causes fluctuations in the signal power and phase. These fluctuations complicate the signal processing required to detect, locate, and identify elevated sources. This presentation evaluates a model that assumes a gamma distribution describes the power and a wrapped normal distribution describes the phase. The gamma distribution approximates a log-normal distribution when the normalized power variance is small compared to one and an exponential distribution when the normalized power variance is one. The wrapped normal distribution approximates a normal distribution when the phase variance is small and a uniform distribution when the phase variance is large. In 2018, we measured vertical and slanted sound propagation at the National Wind Technology Center (NWTC). The ground-borne source emitted tones in the range 600–3500 Hz, and several microphones were mounted to a meteorological tower up to 130 m high. Overall, the gamma power, wrapped normal phase model accurately approximated the NWTC data across a wide range of scattering regimes. This presentation also discusses the largest discrepancies.

Spatial coherence of vertical atmospheric sound propagation

Atmospheric turbulence causes the amplitude and phase of sound waves to fluctuate, which reduces the coherence of acoustic signals. Spatial coherence describes the similarity of two signals at different points in space, and a better understanding of acoustic coherence could lead to improved target detection, tracking, and identification. This presentation compares theoretical predictions and measurements of the acoustic spatial coherence. The theoretical coherence is derived by combining sound propagation theory with turbulence models that include the effects of atmospheric shear and buoyancy instabilities. To be applicable to vertical and slanted propagation, the turbulence models use height-dependent variances and length scales for the fluctuations in temperature, shear-produced velocity, and buoyancy-produced velocity. Instrumentation on a 135-m meteorological tower at the National Wind Technology Center (Boulder, CO) provided the required model input data. The coherence measurements used a ground-based source and nine microphones attached to the same meteorological tower. Overall, the theoretical model accurately approximated the measured spatial coherences, especially when the atmospheric turbulence was fully developed (e.g., in the afternoon on sunny days). The largest disagreement occurred for measurements taken at dawn. For a source frequency of 3.4 kHz and microphones that are 130 m high and 1.5 m apart, the measured coherence was 0.3–0.6 for sunny conditions and 0.8–0.9 for cloudy conditions.

A Combined Model Based on the Social Cognitive Optimization Algorithm for Wind Speed Forecasting

The use of wind power generation can reduce the pollution in the environment and solve the problem of power shortages on offshore islands, grasslands, pastoral areas, mountain areas, and highlands. Wind speed forecasting plays a significant role in wind farms. It can improve economic and social benefits and make an operation schedule for wind turbines on large wind farms. This paper proposes a combined model based on the existing artificial neural network algorithms for wind speed forecasting at different heights. We first use the wavelet threshold method with the original wind speed dataset for noise reduction. After that, the three artificial neural networks, extreme learning machine (ELM), Elman neural network, and Long Short-term Memory (LSTM) neural network, are applied for wind speed forecasting. In addition, the variance reciprocal method and social cognitive optimization (SCO) algorithm are used to optimize the weight coefficients of the combined model. In order to evaluate the forecasting performance of the combined model, we select wind speed data at three heights (20 m, 50 m and 80 m) at the National Wind Technology Center M2 Tower. The experimental results show that the forecasting performance of the combined model is better than the single model, and it has a good forecasting performance for the wind speed at different heights.

Vertical and slanted sound propagation in the near-ground atmosphere: Coherence and distributions.

Atmospheric turbulence causes acoustic signals to fluctuate and diminishes their coherence. These phenomena are important in applications such as source localization and sonic boom propagation. This article provides formulations for the spatial, cross-frequency, and temporal coherences of narrowband acoustic signals propagating over vertical and slanted paths in the atmosphere. Formulations for single- and two-point distributions of acoustic signals are also overviewed. The theoretical formulations are compared with data from a comprehensive sound propagation experiment carried out in 2018 at the National Wind Technology Center (Boulder, CO). The theories for sound propagation in a turbulent atmosphere, when combined with turbulence models incorporating shear and buoyancy instabilities, correctly predict the measured spatial coherence, which is primarily affected by small-scale isotropic turbulence. For relatively small coherence times, this approach also correctly predicts the temporal coherence. However, the approach underpredicts the cross-frequency coherence and temporal coherence for relatively large coherence times, which are affected by large-scale anisotropic buoyancy-driven velocity fluctuations. For different regimes ranging from unsaturated to fully saturated scattering, the measured distributions agree well with the theoretical predictions.

