Cost-Effective Adaptive Predictor for Large-Scale Wind Signal

This study introduces a large-scale short-term wind signal predictor via the online censoring-based complex-valued least mean kurtosis (OC-CLMK) algorithm. This predictor owing to its OC-CLMK algorithm overcomes the overwhelming computational challenges resulting from large-scale short-term wind data by censoring the noninformative data during the prediction process. The proposed predictor also simultaneously predicts the amplitude and phase information of the wind signal because of its ability to work on the complex domain. Moreover, the proposed predictor provides satisfactory prediction performance when compared to other predictors based on existing algorithms. The simulation results on real-world large-scale wind signal indicate that the proposed predictor can predict short-term wind signals with high accuracy and low runtime.