Wind power forecasting has become an essential tool for energy trading and the operation of the grid due to the increasing importance of wind energy. Therefore, estimating the forecast accuracy of a WPF model and understanding how the accuracy is calculated are necessary steps to appropriately validate WPF models. The present study gives an extensive overview of the performance evaluation methods used for assessing the forecast accuracy of short-term statistical wind power forecast estimates, and the concept of robustness is introduced to determine the validity of a model over different wind power generation scenarios over the testing set. Finally, a numerical study using decomposition-based hybrid models is presented to analyse the robustness of the performance evaluation metrics under different conditions in the context of wind power forecasting. Data from Ireland are employed using two different resolutions to examine its influence on the forecast accuracy.

由于风能的重要性日益提高，风能预测已成为能源交易和电网运营的重要工具。因此，估计WPF模型的预测准确性并了解如何计算准确性是正确验证WPF模型的必要步骤。本研究对用于评估短期统计风电预测估计值的预测准确性的性能评估方法进行了广泛的概述，并引入了稳健性的概念来确定模型在不同风电情景下的有效性。测试集。最后，利用基于分解的混合模型进行了数值研究，以分析风电预测情况下不同条件下性能评估指标的鲁棒性。