Large-span Cable-stayed Bridge Model Updating Algorithm Based On Fish-swarm Algorithm

As a hot topic in the research of structural health monitoring，the significance ofmodel updating technology is profound. This dissertation studies the model updatingbased on an improved fish-swarm algorithm. The main research contents include:Firstly, the optimization mechanisms of the artificial fish-swarm algorithm areanalyzed; the behaviors of fish-swarm algorithm are explained in depth. The algorithmis tested through mathematical examples, it is proved that the algorithm has a fast andstrong global searching ability, and has an effective searching ability in flat regions.For the low convergence speed and low precision of the algorithm in the latter part ofthe search, improvements are proposed, an improved artificial fish-swarm algorithmare obtained (U-AFSA).Secondly, using a large-span cable-stayed bridge as an example, the finite elementmodel of the bridge is built through Midas Civil, and the dynamic and staticcharacteristics of the bridge are analyzed preliminarily. Three kinds of errors areanalyzed. The existence of the structure errors is described; the influence of severaldifferent generations of the finite element mesh is discussed; for parameter errors, thesensitiveness of some aspects of the errors, which could influence the modalcharacteristics, is analyzed through orthogonal experimental.Thirdly, two parameters of a multi-span statically determinate beam model aremodified by the using of U-AFSA, and good results are obtained, the effectiveness ofthe algorism is proved.Finally, considering the long computing time of the large finite element model,numerical equivalent model of large-span cable-stayed bridge is studied，and the theoryderivation formulas are obtained. As an example, the model of Qing Shui Pu Bridge ismodified through the algorism in this paper. The feasibility and validity of thisalgorism are proved through the results.