基于综合特征矩阵的配电网故障判别方法

配电网短路故障类型的快速识别是保证故障精准可靠切除的基础。提出了一种基于综合特征矩阵的配电网故障判别方法，该方法综合考虑长距离输电、非线性负载投切、数据量残缺、噪声和高阻抗接地等情况下的配电网短路故障类型识别的难点，基于Hilbert变换得到的电压均方根（RMS）值、离散傅立叶变换得到的电压谐波信息及Hilbert-Huang变换得到的电压突变信息构造特征值进行故障类型判别。通过搭建含3条馈线的10 kV辐射配电网络Simulink仿真测试系统，对所提的综合特征矩阵方法进行了验证，并与经典故障检测方法进行了对比分析，论证了所提方法的精度优势。

Distribution Network Fault Type Identification Method Based on Feature-summarizing Matrix

Fast identification of short-circuit fault types of distribution network is the basis for accurate fault removal. A method for detecting short-circuit fault types in distribution networks is proposed based on comprehensive feature-summarizing matrix. This method comprehensively considers the difficulties in short-circuit fault-type identification of distribution networks in the case of long-distance transmission, non-linear load switching, incomplete data volume, noise and high-impedance grounding, and the fault types are identified based on the voltage RMS value obtained by Hilbert transform, the voltage harmonics obtained from the Discrete Fourier transform and the converted voltage information by the Hilbert-Huang transform. The proposed feature-summarizing matrix method was verified through a Simulink simulation test system of a 10kV radial network consisting of three feeders, and compared with the traditional methods, which has proved the accuracy advantages of the proposed method.