| In a cable-stayed bridge,the cable force has a greater impact on the safety of the entire bridge structure,but the number of cables is large,and the cost of installing a sensor on each cable for monitoring is high,which is not economical and unsuitable.In addition,the data obtained by the health monitoring system cannot directly obtain the safety state of the bridge structure.It is necessary to evaluate and analyze the stress state of the bridge in combination with the benchmark finite element model.In this paper,based on BP neural network,combined with the finite element benchmark model of cable-stayed bridge,the cable force inversion model and parameter inversion model are constructed.Through nine groups of real-time monitoring cable force values,the cable force values without sensors,the elastic modulus of cable and the elastic modulus of main girder concrete are inversed,and the original cable force values of cable are used to verify.The specific research results are as follows:(1)Under constant load,the ratio of the sum of the ratio of the calculated value and the measured value of the cable force of 75 cables without sensors and the ratio of the sum of the calculated value of the cable force of the 21 cables with sensors and the measured value is approximately 0.28;Compared with the ratio of calculated and measured cable forces of 75 cables without sensors,the ratio of calculated and measured cable forces of each cable is approximately 0.013.(2)According to the force values of 21 stayed cables measured in 9 groups,the force values of 75 stayed cables were retrieved by using the optimized inversion model of the force values of stayed cables without sensor.Based on the bridge cable force values,the average variation rates of 21 measured cable forces and the average variation rates of 75 cable forces are compared.The maximum difference between the two is 0.05%,and the inversion results are accurate.(3)Using BP neural network algorithm to invert the cable force values on June 16 and September 19 without sensors,compared with the cable force inverse values under the traditional method,the error between the two is larger at 12LS1 and 12RS1,and the error is more than 2%;The force errors of the remaining 74 cables are all about 1%,indicating that this algorithm can be used for the inversion of the cable force values of the stay cables without sensors.(4)Relative to the influence of changes in temperature and overall humidity of the environment on the cable force of stay cables,the change in the elastic modulus of stay cables and the elastic modulus of main beam concrete by ±5% has greater influence on the cable force of stay cables.The optimized parameter inversion model,inverted the stay cable elastic modulus and the main beam concrete elastic modulus.According to the correction coefficient,by constructing the relationship between the new elastic modulus X and the changed elastic modulus Y,the verification value of the stay cable elastic modulus and the main beam concrete elastic modulus is the corrected value,which is contrary to the parameters.Compared with the simulated values,the maximum error value of the stay cable elastic modulus is 11.63%,and the maximum error value of the main beam concrete elastic modulus is 3.29%,indicating that the constructed parameter inversion model needs to be further improved.(5)Using the revised elastic modulus of stay cables and the concrete elastic modulus of the main girder to calculate the cable force value of the stay cable,and according to the cable force prediction formula,the cable force value of the stay cable is predicted to be compared with the original cable force.Compared with the values,the maximum error reduction is1.30%,and the remaining cable force error values are all below 1%,indicating that the modified inversion parameters can be used to evaluate the safe operation of cable-stayed bridges. |
