An Overview of Data Preprocessing for Short-Term Wind Power Forecasting

Wind power generation takes on an increasingly vital role in the power grid due to its environmental and economic benefits. However, the primary challenges that are related to the integration of wind power into power systems include variability, uncertainty. An accurate forecasting reduces operating costs and enhances power system stability. Wind power forecasting include many steps, including data collection, data preprocessing, the construction and training for models, and error calculation. Among them, data preprocessing plays an important role on the process of wind power forecasting since the inputs of the forecasting model would be sensitive to the quality of data. As a result, this paper presents a survey on the methods for wind-data processing. These methods aim to preprocess and extract suitable features from numerical weather prediction (NWP) wind speeds and measured wind power data. Finally, this paper used a case study to demonstrate the important of the preprocessing step on wind power forecasting.