基于风电出力特性和Copula理论的风电时间序列生成

发布时间：2018-03-15 00:26

本文选题：风电　切入点：时间序列　出处：《华中科技大学》2015年硕士论文　论文类型：学位论文

【摘要】：随着全经济的发展,不可再生能源逐渐减少。为了应对这一问题,新能源成为了发展和研究的热门领域。风能由于清洁和可再生成为了研究的重点,但是风电的随机性对电力系统带来了重大的影响,同时我国风资源分布不均,大型风电生产基地都在远离负荷中心的位置,造成了输送的困难,弃风严重。为此必须对风电出力的特性进行详细的分析,根据风电特性制定相应的电力系统建设与控制策略。目前对于风电出力特性的研究大都停留在定性研究阶段,无法全面的反映风电出力特性,相关工作急需开展。风电出力时间序列对于风电接入电力系统的研究必不可少,目前已经有大量关于短期风电功率预测的研究,但此类研究计算方法复杂,预测时间大都不能超过3天,对于长时间尺度的风电出力时间序列,主要有风速和时间序列法,物理法根据风速模型结合风机出力曲线得到风电功率,需要考虑尾流效应等多种风电场现实因素,计算复杂;时间序列法包括传统的MCMC(蒙特卡洛马尔科夫链)法和AMRA(自回归滑动平均模型)法,它们都有着明显的缺陷,对风电出力特性的刻画也不够全面。为了解决以上两个问题,本文对国内外24个风电场/群的历史数据进行了研究,总结了风电出力的特性,包括均值方差,自相关系数、不同时间尺度的出力概率分布,波动特性等,并对不同年份的出力概率分布进行了对比。基于风电出力的概率分布特性、自相关特性与波动特性,提出了使用Copula理论的时间序列生成方法,对该方法进行了大量仿真分析,对比了国内外风电场原始序列以及MCMC方法生成的时间序列,验证了该方法的精度,该方法能够很好的反映风电出力特性。
[Abstract]:With the development of the whole economy, non-renewable energy is gradually decreasing. In order to deal with this problem, new energy has become a hot area of development and research. Wind energy has become the focus of research because of clean and renewable energy. However, the randomness of wind power has a great impact on the power system. At the same time, the distribution of wind resources in China is uneven, and the large wind power production bases are located far from the load center, resulting in the difficulty of transmission. For this reason, the characteristics of wind power output must be analyzed in detail, and the corresponding power system construction and control strategies should be formulated according to the wind power characteristics. At present, the research on wind power output characteristics mostly stays at the stage of qualitative research. The wind power generation time series is indispensable for wind power to be connected to the power system, and there has been a lot of research on short-term wind power prediction. However, this kind of research and calculation methods are complex, and most of them can not be predicted for more than 3 days. For a long time scale wind power generation time series, there are mainly wind speed and time series methods. According to the wind speed model combined with the wind force curve, the physical method can get the wind power, which needs to consider the wake effect and other practical factors of wind farm, so the calculation is complicated. The time series methods include the traditional MCMC (Monte Carlo Markov chain) method and the AMRA( autoregressive moving average model) method. Both of them have obvious defects, and the characterization of wind power output characteristics is not comprehensive. In order to solve the above two problems, In this paper, the historical data of 24 wind farms / clusters at home and abroad are studied, and the characteristics of wind power output are summarized, including mean variance, autocorrelation coefficient, probability distribution of output force at different time scales, fluctuation characteristics, etc. Based on the probability distribution characteristics of wind power output, autocorrelation and fluctuation characteristics, a time series generation method using Copula theory is proposed, and a large number of simulations are carried out. By comparing the original sequence of wind farm at home and abroad and the time series generated by MCMC method, the accuracy of the method is verified, and the method can well reflect the characteristics of wind power output.
【学位授予单位】：华中科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TM614

【相似文献】