Abstract A new local current-based fast high impedance fault (HIF) detection scheme for DC microgrid clusters using mathematical morphology (MM) is proposed in this paper. The proposed strategy consists of two MM based parts. The first part is MM erosion filtering to extract the current signals and its components to extract the differential feature vector. The second part is MM regional maxima, for defining a determinative value to detect faults in a line segment by the lowest possible time. This scheme also uses local measured values to eliminate the need for communication channels, which provide a low cost, reliable, and fast fault detection method for DC microgrid clusters. Moreover, to provide an accurate HIF detection method, the accurate HIF model in DC systems is presented and used in the proposed method. For demonstrating the efficiency, authenticity, and compatibility of the proposed method, digital time-domain simulations are carried out in MATLAB/Simulink environment under different scenarios such as overload, noise, low and HIFs to distinguish between overloads and HIFs, and the results are compared with several reported algorithms. The obtained simulation results are verified by experimental tests, which validate the proposed strategy's accuracy and speed under different conditions.

中文翻译：

---

基于数学形态学的直流微电网集群局部故障检测

摘要 本文提出了一种基于局部电流的基于数学形态学（MM）的直流微电网群快速高阻抗故障（HIF）检测方案。所提出的策略由两个基于 MM 的部分组成。第一部分是MM侵蚀滤波，提取当前信号及其分量，提取差分特征向量。第二部分是MM区域最大值，用于定义一个确定值，以尽可能短的时间检测线段中的故障。该方案还使用本地测量值来消除对通信通道的需求，为直流微电网集群提供了一种低成本、可靠和快速的故障检测方法。此外，为了提供准确的 HIF 检测方法，提出了直流系统中的准确 HIF 模型并将其用于所提出的方法。为了证明效率，为验证所提方法的真实性和兼容性，在 MATLAB/Simulink 环境下在过载、噪声、低和 HIFs 等不同场景下进行数字时域仿真，以区分过载和 HIFs，并将结果与​​几种已报道的算法进行比较. 通过实验测试验证了所获得的仿真结果，验证了所提出策略在不同条件下的准确性和速度。