Article Details

Abstract

Aiming at the problems of shallow knowledge mining and difficult fault location in the fault field of railway signaling equipment in mining areas, a fault information identification method based on knowledge graphs is proposed, which aims to quickly locate fault causes and shorten fault recovery time based on historical fault data. Firstly, fault text information related to the signaling equipment is collected, and nodes and association relationships are constructed. At the same time, a fault extraction model of the YOLOv5 turnout action curve is constructed. After model training, the accuracy rate reaches 88.32%, the recall rate reaches 90.71%, the F1 value reaches 89.52%, and the mAP@0.5 reaches 92.43%, which effectively solves the problem of judging the abnormal fault category of turnout curves. Finally, entities and their relationships are extracted from the signaling equipment fault corpus, and a knowledge representation model based on triples is constructed. With the help of the Neo4j graph database, the knowledge is stored, updated and improved. On this basis, fault phenomenon phrases are matched in the knowledge graph to locate similar fault phenomenon nodes, so as to quickly obtain relevant fault knowledge and meet the needs of information-based fault diagnosis.

Keywords

railway signaling equipment; knowledge graph; Target Detection; fault diagnosis; Neo4j graph database

References

1. [1] Chen Lidong, Wang Lushui, Ding Ruo Railway Signal Cable Monitoring and Intelligent Maintenance Management Platform Based on Digital Twins [J]. Railway Communication Signal, 2026, 62 (3): 32-37
2. [2] Cai Xueli Construction Technology Application of Railway Signal Interlocking Transformation Project in Mining Areas [J]. China Science and Technology Information, 2022 (14): 75-78
3. [3] Shi Bin, Chen Ting, Zhu Changqun, etc Research on the Construction of a Fine grained Knowledge Graph of Tobacco Diseases and Pests [J]. China Agricultural Machinery Chemistry Journal, 2025, 46 (12): 40-48102
4. [4] Shi H B，Huang D，Wang L，et al. An Information Integration Approach to Spacecraft Fault Diagnosis [J].Enterprise Information Systems，2019，15（8）：1128-1161.
5. [5] Yun F，Feng Z，Li B F，et al. Research on Intelligent Fault Diagnosis of Power Acquisition Based on Knowledge Graph [C]//2019 3rd International Conference on Electronic Information Technology and Computer Engineering (EITCE)，2019：1737-1740.
6. [6] Shi Z，Zhong Q W，Zheng S B，et al. Analysis of Fault Events in Rail Transit Vehicle Traction Systems Based on Knowledge Graph Reasoning [J/OL].Eksploatacja i Niezawodnosc-Maintenance and Reliability，2025，27（1）.http://doi.org/10.17531/ein/192171.
7. [7] Jiao Shuangsuo, Guo Haodong Research on the Construction and Application of Fault Knowledge Graph for Railway Signal Equipment [J]. Digital Communication World, 2025 (4): 144-148
8. [8] Sun Ting, Jenny, Zhang Yang, etc Construction Method of Railway Line Disease Knowledge Graph with Three Domain Correlation [J/OL]. Railway Standard Design, 2025:1-10 https://doi.org/10.13238/j.issn.1004-2954.202505280004.
9. [9] Guo Haodong Research on the Construction Technology of Fault Diagnosis Knowledge Graph for Railway Signal Equipment [J]. Transportation World, 2025 (19): 26-30
10. [10] Li Fenglin, Liu Fang, Li Yujie Analysis and Application Research of Knowledge Fusion Mode [J]. Information Science, 2021, 39 (4): 3-8
11. [11] Han Chen, Zhen Yuyang, Feng Qi, etc Construction method of knowledge graph for railway signal equipment operation and maintenance status [J]. Railway Building Technology, 2025 (6): 191-196
12. [12] Han Ziwei Research on Key Technologies and Applications of Fault Analysis for High Speed Train Based on Knowledge Graph [D]. Beijing: China Academy of Railway Sciences, 2025
13. [13] Yang Shengli, Lao Pengfei, Deng Zilong Research on Fault Diagnosis of Turbine Runner Based on VMD-SE-TCN [J/OL]. Journal of Hydroelectric Energy Science, 2026 (9): 188-193
14. [14] Li Yanshu, Wang Senwang Fault diagnosis of variable frequency motor based on GADF-CNN-DOA-LSSVM [J/OL]. Journal of Railway Science and Engineering, 2026:1-12 https://doi.org/10.19713/j.cnki.43-1423/u.T20251792.
15. [15] Gao W，Zong C，Wang M，et al. Intelligent Identification of Rice Leaf Disease Based on YOLO V5-EFFICIENT [J/OL].Crop Protection，2024，183：106758.https://doi.org/10.1016/j.cropro.2024.106758.
16. [16] Wang Yuchen, Feng Hailong, Liu Boqi A bolt missing detection system for steel structure canopy of high-speed railway passenger stations based on YOLO algorithm [J]. Railway Computer Application, 2022, 31 (6): 1-5
17. [17] Xu L，Li P，Peng Q，et al. Vehicle Detection on Roads Based on YOLOv5 with Multi-scale Feature Fusion [J/OL].Array，2025：100522.https://doi.org/10.1016/j.array.2025.100522.
18. [18] Song Runqing Event Extraction Based on Word Vectors and Text Features [D]. Xi'an: Xi'an University of Electronic Science and Technology, 2015

About the authors

Shen Xiaoyu (1998-), male, Han nationality, from Huaibei, Anhui Province, assistant engineer, master's student, research direction: knowledge graph, object detection.