谐振接地系统发生单相接地故障时，因其故障电流微弱而不易检测，并且故障特征量的大小受故障合闸角、接地电阻、故障位置以及系统补偿方式的影响，使得基于单一故障特征区段定位方法的可靠性和适用性受到限制。基于此，本文在对谐振接地系统故障特征量研究的基础上，提出了基于层次聚类的多特征量融合的故障区段定位新方法，该方法可以充分地利用故障特征量，提高故障区段定位的准确率。 分析了在谐振接地系统中发生单相接地故障时，零序电流五次谐波幅值分量与相位分量、瞬时正序实部电流工频幅值以及零序电流本征模态能量这四个特征量的物理特点及其区段定位原理，并对这四个特征量在不同故障条件下进行区段定位的有效性进行了研究。经分析可得，这四个特征量在不同的故障条件下具有优势互补性，融合这四个特征量进行区段定位，可以提高区段定位的准确率，为基于融合特征量的故障区段定位方法的提出奠定了基础。 提出了基于层次聚类的多特征量融合的故障区段定位方法。论文采用原理简单、易于实现的层次聚类作为融合算法，对零序电流五次谐波幅值、零序电流五次谐波相位、瞬时正序实部电流工频幅值以及零序电流本征模态能量这四个特征量进行融合；将这四个特征量作为检测点的特征向量，通过欧氏距离对各个检测点之间相关性的度量，实现了阈值的合理设置，从而构建了故障区段定位的判据，实现了基于层次聚类多特征量融合的故障区段定位。 为了验证本论文提出的基于层次聚类多特征量融合的故障区段定位方法的准确性，本论文设定了不同的故障条件，进行了大量的仿真验证及实验验证，结果表明，该故障区段定位方法准确可靠，具有有效性与适应性。

When a single-phase ground fault occurs in a resonant grounding system, the fault current is weak and not easy to detect, and the magnitude of the fault characteristic is affected by the fault closing angle, grounding resistance, fault location, and system compensation mode, so that the single fault characteristic section is based on The reliability and applicability of the positioning method are limited. Based on this, based on the research on the fault feature quantity of the resonant grounding system, a new fault segment location method based on hierarchical clustering for multi-feature fusion is proposed. The method can fully utilize the fault feature quantity and improve the accuracy of the fault segment location. The physical characteristics and its segmentation principle of the four characteristics of the fifth harmonic amplitude component and phase component, the instantaneous positive sequence real current frequency amplitude and the zero sequence current eigenmode energy are analyzed when the single phase ground fault occurs in the resonant grounding system. And the effectiveness of segmentation of these four feature quantities under different fault conditions is studied. According to the analysis, the four feature quantities have complementary advantages under different fault conditions. Combining these four feature quantities for segment location can improve the accuracy of segment location, which lays the foundation for the proposed fault location method based on the fusion feature quantity. A fault segment location method based on hierarchical clustering for multi-feature fusion is proposed. The paper adopts the hierarchical clustering with simple principle and easy implementation as the fusion algorithm, and the four characteristic quantities of the zero-order current fifth-order harmonic amplitude, the zero-sequence current fifth-order harmonic phase, the instantaneous positive-sequence real-time power frequency amplitude, and the zero-sequence current eigenmode energy are fused; A reasonable setting of the threshold is achieved by measuring the correlation between the individual detection points by the Euclidean distance, and the criterion of fault segment location is constructed. The fault segment location based on hierarchical clustering multi-feature fusion is realized. In order to verify the accuracy of the fault segment location method based on hierarchical clustering multi-feature fusion, this paper has carried out a lot of simulation verification and experimental verification. The results show that the fault segment method is accurate and reliable, and it is effective and adaptable.