Structural Health Diagnosis Method Based On Statistical Modal Analysis For Existing Urban Continuous Bridge

Bridge, a key element in traffic network, is essential lifeline system for social stability and economic development. After years since its being run, bridge is always vulnerable because of bad environments and poor management for structural element performance. Once a bridge loses its function, especially to a key bridge, a paralysis of traffic seriously direct and indirect economic losses will be brought. Therefore, it is very necessary to set up a system to servicing real-time healthy detection, damage identification and health diagnosis for a bridge.Nowadays, physical change of structural response information such as strain and displacement mode shape has been widely utilized in health diagnosis problems and monitoring those parameters is getting more and more attention to ensure the structural integrity. Strain mode is a modal parameter that can response the changes of the local characteristic of the structure. It has been proved that strain mode has much more sensitization than frequency to reflect local structural variable characteristics, especially to continuous bridges. Therefore, in this thesis, the peak of strain mode curve diagram is used to locate damages. Then, finite element analysis is applied to formulate a quantitative damage extent analysis method and a checking test for this method is done to prove the method is feasible.Concerning that the bridge system is generally under unknown random ambient excitations and the monitoring measurements are generally contaminated by different sources of noises and influenced by some non-damage factors (such as temperature, rainfall, pier settlement and etc.), the generally used techniques which work when structural response to static and dynamical loads are deterministic are not efficient for those applications directly. Therefore, statistical modal analysis method is adopted for minishing negative effects of measurement noise, and in order to assess the state of global health of bridge structures using the modal parameters with many uncertainties, hypothesis test theory is adopted herein.In addition, a statistical modal analysis method for structural health diagnosis of continuous bridge based on Bayes statistical methodology is presented and Gaussian distribution assumption is applied to simulate possible uncertainties in the procedure of structural health diagnosis. The Bayes factor is treated as a random variable. In order to prove that structural health diagnosis of continuous bridge can be done by use of the proposed methods, a finite element model of a real continuous bridge structure for simulated damage location and damage extent bases on ANSYS and MATLAB software is set up to test feasibility of the proposal method. The result shows that the proposal methods are correct, reliable and suitable for engineering application.At last, the main contributions and conclusions of the thesis are summarized, and some further research topics concerning are brought forward.