Abstract
In this paper, we propose nonlinear wind prediction methods for predictive control of wind power generation, based on techniques of nonlinear time series analysis. The proposed methods provide prediction of wind at an observational equipment in contiguity with a windmill, and achieve local wind prediction suitable for the time-space scale of windmill by the reconstruction of a local wind dynamical system only from observed time series data. We demonstrates that the methods have predictability of real wind time series data.
