Neural Network Classifiers for Local Wind Prediction

Kretzschmar, R.; Eckert, Pierre; Cattani, Daniel; Eggimann, Fritz

doi:10.1175/2057.1

Cited by 45 publications

(23 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Traditionally, short-range forecasts have utilized on-site observations with various persistence based statistical forecast models, including autoregressive time series techniques (Brown, Katz and Murphy 1984) and neural network methodologies (Kretzschmar, Eckert, Cattani and Eggimann 2004). Giebel (2003) surveys the literature on short-range wind speed and wind power forecasting.…”

Section: Introductionmentioning

confidence: 99%

Calibrated Probabilistic Forecasting at the Stateline Wind Energy Center: The Regime-Switching Space-Time (RST) Method

Gneiting¹,

Larson²,

Westrick³

et al. 2004

150

View full text Add to dashboard Cite

With the global proliferation of wind power, accurate short-term forecasts of wind resources at wind energy sites are becoming paramount. Regime-switching space-time (RST) models merge meteorological and statistical expertise to obtain accurate and calibrated, fully probabilistic forecasts of wind speed and wind power. The model formulation is parsimonious, yet takes account of all the salient features of wind speed: alternating atmospheric regimes, temporal and spatial correlation, diurnal and seasonal non-stationarity, conditional heteroscedasticity, and non-Gaussianity. The RST method identifies forecast regimes at the wind energy site and fits a conditional predictive model for each regime. Geographically dispersed meteorological observations in the vicinity of the wind farm are used as off-site predictors.The RST technique was applied to 2-hour ahead forecasts of hourly average wind speed at the Stateline wind farm in the US Pacific Northwest. In July 2003, for instance, the RST forecasts had root-mean-square error (RMSE) 28.6% less than the persistence forecasts. For each month in the test period, the RST forecasts had lower RMSE than forecasts using state-of-the-art vector time series techniques. The RST method provides probabilistic forecasts in the form of predictive cumulative distribution functions, and those were well calibrated and sharp. The RST prediction intervals were substantially shorter on average than prediction intervals derived from univariate time series techniques. These results suggest that quality meteorological data from sites upwind of wind farms can be efficiently used to improve short-term forecasts of wind resources. It is anticipated that the RST technique can be successfully applied at wind energy sites all over the world. Report Documentation PageForm Public reporting burden for the collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington VA 22202-4302. Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to a penalty for failing to comply with a collection of information if it does not display a currently valid OMB control number.

show abstract

Section: Introductionmentioning

confidence: 99%

Calibrated Probabilistic Forecasting at the Stateline Wind Energy Center: The Regime-Switching Space-Time (RST) Method

Gneiting¹,

Larson²,

Westrick³

et al. 2004

150

View full text Add to dashboard Cite

show abstract

“…Additionally, the space-time dependencies of wind are known to be complex, and a localized understanding of this structure may provide more targeted information to improve forecasts. A noted difficulty in characterizing the spatial dependence among many locations is that the location that is upwind of the site of interest changes with the wind direction, so no single off-site location will have a consistently high correlation with the prediction site (Kretzschmar et al 2004). However, with the regimes identified by the GMM( k ) method, we can characterize the spatial correlation among sites within each regime and have demonstrated with the case study that correlations can be different when restricting computations to observations within a regime.…”

Section: Discussionmentioning

confidence: 99%

Assessing the Performance of Model-Based Clustering Methods in Multivariate Time Series with Application to Identifying Regional Wind Regimes

Kazor

Hering

2015

JABES

View full text Add to dashboard Cite

The desire to group observations generated from multivariate time series is common in many applications with the goal to distinguish not only between differences in the means of individual variables but also changes in their covariances and in the temporal dependence of observations. In this analysis, we compare ten model-based clustering methods in terms of their ability to identify such features under four scenarios in which data are simulated with varying levels of variable and temporal dependence. To consider these methods in a realistic environment, we focus our analysis on wind data, where observations are often strongly correlated in time, and the dependence of variables is known to vary across different regional weather patterns. In particular, we assess each method's performance when applied to wind data simulated under a realistic two-regime Markov-switching vector autoregressive (VAR) model with a diurnally varying mean. A Gaussian mixture model and a basic Markov-switching model outperform the other methods considered in terms of misclassification rates and number of clusters identified. These two methods and an additional Markov-switching VAR model are then applied to one year of averaged hourly wind data from twenty meteorological stations, and we find that the methods can identify very different features in the data.Supplementary materials accompanying this paper appear on-line.

show abstract

“…• the Box-Jenkins methodology (Box and Jenkins, 1976) -for examples of application see Brown et al (1984); Sfetsos (2000); • spectral analysis based methods -see, for example, Tilley and McBean (1973); Trivikrama et al (1976); Ghil et al (2002), and, • ANN methods -see for example Sfetsos (2000); Kretzschamar et al (2004); Potter and Negnevitsky (2006).…”

Section: Statistical Methods and Artificial Neural Network (Ann)mentioning

confidence: 99%

The similar days method for predicting near surface wind vectors

Klausner

Kaplan

Fattal

2009

Meteorological Applications

View full text Add to dashboard Cite

A reliable forecast of the wind vector at a given site is an important problem with a broad scope of application. Many uses are in the field of wind power in which prediction of strong winds is important; other uses are in the field of civil engineering, where wind gusts are important to structural integrity. For predicting air pollution events or the hazard zone in case of toxic gas accidents the low wind scenarios are of the highest importance, especially under stable conditions. These low wind scenarios are the situations which meso-scale numerical models encounter difficulties or are often neglected when dealing with wind power tasks. Additional drawbacks of meso-scale numerical models in this context are their lack of site-specific detail and the fact that they are very demanding numerically.The 'similar days' method presented here is a method for the prediction of winds at a given location which works well also for low and variable wind conditions. The method is very efficient numerically: typically, runtime is less than a minute on a single PC. This method is based on using the last few hours' measurements at the given site and comparing it to an historical database. A set of criteria is defined to determine the time series similarity. Those similar days are used to construct the forecast.The similar days method was applied to a total of 5 years of measurements collected in two different locations. Very good agreement was achieved between prediction and measurements.

show abstract

Neural Network Classifiers for Local Wind Prediction

Cited by 45 publications

References 9 publications

Calibrated Probabilistic Forecasting at the Stateline Wind Energy Center: The Regime-Switching Space-Time (RST) Method

Calibrated Probabilistic Forecasting at the Stateline Wind Energy Center: The Regime-Switching Space-Time (RST) Method

Assessing the Performance of Model-Based Clustering Methods in Multivariate Time Series with Application to Identifying Regional Wind Regimes

The similar days method for predicting near surface wind vectors

Contact Info

Product

Resources

About