Enhanced fault localization in multi-terminal transmission lines using novel machine learning

• Published in: International journal for simulation and multidisciplinary design optimization Vol. 15; p. 15
• Main Authors: Wang, Yingyun, Qi, Xiaoxia, Chen, Yang
• Format: Journal Article
• Language: English
• Published: Les Ulis EDP Sciences 2024
• Subjects: Configurations; detection system; Electric power grids; Electric power systems; Electricity distribution; Fault detection; Fault location; Localization; Machine learning; Measuring instruments; phasor measurement unit; Phasors; Reliability; tapped lines; transmission line; Transmission lines
• ISSN: 1779-6288; 1779-6288
• DOI: 10.1051/smdo/2024010

Accurate fault location on transmission lines is paramount for ensuring the reliable and efficient operation of the electricity grid, which underpins every aspect of modern society. Existing fault localization methods for transmission lines often face shortcomings, particularly in scenarios involving multi-terminal transmission lines, where complexities arise due to dispersed generations and intricate network configurations. Traditional approaches may struggle to provide accurate fault localization, impacting the reliability and efficiency of the electricity grid. Research provide a unique fault localization technique in this article depends on Phasor Measurement Units (PMUs) and Bidirectional Gradient Boost Random Forest (BDGB-RF) machine learning technique to address these challenges. The proposed method offers several advantages over traditional methods, including enhanced accuracy and efficiency. By leveraging PMU data and BDGB-RF, the method provides a two-phase fault localization strategy, incorporating fault line selection based on nodal current imbalance and subsequent fault distance determination. Simulation results demonstrate the effectiveness of the proposed approach, achieving up to 97% accuracy in the majority of studied scenarios, even under different tapping configurations. The adoption of this approach could significantly impact the practices of experts in the field, facilitating more reliable fault detection and localization in complex transmission line networks. This, in turn, can contribute to the resilience and stability of power systems, ultimately improving grid reliability and minimizing downtime.