TY - GEN
T1 - Residual Super-Resolution-Based Structural Field Rapid Prediction Method for BWBUG
AU - Liu, Junchang
AU - Li, Jinglu
AU - Long, Wenyi
AU - Li, Jiale
AU - Dong, Huachao
N1 - Publisher Copyright:
© The Chinese Mechanical Engineering Society 2026.
PY - 2026
Y1 - 2026
N2 - The conventional measurement methods, constrained by the limited spatial capacity, predominantly rely on finite sensors, consequently exhibiting a deficiency in global sensing capacity for structural field responses. Meanwhile, existing reduced-order model (ROM) algorithms, while capable of achieving full-field reconstruction through finite sensors, exhibit limitations in addressing complex structural mechanical loads. To enhance the global sensing capacity for structural field responses, this paper proposes a structural field reconstruction method based on residual super-resolution, which establishes a mapping relationship between finite inputs and a high-fidelity simulated strain field. When applied to blended-wing-body underwater glider (BWBUG) structural field prediction, validation using finite element analysis datasets demonstrates that the proposed method maintains superior prediction accuracy compared to conventional data-driven ROMs and state-of-the-art network-based reconstruction methods, showing promising potential for further BWBUG digital twin applications.
AB - The conventional measurement methods, constrained by the limited spatial capacity, predominantly rely on finite sensors, consequently exhibiting a deficiency in global sensing capacity for structural field responses. Meanwhile, existing reduced-order model (ROM) algorithms, while capable of achieving full-field reconstruction through finite sensors, exhibit limitations in addressing complex structural mechanical loads. To enhance the global sensing capacity for structural field responses, this paper proposes a structural field reconstruction method based on residual super-resolution, which establishes a mapping relationship between finite inputs and a high-fidelity simulated strain field. When applied to blended-wing-body underwater glider (BWBUG) structural field prediction, validation using finite element analysis datasets demonstrates that the proposed method maintains superior prediction accuracy compared to conventional data-driven ROMs and state-of-the-art network-based reconstruction methods, showing promising potential for further BWBUG digital twin applications.
KW - BWBUG
KW - Digital twin
KW - Residual super-resolution
KW - Structural field prediction
UR - https://www.scopus.com/pages/publications/105041226092
U2 - 10.1007/978-981-95-7342-4_95
DO - 10.1007/978-981-95-7342-4_95
M3 - 会议稿件
AN - SCOPUS:105041226092
SN - 9789819573417
T3 - Mechanisms and Machine Science
SP - 1353
EP - 1362
BT - Advances in Mechanical Design - Proceedings of the 2025 International Conference on Mechanical Design ICMD 2025
A2 - Tan, Jianrong
A2 - Liu, Zhenyu
A2 - Hu, Weifei
PB - Springer Science and Business Media B.V.
T2 - International Conference on Mechanical Design, ICMD 2025
Y2 - 9 May 2025 through 11 May 2025
ER -