Improvement And Application Of Structuraldamage Identification Method Based On Bayesian Model Updating

The structural earthquake damage identification and evaluation is one of the most important research areas in civil engineering. The structural damage identification is a high indeterminate problem under the influences of some uncertainty factors such as observed data, structural model and so on. Therefore, the structural damage identification approach based on probabilistic and statistical analysis needs to be developed and improved urgently at present. In this paper, according to the outstanding problems of the application of the existing Bayesian approach in the structural damage identification, the damage identification method based on Bayesian model updating will be presented and improved by improving the sampler technique. Furthermore, the cumulative seismic damage model based on displacement and hysteretic energy proposed by Park and Ang will be introduced into the Bayesian model updating to compute the probability distribution of the quantitative damage index, by synthesizing structural nonlinear seismic response and the new sampler algorithm. Then, the structural damage will be identified and estimated quantitatively in term of probability meanings. At last, the efficiency and accuracy of the new method will be verified by using theoretical analysis combined with structural numerical simulation and shaking table test. The main content of the paper are as follows:1、 Improvement of Markov chain Monte Carlo(MCMC) sampler methodsThe MCMC sampling method is one of the most important means to achieve the Bayesian model updating. According to the outstanding problems such as low computational efficiency, slow convergence speed and only suitable for low dimensional problems of the current sampling algorithm exist in practical engineering applications, the improved single component Metropolis-Hasting(MH) sampler technique and algorithm is presented based on the optimal proposal distribution with changing variance, to enhance the sampling efficiency, stability and reliability. And the validity and reliability of the sampling algorithm is verified by a numerical example.2、 Bayesian identification method of structural physical parameters based onthe time domain responseAccording to the uncertainly problems of the structural damage identification and the modal parameters dependence problems of the current two-stage structural physical parameters identification approach based on Bayesian model updating, a new Bayesian identification method based on the structural time domain response is given. Treated the measured structural time domain response as the observed data, then the posterior joint probability distribution of the structural physical parameters is deduced according to the Bayesian theory. The marginal probability distribution and the optimal estimation values of the structural physical parameters can be obtained by sampling from the posterior joint probability distribution using the improved single component MH sampler technique. Numerical studies for a five-story shear structures show that: the given method can not only identify the physical parameters and its changes accurately, but also reduce the uncertainties of the parameters effectively.3、Damage identification of a Reinforced Concrete(RC) frame tested on the shake tableThe time domain response based Bayesian model updating method is applied for a three-story RC building tested on the IEM shake table to identify the physical parameters and damage. The shake table tests were designed so as to damage the building progressively through a sequence of earthquake records reproduced on the shake table. The results further verified the efficiency, robustness and accuracy of the new method. The comparison with the identification results obtained by the modal parameters based method showed that: the identification results obtained by the time domain response based method were more reliable.4、Probabilistic assessment of the structural seismic damage levelThe cumulative seismic damage model based on displacement and hysteretic energy proposed by Park and Ang is introduced into the Bayesian model updating to compute the probability distribution of the quantitative damage index, by synthesizing structural nonlinear seismic response and the modified sampler algorithm, and a new probabilistic identification and evaluation method for structural seismic damage level is presented. The new method is applied for the three-story RC building above, by comparison with the experimental phenomena, the validity and rationality of the method are verified.