A Method for Analyzing Wind Turbine Dynamic Response Test Data

1988 ◽  
Vol 110 (4) ◽  
pp. 335-339 ◽  
Author(s):  
A. C. Hansen
Keyword(s):  
Time Series ◽  
Dynamic Response ◽  
Wind Turbine ◽  
Test Data ◽  
Fourier Coefficients ◽  
Structural Response ◽  
Poor Correlation ◽  
Data Set ◽  
Response Data ◽  
Original Time

Correlation of wind turbine structural response with ambient wind conditions is an essential but expensive and difficult task. The numbers of variables involved in a typical atmospheric test, the poor correlation between measured instantaneous winds and the actual wind across a rotor disc, and the range of input and response time scales involved all make the correlation task formidable. This paper describes a method which has proven effective for analyzing test data and gaining insight into wind turbine behavior. The method basically consists of representing the dynamic response data in terms of its Fourier Series. A time-series of Fourier coefficients is then created to replace the original time-series raw data. The entire data set, consisting of thousands of rotor revolutions is subdivided into hundreds of sets, each consisting of the azimuth average of (typically) two to ten revolutions. One set of Fourier coefficients (magnitudes and phases of response) is calculated for each azimuth average. The resulting reduced data has a greatly compressed volume with virtually no loss of information. The result is greater insight and a manageable data set size. This new technique is demonstrated for two different wind turbines, an ESI-80 and a Hamilton Standard WTS-4.

scholarly journals A Surface Ice Module for Wind Turbine Dynamic Response Simulation Using FAST

2016 ◽  
Vol 138 (5) ◽  
Author(s):  
Bingbin Yu ◽  
Dale G. Karr ◽  
Huimin Song ◽  
Senu Sirnivas
Keyword(s):  
Dynamic Response ◽  
Wind Turbine ◽  
Wind Turbines ◽  
Structural Response ◽  
Simulation Software ◽  
Offshore Wind ◽  
Foundation Soil ◽  
Offshore Wind Turbines ◽  
Ice Failure ◽  
Lock In

Developing offshore wind energy has become more and more serious worldwide in recent years. Many of the promising offshore wind farm locations are in cold regions that may have ice cover during wintertime. The challenge of possible ice loads on offshore wind turbines raises the demand of modeling capacity of dynamic wind turbine response under the joint action of ice, wind, wave, and current. The simulation software FAST is an open source computer-aided engineering (CAE) package maintained by the National Renewable Energy Laboratory. In this paper, a new module of FAST for assessing the dynamic response of offshore wind turbines subjected to ice forcing is presented. In the ice module, several models are presented which involve both prescribed forcing and coupled response. For conditions in which the ice forcing is essentially decoupled from the structural response, ice forces are established from existing models for brittle and ductile ice failure. For conditions in which the ice failure and the structural response are coupled, such as lock-in conditions, a rate-dependent ice model is described, which is developed in conjunction with a new modularization framework for FAST. In this paper, analytical ice mechanics models are presented that incorporate ice floe forcing, deformation, and failure. For lower speeds, forces slowly build until the ice strength is reached and ice fails resulting in a quasi-static condition. For intermediate speeds, the ice failure can be coupled with the structural response and resulting in coinciding periods of the ice failure and the structural response. A third regime occurs at high speeds of encounter in which brittle fracturing of the ice feature occurs in a random pattern, which results in a random vibration excitation of the structure. An example wind turbine response is simulated under ice loading of each of the presented models. This module adds to FAST the capabilities for analyzing the response of wind turbines subjected to forces resulting from ice impact on the turbine support structure. The conditions considered in this module are specifically addressed in the International Organization for Standardization (ISO) standard 19906:2010 for arctic offshore structures design consideration. Special consideration of lock-in vibrations is required due to the detrimental effects of such response with regard to fatigue and foundation/soil response. The use of FAST for transient, time domain simulation with the new ice module is well suited for such analyses.

Probabilistic Methods for Predicting Wind Turbine Design Loads

2003 ◽  
Author(s):  
Patrick J. Moriarty ◽  
William E. Holley ◽  
Sandy Butterfield
Keyword(s):  
Time Series ◽  
Wind Turbine ◽  
Extreme Values ◽  
Probabilistic Methods ◽  
Exceedance Probability ◽  
Wind Condition ◽  
Data Set ◽  
Out Of Plane ◽  
Blade Loads

Further study of probabilistic methods for predicting extreme wind turbine loading was performed on two large-scale wind turbine models with stall and pitch regulation. Long-term exceedance probability distributions were calculated using maxima extracted from time series simulations of in-plane and out-of-plane blade loads. It was discovered that using a threshold on the selection of maxima increased the accuracy of the fitted distribution in following the trends of the largest extreme values for a given wind condition. The optimal threshold value for in-plane and out-of-plane blade loads was found to be the mean value plus 1.4 times the standard deviation of the original time series for the quantity of interest. When fitting a distribution to a given data set, the higher-order moments were found to have the greatest amount of uncertainty and also the largest influence on the extrapolated long-term load’s. This uncertainty was reduced by using large data sets, smoothing of the moments between wind conditions and parametrically modeling moments of the distribution. A deterministic turbulence model using the 90th percentile level of the conditional turbulence distribution given mean wind speed was used to greatly simplify the calculation of the long-term probability distribution. Predicted extreme loads using this simplified distribution were equal to or more conservative than the loads predicted by the full integration method.

scholarly journals A Fast Data Processing Technique for Continuous Gravitational Wave Searches

Universe ◽  
2021 ◽  
Vol 7 (12) ◽  
pp. 486
Author(s):  
Massimo Tinto
Keyword(s):  
Time Series ◽  
Data Processing ◽  
Gravitational Wave ◽  
Signal To Noise Ratio ◽  
False Alarm Probability ◽  
Processing Technique ◽  
Data Set ◽  
Original Time ◽  
Signal Search ◽  
Data Processing Technique

This article discusses the potential advantages of a data processing technique for continuous gravitational wave signals searches in the data measured by ground-based gravitational wave interferometers. Its main advantage over other techniques is that it does not need to search over the signal’s direction of propagation. Although it is a “coherent method” (i.e., it coherently processes year-long data), it is applied to a data set obtained by multiplying the original time-series with a (half-year) time-shifted copy of it. As a result, the phase modulation due to the interferometer motion around the Sun is automatically canceled in the signal of the synthesized time-series. Although the resulting signal-to-noise ratio is not as high as that of a coherent search, it equals that of current hierarchical methods. In addition, since the signal search is performed over a parameters space of smaller dimensionality, the associated false-alarm probability should be smaller than those characterizing hierarchical methods and result in an improved likelihood of detection.

Denoising of structural health monitoring data: method and coding

2021 ◽  
Author(s):  
Peng Ni ◽  
Ye Xia ◽  
Wanheng Li ◽  
Hanyong Liu ◽  
Limin Sun
Keyword(s):  
Structural Health Monitoring ◽  
Health Monitoring ◽  
Structural Response ◽  
Real Data ◽  
Data Set ◽  
Structural Health ◽  
Response Data ◽  
System Data ◽  
Health Monitoring Systems ◽  
Tied Arch Bridge

<p>Numerous denoising approaches have already been presented to handle the noise in measured data of structural health monitoring systems. However, the performances and features of these existing methods applied in real data-set are not clear enough yet, where the noise is not known in advance. Therefore, based on the measured structural response data from a tied-arch bridge in China, six common data denoising methods are selected for a comparative study. The denoising effects are evaluated based on spectrums. Conclusions on the applicable situations and robustness of involved methods are given. A corresponding program is also developed. This study can provide references for applying the denoising methods in real structural health monitoring system data-set.</p>

Statistical Analysis of Wind Turbine Inflow and Structural Response Data from the LIST Program

