Wind Power Time Series Simulation Model Based on MCMC Algorithm and Daily Fluctuation Characteristics

Wind power time series modeling plays a pivotal role in enhancing microgrid operational stability and optimizing renewable energy planning. This study proposes an improved Markov Chain Monte Carlo (MCMC) model for simulating hourly annual wind power data. By incorporating daily periodic and short-term fluctuation characteristics into the state transition matrix, the model effectively captures the long-term periodicity and intra-day variations of wind power. Additionally, the integration of wavelet transform techniques suppresses high-frequency noise, significantly improving the accuracy of short-term fluctuation fitting. The number of states and the rejection factor are optimized using Genetic Algorithm, enabling efficient exploration of complex parameter spaces and convergence toward global optima. Compared with traditional modeling methods, the proposed MCMC model demonstrates superior performance in fitting probability distribution functions and autocorrelation functions. It also effectively addresses the nonlinear and stochastic nature of wind power data, showcasing remarkable capability in generating high-quality synthetic time series.