Research On State Evaluation Method Of Cable Group Based On Spatial Correlation

In recent years,with the development of society and economy,China is in a period of large-scale infrastructure construction,and many long span bridges have been built or are in the process of planning and construction.Cable-stayed bridge has become one of the main options for long-span bridges because of its strong spanning ability,stiffness and wind resistance.As an important component of the cable-stayed bridge,in the course of service for decades or even a hundred years,the cable is easy to suffer from fatigue damage and corrosion damage under the reciprocating effect of the external environment and vehicle load,which can reduce the strength of the cable and reduce the life of the cable,thus endangering the safety of the bridge structure.With the development of structural health monitoring technology,the load response and variation trend of cable force are monitored.It provides the basis for cable state evaluation and bridge structure safety,and has become an important part of the cable-stayed bridge health monitoring system.However,considering the economy of the structural health monitoring system,the monitoring sensors are usually arranged on a small number of cables,and timevarying cable forces of cable with no sensors cannot be obtained for state evaluation to ensure the safety of the structure.In this paper,the author make full use of the monitoring data of cable tension of large span cable-stayed bridge,systematically studied the condition assessment method of cable group based on spatial correlation.The quantizing and modeling methods for spatial correlation of cable environmental load effect are studied,the optimization algorithm of cable force monitoring location to minimize spatial correlation model error and the prediction method of environmental load effect in nonsensor cable are proposed,and the vehicle load and spatial distribution recognition method based on the spatial correlation of cable load effect are proposed.Based on the prediction of time-varying cable forces,the fatigue damage and ultimate bearing capacity of cables in design life are evaluated respectively.The main research contents include:First,according to the different characteristics of cable force variation under multiple types of loads,the cable force is decoupled into cable force trend due to constant load and temperature,and the peak value of cable force caused by random load effect(vehicle load and wind speed).A correlation modeling method of cable force trend is proposed based on support vector regression,and the optimal placement method of sensor based on genetic algorithm and information entropy is proposed.Based on the optimal cable force monitoring position,the trend of cable tension of non-sensor cables are predicted by using the spatial correlation model,and the results are compared with the measured data to verify the accuracy of the proposed method.Then,a vehicle load identification algorithm based on the genetic algorithm and the mixture model clustering is proposed by using the spatial correlation of the upstream and downstream cable forces,and the vehicle load information and the lateral distribution are obtained.Using the vehicle load of each lane as the input load,the peak value of the cable force is identified.Finally,based on the prediction of environmental load and vehicle load effect,the extreme value distribution model of cable load effect and the degradation model of cable service performance are established,and the fatigue damage and ultimate bearing capacity of the cable group are evaluated in the design life.