A Hierarchical Minimum Hitting Set Calculation Method for Multiple Multiphase Faults in Power Distribution Networks

In this article, model-based diagnosis (MBD) is an important fault diagnosis method for power distribution network faults. The minimum hitting set (MHS) calculation, as one of the steps of MBD, has a significant impact on the performance of MBD. However, the efficiency and accuracy of MHS calculation limit the performance of MBD when the distribution network has multiple multiphase faults. To overcome these limitations, a hierarchical MHS calculation method for multiple multiphase faults is proposed. This method first converts a single high-dimensional MHS calculation into multiple low-dimensional MHS calculations by establishing a hierarchical MBD. In the low-dimensional MHS calculations, the fitness function and MHS criterion of the MHS calculation are improved to enhance the versatility of the method for different minimum conflict sets, and the exhaustive method is used instead of the intelligent algorithm to realize the search process. Finally, a 14-node distribution network is taken as an example to verify the advantages of the hierarchical MHS method in terms of calculation efficiency and accuracy under multiple multiphase faults.

I. Introduction

Distribution networks distribute power in power systems, and if one breaks down, people's daily lives will be seriously affected [1]. To maintain the stability and reliability of a distribution network, it is important to locate and isolate faults quickly and accurately in power distribution systems [2]. There are mainly two types of methods for fault diagnosis in distribution networks. One is explicit methods, such as the analytical model-based method [3], [4], traveling wave-based method [5], [6], and impedance-based method [7], [8]. The other is implicit methods, such as multiagents [9], fuzzy set theory [10], deep learning [11]–[13], expert systems [14], [15], and artificial neural networks [16], [17]. Although these methods provide a powerful solution to different aspects of fault diagnosis, there are still some shortcomings.