A new prediction method based on VMD-PRBF-ARMA-E model considering wind speed characteristic

Abstract

With the characteristics of randomness, fluctuation, nonlinearity and uncertainty, wind speed affects the stability of wind power system. In order to improve the safety and stability of wind power system, accurate and effective wind speed prediction is essential. In the paper, a novel wind speed prediction method based on wind speed characteristics is proposed. Firstly, VMD is used to decompose wind speed into the nonlinear part, the linear part and the noise part. Nonlinear part reflects the nonlinear characteristic of wind speed, linear part embodies the linear process of wind speed formation, noise part is the error (ER) sequence decomposed wind speed by VMD. According to the characteristics of different parts, different models are built, PCA-RBF model is built for the nonlinear part, ARMA model under the MCMC framework is built for the linear part, and probability distribution is fitted for the noise part. These three parts are combined to establish VMD-PRBF-ARMA-E model to make off-line deterministic prediction and uncertainty prediction. Then the superiority of VMD-PRBF-ARMA-E model is verified by comparing with other nonlinear models and time series models. At last, based on off-line scheme, VMD-PRBF-ARMA-E model is used to make real-time wind speed prediction. The deterministic prediction of VMD-PRBF-ARMA-E model has high accuracy, and can reflect the characteristics of wind speed well and truly, which can provide a scientific basis for the power grid dispatching department, and help to ensure the stability of wind power system.