基于场景分区的半不变量法概率电压稳定评估

Probabilistic Voltage Stability Assessment Using Scenario Partition Based Cumulant Method

针对高比例可再生能源电力系统采用半不变量法进行电压稳定负荷裕度计算中存在的由于随机变量波动范围大造成精度低的问题,提出一种基于场景分区的半不变量法负荷裕度概率评估方法。首先采用Wasserstein距离指标和轮盘赌选择算法建立可再生能源随机出力多场景概率模型,形成初始场景集;其次基于改进K-means聚类算法对初始场景集进行分区;继而采用半不变量法和Gram-Charlier级数展开法对每个分区分别计算负荷裕度的概率分布,最后应用全概率公式获得场景全集中负荷裕度的整体概率分布。该文对工程应用中场景分区数目选取策略进行探讨,并对IEEE118节点系统进行仿真分析,验证了所提方法的有效性。

Aiming at the low accuracy of the cumulant method caused by large fluctuation range of random variables in the calculation of load margin in high proportion renewable energy power systems,a scenario partition based cumulant method for probabilistic voltage stability assessment is proposed.Firstly,The Wasserstein distance index and roulette selection algorithm are used to establish the multi-scenario probability model of stochastic output power of renewable energy,from which the initial scenario set is obtained.Secondly,the initial scenario set is partitioned based on the improved K-means clustering algorithm.Then,the cumulant method and Gram-Charlier series method are used to calculate the probabilistic load margin for each partition.Finally,the overall distribution of load margin is obtained by using the total probability formula.The selection strategy of the number of scenario partitions for engineering applications is also discussed in this paper.The numerical results on IEEE-118 system show that the proposed method is effective.