Research On Health State Warning And BIM Visualization Of Concrete Composite Box-girder Bridge

With the increasing number of bridges in China and the increasingly complex service environment,more and more attention has been paid to the maintenance and management of Bridges.The emergence of bridge health monitoring system has brought great help to the maintenance and management of bridges,and provided strong support for making maintenance decisions.The massive data collected by the health monitoring system brings great convenience to various scientific researches on bridges.Taking Lieshi river bridge along the jiangsu coastal expressway as the engineering background,this thesis lays emphasis on the research of health status warning and BIM visualization method of existing concrete composite box-girder bridge based on the data from bridge health monitoring system.The main contents and conclusions include:1.Explore the temperature effect of concrete composite box-girder bridge.Firstly,the temperature field of Lieshi river bridge is analyzed,and it is found that the temperature distribution of the structure has typical seasonal characteristics and symmetry characteristics.The temperature distribution of the box girder section has horizontal and vertical temperature difference,and these two temperature differences are kinds of random process independent of time.The probability density of the positive and negative parts of the horizontal and vertical temperature difference of the box-girder bridge’s section is fitted.The probability density distribution curve fitting formulas of horizontal and vertical temperature difference existing at the top plate and the base plate are given,which pass the goodness of fit test of significance level=0.1.After that,using the wavelet packet decomposition method to separate the strain data into thermo-strain and vehicular-strain.It was found that there is a time-delay effect between the structure temperature and thermo-strain.Due to this time-delay effect,the scatter diagram of the structure temperature and thermo-strain has obvious ring characteristics,and the correlation coefficient also has a large deviation.Two methods are proposed to reduce the time-delay effect,namely,the translation algorithm of single-temperature data based on Fourier series and the linear fitting algorithm based on multi-temperature data.Validated by the measured data,twomethods can effectively reduce time-delay effect,and significantly improve correlation between the temperature data and thermo-strain.Finally,using the finite element modeling（FEM）simulation method to verify the sensitivity of temperature and thermo-strain time-delay effect features about structural damage,and taking it as a bridge health stauts evaluation index.2.A dynamic reliability early-warning method for multi-source load effect of concrete composite box-girder bridge is developed.Firstly,the dynamic prediction equations of bridge strain maxima caused by temperature load and vehicle load were established by using the bayesian dynamic model.The daily extreme values of thermo-strain and vehicular-strain obtained from wavelet packet separation are selected and studied.It is found that the daily extreme values of thermo-strain and vehicular-strain are log-normal distribution.Given that the needed priori information of data is difficult to get when using Bayesian forecasting model,this thesis studies three methods to build the Bayesian dynamic model without the priori information,named estimation via equation of state method,maximum likelihood estimation method and prior point estimation method respectively.The results show that all the three methods can establish bayesian dynamic model effectively.Estimation via equation of state method is slightly better than maximum likelihood estimation method in the term of mean square error index,while slightly worse in the term of mean absolute error index.The predicted confidence intervals are narrower using estimation via equation of state method,which shows more strict in data waring.This thesis also studies the selection of the shape parameter n₀/2 and the scale parameter d₀/2 when using prior point estimation method.The results show that the prediction effect is better when the ratio of the shape parameter n₀/2 and the scale parameter d₀/2 is close to variance of sample data minus the trend term of equation of state or variance of sample data itself.Finally,parameter of probability distribution can be dynamic updated via the prediction model of thermo-strain and vehicular-strain.Importing dynamic updated parameters of load probability distribution in the limit state equation,the JC method is carried out to calculated time-varying reliability under multi-source load effect.The trend of reliability index and the upper and lower forecast range is also given.3.The visualization method of bridge health condition warning is studied.Firstly,a component table of concrete composite box-girder bridge is compiled which is suitable for being carrier of BIM information.A set of scientific and integrated information classification method is established under dimension requests of main body,elements,time,and type.Based this method,a three-level coding system is compiled in the form of parallel encoding to connect BIM model and the information in database.Secondly,a hierarchical structure modeling method is set up,according to the component table of concrete composite box-girder bridge which is suitable for being carrier of BIM information and using INVENTOR.The model established by this method can meet the requirements of hierarchical and being able to carry the information.This lays a foundation for the the combination of bridge health monitoring,inspection and evaluation and BIM technology.Finally,the result of early-warning graded by extreme value of bridge health monitoring,the dynamic reliability index and the result of inspection and evaluation,after the model lightheaded by HOOPS platform,is visualized,based on the B/S framework.The visualized result not only displays the health state of the bridge intuitively,but also paves the way for further exploring the evolution rules of the bridge state.