Autonomous Inspection Method of UHV Substation Robot Based on Deep Learning in Cloud Computing Environment

• Published in: Journal of circuits, systems, and computers Vol. 33; no. 5
• Main Authors: Wang, Zhao-lei, Meng, Rong, Zhao, Zhi-long
• Format: Journal Article
• Language: English
• Published: Singapore World Scientific Publishing Company 30.03.2024; World Scientific Publishing Co. Pte., Ltd
• Subjects: Algorithms; Ant colony optimization; Automation; Cloud computing; Data analysis; Deep learning; Error analysis; Fault diagnosis; High voltages; Inspection; Machine learning; Robots; Substations; UHV substation robot; deep learning algorithm; fault diagnosis; Cloud computing environment; improved ant colony algorithm; self-inspection; association mining
• ISSN: 0218-1266; 1793-6454
• DOI: 10.1142/S0218126624500889

Aiming at the problem that substation robots cannot automatically find and analyze the fault equipment and carry out patrol inspection, a method of autonomous patrol inspection for ultra-high voltage (UHV) substation robots based on deep learning in cloud computing environment is proposed. Firstly, based on the cloud computing environment, an autonomous patrol system is designed to upload the data obtained by robots to the cloud platform for processing, so as to complete data analysis quickly and with high quality. Then, the deep learning algorithm (DL) is used for fault diagnosis, and the FP-growth algorithm is combined to realize the association mining of fault data, so as to clarify the patrol order of fault-related nodes. Finally, the improved ant colony algorithm (IAC) is used to optimize the path of the robot to complete the reliable inspection of the substation in the shortest time. Based on the selected UHV substation, the experimental analysis of the proposed method shows that the fault diagnosis error rate and time are about 4.3% and 14.2 s, respectively, and the patrol path is only 180.351 m, and the patrol time is 19.708 s, which can realize the optimal patrol of the robot.