论文部分内容阅读
用少量的代表性风电功率序列场景来准确刻画风电随机特征，对含有风电电力系统的规划和运行具有重要意义。然而，代表性风电功率序列场景的生成，目前方法难以实现从庞大的发生空间中选择有效的代表场景，场景模拟的质量有待提高。为此，提出一种纵横双向优化的方法以生成日风电功率序列场景。纵轴方向，基于历史的日风电功率序列数据，采用最优消减技术，产生每个时段的代表场景；横轴方向，采用禁忌搜索方法，有选择地连接每个时段的代表场景从而形成所需的日风电功率序列代表场景。该方法无需预先知道风电功率的解析概率分布函数，仅需基于已有的历史序列数据，通过纵横双向优化，自动生成满足风电随机概率特征的日序列代表场景。以爱尔兰风电场数据为例，对所产生的单时段代表场景，在均值、方差、偏态和峰度4个指标上具有与历史数据相近的统计特性；将这些场景应用于含有风电电力系统的多时段最优潮流问题，从稳定性和准确性两个方面，验证了所提出的双向优化算法的有效性。“,”Generating a set of scenarios that represent the stochastic characteristics of the time series of wind farm output is very important for power system planning and running. However, the existing methods for producing wind power time series scenarios are difficulty in selecting effective representing scenarios from hugeness space and the accuracy of the scenarios are not satisfied. In light of this, this paper proposed a two-dimension optimization method to generate daily wind power time series scenarios through vertical and horizontal direction. Representative scenarios of every time period were obtained by reducing the historical daily wind power series in the longitudinal direction. Tabu search algorithm was used to choose a scenario for each time period to form a representative wind power time series scenario. The proposed method can obtain the representative wind power time series scenarios automatically through two-dimension optimization without knowing the probability distribution function of wind power, which is hard to obtain. Based on the wind farm data of Ireland, the reduced scenarios in each time period were proved to have similar statistical characteristics with the history data when comparing the indexes of mean, variance, skewness and kurtosis. The stability and accuracy of the proposed method is verified by a number of cases solved using multi-stage optimal power flow.