2003 ◽  
Vol 125 (4) ◽  
pp. 541-550 ◽  
Author(s):  
Luke D. Nelson ◽  
Lance Manuel ◽  
Herbert J. Sutherland ◽  
Paul S. Veers
Keyword(s):  
Wind Speed ◽  
Wind Turbine ◽  
Test List ◽  
Structural Response ◽  
Vertical Shear ◽  
Horizontal Wind ◽  
Parameter Vector ◽  
Response Data ◽  
Extreme Loads ◽  
Extreme Load

The Long-Term Inflow and Structural Test (LIST) program is gathering inflow and structural response data on a modified version of the Micon 65/13 wind turbine at a test site near Bushland, Texas. Data from 491 ten-minute time data records are analyzed here to determine the dependency of fatigue and extreme loads on inflow parameters. Flap and edge bending moment ranges at a blade root are chosen as the structural response variable, z. Various parameters related to the inflow (including, for example, primary parameters such as the mean and standard deviation of the hub-height horizontal wind speed, and secondary parameters such as Reynolds stresses, vertical shear exponent, etc.) are each considered in an inflow parameter vector, x. Time series for the structural response, z, are processed in order to obtain a structural response parameter, y, where in separate statistical studies, y is taken to be either an equivalent fatigue load or an extreme load. This study describes a procedure by which the important “dependencies” of y on the various variables contained in the inflow parameter vector, x, may be determined considering all the available data. These dependencies of y on x are then recomputed using only the data with above-rated mean wind speeds (taken to be approximately 13 m/s). The procedure employed is similar to other studies, but we do not bin the data sets by wind speed since dependencies in one wind speed bin may be different from those in other bins. Also, our procedure, in sharp contrast to previous studies, examines each inflow parameter in the vector, x, in a sequential analysis, rather than by using multivariate regression. Results from the present study suggest that the primary inflow parameters have a small amount of predictive power in establishing fatigue and extreme loads. In addition, large correlations that exist between several of the secondary parameters individually and each of the primary parameters make it difficult for the secondary parameters to provide any additional explanation of turbine response once the primary parameters have been accounted for.

scholarly journals Damping Measurements Using Operational Data

1996 ◽  
Vol 118 (3) ◽  
pp. 190-193 ◽  
Author(s):  
G. H. James ◽  
T. G. Carne ◽  
P. S. Veers
Keyword(s):  
Wind Turbine ◽  
Strain Gauge ◽  
Test Bed ◽  
Data Set ◽  
Response Data ◽  
Modal Damping ◽  
Gauge Data ◽  
Prediction And Control ◽  
And Control ◽  
Operational Data

We have measured modal damping using strain-gauge data from an operating wind turbine. This new technique for measuring modal damping is easier and less expensive than previously used methods. Auto-correlation and cross-correlation functions of the strain-gauge data have been shown to consist of decaying sinusoids which correspond to the modal frequencies and damping ratios of the wind turbine. We have verified the method by extracting damping values from an analytically generated data set. Actual operating response data from the DOE/Sandia 34-m Test Bed has been used to calculate modal damping ratios as a function of rotor rotation rate. This capability will allow more accurate fatigue life prediction and control.

scholarly journals Modelling and Monitoring Erosion of the Leading Edge of Wind Turbine Blades

2021 ◽  
Author(s):  
Gregory Duthé ◽  
Imad Abdallah ◽  
Sarah Barber ◽  
Eleni Chatzi
Keyword(s):  
Time Series ◽  
Deep Learning ◽  
Dynamic Response ◽  
Wind Turbine ◽  
Leading Edge ◽  
Surface Erosion ◽  
Degradation Processes ◽  
Aerodynamic Properties ◽  
Proposed Model ◽  
Spatio Temporal

