为了能够生成与已有风电功率序列数据特性一致的风特性的改进马尔可夫链蒙特卡罗(Markov chain Monte Carlo,PV-MC)法,即持续与波动蒙特卡罗(persistence and variation-Monte Carlo,PV-MC)法。该方法基于风电功率状态,首先生成满足状态跳变率矩阵的状态序列;而后,利用风电功率状态的持续特性,确定状态序列中状态的持续时间,得到满足持续特性的状态序列;最后,基于波动特性,将状态序列转换为风电功率序列。利用PV-MC方法与传统的MCMC法分别对全球6个不同地区共26座风电场生成风电功率序列,并与原始风电功率序列进行特性对比分析,结果表明:无论在基本统计特性(均值、标准差、概率密度函数和自相关系数)还是在时域特性(持续性和波动性)上,PV-MC法生成的风电功率序列都优于传统的MCMC法所生成的序列。
In order to generate a new wind power time series capturing the time domain features as same as the original wind power time series, an improved Markov Chain Monte Carlo (MCMC) method, referred to as persistence and variation-Monte Carlo(PV-MC) method, was proposed in this paper. The method considered the persistence and variation characteristics of wind power. Firstly, the wind power state series was generated to meet the state transition matrix based on the definition of wind power state. Then the duration time of every state in the series was determined by its duration character. Finally, the variation characteristic was used to convert the state series to wind power series. PV-MC method and the traditional MCMC method had been used to generate wind power time series based on the original wind power time series obtained from 26 wind farms located in 6 various places all over the world. The characteristics of the new time series have been compared with that of the original wind power time series respectively. The results show that PV-MC method is superior to the traditional MCMC method in the conventional statistic characteristics (mean value, variance, probability density function, autocorrelation function) and time domain features (persistence characteristics as well as the variation characteristics).