Short-term wind power forecasting using wavelet-based neural network

Wind power generation highly depends on the atmospheric variables which itself depend on the time of the day, months and seasons. The intermittency of wind hinders the accuracy of wind forecasting, which is important for safe operation and reliability of future power grid. One way to address this problem is to consider all these atmospheric variables which can be obtained from Numerical Weather Prediction (NWP) models. However, using NWP parameters increases the complexity of the forecast model and it requires a large amount of historic data. Additionally, different models are required for different seasons or months. This paper presents a wavelet-based neural network (WNN) forecast model which is robust enough to predict the wind power generation in short-term with significant accuracy, and this model is applicable to all seasons of the year. With reduced complexity, the model requires less historic data as compared to that in available literatures.