Leading edge surface erosion is an emerging issue in wind turbine blade reliability, causing reduction in power performance, aerodynamic loads imbalance, increased noise emission and ultimately additional maintenance costs, and if left untreated, leads to the compromise of the functionality of the blade. In this work, we first propose an empirical spatio-temporal stochastic model for simulating leading edge erosion, to be used in conjunction with aeroelastic simulations, and subsequently propose a deep learning model trained on simulated data, which aims to monitor leading edge erosion by detecting and classifying the degradation severity. The main ingredients of the model include a damage process that progresses at random times, across multiple discrete states characterized by a non-homogeneous compound Poisson process, which is used to describe the random and time-dependent degradation of the blade surface, thus implicitly affecting its aerodynamic properties. The model allows for one, or more, zones along the span of the blades to be independently affected by erosion. The proposed model accounts for uncertainties in the local airfoil aerodynamics via parameterization of the lift and drag coefficients curves. The proposed model is used to generate a stochastic ensemble of degrading airfoil aerodynamic polars, for use in forward aero-servo-elastic simulations, where we compute the effect of leading edge erosion degradation on the dynamic response of a wind turbine under varying turbulent input inflow conditions. The dynamic response is chosen a defining output as this relates to the output variable that is most commonly monitored under a Structural Health Monitoring (SHM) regime. In this context, we further propose an approach for spatio-temporal dependent diagnostics of leading erosion, namely, a deep learning attention-based Transformer, which we modify for classification tasks on slow degradation processes with long sequence multivariate time-series as inputs. We perform multiple sets of numerical experiments, aiming to evaluate the Transformer for diagnostics and assess its limitations. The results reveal Transformers as a potent method for diagnosis of such degradation processes. The attention-based mechanism allows the network to focus on different features at different time intervals for better prediction accuracy, especially for long time-series sequences representing a slow degradation process.

scholarly journals A Fast Data Processing Technique for Continuous Gravitational Wave Searches

2021 ◽  
Author(s):  
Massimo Tinto
Keyword(s):  
Time Series ◽  
Data Processing ◽  
Gravitational Wave ◽  
Signal To Noise Ratio ◽  
False Alarm Probability ◽  
Processing Technique ◽  
Data Set ◽  
Original Time ◽  
Signal Search ◽  
Data Processing Technique

This article discusses the potential advantages of a data processing technique for continuous gravitational wave signals searches in the data measured by ground-based gravitational wave interferometers. Its main advantage over other techniques is that it does not need to search over the signal&rsquo;s direction of propagation. Although it is a &ldquo; coherent method&rdquo; (i.e. it coherently processes year-long data), it is applied to a data set obtained by multiplying the original time-series with a (half-year) time-shifted copy of it. As a result, the phase modulation due to the interferometer motion around the Sun is automatically canceled in the signal of the synthesized time-series. Although the resulting signal-to-noise ratio is not as high as that of a coherent search, it equals that of current hierarchical methods. In addition, since the signal search is performed over a parameters space of smaller dimensionality, the associated false-alarm probability should be smaller than those characterizing hierarchical methods and result in an improved likelihood of detection.

scholarly journals Modeling and Monitoring Erosion of the Leading Edge of Wind Turbine Blades

Energies ◽  
2021 ◽  
Vol 14 (21) ◽  
pp. 7262
Author(s):  
Gregory Duthé ◽  
Imad Abdallah ◽  
Sarah Barber ◽  
Eleni Chatzi
Keyword(s):  
Time Series ◽  
Deep Learning ◽  
Dynamic Response ◽  
Wind Turbine ◽  
Leading Edge ◽  
Surface Erosion ◽  
Maintenance Costs ◽  
Degradation Processes ◽  
Proposed Model ◽  
Spatio Temporal

