Proceedings of the Computational Structural Engineering Institute Conference (한국전산구조공학회:학술대회논문집)
- 2011.04a
- /
- Pages.57-60
- /
- 2011
MINNs for FE model updating of a steel box girder bridge
강박스 거더교의 FE 모델 개선을 위한 평균값 반복 신경망
- Vu, Thuy Dung (Kunsan National University) ;
- Cui, Jintao (Kunsan National University) ;
- Kim, Doo-Kie (Kunsan National University) ;
- Koo, Ki-Young (University of Sheffield)
- Published : 2011.04.14
Abstract
Updating model parameters are required in order to simulate the actual behavior of the dynamic structure. A new strategy, mean-iterative neural networks (MINNs) has been proposed in this paper for model parameter updating of a steel box girder bridge. With new strategy for structural dynamic model updating, it offers many advantages such as potential savings of computational effort, more consistent in reaching convergence. The dynamic response obtained from the experimental test on a two span continuous bridge is used as the target for model updating. And the presented algorithm is applied to update the model parameters. These results show a good possible of using MINNs in practice for dynamic model updating.