|
| A study on the uncertainty of model parameters by Bayesian method |
| Revised:December 03, 2004 |
| View Full Text View/Add Comment Download reader |
| DOI：10.7511/jslx20066131 |
| KeyWord:modal parameter,Bayesian method,maximal posterior probability,RC frame structure, |
| YI Wei-jian WU Gao-lie XU Li |
|
| Hits: 1544 |
| Download times: 8 |
| Abstract: |
| In structural damage diagnosis and parameters identification,errors are inevitably intermixed in the measured model parameters of a structure.In this paper the model parameters are considered as the random variables,and the uncertainty of model parameters are analyzed by the Bayesian method.The Gaussian joint probability density function(PDF) is selected as a prior PDF in the analysis.By several independent measurements on the transfer function of a structure,the conditional PDF of the transfer function and then the formula of the posterior PDF of model parameters can be obtained.The integral of the margin PDF is approximated by using Laplace's method for asymptotic approximation,and the estimation of model parameters with maximal posterior probability can be found.In an experimental model analysis of a reinforced concrete frame structure,the estimation of model parameters is computed with this method and good convergence is shown by the results. |
|
|
|