Gaussian models for probabilistic and deterministic Wind Power Prediction: Wind farm and regional

Due to the unstable situation of the atmosphere, the wind power forecasting of the wind farms (WFs) will be so complex. This unsteady provides an opportunity in order to investigate the wind power probabilistic prediction. This paper proposed a novel method based on Gaussian Processes (GPs) to improve the probabilistic prediction of a group or regional of WFs. Covariance Functions (CFs), known as Kernel, are the key ingredient in using GPs. One of the most usually of these functions is Squared Exponential (SE), which is applied with other functions to the model of the proposed method. Thus, different groupings of CFs are investigated comprehensively. Additionally, evaluating the accuracy of the prediction is carried out. This study went through two types of comparisons of dynamic and static GP as well as direct and indirect prediction plan. The result of the comparison between dynamic and static GP revealed that the dynamic GP generates keen Prediction Intervals (PIs). Besides, comparing the accuracy of direct and indirect prediction plan, it shows that indirect prediction strategy brings about wider PIs with higher coverage probability on the part of net demand prediction. Moreover, the proposed model provides precise results of forecasted energy in every time step. (C) 2020 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.