Research On Deflection Separation And Cable Damage Identification In Long-term Monitoring Of Cable-stayed Bridges

The application of health monitoring system to long-span bridges provides the possibility and convenience for real-time control of structural status,the deflection of the main beam is the intuitive and easy-to-measure overall response of the bridge under various operating factors during operation.Accurately grasping the deflection of the main girder of the bridge under the action of these factors alone is conducive to the evaluation of the structural state.Therefore,this article is based on the National Natural Science Foundation of China(51408452),based on the deflection monitoring data of a cable-stayed bridge in Wuhan under the health monitoring system.Based on the deflection monitoring data of a Wuhan cable-stayed bridge under the health monitoring system,this paper focuses on the separation of the components of the main beam monitoring deflection in the long-span cable-stayed bridge and the damage identification of the cables of the cable-stayed bridge.The work results are as follows:1)Using Complete Ensemble Empirical Mode Decomposition with Adaptive Noise can decompose single-channel mixed deflection signals into multi-channel signal according to different time scale.However,there are problems of false signal components and modal aliasing in the obtained multi-channel signal.Aiming at the problem of false signal component rejection,a false component signal rejection method based on permutation entropy(PE)and K-L divergence is proposed;the result of processing analog signals shows that the proposed algorithm has less error than the result of removing false components based on K-L divergence directly.2)Combined with the application of single-channel blind source separation algorithm in the separation of bridge’s monitoring deflection,a single-channel blind source separation algorithm based on CEEMDAN-SOBI is proposed to separate the low-frequency terms of bridge monitoring deflection.CEEMDAN is used as a multiresolution decomposition algorithm for constructing virtual channels,the second-order blind source identification algorithm uses the independent characteristics of the original signals to separate the original signals from the aliased signals.The separation result of the simulated deflection signal and the measured deflection signal of the main beam of the cable-stayed bridge shows that the algorithm can effectively separate the various effect values in the deflection monitoring signal,which can provide data reference for practical engineering3)Using the time-varying global temperature difference as the external excitation of the finite element model of a cable-stayed bridge,a method for damage identification of the cable using the temperature deflection of each measuring point of the main beam is proposed.This method uses the temperature deflection curve of each two adjacent measuring points to construct a regression model to obtain regression parameters,which are used as an inherent property of the structure,and the damage of the cables is identified through the change of the regression parameters;Simulation results of cable damage identification of finite element model of cable-stayed bridge show that the method can effectively identify cable damage;Considering the influence of measurement error and noise,the definition of damage probability is introduced,and the effect of the method proposed in this paper on the damage identification of cables under the influence of different signal-to-noise ratios is analyzed,the results show that the method proposed in this paper has certain anti-noise performance.4)For the damage identification of the cables of an actual cable-stayed bridge,it’s assumed that the daily temperature difference deflection of the main girder of a cablestayed bridge in Wuhan from 1am to 8am every day includes the deflection generated by the overall temperature difference and the deflection generated by the constant beam section temperature difference.This assumption is verified by the fitting degree of the regression model constructed by the temperature deflection of each two adjacent measuring points,and the results show that a good fitting degree can be obtained.Taking the daily temperature difference deflection of a cable-stayed bridge in Wuhan from 1 am to 8 am as the data basis for damage identification,using the CEEMDANSOBI method combined with the definition of damage probability to calculate the damage probability of the cables near the middle span of the main span of the cablestayed bridge.The final recognition result is in good agreement with the results of manual inspection and static load test of the cable-stayed bridge,which verifies the validity of the method proposed in this paper in practical engineering.