Stochastic Analysis About Creep Effects Of Long-Span Concrete-Filled Steel Tube Arch Bridge Considering Spatial Correlation

The stochastic analysis of creep effect usually would choose a single random variables to simulate a random factor of bridge structure. In fact, Concrete beam body is not a pouring completed, the using is not the same batch of concrete, the material parameters of different position of the bridge (such as elastic modulus, compressive strength of concrete and so on)may have different under the bridge construction, especially the long-span bridge. That is to say, the random parameters of bridge structure have the variability in space. In this paper, the first Hejiang Yangtze River Bridge as the engineering background, studied the long-term creep behavior of long-span concrete-filled steel tube arch bridge by using the stochastic finite element method and considering the spatial correlation of material. The main content of this paper includes the following parts:1. This paper discussed the research situation of randomness of creep effect and the random field theory, and introduced the degenerated beam element theory and the finite element analysis software CSBNLA.2.This paper reviewed the analysis theory of concrete creep and creep prediction models, discussed the influence of main random variables and random effects on concrete creep, summarized the stochastic analysis method of concrete creep.And the Monte Carlo sampling technology based on response surface is introduced into the random analysis of creep effect, the long-term deformation of the first Hejiang Yangtze River Bridge is predicted based on probability significance without considering the spatial correlation. The results showed that, the creep effect has obvious randomness, suggested that considering random effect of long-term deformation of structure in the design.3.Introduced the random field theory, used Karhunen-Loeve decomposition method to consider the spatial correlation of the creep randomness, and combined the Monte Carlo sampling technology based on response surface, studied the random response problems of creep deformation of the first Hejiang Yangtze River Bridge. Analysis results showed that, considering the spatial correlation has a certain influence on the random analysis of creep effect.4. On the basis of response surface, defined the random variable sensitivity index and analyzed the random variables sensitively. Analysis results showed that, the creep model uncertainty is the most sensitive to displacement and stress, and the correlation length is longer, the sensitivity of creep model is increasing.