This article proposes a method based on data clustering for the optimal placement and sizing of wind farms (WFs) in radial distribution systems considering the wind uncertainty and the wake effect. The stochastic output of wind turbines makes the optimization problem more challenging. Using the probabilistic methods for such analyses proves efficient to yield more reliable results. The data clustering-based Monte Carlo simulation is used to bunch the output powers of wind farm into clusters. The network loss and voltage profile are then evaluated to achieve the optimal candidate bus for the placement of WF with the optimal number of turbines. To accelerate the procedure, a technical constraint is used so as to eliminate the nonimportant samples in WF size evaluation, which results in a reduction in computational burden. In addition, two methods for load flow calculations are investigated, namely the direct load flow and indirect backward/forward sweep load flow methods to evaluate the time burden of load flow method on the proposed problem. The effectiveness of the proposed methodology is illustated by conducting case studies.

中文翻译：

---

考虑尾流效应和时间加速约束的径向配电系统中基于聚类的风电场布局方法


本文提出了一种基于数据聚类的方法，用于考虑风的不确定性和尾流效应，在径向分布系统中优化风电场（WF）的布局和规模。风力发电机的随机输出使得优化问题更具挑战性。事实证明，使用概率方法进行此类分析可以有效地产生更可靠的结果。采用基于数据聚类的蒙特卡罗模拟将风电场的输出功率进行聚类。然后评估网络损耗和电压分布，以获得最佳候选母线，用于放置具有最佳数量涡轮机的 WF。为了加速该过程，使用技术约束来消除WF大小评估中的不重要样本，从而减少计算负担。此外，还研究了两种潮流计算方法，即直接潮流和间接向后/向前扫描潮流方法，以评估潮流方法对所提出问题的时间负担。通过案例研究说明了所提出方法的有效性。