Kabot Ondřej, Klein Lukáš, Prokop Lukáš, Walendziuk Wojciech
ENET Centre-CEET, VSB-Technical University of Ostrava, 708 00 Ostrava, Czech Republic.
Department of Computer Science, VSB-Technical University of Ostrava, 708 00 Ostrava, Czech Republic.
Sensors (Basel). 2023 Oct 10;23(20):8353. doi: 10.3390/s23208353.
This study introduces an innovative approach to enhance fault detection in XLPE-covered conductors used for power distribution systems. These covered conductors are widely utilized in forested areas (natural parks) to decrease the buffer zone and increase the reliability of the distribution network. Recognizing the imperative need for precise fault detection in this context, this research employs an antenna-based method to detect a particular type of fault. The present research contains the classification of fault type detection, which was previously accomplished using a very expensive and challenging-to-install galvanic contact method, and only to a limited extent, which did not provide information about the fault type. Additionally, differentiating between types of faults in the contact method is much easier because information for each phase is available. The proposed method uses antennas and a classifier to effectively differentiate between fault types, ranging from single-phase to three-phase faults, as well as among different types of faults. This has never been done before. To bolster the accuracy, a stacking ensemble method involving the logistic regression is implemented. This approach not only advances precise fault detection but also encourages the broader adoption of covered conductors. This promises benefits such as a reduced buffer zone, improved distribution network reliability, and positive environmental outcomes through accident prevention and safe covered conductor utilization. Additionally, it is suggested that the fault type detection could lead to a decrease in false positives.
本研究介绍了一种创新方法，以增强用于配电系统的交联聚乙烯（XLPE）包覆导体中的故障检测。这些包覆导体广泛应用于林区（自然公园），以减少缓冲区并提高配电网的可靠性。认识到在此背景下精确故障检测的迫切需求，本研究采用基于天线的方法来检测特定类型的故障。目前的研究包含故障类型检测的分类，此前该分类是使用一种非常昂贵且安装具有挑战性的电接触方法完成的，而且程度有限，该方法无法提供有关故障类型的信息。此外，在接触法中区分故障类型要容易得多，因为每一相的信息都是可用的。所提出的方法使用天线和分类器来有效区分故障类型，范围从单相故障到三相故障，以及不同类型的故障之间。这是以前从未做过的。为了提高准确性，实施了一种涉及逻辑回归的堆叠集成方法。这种方法不仅推进了精确的故障检测，还鼓励更广泛地采用包覆导体。这有望带来诸如减少缓冲区、提高配电网可靠性以及通过预防事故和安全使用包覆导体带来积极的环境效益等好处。此外，有人认为故障类型检测可能会减少误报。
