跳到主要导航 跳到搜索 跳到主要内容

Multi-Source Time-Frequency Vision Transformer for Fault Diagnosis of Electro-Hydrostatic Actuators

  • Northwestern Polytechnical University Xian

科研成果: 书/报告/会议事项章节 › 会议稿件 › 同行评审

摘要

Reliable fault diagnosis of electro-hydrostatic actuators (EHAs) remains a critical yet difficult task, owing to the limited availability of real fault data and the degradation of diagnostic performance under varying operating conditions. This paper proposes SE-CWTViT, a multi-source diagnosis framework that integrates continuous wavelet transform, SE-based channel reweighting, and Vision Transformer feature learning. Multiple sensing signals are fused into a nine-channel time-frequency tensor, and fault-sensitive channels are adaptively enhanced before classification. Experiments are conducted on a high-fidelity EHA simulation platform under square-wave and sinusoidal excitations. Comparison experiments, noise robustness analysis, and ablation studies demonstrate that SE-CWTViT provides accurate and robust EHA fault diagnosis.

源语言英语
主期刊名2026 the 8th International Conference on System Reliability and Safety Engineering, SRSE 2026
出版商Institute of Electrical and Electronics Engineers Inc.
页152-157
页数6
ISBN(电子版)9798319515704
DOI
出版状态已出版 - 2026
活动8th International Conference on System Reliability and Safety Engineering, SRSE 2026 - Xi'an, 中国
期限: 7 8月 2026 → 10 8月 2026

丛书

姓名2026 the 8th International Conference on System Reliability and Safety Engineering, SRSE 2026

会议

会议8th International Conference on System Reliability and Safety Engineering, SRSE 2026
国家/地区中国
市Xi'an
时期7/08/26 → 10/08/26

学术指纹

探究 'Multi-Source Time-Frequency Vision Transformer for Fault Diagnosis of Electro-Hydrostatic Actuators' 的科研主题。它们共同构成独一无二的学术指纹。

引用此