Research On Bridge Damage Identification Based On Bayesian Statistical Approach

More long-span bridges have been built with the rapid development of the human society, techniques and communication, more attention has been drawn to the safety, durability and normal functions of these bridges. Health monitoring systems and intelligent control technologies have been introduced into these long span bridges and are developing rapidly. Proceeding from this background, the thesis depicts the research states and the trend in the field of bridge damage identification, especially the methods based on the probability and statistic, and then puts forward the main research of the study.Combining the assumption that the damage is always local, counting on the theoretical foundation of structure finite element, the perturbation theory and Bayesian statistical framework , the thesis developed a sort of damage statistical identification and a new strategy of structural health monitoring.The optimization equation constructed through the Bayesian theory can not only identify damage degree, but can estimate the damage probability. This method can describe structure damage more accurately. To the instance that not predicting the damage in supposed position, the thesis puts forward a tactics of iterative damage position.The health monitoring strategy based on Bayesian theory, assumes that the damage position is relatively confirmed and can be classified according to its reliable index. Combining concept of the damage probability and damage alarm, detecting the element with higher probability, we realize the supervision and control of the bridge and its component.With the theory and the method of this paper, three programs were written to analyses bridge characteristic, first-order perturbation and damage identification on MATLAB flat. After a comparatively successful application on a continuous box-gird beam's damage identification research, the paper probed into the cable-stay bridge tentatively. At the end the article puts forward the relevant problems remained to study further.