• Session Materials

Research on short-term wind power volatility characteristics based on high-order Markov chain

Ref C2_12519_2026 • 2026
  • This publication is available for download to delegates only.
    A delegate is someone registered for the Session.

Summary

Random meteorological variations cause stochastic wind power fluctuations that challenge stable state-transition modeling and grid operation. This paper proposes a probabilistic framework combining high-order Markov chains and Weibull state distributions to characterize short-term volatility. First-order differencing and exponential-similarity-based FCM are used to construct discrete power states with similar fluctuation features. A high-order Markov chain then captures multi-step temporal dependence and cumulative transition behavior, while

Weibull distributions describe state occurrence and extreme tendencies. Using six months of

SG5.0-145 turbine data, the method outperforms Elman and stacked autoencoder baselines in fluctuation coefficient, peak factor, MAE, and RMSE, with tests confirming its robustness.

Additional informations

Publication type Session Materials
Reference C2_12519_2026
Publication year
Publisher CIGRE
Country China, People's Republic of
Study committees
File size 880 KB
Price for non member 30 €
Price for member 30 €

Authors

MO Ruohui - Hainan Power Grid Corporation; HE Yongqi - Hainan Power Grid Corporation; WANG Jun - Hainan Power Grid Corporation; ZOU Minglong - Hainan Power Grid Corporation; SONG Jialin - Hainan Power Grid Corporation; MAO Lifan - Hainan Power Grid Corporation; DAI Yangyu - Hainan Power Grid Corporation

Keywords

High-order Markov chain; Wind power output; First-order difference; Index similarity coefficient; Maximum likelihood estimation

Top of the page

Research on short-term wind power volatility characteristics based on high-order Markov chain