Multi‐fidelity digital twin structural model for a sub‐scale downwind wind turbine rotor blade

AbstractThis paper presents the development of a multi‐fidelity digital twin structural model (virtual model) of an as‐built wind turbine blade. The goal is to develop and demonstrate an approach to produce an accurate and detailed model of the as‐built blade for use in verifying the performance of the operating two‐bladed, downwind rotor. The digital twin model development methodology, presented herein, involves a novel calibration process to integrate a wide range of information including design specifications, manufacturing information, and structural testing data (modal and static) to produce a multi‐fidelity digital twin structural model: a detailed high‐fidelity model (i.e., 3D finite element analysis [FEA]) and consistent beam‐type models for aeroelastic simulation. A key element is that the multi‐fidelity structural digital twin method follows the rotor from the stages of design, to manufacturing, then to the ground testing and field operation. The result of this comprehensive approach is an accurate multi‐fidelity digital twin structural model for the geometric, structural, and structural dynamic properties of the as‐built blade within a 1% match in mass properties, 3.2% in blade frequencies, and 6% in deflection. The different stages of processing this information within the methodology are discussed. The rotor examined is the SUMR‐Demonstrator (SUMR‐D), which was installed on the Controls Advanced Research Testbed (CART‐2) wind turbine at the National Wind Technology Center. The digital twin model developed here was utilized to design controllers to safely operate SUMR‐D in field tests, which are providing additional data for further evaluation and development of the multi‐fidelity digital twin structural model.

Vertical and slanted sound propagation in the near-ground atmosphere: Amplitude and phase fluctuations.

Sound propagation along vertical and slanted paths through the near-ground atmosphere impacts detection and localization of low-altitude sound sources, such as small unmanned aerial vehicles, from ground-based microphone arrays. This article experimentally investigates the amplitude and phase fluctuations of acoustic signals propagating along such paths. The experiment involved nine microphones on three horizontal booms mounted at different heights to a 135-m meteorological tower at the National Wind Technology Center (Boulder, CO). A ground-based loudspeaker was placed at the base of the tower for vertical propagation or 56 m from the base of the tower for slanted propagation. Phasor scatterplots qualitatively characterize the amplitude and phase fluctuations of the received signals during different meteorological regimes. The measurements are also compared to a theory describing the log-amplitude and phase variances based on the spectrum of shear and buoyancy driven turbulence near the ground. Generally, the theory correctly predicts the measured log-amplitude variances, which are affected primarily by small-scale, isotropic turbulent eddies. However, the theory overpredicts the measured phase variances, which are affected primarily by large-scale, anisotropic, buoyantly driven eddies. Ground blocking of these large eddies likely explains the overprediction.

Analysis of blade fragment risk at a wind energy facility

AbstractAn analysis was performed to determine the risk posed by wind turbine fragments on roads and buildings at the National Wind Technology Center at the National Renewable Energy Laboratory. The authors used a previously developed model of fragment trajectory and took into account the wind speed/direction distribution at the site and the probability of rotor failure. The site‐specific risk was assessed by determining the likelihood of impact and related consequences. For both the roads and buildings, the risk varied from low to routine, which was considered acceptable.

