A Novel Hammerstein Wind Power Forecasting Model

This paper proposed a novel Hammerstein wind power forecasting model. For the wind power system with strong nonlinear characteristics, it combines the advantages of Hammerstein model and fuzzy neural network. Moreover, in the part of the proposed model parameter learning, a weight adjustment method based on Lyapunov's global convergence is proposed. The performance of the novel Hammerstein wind power forecasting model is assessed by contrast experiment, including the auxiliary model recursive least squares (AMRLS) Hammerstein wind power forecasting model and a multiple input single output fuzzy neural network wind forecasting model. Results show that the model is superior in prediction accuracy. So, the new model is beneficial to various applications of integrated energy and renewable energy, especially power scheduling.