Simulation of wind power time series based on the MCMC method

Building an accurate and reasonable time-series model of wind power is of great significance for the power system operation and planning. This paper proposes a Markov chain Monte Carlo (MCMC) method to simulate the time series of the wind power. Taking into account the nonstationarity and stochastics of the wind power, this model constructs Markov chain for time series of the wind power to keep the stochastics and utilize the Gibbs sampling to realize the probability transition matrix. A numerical case study with one-year real measurement of the wind power from one large wind farm in Gansu province of China is applied to illustrate the validity the proposed model: the simulated wind-power time series of arbitrary length that accurately reproduce the statistical behavior of the original series.