Experimental studies of sound propagation over vertical and slanted paths in a turbulent atmosphere

Sound propagation over vertical and slanted paths differs from horizontal paths due to the height dependence of turbulence parameters and atmospheric stratification. Vertical and slanted paths are important for several applications, such as localization of unattended aerial systems (UASs) with ground-based acoustic microphone arrays, detection of ground-based sources with elevated arrays, and the effect of atmospheric turbulence on sonic booms. This paper describes a comprehensive experiment on near-vertical sound propagation conducted over five days in September 2018 at the National Wind Technology Center (NWTC), located near Boulder, CO. The experiment involved the NWTC 135-m meteorological tower with meteorological instruments installed at 17 heights. Nine microphones were located on three horizontal booms attached to the tower at different heights, and a speaker was placed on the ground. Two reference microphones where located near the speaker. About 65% of the time, the speaker transmitted nine or twelve tones. During the remaining time, 6-ms chirps were transmitted. A preliminary analysis of the signal statistics is presented and compared with theoretical predictions.

Wind farm wake modeling using NWTC design codes

Wind turbines located in a wind farm are subject to a wind field that is substantially modified when compared to the ambient wind field due to wake effects. It is clear that in addition to the power, a tool that could model the loads of the waked turbine is needed. In this work, a wind farm wake model is created using a single wake model based on the dynamic wake meandering model to systematically model the wake effects and thus the power and loads of an entire wind farm with an arbitrary wind turbine layout and wind condition. This wind farm wake model is incorporated into the National Wind Technology Center design codes. Preliminary results demonstrate that this wind farm tool yields satisfactory results when compared with Simulator for Wind Farm Applications, a high fidelity computational fluid dynamics Large eddy simulation model, and with experimentally-obtained field data. The integration of the wind farm model with the National Wind Technology Center design codes is described in detail and validation results of the tool are provided.

Wind turbine power production and annual energy production depend on atmospheric stability and turbulence

Abstract. Using detailed upwind and nacelle-based measurements from a General Electric (GE) 1.5sle model with a 77 m rotor diameter, we calculate power curves and annual energy production (AEP) and explore their sensitivity to different atmospheric parameters to provide guidelines for the use of stability and turbulence filters in segregating power curves. The wind measurements upwind of the turbine include anemometers mounted on a 135 m meteorological tower as well as profiles from a lidar. We calculate power curves for different regimes based on turbulence parameters such as turbulence intensity (TI) as well as atmospheric stability parameters such as the bulk Richardson number (RB). We also calculate AEP with and without these atmospheric filters and highlight differences between the results of these calculations. The power curves for different TI regimes reveal that increased TI undermines power production at wind speeds near rated, but TI increases power production at lower wind speeds at this site, the US Department of Energy (DOE) National Wind Technology Center (NWTC). Similarly, power curves for different RB regimes reveal that periods of stable conditions produce more power at wind speeds near rated and periods of unstable conditions produce more power at lower wind speeds. AEP results suggest that calculations without filtering for these atmospheric regimes may overestimate the AEP. Because of statistically significant differences between power curves and AEP calculated with these turbulence and stability filters for this turbine at this site, we suggest implementing an additional step in analyzing power performance data to incorporate effects of atmospheric stability and turbulence across the rotor disk.

Temporal Coherence: A Model for Non-stationarity in Natural and Simulated Wind Records

We present a novel methodology for characterizing and simulating non-stationary stochastic wind records. In this new method, non-stationarity is characterized and modelled via temporal coherence, which is quantified in the discrete frequency domain by probability distributions of the differences in phase between adjacent Fourier components. Temporal coherence can also be used to quantify non-stationary characteristics in wind data. Three case studies are presented that analyze the non-stationarity of turbulent wind data obtained at the National Wind Technology Center near Boulder, Colorado, USA. The first study compares the temporal and spectral characteristics of a stationary wind record and a non-stationary wind record in order to highlight their differences in temporal coherence. The second study examines the distribution of one of the proposed temporal coherence parameters and uses it to quantify the prevalence of nonstationarity in the dataset. The third study examines how temporal coherence varies with a range of atmospheric parameters to determine what conditions produce more non-stationarity.

