https://doi.org/10.1140/epjs/s11734-026-02203-3
Regular Article
CNN-Transformer-based lifetime prediction and degradation modeling of electromagnetic relay contacts for aerospace materials
1
College of Automation Engineering, Jiangsu University of Science and Technology, 666 Changhui Road, 212100, Zhenjiang, Jiangsu, China
2
School of Electronic and Electrical Engineering, University of Leeds, Woodhouse Lane, LS2 9JT, Leeds, UK
a
This email address is being protected from spambots. You need JavaScript enabled to view it.
Received:
29
July
2025
Accepted:
12
February
2026
Published online:
18
February
2026
Abstract
Ensuring the long-term reliability of electromechanical relays is critical in spaceborne systems, where components operate under extreme thermal, mechanical, and electrical stresses. To address the challenges of multi-scale degradation modeling in aerospace-grade contact materials, this study proposes a convolutional neural network-transformer (CNN-Transformer) hybrid architecture for the predictive assessment of electromagnetic relay lifespan. The model leverages parallel dilated convolutions to extract micro-scale transient features caused by arc discharges (
s) and axial attention to capture macro-scale wear dynamics over extended service cycles (10
s). A physics-guided positional encoding (PGPE) embeds contact cycles and material characteristics to enhance interpretability, while a gated fusion mechanism dynamically adjusts feature weighting across degradation stages. The proposed method is validated using a surge-load durability data set based on 
contact materials, commonly used in aerospace switching applications. The model achieves a 23.0% reduction in Mean Absolute Error (MAE) compared to baseline Transformers and maintains a Coefficient of Determination (
) of 0.898, even under 20 dB noise conditions. Real-time feasibility is demonstrated with an inference speed of 8.7 ms/sample, enabling deployment in embedded systems for in situ relay monitoring. This work provides a scalable and interpretable framework for space-grade material reliability assessment, supporting early failure warning, contact design optimization, and predictive maintenance in space manufacturing environments. The approach can be extended to other mission-critical electromechanical systems operating in harsh space conditions.
Copyright comment Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
Tianyang He and Zhaobin Wang contributed equally to this work.
© The Author(s), under exclusive licence to EDP Sciences, Springer-Verlag GmbH Germany, part of Springer Nature 2026
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
