Wind power prediction using a hybrid approach with correction strategy based on risk evaluation

With the rapid increase of renewable-energy capacities, the management of grid-connected wind farms is becoming more and more important. In this paper, a very short-term wind power prediction (VSTWPP) method with hybrid strategy based on risk evaluation is proposed. The VSTWPP is essential for both producers and consumers in the electricity market, because it can reduce uncertainties of wind power fluctuation and thus maintain power balance, security and quality of the system. This paper focuses on a hybrid approach with correction (HWC) strategy for the VSTWPP method, in which the Gaussian model is applied to calculate the probability distributions of wind power value and its error during different time periods and different methods. The WPP process includes: 1) Wind power ratios are predicted using the hybrid approach of multiple linear regression and least squares; 2) Transformation of these ratios is performed to obtain predicted wind power values; 3) Correction strategy is implemented to obtain the final results of WPP. Besides, in order to observe the prediction performance, WPP model with HWC, with the hybrid approach without correction (HWoC), with autoregressive moving average (ARMA) and with autoregressive integrated moving average (ARIMA) are examined respectively. The results confirm the accuracy and validity of the proposed HWC-based VSTWPP method, and show great promise for the prediction within intricate time series, which are highly volatile, irregular and uncertain. The obtained results confirm an observable accurate for the prediction validity of the proposed hybrid approach with correction strategy.