Monitoring of Wind Turbine Gearbox Condition through Oil and Wear Debris Analysis: A Full-Scale Testing Perspective

ABSTRACTDespite the wind industry's dramatic development during the past decade, it is still challenged by premature turbine subsystem/component failures, especially for turbines rated above 1 MW. Because a crane is needed for each replacement, gearboxes have been a focal point for improvement in reliability and availability. Condition monitoring (CM) is a technique that can help improve these factors, leading to reduced turbine operation and maintenance costs and, subsequently, lower cost of energy for wind power. Although technical benefits of CM for the wind industry are normally recognized, there is a lack of published information on the advantages and limitations of each CM technique confirmed by objective data from full-scale tests. This article presents first-hand oil and wear debris analysis results obtained through tests that were based on full-scale wind turbine gearboxes rated at 750 kW. The tests were conducted at the 2.5-MW dynamometer test facility at the National Wind Technology Center at the National Renewable Energy Laboratory. The gearboxes were tested in three conditions: run-in, healthy, and damaged. The investigated CM techniques include real-time oil condition and wear debris monitoring, both inline and online sensors, and offline oil sample and wear debris analysis, both onsite and offsite laboratories. The reported results and observations help increase wind industry awareness of the benefits and limitations of oil and debris analysis technologies and highlight the challenges in these technologies and other tribological fields for the Society of Tribologists and Lubrication Engineers and other organizations to help address, leading to extended gearbox service life.

Effects of power reserve control on wind turbine structural loading

AbstractAs the penetration of wind energy in worldwide electrical utility grids increases, there is a growing interest in the provision of active power control (APC) services from wind turbines and power plants to aid in maintaining grid stability. Recent research has focused on the design of active power controllers for wind turbines that can provide a range of APC services including inertial, primary frequency and secondary frequency control. An important consideration for implementing these controllers in practice is assessing their impact on the lifetime of wind turbine components.In this paper, the impact on the structural loads of a wind turbine providing a power reserve is explored by performing a load suite analysis for several torque‐based control strategies. Power reserve is required for providing those APC services that require the ability of the wind turbine to supply an increase in power. To study this, we performed a load suite on a simulated model of a research turbine located at the National Wind Technology Center at the National Renewable Energy Laboratory. Analysis of the results explores the effect of the different reserve strategies on turbine loading. In addition, field‐test data from the turbine itself are presented to augment and support the findings from the simulation study results. Results indicate that all power‐reserve strategies tend to decrease extreme loads and increase pitch actuation. Fatigue loads tend to be reduced in faster winds and increased in slower winds, but are dependent on reserve‐controller design. Copyright © 2015 John Wiley & Sons, Ltd.

Power Performance Measurements of the NREL CART-2 Wind Turbine Using a Nacelle-Based Lidar Scanner

Abstract Different certification procedures in wind energy, such as power performance testing or load estimation, require measurements of the wind speed, which is set in relation to the electrical power output or the turbine loading. The wind shear affects the behavior of the turbine as hub heights and rotor diameters of modern wind turbines increase. Different measurement methods have been developed to take the wind shear into account. In this paper an approach is presented where the wind speed is measured from the nacelle of a wind turbine using a scanning lidar system. The measurement campaign was performed on the two-bladed Controls Advanced Research Turbine (CART-2) at the National Wind Technology Center in Colorado. The wind speed of the turbine inflow was measured and recalculated in three different ways: using an anemometer installed on a meteorological mast, using the nacelle-based lidar scanner, and using the wind turbine itself. Here, the wind speed was recalculated from turbine data using the wind turbine as a big horizontal anemometer. Despite the small number of useful data, the correlation between this so-called rotor effective wind speed and the wind speed measured by the scanning nacelle-based lidar is high. It could be demonstrated that a nacelle-based scanning lidar system provides accurate measurements of the wind speed converted by a wind turbine. This is a first step, and it provides evidence to support further investigations using a much more extensive dataset and refines the parameters in the measurement process.

