논문 상세보기
This study proposes an improved method for updating finite element models (FEM) by incorporating the random field characteristics of concrete material properties in reinforced concrete structures. Traditional FEM often assumes homogeneous material properties, which can lead to significant discrepancies between predicted and actual dynamic responses, especially in structures where the Young’s modulus (E) of concrete varies due to factors like curing conditions, material composition, and construction methods. We employed a Gaussian random field model and a system identification (SI) technique to address this limitation to optimize sensor placement. We developed an FEM updating method that incorporates the spatial variability of concrete elasticity. This optimization allowed for a more accurate capture of dynamic properties across various structural locations, resulting in FEM updates that reflect concrete’s inherent heterogeneity. The proposed method was validated through numerical examples, comparing dynamic response accuracy in models before and after updating. Results demonstrated that error values, measured in terms of maximum value error and normalized root mean squared Error (NRMSE), were significantly reduced in the updated models compared to the pre-update model. This approach effectively addresses the limitations of homogeneous assumptions in FEM.
1. 서 론
2. 콘크리트 무작위장
3. 콘크리트 무작위장에 대한 시스템 식별 및 유한요소 모델 업데이트 절차
3.1 시스템 식별을 위한 최적 센서 배치
3.2 시스템 식별
3.3 Universal Kriging 모델을 활용한 유한요소모델 업데이트
4. 수치예제를 통한 검증
4.1 콘크리트 무작위장 모델링 및 대상구조물
4.2 무작위장에 대한 시스템 식별
4.3 유한요소모델 업데이트
5. 결 론
/ 감사의 글 /
/ REFERENCES /
