Analysis Of Finite Element Model Updating Based On The Frequency Response Function And Experiment Investigation

In recent decades, the finite element model updating, which is used to improve the accuracy ofthe finite element simulation, has been great interest of researchers and has developed rapidly. Basedon the frequency response function sensitivity, this paper studied how to identify the dampingparameters and how to screen the test data in the process of finite element model updating. Thecontent is as followed.This paper describes several classic element model updating methods and is compared with eachother. The advantages and disadvantages of those methods are discussed and proposed basic principlesof the selection element model updating methods. As the test datum was mostly the complex numberand damping models were lack of accuracy, the finite element model updating based on frequencyresponse function, which is based on the three-parameter damping model, is carried out. That issubstituting the three-parameter damping model for the Rayleigh damping model which is commonlyused. The numerical example shows that, the three parameters damping model updating method iswith great accuracy on the recognition of system damping.In the process of updating, the updating equations were indeterminate, and the sensitivity matrixwas prone to be morbid. In order to make sure the accuracy, the selection of small amount testfrequency response datum which was used in the updating equations was discussed, and the testdatum screening criterion for finite element model updating based on frequency response functionwas summarized. Through the verification of simulation, on the basis of ensuring the accuracy ofupdating, finite element model updating is carried out by using the priciple of screening the test data.Results show that the efficiency is improved notably.In this thesis, new modelling method is applied to a structure. Results show that this methodsatisfies the requests of the engineering stucture dynamic response analysis. The updating resultsshow that not only the updating model can be in accordance with measured data used in the modelupdating, but also in accordance with measured data unused in the model updating. Therefore, themethod proposed in this paper can be used to ensure the accuracy of the correction, and is of value forprediction.