The Research On The Bridge Structural Damage Identification Method Based On Mutli-modal Parameters

With the development of economy and science technology, many novel and unique structures and a number of large-scale bridges crossing sea and river have appeared in the transportation facilities, at the same time, the number of bridges are growing little by little. Since bridges are affected by some force majeure factors during the course of design, construction and operation, these factors will result in various degrees of damage. Once the damage reaches a certain level, it will lead to destruction of the bridges, which will seriously affect national economy and social enviornment. In order to fully understand the actual operating condition of existing bridges, it is necessary for these structures to carry on the damage detection and determine the location and extent of damage, which will provide reference for later repair and maintenance to avoid the catastrophic accidents. Therefore, the research of damage identification of bridges has important theoretical value and practicality.Based on the broad reading the references of bridge structure damage identification methods and the research achievements of the artificial networks in the field of bridge structure damage identification, based on multi-modal parameters and neural network the bridge structure damage identification method is presented in this paper, which can simultaneously identify the location and extent of structural damage identification. The main work is as follows:(1) Combined with the research background and research status at home and abroad, developments and correlation detection technology, the existing main problems and related trends in the field of bridge structure damage identification are stressed. Bridge structural damage identification methods are systematically summarized and the advantages and disadvantages, scope of application and prospects of various methods are deeply compared and analyzed in detail, which can lay a theoretical foundation of the next bridge structural damage identification method.(2) The research on neural networks method of bridge structure damage identification is carried out and several commonly used neural network models are compared and analyzed, and the basic principles and steps of bridge structure damage identification based on BP neural network and RBF neural network are systematically studied. Combined parameters with curvature mode and flexibility curvature is put forword. Then the bridge structure damage identification method based on neural networks and multi-modal parameters is proposed, which combined parameters with curvature mode and flexibility curvature is acted as input vector of the neural network and damage state of the member is acted as output vector of the neural network. The numerical experiments are done through a simply supported beam. In the general finite element software MIDAS platform, the location and extent of the damage can be effectively identified through the network training for single damage, symmetric damage and asymmetric damage and other damage cases, which verify the validity and reliability of the identification method in this paper.(3) For the steel truss bridge that is used widespread, the numerical simulation and physical model tests are done. According to the characteristics of the steel truss bridge, the damage of some key components are identified to reduce the number of training samples and improve the operating speed of the neural networks. In the numerical simulation tests, the identifying research on single damage and multiple damage of key members are carried out, which the results show that the location can be accurately identified and the extent can be quantified through this method. In the model experiment, intact members are replaced by the prefabricated damage members to form damage structure. For test own limitations, some damage samples can be obtained by the model experiment and these damage samples can be considered as the testing samples of the neural networks. The identification results obtained by analyzing the trained neural network in single damage and multiple damage cases show that the experimental results agree well with the simulation results, which further confirm the identification method proposed in this paper the effectiveness and reliability.(4) The experimental platform of bridge structural damage identification is established by making the test model. Based on the similarity theory of thin-walled structure and according to the characteristics of Jiujiang Yangtze River Bridge steel truss beam flexible arch, static and dynamic similarity relations are derived. Through using the modal analysis theory, the similar ratios of various modal parameters are obtained and the accuracy of similarity relations is verified through numerical test. Then specific design and producing process of reduced scale model are discussed. The dynamic load test program of model bridge is designed and the dynamic test of model bridge is carried out in the laboratory and the dynamic characteristics are determined.(5) The real bridge and model bridge of Jiujiang Yangtze River bridge flexible steel truss arch beam are established by the large-scale general finite element software ANSYS. Through the modal analysis theory the dynamic characteristics of finite element models are conducted. It is indicated that the static and dynamic parameters of the real bridge and model bridge have a good conversion relationship which further verify the validity of the modal parameters’similarity ratio through comparing the theoretical results, numerical simulation results with experimental result, which also reflects that the model bridge can more accurately simulate the dynamic characteristic of the real bridge. Based on the test model of steel truss beam flexible arch and the corresponding damage identification program based on the MATLAB platform, the damage identification research on scale model bridge indicates that damage identification method proposed in this paper can successfully identify the damage of the model bridge and good results are achieved, which can provide a solid theoretical foundation and technical support for the development of the next bridge structural health monitoring system and engineering practical application.