Frequency And Optimization Algorithm-based Model Updating Methods Of The Long-span Bridge

For a long-span bridge designed for long service time,the physical parameters,modal parameters and related mechanical properties of it will change during the service period.So it is necessary to modify the parameters of the long-span bridge.At present,there are many studies focus on parameter identification and model updating for different types of structures,and there are some effective solutions among them.However,as for the multi-parameters optimization of large scale structure,due to the complexity of long-span bridges,there are few studies can satisfy both requirements: speed and accuracy.There is still a great deal of research need to be done on how to effectively update the models of large-span bridges.On the other hand,damage is bound to be formed in long-serving bridges due to harsh environmental effects.While small damages are always difficult to detect because of noise interference and the resolution of the instrument.This paper mainly carries out the following work:(1)A method for detecting small damages and damage degree of stayed cables of cable-stayed bridges based on temporary bar is proposed.With the help of the temporary bar installed in the bottom of a stayed cable,the natural frequency of short cable(the cable is divided into two parts after the installation)can be amplified,and the frequency difference before and after the damage can be amplified to about 10 times or even 100 times,so that the small damage can be detected sufficiently.A stay cable from a cable-stayed bridge is selected to demonstrate the method with numerical study,and experimental and numerical studies of a laboratorial stay cable are also demonstrated in this paper.The results show that with the help of the proposed temporary bar,small damage of the stay cable can be detected efficiently.(2)The artificial fish swarm algorithm has been improved in this paper,the direction storage is introduced into the preying behavior strategy to reduce the computational steps of the preying behavior.As a result,the calculation time of the preying behavior is considerably shortened,and the performance in accuracy is also improved.The De Jong function is used to test the original artificial fish swarm algorithm and the improved artificial fish swarm algorithm.The results show that the improved algorithm has a great improvement in convergence speed and accuracy comparing with the original algorithm.A finite element model updating method based on the improved artificial fish swarm algorithm was proposed and applied to the model updating of the Dongying Yellow River Highway Bridge.During the updating process,17 parameters are selected at the same time.Both of the improved and original algorithms are used to modify the finite element model,and the results show that compared with the original artificial fish swarm algorithm,the calculation speed of improved artificial fish swarm algorithm is greatly improved by about 26% and the accuracy has also increased slightly by about 6%.The proposed method can be applied to the updating of the finite element model of long-span bridges.