Model Updating, Damage Diagnosis And Prewarning Methods Based On Health Monitoring Of Large Cable Stayed Bridges

This dissertation studies some key scientific and technical problems involved in large bridge health monitoring. These problems include: design and implementation of vibration monitoring system, modal parameter identification and effect analysis of temperature on modal parameters, finite element modeling and model updating, damage identification and diagnosis, and extreme accident prewarning. The main research contents of this subject are presented as follows: Firstly, some practical experiences on design and implementation of bridge vibration monitoring system are summarized. And the solution and recommendation aiming at practical problems is presented.Secondly, a time-frequency analysis method of integrally employing multiple short-time stable data under no traffic condition to identify bridge modal parameters is introduced. In this method, wavelet analysis is applied to extract frequency band of each mode and modal parameters are identified directly in time domain. This method can solve the problems of modal parameter identification under free traffic condition to some extent. But modal parameters of cable stayed bridge identified by means of above method still contain temperature influence, therefore the regularity and internal mechanism of temperature influence on measured frequency of this bridge is discussed based on measured data and finite element model. According to above regularity and mechanism, a dynamic matching method is introduced. This method can eliminate the component of linear dependence between temperature and frequency.Thirdly, considering baseline finite element modeling of cable stayed bridge is indispensable precondition and foundation of structural dynamic analysis, damage identification, safety assessment and prewarning, the finite element model of Binzhou cable stayed bridge is established by simulating cantilevered construction process and cable tension adjustment of Binzhou Bridge. The mass and stiffness parameters of bridge are updated based on measured cable tension and frequencies. In this process, the cable tension adjustment algorithm and the practical technique to improve computational efficiency in updating process are introduced.Fourthly, based on updated finite element model, considering the randomicity of noise and the stability of the damage, a three-point linear matching curvature method for girder damage localization of cable stayed bridge is introduced. This mehod has ability to resist noise. The basic idea of this method is described as follows: obtain many curves of curvature difference from the modes identified from no-damage and damage structures, set the location where curvature difference in the group is always positive as the damage location and set the location where curvature difference is sometimes negative as the no-damage location. Extending above idea, a cable stayed bridge damage empirical diagnosis system framework based on damage-symptom database is introduced. This diagnosis system is based on damage-symptom database, and the core of this system is symptom matching mechanism. Under the guide of this framework, the damage-symptom database of Binzhou cable stayed bridge is established based on finite element model, the universal algorithm framework of symptom matching is introduced, and the concrete algorithm of symptom matching of cable damage is presented based on above framework. Because above method is not a rigorous ratiocinative method, but a empirical diagnosis based on partial information, this method has robustness to the limitation of testing accuracy and information.Finally, extreme accident prewarning method of Binzhou cable stayed bridge is investigated. Because the cable stayed bridge will first turn into undercritical state before the extreme accident, identifying undercritical state is the key process of extreme accident prewarning. Based on updated finite element model of Binzhou cable stayed bridge, the change of measured mechanical index is investigated aiming at the five main undercritical states of Binzhou cable stayed bridge, and integrated prewarning method for undercritical states of Binzhou cable stayed bridge is introduced based on the change of mechanical index.