A Case Study Of Time Series Analysis Applied For Construction Monitoring Of Long-span Bridge

The bridge construction monitoring is an important part of bridge construction. It aims at realization of bridge's designed completion. The differences between the actual state and ideal state of the bridge could be obtained through the real-time monitoring during the construction, meanwhile by using modern control theory, errors could be identified,adjusted and predicted in order to make the bridge construction state as closer to the ideal one as possible, ensuring the safety of the bridge during the construction process. This thesis is based on the construction control of Xiao Lan Shui Dao Bridge which is a concrete continuous rigid frame bridge with a main-span of 155m. Time Series Analysis Method combining with the deflection and stress tested data is applied and a comparison between regression analysis method and time series analysis method has been made. The application of time series analysis method in modeling of bridge and prediction of bridge parameters is also included.The main works have been done in this thesis include:1. Introduction of the content and mean of control of bridge construction. Several methods for construction error analysis are introduced based on the study of the developing history of bridge construction monitoring.2. The basis of theory and modeling process are presented including comparison with regression analysis method.3. FE software of Midas/Civil is used to simulate the construction process of Xiao Lan Shui Dao Bridge; the deflection and section stress are obtained which is later used to compare with the test result value and the prediction values using time series analysis method. The applicability and practice are rectified.4. An approximate time series model is developed with the analysis of level and stress data during the bridge construction stages. The model can predict the deflection and stress value in later construction steps. It is found that the time series analysis method may result in a reliable and accurate prediction compared with that by regression analysis.