Abstract
In this paper, a new method for probabilistic identification of structural parameters is proposed to tackle the uncertainty (confidence) in measured modal properties and initial FE model by integrating a perturbation method with Bayesian theorem. It takes into account the statistical correlation between measured modal data and updated structural parameters in the iterative updating and the prior knowledge of structural parameters. This method explores two-stage updating. The first stage updating involves the use of measured modal properties that are assumed as random variables with a normal distribution, to calculate the statistical second moment characteristics of updated structural parameters by using a perturbation procedure in conjunction with the least-square algorithm. The second stage updating refers to the determination of posterior distribution of structural parameters in terms of the obtained distribution from the first stage updating and the prior distribution of structural parameters by using Bayesian theorem. A numerical example is presented to demonstrate this novel method for probabilistic identification of structural parameters.
| Original language | English |
|---|---|
| Title of host publication | Structural Health Monitoring and Intelligent Infrastructure - Proceedings of the 2nd International Conference on Structural Health Monitoring of Intelligent Infrastructure, SHMII 2005 |
| Pages | 1021-1027 |
| Number of pages | 7 |
| Volume | 2 |
| Publication status | Published - 1 Dec 2006 |
| Event | 2nd International Conference on Structural Health Monitoring of Intelligent Infrastructure, SHMII 2005 - Shenzhen, China Duration: 16 Nov 2005 → 18 Nov 2005 |
Conference
| Conference | 2nd International Conference on Structural Health Monitoring of Intelligent Infrastructure, SHMII 2005 |
|---|---|
| Country/Territory | China |
| City | Shenzhen |
| Period | 16/11/05 → 18/11/05 |
ASJC Scopus subject areas
- Civil and Structural Engineering
- Building and Construction