含风电功率时域特性的风电功率序列建模方法

发布时间：2018-03-11 18:48

本文选题：风电功率序列　切入点：马尔科夫链　出处：《中国电机工程学报》2014年22期 　论文类型：期刊论文

【摘要】：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。
[Abstract]:In order to generate the improved Markov chain Monte Carlo Markov chain Monte Monte PV-MCmethod, which can generate the wind characteristics consistent with the existing wind power sequence data, that is, the persistence and volatility Monte Carlo persistence and variation-Monte Carlos PV-MC-based method, this method is based on the wind power state. First, the state sequence satisfying the state jump rate matrix is generated; then, the duration of the state in the state sequence is determined by using the sustained characteristics of the wind power state, and the state sequence satisfying the sustained characteristic is obtained. Finally, based on the fluctuation characteristic, the state sequence is obtained. The state sequence is converted to wind power sequence, and the PV-MC method and the traditional MCMC method are used to analyze the wind power series of 26 wind farms in 6 different regions of the world, and the characteristics of the wind power series are compared with the original wind power series. The results show that: regardless of the basic statistical characteristics (mean, standard deviation), Both the probability density function and the autocorrelation coefficient) or the wind power series generated by PV-MC method are superior to those generated by the traditional MCMC method on the time-domain characteristics (persistence and volatility).
【作者单位】：强电磁工程与新技术国家重点实验室(华中科技大学电气与电子工程学院);ABB中国有限公司集团研究中心;
【基金】：国家863高技术基金项目(2011AA05A112) 国家自然科学基金重点项目(50937002) 中国博士后特别资助项目(2013T60717)~~
【分类号】：TM614

【相似文献】