基于分段Copula函数和高斯混合模型的多段线性化概率潮流计算

Multistage Linearization Probabilistic Power Flow Calculation Based on Piecewise Copula and Gaussian Mixture Model

大规模风电并网以及负荷的随机波动加剧了电网运行的不确定性,为了有效分析新环境下的电网运行特性,提出一种基于风功率分段Copula函数和负荷高斯混合模型的多段线性化概率潮流计算方法。采用分段Copula函数在时间维度上刻画相邻风电场的空间相关性,分析风功率相关性的季节变化。针对实际负荷的不对称、多峰特性,采用改进K-means聚类优化的期望最大化(expectation maximization,EM)算法,准确快速地建立负荷高斯混合模型。在此基础上,采用多段线性化半不变量法进行概率潮流计算,以减小风功率和负荷大范围波动造成的潮流方程线性化误差。对改进的IEEE 14节点系统进行仿真分析,验证了所提方法的准确性、快速性及有效性。

Large-scale integration of wind power into grid and load stochastic volatility increase the uncertainty in power system operation,in order to effectively analyze the system operation features in the new environment,a calculating method based on wind power piecewise Copula and load Gaussian mixture model for multistage linearization probabilistic power flow is proposed. Piecewise Copula is used to establish the spatial correlation model among wind farms on the time dimension considering seasonal variation. For non-normal and multimodal load,expectation maximization（ EM） algorithm is used to establish load Gaussian mixture model,and an improved K-means clustering is proposed to optimize EM algorithm,which can simplify the modeling process. On the premise of these models,calculating probabilistic power flow in the method of multistage linearization cumulant method,fully considering the impact of wind power and load fluctuation on the system operation. The accuracy and efficiency of the proposed probabilistic power flow calculation process is verified through the test on modified IEEE 14-bus system.