Research On The Damage Identification Of Girder Bridge Structures Based On Minimization Of The Matrix Perturbation

Bridge plays an important role in contemporary transportation. Their health conditions are vital to the operation of highway or railway. Recently severe accidents with tremendous casualties caused by collapse of bridges happened at home and abroad make the bridge health monitoring one of the interesting research focuses. In this domain there are not suitable solutions so far due to the complexity of this kind problem.In the first chapter of this thesis, the state-of-art of the study about bridge health monitoring and dynamic-based damage identification is reviewed and some ongoing implementations of bridge monitoring systems are introduced. This thesis will emphasize on the study of the theoretical development of model updating, and the methods of damage identification based on it are also presented. Finally the main research wok is outlined.The theoretical development of model updating and the relation of model updating and parameter identification are introduced in Chapter 2. Finite element (FE) model updating is one of the optimization. The purpose of model updating is to identify and correct uncertain parameters of FE model in order to predict of the dynamic behavior of actual structure more accurately. Then being the theoretical foundation for the whole study of this thesis, the adopted theory about optimal matrix updating is introduced. The minimization process of the stiffness matrix perturbation is also presented, namely, Setting-up of an objective function, selecting updating parameters and using robust optimization algorithm.The optimization algorithm is introduced in Chapter 3 including the relation of optimization and inverse-problem, the classification of optimization and respective solution characteristic, emphasizing on the realization of the optimization algorithm of the minimization of the stiffness matrix perturbation. Then in the process of numerical verifications about the optimization algorithm, some problems, which must be noticed, are introduced.Chapter 4, there are numerical verifications about damage identification methods mentioned in this thesis. Taking a simply supported beam bridge as the example, identification algorithm program based on MATLAB is carried out, making use of ANSYS to obtain modal information of the FE and meanwhile add noise factors into consideration. This identification algorithm program also testifies feasibility, availability and applicability according to algorithms of the minimization of the stiffness matrix perturbation.Finally, the main conclusions of this work are figured out and some damage detection problems are also discussed.