Special condition wind power forecasting based on Gaussian Process and similar historical data

Due to the variability of wind power, it is imperative to accurately and timely forecast wind generation to enhance the flexibility and reliability of the operation and control of real-time power systems. Special events such as ramps and spikes are hard to predict with traditional methods using solely recently measured data. In this paper, a new Gaussian Process model with hybrid training data taken from both the local time and historical dataset is proposed and applied to make short-term predictions from 10 minutes to one hour ahead. A key idea is that the historical data with similar patterns are selected and then embedded in Gaussian Process model to make new predictions. The results of the proposed algorithms are compared to those of standard Gaussian Process model and the persistence model. It is shown that the proposed method not only reduces the magnitude error but also the phase error.