采用Spiked协方差模型与“相变”现象的电网不平衡扰动评估

Unbalanced disturbance evaluation in power grid using Spiked covariance model and phase transition phenomenon

为了提升随机矩阵理论在电网不平衡扰动场景中的适用性,提出了一种基于Spiked协方差模型与"相变"现象的电网扰动识别与定位方法。首先,通过含噪声三相数据源的采集,构造Spiked协方差模型与样本协方差矩阵。然后,利用特征值"相变"现象构建样本最大特征值评价指标及对应的动态阈值。当该指标越过其阈值,即判定电网中有扰动事件发生时,根据Spiked与样本最小特征向量元素改变量的网络位置,结合特征向量"相变"现象以实现电网不平衡扰动的快速定位。最后,借助DIgSILENT和MATLAB R2014a软件,案例分析在一个IEEE 118母线177支路系统和一个德国872母线1 840支路实际配电系统中展开,涉及负荷突变和线路故障等不平衡扰动事件,与传统随机矩阵理论方法的比较结果表明了所提方法的有效性和高效性。

To enhance the application of random matrix theory(RMT) in the scenario of unbalanced disturbance event in power grid, an identifying and positioning method based on the Spiked covariance model and phase transition phenomenon is proposed. Firstly, a Spiked covariance model and a sample covariance matrix are established by a three-phase data source with noise interference. Then, the maximum eigenvalue of the sample covariance matrix(MESCM) is taken as a disturbance identification index. Meanwhile, its corresponding dynamic threshold can be obtained by the phase transition of eigenvalues. According to the network location associated with the changing elements involving both Spiked and sample minimum eigenvector, the unbalanced disturbance event in the power grid can be rapidly located by employing the phase transition of eigenvector if the dynamic threshold for MESCM is violated. The case studies are implemented on an IEEE 118-bus and 177-line system, and a real 872-bus and 1 840-line distribution network system in Germany with the help of DIgSILENT and MATLAB R2014 a software. The unbalanced disturbance events are involved, such as load variations and line faults. Compared with the traditional RMT based method, the proposed method can achieve better effectiveness and efficiency results.