Paper
5 June 2024 Life prediction of railway relays based on XGboost-informer model
Author Affiliations +
Proceedings Volume 13163, Fourth International Conference on Mechanical, Electronics, and Electrical and Automation Control (METMS 2024); 131634N (2024) https://doi.org/10.1117/12.3030154
Event: International Conference on Mechanical, Electronics, and Electrical and Automation Control (METMS 2024), 2024, Xi'an, China
Abstract
Aiming at the problem of how to accurately and efficiently predict the life of railroad relays, a railroad relay electrical life prediction method integrating XGBoost and Informer is proposed. Firstly, a railroad DC relay electrical life test system platform is set up for testing. Various performance characteristic parameters are extracted from it, which characterize the operation of the relay. Secondly, the XGBoost feature selection method is used to perform data statute and screen the optimal feature parameters. Finally, the Informer model is utilized to learn the intra-sequence regularities and long-range dependencies between sequences to improve prediction accuracy and efficiency. A comparison example is established. The analysis of the example shows that the XGBoost-Informer model is obviously better than other single models. It has strong applicability and accuracy. For the state monitoring and fault diagnosis of the relay, it provides an effective way and another way of thinking.
(2024) Published by SPIE. Downloading of the abstract is permitted for personal use only.
Jingjie Nie, Jing Xu, Chaojian Xing, Shuxin Liu, and Yang Liu "Life prediction of railway relays based on XGboost-informer model", Proc. SPIE 13163, Fourth International Conference on Mechanical, Electronics, and Electrical and Automation Control (METMS 2024), 131634N (5 June 2024); https://doi.org/10.1117/12.3030154
Advertisement
Advertisement
RIGHTS & PERMISSIONS
Get copyright permission  Get copyright permission on Copyright Marketplace
KEYWORDS
Relays

Feature selection

Education and training

Performance modeling

Power supplies

Feature extraction

Photovoltaics

Back to Top