The Model Updating Of Large Suspension Bridge Based On Response Surface Methods

A well-defined and precise analytical finite element model of bridge, which is the foundation of optimal sensor placement, damage identification, safety evaluation and state prediction for a well-design of structural health monitoring system, is very important in the bridge structural health monitoring. The response surface method has been widely applied for structural design optimization and structure finite element model updating for simple bridge. It is also the method we want to introduce in the model updating of large suspension bridge in this thesis. The main contents and conclusions are as follows:(1)The requirement of a structural health-monitoring-oriented FE model is discussed, and the relative strategy for establishing the bridge FE model is investigated. To reduce the uncertainties in the process of establishing FE model, the mechanical characteristics of large span suspension bridge, theory of computation for main cable geometric shape and the failure mode of large span suspension bridge are elaborated in this thesis. As a case study, a FE model of a large span suspension bridge is established for its structural health monitoring system, and then give its characteristic structure analysis results.(2)The response surface method, which is based on radial basis functions, is investigated to the finite element model updating with elaboration. A through description of the parameters selection methods of sensitivity analysis, the selection of feature information requirements, the construction of objective function method and the selection of optimization method is presented. These selection methods are performed for a truss model updating based on experimental data of a laboratorial steel truss model to investigate the reasonability. The results illustrate that the parameters of the updated model still keep the mechanical meanings, and the characteristics of the updated model can reflect the characteristics of the real structure.(3) The model updating method based on radial basis function response surface modeling is applied to reach the actual working state of a large suspension bridge. Firstly, the sensitivity analysis method is applied to select the available characteristic parameters and the parameters to be amended. The central composite experimental design method is applied to construct the sample of parameters to be amended at different perturbation levels and samples of characteristic parameters are analyzed by static and dynamic FEM method. Secondly a radial basis function response surface model which approaches the implicit functional relationship between the design parameters and the characteristic parameters of large-scale structural systems is developed. Finally, the optimization algorithm method is taken to make modification on the radial basis function response surface model. Based on the static and dynamic test data, the FEM model of a large suspension bridge is modified. The result indicates that the modified FEM model shows a more realistic static and dynamic response of the bridge. The method is suitable for FE model modification in large span suspension bridge with high computational efficiency and accuracy.(4) Verification of the updated finite element model of a suspension bridge is carried out. The updated finite element model is validated with the measured and the corresponding static response processed information. As a case study, the updated finite element model of the suspension bridge is verified. The Results show that the updated finite element model could simulate the static response accurately.