Probabilistic short-term wind power forecasting based on deep neural networks

Abstract

High-precision wind power forecasting is an essential operation issue of power systems integrated with large numbers of wind farms. In addition to traditional forecasting methods, probabilistic forecasting is recognized as an optimal forecasting solution since it provides a wealth of valuable uncertainty information of wind power. In this paper, a novel approach based on deep neural networks (DNNs) for the deterministic short-term wind power forecasting of wind farms is proposed. DNN models including long short-term memory (LSTM) recurrent neural networks (RNNs) have achieved better results compared with traditional methods. Further, probabilistic forecasting based on conditional error analysis is also implemented. Favorable results of probabilistic forecasting are achieved owing to elaborate division of the conditions set based on cluster analysis. The performance of the proposed method is tested on a dataset of several wind farms in north-east China. Forecasting results are evaluated using different indices, which proves the effectiveness of the proposed method.