| Genetic Algorithm(GA) is a very good optimization tools, which lead it to be widely used in optimization problems in Civil Engineering. Based on summarizing the research achievements of predecessors about the genetic algorithm, this paper proposes a new improved genetic algorithm which is provided with functions of global searching and local searching.This improved algorithm is used well in solving the problem of optimum sensor placement and damage detection for structures in this paper. The main research works of the dissertation are as follows :(1) Improved genetic algorithmAccording to the ideas and characteristics of two thresholds GA, pseudo-parallel GA, adaptive GA and parallel GA, this paper proposed a comprehensive improvement strategy,which is a combination of the above four kind s of improved genetic algorithm. Algorithm testing and application showed that the algorithm improved the convergence speed and had high computing efficiency.(2) The integration of multiple coding methodsWith the aid of MATLAB software, this paper adopted four encoding methods to program for the improved genetic algorithm, and combined four encoding methods into one program, which makes the genetic algorithm possess the ability to deal with different problems.(3) Based on two modal assurance criterions, the paper optimized the sensor plancement of a high-rise concrete filled steel tubular(CFST) structures. Later, a modal test was done..(4) Modal identificationUsing two kinds of modal identifying methods for test data, modal parameters was abtained for re-optimization of sensor placement.(5) The simulation of damage identificationWith the damage simulation of five-span continuous beams and three dimensional frame structures by ANSYS, the improved GA was used to identify the damage by constructing the fitness function of residual force vector. According to these methods, we pointed out some features of residual force vector for damage detection, and gave some suggestions of usage. |
