Wind Power Prediction Using a Hybrid Approach with Correction Strategy Based on Risk Evaluation

Abstract

With the rapid development of renewable energy applications and for managing wind power farms connected to power grid, in this work, is proposed evaluation risk based hybrid strategy for very short-term wind power prediction (VSTWPP) method, it has become more important. VSTWPP is essential in the electricity market for producers and consumers. Power balance of network at any moment must be maintained due to security and quality, and the uncertainty of wind power fluctuation can be reduced by accurate prediction method. This paper focuses on a hybrid approach with correction (HWC) strategy for VSTWPP. In this study, Gaussian model is used to calculate the probability distributions of wind power value and its error for different time periods and different methods. The WPP is achieved in three phases: 1) Ratios of wind power were predicted using hybrid approach of multiple linear regression and least squares; 2) Transformation of these ratios was performed to obtain predicted wind power values; 3) Correction strategy was implemented to obtain the final results of WPP. Comparisons between HWC, hybrid approach without correction (HWoC), autoregressive moving average (ARMA) model and autoregressive integrated moving average (ARIMA) model were done to observe prediction performance. Observation shows that the error of WPP with the proposed HWC was significantly lower than those with other methods.