| With the rapid development of China’s economy,the number of bridges built on major traffic roads is increasing,however,bridge damage and even collapse caused by the decline of bridge characteristics or other reasons often occur,in order to ensure that the bridge structure always operates in a safe state,health monitoring systems are generally installed on large bridges as important monitoring means besides manual inspection,the health monitoring system can monitor the bridge automatically and in real time,so it is very important to use the monitoring data of the bridge monitoring system to evaluate the safety of the bridge.Based on the project funded by national natural science foundation of China(51408452)and the project funded by Hubei key laboratory open fund of China(DQJJ201709),according to the theory of modal decomposition,wavelet decomposition,cross-correlation function and deflection data collected by Beidou Bridge Safety Monitoring System installed on cable-stayed bridge,the methods of deflection noise reduction,temperature effect separation and damage identification of box girder are proposed and studied respectively in this paper.The details are as follows:Firstly,the common methods and application status of bridge deflection temperature effect separation and structural damage identification are summarized,and the basic principles of modal decomposition and wavelet decomposition are introduced.The deflection of cable-stayed bridge is reconstructed by analyzing the components of bridge deflection and using the reference data obtained from the finite element model of cable-stayed bridge.By comparing various noise reduction methods,it is verified that combining waveform extension,pre-noise reduction and wavelet decomposition has better noise reduction effect than the traditional single decomposition algorithm.Through the temperature effect separation simulation of deflection after noise reduction,it is verified that MEEMD decomposition algorithm can separate temperature effects in different periods with high precision.Then,the measured deflection data are denoised by combining waveform extension,pre-noise reduction and wavelet decomposition algorithm,and the temperature effects with different periods are effectively separated by MEEMD decomposition method,the good applicability of the denoising algorithm and the temperature effect separation algorithm with different periods for the measured data is verified.After that,the related concepts and the related damage identification algorithm of cross-correlation are introduced.Based on the measured data,a damage identification algorithm based on internal accumulative volume(IAV)is proposed in this paper.The finite element model simulation of simply supported beam and cablestayed bridge verifies that the damage identification based on IAV index can be carried out and has a certain noise resistance.Last,after preprocessing the measured deflection data,damage identification is carried out by using IAV index in this paper.Aiming at the possible minor damage or non-damage conditions of box girder structure,the deflection data is decomposed into the first four order components,and the damage probability is calculated by fitting the probability density of healthy working conditions and unknown working conditions.A comprehensive structural safety evaluation is given by comparing the damage probabilities of multi-level and upstream and downstream measuring points.In addition,the results of damage identification of this paper and based on Mahalanobis distance cumulant and manual inspection are also compared to verify the effectiveness of the algorithm of this paper. |
