Robust Assessment of Short-Term Wind Power Forecasting Models on Multiple Time Horizons
- PDF / 5,458,294 Bytes
- 15 Pages / 595.224 x 790.955 pts Page_size
- 104 Downloads / 162 Views
(2020) 5:19
ORIGINAL PAPER
Robust Assessment of Short-Term Wind Power Forecasting Models on Multiple Time Horizons Fabrizio De Caro1
· Jacopo De Stefani2 · Gianluca Bontempi2 · Alfredo Vaccaro1 · Domenico Villacci1
Received: 7 February 2020 / Accepted: 24 August 2020 © Springer Nature Singapore Pte Ltd. 2020
Abstract The massive penetration of renewable power generation in modern power grids is an effective way to reduce the impact of energy production on global warming. Unfortunately, the wind power generation may affect the regular operation of electrical systems, due to the stochastic and intermittent nature of the wind. For this reason, reducing the uncertainty about the wind evolution, e.g. by using short-term wind power forecasting methodologies, is a priority for system operators and wind producers to implement low-carbon power grids. Unfortunately, though the complexity of this task implies the comparison of several alternative forecasting methodologies and dimensionality reduction techniques, a general and robust procedure of model assessment still lacks in literature. In this paper the authors propose a robust methodology, based on extensive statistical analysis and resampling routines, to supply the most effective wind power forecasting method by testing a vast ensemble of methodologies over multiple time-scales and on a real case study. Experimental results on real data collected in an Italian wind farm show the potential of ensemble approaches integrating both statistical and machine learning methods. Keywords Wind Power Forecasting · Wind Energy · Robust Forecasting · Ensemble Forecasting Nomenclature Symbols D[N, S] P[N, φ] R[N, ρ]
Matrix of observations of size [N, S] Predictor matrix of size [N, φ] Target matrix of size [N, ρ]
Fabrizio De Caro
[email protected] Jacopo De Stefani [email protected] Gianluca Bontempi [email protected] Alfredo Vaccaro [email protected] Domenico Villacci [email protected] 1
University of Sannio, piazza Roma 21, Benevento, 82100, Italy
2
Universit`e Libre de Bruxelles, Campus de la Plaine ULB CP212, boulevard du Triomphe, 1050 Bruxelles, Belgium
N φ ρ c γ L H Φ X0 Y0 X(v) Y(v) (v) Xtrn (v) Xval r1,...,N RSS T SS A ya , yˆa a th yˆ σ
Number of observations Number of predictors Number of targets Number of lagged time windows Number of smoothed variables Auto-regressive lag Forecasting horizon (number of steps ahead) Number of features after pre-processing Embedded Input Matrix Embedded Output Matrix Matrix X for test case v Matrix Y for test case v Training matrix for test case v Validation matrix for test case v Generic smoothed time series Residual Sum of Squares Total Sum of Squares Number of samples in validation set actual,predicted power in validation set mean actual value in validation set actual power Standard deviation in validation set
Abbreviations Ad. Adaptive Ensemble Forecasting
19
Page 2 of 15
ANN ARIMA ARMA CFD DEM ELM GRU GBM HW-ES HYB LSTM M3 ML MAE MAPE MSE mRMR nMSE NWP PH PCA RF RMSE R2 Av. SCADA SF SVM WPF WTG
T
Data Loading...
