On-line quantile regression in the RKHS (Reproducing Kernel Hilbert Space) for operational probabilistic forecasting of wind power

Gallego-Castillo, Cristóbal; Bessa, Ricardo J.; Cavalcante, Laura; López-García, Óscar

doi:10.1016/j.energy.2016.07.055

Cited by 50 publications

(16 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In [1], the framework for calculating day ahead probabilistic forecast and the framework for its impact evaluation on network operation has been established. In [2,3] the improvement of extreme quantiles forecasting for wind power production has been explored. To help the reader, descriptions of the models developed in [1,2] are presented in Sections 3.1 and 3.2, respectively.…”

Section: State Of the Artmentioning

confidence: 99%

Extreme Quantiles Dynamic Line Rating Forecasts and Application on Network Operation

Dupin

Cavalcante²,

Bessa

et al. 2020

Energies

Self Cite

View full text Add to dashboard Cite

This paper presents a study on dynamic line rating (DLR) forecasting procedure aimed at developing a new methodology able to forecast future ampacity values for rare and extreme events. This is motivated by the belief that to apply DLR network operators must be able to forecast their values and this must be based on conservative approaches able to guarantee the safe operation of the network. The proposed methodology can be summarised as follows: firstly, probabilistic forecasts of conductors’ ampacity are calculated with a non-parametric model, secondly, the lower part of the distribution is replaced with a new distribution calculated with a parametric model. The paper presents also an evaluation of the proposed methodology in network operation, suggesting an application method and highlighting the advantages. The proposed forecasting methodology delivers a high improvement of the lowest quantiles’ reliability, allowing perfect reliability for the 1% quantile and a reduction of roughly 75% in overconfidence for the 0.1% quantile.

show abstract

Section: State Of the Artmentioning

confidence: 99%

Extreme Quantiles Dynamic Line Rating Forecasts and Application on Network Operation

Dupin

Cavalcante²,

Bessa

et al. 2020

Energies

Self Cite

View full text Add to dashboard Cite

show abstract

“…The larger the interval coverage rate, the smaller the average width, which is the best performance of the prediction model. The sensitivity index ζ α mean is defined as (25).…”

Section: Prediction Effect Evaluationmentioning

confidence: 99%

“…Haque et al [24] proposed the firefly algorithm to optimize the combined prediction algorithm, using the quantile regression method to achieve probability prediction. In Reference [25], an online quantile regression method based on Reproducing Kernel Hilbert Space was proposed, which enabled online learning and online calibration. The KDE method can provide the probability density function (PDF) of wind power prediction error.…”

Section: Introductionmentioning

confidence: 99%

Wind Power Prediction Based on LSTM Networks and Nonparametric Kernel Density Estimation

et al. 2019

View full text Add to dashboard Cite

Wind energy is a kind of sustainable energy with strong uncertainty. With a large amount of wind power injected into the power grid, it will inevitably affect the security, stability and economic operation of the power grid. High-precision wind power spot prediction and fluctuation interval information can provide more adequate decision-making support for grid scheduling and optimization. Hence, this paper proposes a K-Means-long short-term memory (K-Means-LSTM) network model for wind power spot prediction, and a nonparametric kernel density estimation (KDE) model with bandwidth optimization for wind power probabilistic interval prediction. The long short-term memory (LSTM) network has a strong memory function, which can establish the correlation between the data before and after, so as to improve the prediction accuracy. The K-Means clustering method forms different clusters of wind power impact factors to generate a new LSTM sub-prediction model. The optimization of the bandwidth in the nonparametric KDE is implemented by the mean integrated squared error criterion. In addition, a part of the dataset is deliberately demarcated from the wind power historical dataset to generate reasonable wind power prediction errors. The simulation results show that the proposed K-Means-LSTM network model has higher prediction accuracy than the back propagation (BP) neural networks, Elman neural networks, support vector regression (SVR) and LSTM network models. Compared with the KDE model with random bandwidth and the Gaussian distribution model, the bandwidth optimization model proposed in this paper has more narrow prediction intervals with higher interval coverage rates. INDEX TERMS Wind power prediction, Kernel density estimation, long short-term memory, K-means clustering, probabilistic interval prediction.

show abstract

“…Reliability diagram for the probabilistic forecasting results at seven locations is the uncertainty conveyed by the probabilistic forecasts, which can be computed as the average interval size of different confident levels[54]. These two metrics are visualized by the reliability diagram and δ-diagram, respectively.In our case, the reliability diagrams of all seven locations are depicted inFig.…”

mentioning

confidence: 99%

A data-driven multi-model methodology with deep feature selection for short-term wind forecasting

et al. 2017

View full text Add to dashboard Cite

With the growing wind penetration into the power system worldwide, improving wind power forecasting accuracy is becoming increasingly important to ensure continued economic and reliable power system operations. In this paper, a data-driven multi-model wind forecasting methodology is developed with a two-layer ensemble machine learning technique. The first layer is composed of multiple machine learning models that generate individual forecasts. A deep feature selection framework is developed to determine the most suitable inputs to the first layer machine learning models. Then, a blending algorithm is applied in the second layer to create an ensemble of the forecasts produced by first layer models and generate both deterministic and probabilistic forecasts. This two-layer model seeks to utilize the statistically different characteristics of each machine learning algorithm. A number of machine learning algorithms are selected and compared in both layers. This developed multi-model wind forecasting methodology is compared to several benchmarks. The effectiveness of the proposed methodology is evaluated to provide 1-hour-ahead wind speed forecasting at seven locations of the Surface Radiation network. Numerical results show that comparing to the single-algorithm models, the developed multi-model framework with deep feature selection procedure has improved the forecasting accuracy by up to 30%.

show abstract

On-line quantile regression in the RKHS (Reproducing Kernel Hilbert Space) for operational probabilistic forecasting of wind power

Cited by 50 publications

References 41 publications

Extreme Quantiles Dynamic Line Rating Forecasts and Application on Network Operation

Extreme Quantiles Dynamic Line Rating Forecasts and Application on Network Operation

Wind Power Prediction Based on LSTM Networks and Nonparametric Kernel Density Estimation

A data-driven multi-model methodology with deep feature selection for short-term wind forecasting

Contact Info

Product

Resources

About