Physical Parameters and Damage Location Identification using the Gibbs Sampling

Abstract eng:
A new structural physical parameters identification approach is presented for linear structural models. The approach uses a sequence of identified modal parameter data sets to identify and update the structural stiffness parameters continually, using the Gibbs sampling based on the Markov Chain Monte Carlo method. At first, the linear structural identification model is obtained based on a series of conversions of the dynamic characteristic equation, and then the posterior distribution of the model is achieved by using the Bayesian updating theory. Utilize the structural modal parameters, and take their randomness into consideration, the samples of the structural stiffness parameters from the conditional posterior distribution of the linear structural identification model is achieved. During the process, the Gibbs sampling based on the Markov Chain Monte Carlo method is taken. The approach also inherits the advantages of Bayesian techniques: it not only updates the optimal estimate of the structural parameters but also updates the associated uncertainties. So the probability that the continually updated structural stiffness parameters are less than a specified fraction of the corresponding initial structural stiffness parameters could be easily computed. The proposed approach is illustrated by applying it to a 3-DOF linear shear building to detect and quantify the damage using modal data obtained from small-amplitude vibrations measured before and after a severe loading event, such as an earthquake or explosion. The results show that the proposed approach cannot just identify the damage degree and locations in different ways with little error, but interpret the identified values from a probability point of view.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket