Probabilistic damage identification of a designed 9-story building using modal data in the presence of modeling errors

详细信息    查看全文

• 作者：Iman Behmanesh^a ; Babak Moaveni^b ; ^{babak.moaveni@tufts.edu} ; Costas Papadimitriou^c
• 关键词：Probabilistic damage identification ; Modeling error ; Bayesian FE model updating ; Bayesian model class selection
• 刊名：Engineering Structures
• 出版年：2017
• 出版时间：15 January 2017
• 年：2017
• 卷：131
• 期：Complete
• 页码：542-552
• 全文大小：1243 K
• 卷排序：131

文摘

Effects of modeling errors on damage identification (ID) results are studied. Models validity within the frequency domain is non-uniform. A likelihood function is proposed for damage ID without calibrating a reference model. Optimal subset of modes are selected though a Bayesian model class selection approach. Bayesian model averaging technique is used to account for different weight factors.