Short-term prediction of offshore wind power based on an improved N-BEATSX model

Short-term forecasting of offshore wind power plays an essential role in ensuring the safe and stable operation of the grid. Due to the high randomness and variability of offshore wind power, it is difficulty to acquire good results by traditional trend extrapolation of time series of historical power data. In this regard, a novel N-BEATSX model improved by Copula entropy is proposed as a short-term prediction method for offshore wind power. First, the Copula entropy metrics are used to filter predictors to obtaining factors highly relevant to the active power output of offshore wind. Secondly, the auxiliary and historical power data from the wind farms are imported into the N-BEATSX model for future wind power forecasting. Finally, a numerical example is given to demonstrate the feasibility of the developed method. Comparing the prediction results of the proposed model with those of other models, it is demonstrated that the proposed method improves the accuracy of short-term prediction of offshore wind power.