Hybrid Probabilistic Wind Power Forecasting Using Temporally Local Gaussian Process

Abstract

The demand for sustainable development has resulted in a rapid growth in wind power worldwide. Although various approaches have been proposed to improve the accuracy and to overcome the uncertainties associated with traditional methods, the stochastic and variable nature of wind still remains the most challenging issue in accurately forecasting wind power. This paper presents a hybrid deterministic-probabilistic method where a temporally local "moving window" technique is used in Gaussian process (GP) to examine estimated forecasting errors. This temporally local GP employs less measurement data with faster and better predictions of wind power from two wind farms, one in the USA and the other in Ireland. Statistical analysis on the results shows that the method can substantially reduce the forecasting error while it is more likely to generate Gaussiandistributed residuals, particularly for short-term forecast horizons due to its capability to handle the time-varying characteristics of wind power.

Advanced Economic Control of Electricity-Based Space Heating Systems in Domestic Coalitions with Shared Intermittent Energy Resources

“…4.4 we also evaluate the significance of our hybrid approach within AdaHeat+. It is worth noting that the performance of our simple hybrid GP-based approach is in alignment to the literature [Chen et al 2014] and also outperforms far more complex GP-based approaches (in terms of RMSE throughout the predictive horizon examined) evaluated in the proximity of our region of interest (and, in particular, in Ireland) [Yan et al 2016;Yan et al 2014]. …”

Coalitional Game-Based Cost Optimization of Energy Portfolio in Smart Grid Communities

“…, T ], produced by those households m ∈ M owning RESs is considered predicted with sufficient accuracy over the whole period T , hence it is considered known as well. Methods for prediction of residential electricity demand and of renewable energy generation are proposed, for example, in [16], [17] and [18], [19], respectively. Note that for those households from the set M that do not own RESs, but only ESSs, the renewable energy vector is zero, i.e.…”

“…The final set of constraints for this problem is (1)- (3), (5)- (8), (17) and (18). The set of solution variables is {b n , r m , s m , a n } M m=1 , N n=1 .…”

Review of distributed generation (DG) system planning and optimisation techniques: Comparison of numerical and mathematical modelling methods

“…Specifically, in a comparative study, Conditional Parametric Quantile Regression (CP/QR) model was found to outperform Censored Normal (CN) distribution-based Conditional Parametric Autoregressive Conditional Heteroscedasticity (CP/ARCH) model for five-hourahead stochastic wind forecasting [234]. Moreover, time-adaptive Kernel Density Estimation (KDE) [235], Temporarily Local Gaussian Process (TLGP) [236], two-parameter Weibull distribution [237], Generalised Pareto Distribution (GPD) [238], and sparse Bayesian classification (SBC) with Dempster-Shafer theory (DSF) [239] have been introduced with satisfactory wind forecasting performance level.…”