FNO for vehicle bridge interaction system and damage prediction

 

氏名:金 哲佑

所属:京都大学大学院工学研究科 社会基盤工学専攻

概要:
(目的)
•Propose FNO (Fourier Neural Operator) to learn mappings between damage field and structural responses, which only utilize data from a bridge at the healthy state to predict the bridge at the damaged state, and
•Examine VINO numerically and experimentally in structural simulation and SHM.
(内容)
•Establish Vehicle-Bridge Interaction (VBI) numerical and experimental dataset,
•Train and test the FNO by numerical data,
•Fine-tune by experimental healthy data, and validate by experimental damage data.
(結果)
•Forward FNO for prediction of structural responses was more accurate and faster than FEA (Finite Element Analysis) once it is trained, and
•Inverse FNO for damage prediction could determine, localize, and quantify damages from structural responses.

 




Posted : 2023年03月29日