Research On Damage Detection Of Cable-stayed Bridge Based On Intelligent Algorithm

The problem of damage identification can be transformed into a multimodal optimization problem or pattern classification problem, which making use of the change of structural response to detect the value of structural parameters, so structural damage identification belongings to the back analysis of mechanics. Compared with traditional optimization methods, artificial intelligence algorithm has a strong advantage in solving complex optimization problem, pattern recognition and large-scale nonlinear problem, so it can be used to solve the problem of structural damage identification usually. As typical representatives in artificial intelligence algortithm, Genetic Algorithm (GA) and Supprot Vector Machine (SVM) method are most widely used in damage identification. In order to solving the problem of damage identification for cable-stayed bridge, a traditional GA method and an improved GA method were discussed in this thesis, and a new method combined with the improved GA and SVM was proposed to detect the whole structural damage of cable-stayed bridge. The main research work is as follows:(1)The research development of structural damage identification is surveyed, which includes the damage detection method for bridge structure based on SVM and GA.(2)The basic principle and computing processes of genetic algorithm are introduced, and micro-genetic algorithm is chosen to detect the damage of main girder for cable-stayed bridge, which can reduce the number of iteration in the process of calculation. Through the numerical simulation for the cable-stayed bridge in the laboratory, the effectiveness of this method is verified. It is shown that micro-genetic algorithm would greatly enhance the speed of damage identification, but the phenomenon of premature convergence in the process of optimization cann’t be avoided.(3) A new improved GA method, the hierarchical genetic algorithm, has been proposed which promoted the performance of micro genetic algorithm as following aspect:the form of fitness function, algorithm structure, the composition of operators and the strategy of optimal search. A numerical example for a test model of a single-tower cable-stayed bridge is provided to verify the feasibility of the method. It is shown that the hierarchical genetic algorithm can reduce the probability of premature convergence and enhance the local search ability of the algorithm.In the case of small noise pollution, the better result of damage identification will be obtained by using this method.(4)To achieve the whole structure of cable-stayed bridge’s damage identification, a new kind of joint optimization algorithm which combined with hierarchical genetic algorithm and SVM is proposed. Firstly, according to the material characteristics of cable-stayed bridge, the sturcture would be divided into three substructure, such as main girder, pylon and cable, the SVM method could be used to detect which substructure probably occurs damage; Secondly, the element belonging to damage substructure was further identified by the method,which combined with hierarchic genetic algorithm and substructure method to determinate the exact location and extent of damage. The shell model of single-tower cable-stayed bridge in laboratory is built and the numerical simulation is made. It is shown that SVM method can identify the source of damage exactly and hierarchic genetic algorithm can quickly and effectively complete the identification of the element’s damage in the substructure and the whole structure of cable-stayed bridge’s damage detection can be achieved step by step. The number of training samples of SVM and the size of initial population of hierarchic genetic algorithm are reduced and the efficiency optimization is enhanced by this method.