Inspection and monitoring of wind turbine blade-embedded wave defects during fatigue testing

The research presented in this article focuses on a 9-m CX-100 wind turbine blade, designed by a team led by Sandia National Laboratories and manufactured by TPI Composites Inc. The key difference between the 9-m blade and baseline CX-100 blades is that this blade contains fabric wave defects of controlled geometry inserted at specified locations along the blade length. The defect blade was tested at the National Wind Technology Center at the National Renewable Energy Laboratory using a schedule of cycles at increasing load level until failure was detected. Researchers used digital image correlation, shearography, acoustic emission, fiber-optic strain sensing, thermal imaging, and piezoelectric sensing as structural health monitoring techniques. This article provides a comparison of the sensing results of these different structural health monitoring approaches to detect the defects and track the resultant damage from the initial fatigue cycle to final failure.

Quantifying Wind Turbine Wake Characteristics from Scanning Remote Sensor Data

Abstract Because of the dense arrays at most wind farms, the region of disturbed flow downstream of an individual turbine leads to reduced power production and increased structural loading for its leeward counterparts. Currently, wind farm wake modeling, and hence turbine layout optimization, suffers from an unacceptable degree of uncertainty, largely because of a lack of adequate experimental data for model validation. Accordingly, nearly 100 h of wake measurements were collected with long-range Doppler lidar at the National Wind Technology Center at the National Renewable Energy Laboratory in the Turbine Wake and Inflow Characterization Study (TWICS). This study presents quantitative procedures for determining critical parameters from this extensive dataset—such as the velocity deficit, the size of the wake boundary, and the location of the wake centerline—and categorizes the results by ambient wind speed, turbulence, and atmospheric stability. Despite specific reference to lidar, the methodology is general and could be applied to extract wake characteristics from other remote sensor datasets, as well as computational simulation output. The observations indicate an initial velocity deficit of 50%−60% immediately behind the turbine, which gradually declines to 15%−25% at a downwind distance x of 6.5 rotor diameters (D). The wake expands with downstream distance, albeit less so in the vertical direction due to the presence of the ground: initially the same size as the rotor, the extent of the wake grows to 2.7D (1.2D) in the horizontal (vertical) at x = 6.5D. Moreover, the vertical location of the wake center shifts upward with downstream distance because of the tilt of the rotor.

Wind turbine condition monitoring

Wind turbine condition monitoring

On damage diagnosis for a wind turbine blade using pattern recognition

With the increased interest in implementation of wind turbine power plants in remote areas, structural health monitoring (SHM) will be one of the key cards in the efficient establishment of wind turbines in the energy arena. Detection of blade damage at an early stage is a critical problem, as blade failure can lead to a catastrophic outcome for the entire wind turbine system. Experimental measurements from vibration analysis were extracted from a 9 m CX-100 blade by researchers at Los Alamos National Laboratory (LANL) throughout a full-scale fatigue test conducted at the National Renewable Energy Laboratory (NREL) and National Wind Technology Center (NWTC). The blade was harmonically excited at its first natural frequency using a Universal Resonant EXcitation (UREX) system. In the current study, machine learning algorithms based on Artificial Neural Networks (ANNs), including an Auto-Associative Neural Network (AANN) based on a standard ANN form and a novel approach to auto-association with Radial Basis Functions (RBFs) networks are used, which are optimised for fast and efficient runs. This paper introduces such pattern recognition methods into the wind energy field and attempts to address the effectiveness of such methods by combining vibration response data with novelty detection techniques.

LPV Identification of Wind Turbine Rotor Vibrational Dynamics Using Periodic Disturbance Basis Functions

This brief presents an identification experiment performed on the coupled dynamics of the edgewise bending vibrations of the rotor blades and the in-plane motion of the drivetrain of three-bladed wind turbines. These dynamics vary with rotor speed, and are subject to periodic wind flow disturbances. This brief demonstrates that this time-varying behavior can be captured in a linear parameter-varying (LPV) model with the rotor speed as the scheduling signal, and with additional sinusoidal inputs that are used as basis functions for the periodic wind flow disturbances. By including these inputs, the predictor-based LPV subspace identification approach (LPV PBSIDopt) was tailored for wind turbine applications. Using this tailor-made approach, the LPV model is identified from data measured with the three-bladed Controls Advanced Research Turbine (CART3) at the National Renewable Energy Laboratory's National Wind Technology Center.