The optimal state number selection of wind power series generated through MCMC method

Wind power series generation is an effective method to solve the problem of finite time length of wind power output data in simulation. It can provide the necessary data support for the plan and operation of the power system with wind power. Hence, it is of great significance. Markov Chain Monte Carlo (MCMC) method as its specific technology is widely used. But how to select reasonable number of state division is always a main problem existing in the application process of MCMC method. Related research mainly relies on experience to select state number, which is lack of corresponding basis. In order to solve the problem, this paper put forward using the cumulative distribution function (CDF) within the power scope of each state to sample and generate specific wind power randomly combined with MCMC method, which will ensure the overall distribution characteristics of the wind power generated are not changed with the selection of state number. Then this paper proposed a method to select the optimal state number. This method takes the minimum residual sum of square (RSS) of the autocorrelation function (ACF) of the generated series and the original series as the principle to select the state number. Respectively using the method proposed in this paper and the traditional method to generate wind power series for actual wind farms, the results show that the wind power series generated by the improved MCMC method is much better than that generated by the traditional MCMC method in terms of fitting both the distribution and autocorrelation characteristics of original wind power series.