Adaptive ultra-short-term wind power prediction based on risk assessment

Abstract

A risk assessment based adaptive ultra-short-term wind power prediction (USTWPP) method is proposed in this paper. In this method, features are first extracted from the historical data, and then each wind power time series (WPTS) is split into several subsets defined by their stationary patterns. A WPTS that does not match any of the stationary patterns is then included in a subset of non-stationary patterns. Each WPTS subset is then related to a USTWPP model that is specially selected and optimized offline based on the proposed risk assessment index. For online applications, the pattern of the last short WPTS is first recognized, and the relevant prediction model is then applied for USTWPP. Experimental results confirm the efficacy of the proposed method.