Multi-Scale Wind Power Time Series Modeling Method Based on Mathematical Morphology

A novel Wind Power Time Series (WPTS) model considering wind power fluctuations on different time scales is proposed in this paper. Weighted mathematical morphological filter (WMMF) is introduced to decompose WPTS into components of three time scales. All the components are modeled with the correlation between different components being considered. The Markov chain is adopted to simulate the long-term fluctuation of WPTS. Compared to the traditional Markov Chain Monte Carlo (MCMC) model, the state of Markov chain in this model is the fluctuation pattern rather than the discrete wind power value, which enables to simulate fluctuation processes with the information of duration, frequency and intensity etc. The simulation results prove that the proposed WMMF model offers excellent fit for the statistical characteristics of actual WPTS, and it outperforms the traditional MCMC model according to the accuracy testing in different time horizons.