Long-term Deformation Spectrum Analysis About The Reinforced Concrete Arch Bridge Considering Spatial Correlation

Taking Fuxijiang bridge as engineering example, the long-term deformation spectrum of the reinforced concrete arch bridge is predicted in the thesis. Creep and shrinkage of concrete are main factors which make long-term deformation different from short -term deformation. But it is hard to predict creep and shrinkage because of the uncertainty of environment and materials. So deterministic analysis given in all kinds of bridge design code can’t get accurate value of creep and shrinkage. In order to get the long-term deformation spectrum of Fuxijiang bridge, a stochastic analysis was carried out in the thesis instead of deterministic analysis. Moreover, the effect of spatial correlation of creep model randomness on the random analysis of long-term deformation spectrum was also discussed in the paper.First, in order to get the probability density curve of long-term deformation,a hybrid approach combining the Response Surface Method (RSM) and Monte Carlo samping (MC) is introduced into random analysis. The results showed that long-term deformation exhibits strong uncertainty after considering the uncertainty of the material parameters. Ten years after the completion of the bridge, the coefficient of variation of long-term deformation at the vault reaches 0.2. Meanwhile, the standard deviation of long-term deformation gets bigger and bigger with the time going.Then, random field theory is introduced to analyze the effect of spatial correlation of creep model randomness on the long-term deformation. And Karhunen-Loeve decomposition method was used for discretization of the random field. According to the results, the average of long-term deformation had little change after considering spatial correlation of creep model randomness, but the standard deviation of long-term deformation was significantly reduced.Finally, the sensitivity of material parameters to long-term deformation is studied in the thesis. And according to the result, creep model has most influence on long-term deformation. Next one is dead load, and the third one comes to compressive strength and shrinkage model. Moreover, if considering the effect of spatial correlation of creep model randomness, the absolute value of the sensitivity coefficient of creep model will get smaller. And the absolute values of the sensitivity coefficient of dead load, compressive strength and shrinkage model will get bigger.