Prediction Of Long-term Deformation Of CFST Stiffened Skeleton Arch Bridge Based On Bayesian Theory

CFST stiffened skeleton arch bridges became the most competitive structural type of the long span reinforced concrete arch bridge because CFST stiffened skeleton with good compression performance is applied sothat its mechanical performance is more superior.The increase of the span makes this kind of bridge very sensitive to shrinkage and creep,which makes it more difficult to predict the long-term deformation of the structure.According to the existing literature,the difference between the actual value of the midline deflection and the calculated value obtained by the deterministic calculation method is more than 30%,and there is a significant difference between the theoretical calculation and the actual stress state of the structure.This shows that the deterministic analysis can not meet the needs of the engineering practice.The modern statistical method is used for stochastic analysis of long-term deformation of bridge structures.The most scientific way to solve this problem is to analyze and study bridge structure based on stochastic analysis.The prior distribution can not take into account the uncertainty in the actual operation of the structure,and inevitably makes the prediction result deviate from the actual value.The Bayesian method provides a solution.It updates the prior distribution by using the likelihood function made up of short-term deformation observations,not only can get the posterior updated probability distribution of variables,but also get the result of long term deformation updated posterior probability distribution.Relying on shrinkage and creep experiment and model bridge test of Beipanjiang bridge,Long-term deformation of Beipanjiang bridge was analysed in this paper by using the stochastic analysis method based on the Bayesian theory.The main research contents and conclusions are as follows:(1)A summary of the research on the development of CFST stiffened skeleton arch bridges in China.At present,the development of CFST stiffened skeleton arch bridges in China has not been systematically discussed.Many researchers often confuse it with CFST arch bridge,which is very unfavorable for the research of this structral type.Research literature of CFST stiffened skeleton arch bridges were collected to systematically expound the trends of design theory,construction method and the development of the span.In order to provide some advice for the development of CFST stiffened skeleton arch bridges,some measures to increase the span were present.(2)The establishment of shrinkage and creep time-varying model of Beipanjiang bridge.Relying on shrinkage and creep experiment of Beipanjiang bridge,ACI shrinkage and creep model was modified,so that it can better describe the development of tube concrete and outer concrete shrinkage and creep.(3)Prior prediction of long-term deformation of Beipanjiang bridge.A certain confidence probability interval of long-term deformation was obtained by using the stochastic analysis method.In order to understand the effect of the random variables on the long-term deformation of the structure,the long term deformation sensitivity analysis was carried out.The sensitivity analysis showed that the dead load is very sensitive to the uncertainty of long-term deformation of Beipanjiang bridge,the sensitivity of C60 concrete creep model uncertainty is much higher than that of C80 concrete.(4)Posterior prediction of long-term deformation of Beipanjiang bridge.Based on Bayesian theory,Poterior probability distribution of structure deformation was obtained by using likelihood function to update the prior prediction wih the help of short-term deformainton observations.The short-term deformation observation was obtained by the model test.According to different short-term deformation observation values,7 kinds of working conditions are analyzed,and the influence of short-term deformation observation on long-term deformation prediction under Bayesian theory system is obtained.The analysis of the posterior prediction shows that the introduction of more short-term deformation observations is helpful to improve the accuracy of the results of the posterior prediction.