The Detection Of Multiple Damages In The Structures Based On Genetic Algorithms And Topology Optimization

The disadvantages of using neural networks, genetic algorithms, gray theory, wavelet analysis and other optimization methods to identify the damages of structures are discussed comprehensively, and the application and development of the genetic algorithms and topology optimization for structural damage identification are introduced detail at the start of this paper. Based on the work of predecessors, Three methods: genetic algorithm, finite element method and topology optimization are combined to develop a new method for two-dimensional structure damages identification. In view of the respective advantages of the genetic and topology optimization in structural damages identification, In this method, the objective function and constraint equations of topology optimization are linked with the fitness function of GA; the design variables of topology optimization and parameters of genetic algorithm were unified; and while the constraint equations of topology optimization were set as a condition to control the whole genetic computing loop. Instead of using continuous variables topology optimization, discrete variables topology optimization is used in multi-damage identification of two-dimensional structure with the hole damages to avoid the possible negative impaction which is brought by the threshold parameter determination in using continuous variables topology optimization. The results of two models simulation examples show that this method is feasible in two-dimensional structure damages detection, the good agreements with the actual situation can also obtained which is difficult to get by using continuous variable topology optimization method.