TY - JOUR
T1 - Nonlinear response prediction of stratospheric airship envelope differential pressure via multimodal coupling and multiscale decomposition
AU - Li, Yongkao
AU - Liang, Haoquan
AU - Wang, Yan
AU - Liu, Dongxu
N1 - Publisher Copyright:
© 2026 Elsevier Masson SAS.
PY - 2026/11
Y1 - 2026/11
N2 - Accurate analysis and prediction of the differential pressure between the interior and exterior environments of the envelope—a critical component of stratospheric airships—are essential for the reliable assessment of their remaining useful life. However, conventional models based on fundamental physical laws, frequently involving significant simplifications and introducing static physical parameters as inputs, cannot adequately represent the complex and extreme conditions encountered in the stratosphere. Accurate prediction of envelope differential pressure relies on the integration of data from various onboard sensors. Moreover, the envelope exhibits distinct dynamic behaviors governed by varying physical mechanisms. The dominant factors driving pressure evolution are time-variant and shift across multiple time scales. To address the aforementioned issues, a data-driven modeling methodology is adopted. In this study, we propose an interpretable analytical framework—Multimodal Coupling and Multiscale Decomposition (MCMD)—for the first application in predicting the differential pressure of stratospheric airship envelopes. Specifically, a multimodal analysis module is employed to capture and quantify the intricate interdependencies among multiple variables, while a multi-scale mixture of experts module is implemented to effectively integrate dominant factors across different temporal scales. To assess the generalizability of the method, confirmatory experiments are conducted on several publicly available datasets. Subsequently, the performance of the proposed method is evaluated through both complete-process simulations of stratospheric airship systems and ground-based experiments on scale models. Comparative analyses with other prevailing state-of-the-art baseline methods consistently demonstrate the superiority and effectiveness of the proposed approach, proving its significant practical value for predicting the envelope differential pressure in stratospheric airships.
AB - Accurate analysis and prediction of the differential pressure between the interior and exterior environments of the envelope—a critical component of stratospheric airships—are essential for the reliable assessment of their remaining useful life. However, conventional models based on fundamental physical laws, frequently involving significant simplifications and introducing static physical parameters as inputs, cannot adequately represent the complex and extreme conditions encountered in the stratosphere. Accurate prediction of envelope differential pressure relies on the integration of data from various onboard sensors. Moreover, the envelope exhibits distinct dynamic behaviors governed by varying physical mechanisms. The dominant factors driving pressure evolution are time-variant and shift across multiple time scales. To address the aforementioned issues, a data-driven modeling methodology is adopted. In this study, we propose an interpretable analytical framework—Multimodal Coupling and Multiscale Decomposition (MCMD)—for the first application in predicting the differential pressure of stratospheric airship envelopes. Specifically, a multimodal analysis module is employed to capture and quantify the intricate interdependencies among multiple variables, while a multi-scale mixture of experts module is implemented to effectively integrate dominant factors across different temporal scales. To assess the generalizability of the method, confirmatory experiments are conducted on several publicly available datasets. Subsequently, the performance of the proposed method is evaluated through both complete-process simulations of stratospheric airship systems and ground-based experiments on scale models. Comparative analyses with other prevailing state-of-the-art baseline methods consistently demonstrate the superiority and effectiveness of the proposed approach, proving its significant practical value for predicting the envelope differential pressure in stratospheric airships.
KW - Envelope differential pressure
KW - Multimodal
KW - Multiscale
KW - Prediction
KW - Stratospheric airship
UR - https://www.scopus.com/pages/publications/105036187602
U2 - 10.1016/j.ast.2026.112380
DO - 10.1016/j.ast.2026.112380
M3 - 文章
AN - SCOPUS:105036187602
SN - 1270-9638
VL - 178
JO - Aerospace Science and Technology
JF - Aerospace Science and Technology
M1 - 112380
ER -