基于K-means聚类技术改进的多线性蒙特卡洛概率能流算法

可再生能源的间歇性出力以及负荷的波动给综合能源系统(integrated energy systems,IES)引入了大量不确定性因素。提出了一种基于K-means聚类技术改进的多线性蒙特卡洛概率能流计算方法。首先,引入输入随机变量整体灵敏度系数概念,并以权重系数的形式修正输入随机变量样本,改进K-means聚类效果,确保各聚类簇均具有较小的波动范围。然后,采用多线性化求解思路进行概率能流计算,即对聚类中心进行确定性能流计算,而各聚类簇中输入随机变量样本利用同一簇聚类中心处得到的状态变量和雅可比矩阵进行线性化能流求解,从而减少了迭代过程,提高计算效率。以IEEE57节点电力系统和14节点天然气网络构成的IES为算例,验证了所提方法比传统蒙特卡洛法具有更高的计算效率,相比现有多线性蒙特卡洛算法具有更高的准确性和计算效率。

Improved Multi-Linear Monte Carlo Probabilistic Energy Flow Calculation Method Based on K-Means Clustering Technique

Intermittence of renewable energy resources and fluctuation of loads introduce uncertainties into integrated energy systems.This paper proposes an improved multi-linear Monte Carlo simulation for probabilistic energy flow computation based on K-means clustering technique.Firstly,the overall sensitivity coefficients of input random variables are introduced to modify input random variable samples to achieve better clustering result and ensure that input random variable samples in each cluster have smaller fluctuations than original total samples.Secondly,based on the clustering results,multi linearization method is used to solve probabilistic energy flow.Namely,the deterministic energy flow method is performed at each cluster center to obtain corresponding state variables and Jacobian matrix.The state variables and Jacobian matrix obtained at each cluster center are utilized to approximate state variables for other samples within the same cluster,thus reducing iteration and improving computational efficiency.Finally,an integrated energy test system comprised of IEEE 57-node power system and 14-node natural gas network is used to confirm the proposed method.The proposed method has higher computational efficiency than Monte Carlo method and has higher accuracy and computational efficiency than existing multi-linear Monte Carlo algorithm.