Leading edge surface erosion is an emerging issue in wind turbine blade reliability, causing a reduction in power performance, aerodynamic loads imbalance, increased noise emission, and, ultimately, additional maintenance costs, and, if left untreated, it leads to the compromise of the functionality of the blade. In this work, we first propose an empirical spatio-temporal stochastic model for simulating leading edge erosion, to be used in conjunction with aeroelastic simulations, and subsequently present a deep learning model to be trained on simulated data, which aims to monitor leading edge erosion by detecting and classifying the degradation severity. This could help wind farm operators to reduce maintenance costs by planning cleaning and repair activities more efficiently. The main ingredients of the model include a damage process that progresses at random times, across multiple discrete states characterized by a non-homogeneous compound Poisson process, which is used to describe the random and time-dependent degradation of the blade surface, thus implicitly affecting its aerodynamic properties. The model allows for one, or more, zones along the span of the blades to be independently affected by erosion. The proposed model accounts for uncertainties in the local airfoil aerodynamics via parameterization of the lift and drag coefficients’ curves. The proposed model was used to generate a stochastic ensemble of degrading airfoil aerodynamic polars, for use in forward aero-servo-elastic simulations, where we computed the effect of leading edge erosion degradation on the dynamic response of a wind turbine under varying turbulent input inflow conditions. The dynamic response was chosen as a defining output as this relates to the output variable that is most commonly monitored under a structural health monitoring (SHM) regime. In this context, we further proposed an approach for spatio-temporal dependent diagnostics of leading erosion, namely, a deep learning attention-based Transformer, which we modified for classification tasks on slow degradation processes with long sequence multivariate time-series as inputs. We performed multiple sets of numerical experiments, aiming to evaluate the Transformer for diagnostics and assess its limitations. The results revealed Transformers as a potent method for diagnosis of such degradation processes. The attention-based mechanism allows the network to focus on different features at different time intervals for better prediction accuracy, especially for long time-series sequences representing a slow degradation process.

EXTREME VALUES OF SEA SURFACE TEMPERATURE ASSOCIATED WITH LONG-PERIOD PHENOMENA OCCURRED DURING 1960-2015 IN THE COLOMBIAN PACIFIC OCEAN

2017 ◽  
Author(s):  
Diaz Juan Navia ◽  
Diaz Juan Navia ◽  
Bolaños Nancy Villegas ◽  
Bolaños Nancy Villegas ◽  
Igor Malikov ◽  
...  
Keyword(s):  
Time Series ◽  
Pacific Ocean ◽  
Surface Temperature ◽  
Sea Surface Temperature ◽  
Extreme Values ◽  
Sea Surface ◽  
Absolute Maximum ◽  
Data Set ◽  
Arctic Oscillation Index ◽  
Colombian Pacific

Sea Surface Temperature Anomalies (SSTA), in four coastal hydrographic stations of Colombian Pacific Ocean, were analyzed. The selected hydrographic stations were: Tumaco (1°48'N-78°45'W), Gorgona island (2°58'N-78°11'W), Solano Bay (6°13'N-77°24'W) and Malpelo island (4°0'N-81°36'W). SSTA time series for 1960-2015 were calculated from monthly Sea Surface Temperature obtained from International Comprehensive Ocean Atmosphere Data Set (ICOADS). SSTA time series, Oceanic Nino Index (ONI), Pacific Decadal Oscillation index (PDO), Arctic Oscillation index (AO) and sunspots number (associated to solar activity), were compared. It was found that the SSTA absolute minimum has occurred in Tumaco (-3.93°C) in March 2009, in Gorgona (-3.71°C) in October 2007, in Solano Bay (-4.23°C) in April 2014 and Malpelo (-4.21°C) in December 2005. The SSTA absolute maximum was observed in Tumaco (3.45°C) in January 2002, in Gorgona (5.01°C) in July 1978, in Solano Bay (5.27°C) in March 1998 and Malpelo (3.64°C) in July 2015. A high correlation between SST and ONI in large part of study period, followed by a good correlation with PDO, was identified. The AO and SSTA have showed an inverse relationship in some periods. Solar Cycle has showed to be a modulator of behavior of SSTA in the selected stations. It was determined that extreme values of SST are related to the analyzed large scale oscillations.

Sign in / Sign up

Export Citation Format

Share Document