Comparative Study On Uncertainty Identification Methods For Operational Modal Parameters Of Bridge Structures

In the study of bridge operational modal parameter identification,uncertainties are introduced by the factors such as environmental interference,model simplification,test error and identification algorithm;thus there will be deviations between the identification value and the true value of bridge modal parameters inevitably.In order to improve the credibility and robustness of the identification results of bridge modal parameter,it is necessary to study the uncertainty identification of bridge modal parameters.Commonly used structural modal parameter uncertainty identification methods can be divided into two categories,the classical frequency method and the Bayesian method.Due to its perfect theory and high efficiency,the FB-FFT method that is in the scope of bayesian method is gradually popularized.However,the characteristics of long-span complex bridges,such as dense modes,low signal-to-noise ratio responses and non-stationary responses,which will directly restrict the application of FB-FFT.Aiming at the inadequacies in the research of the modal parameter uncertainty identification of long-span complex bridges,a comparative study of FB-FFT method and FB-FFT based sampling method was carried out by using the combination of theory and application.The modal parameter uncertainties of a numerical simple supported beam bridge and a real long-span cablestayed bridge were identified by adopting the two methods respectively,and the identification results were analyzed and compared.The main work accomplished in this study are as follows:(1)Based on the widely review of the literature related on structural modal parameter uncertainty identification,the present status of the relevant fileds were introduced and the inadequacies of the current researches were pointed out;and the research mentality and content were drafted.(2)The theory and method related to structural modal parameter uncertainty identification,such as data quality evaluation,frequency spectrum viewing and uncertainty identification,were introduced;in order to cope with the problems of adopting FB-FFT method in identifying the uncertainty of modal parameters of non-stationary structures,the FB-FFT based sampling method was proposed.(3)The constant model and the random model of a numerical simple supported beam bridge were established;their acceleration responses under environmental excitation were simulated,and their modal parameter uncertainties were identified by using the FB-FFT method and FBFFT based sampling method respectively.Based on the constant model,the effects of noise level on the identification results of the two methods were investigated;and based on the random model,the effects of structural parameter changing on the identification results of the two methods were investigated.It was pointed out that the FB-FFT method may underestimate the uncertainties of non-stationary structures based on the comparison results.(4)By taking a real long-span cable-stayed bridge as the research object,the application effects of data quality evaluation,frequency spectrum viewing,FB-FFT method and FB-FFT based sampling method were verified.The differences between the identification results of FBFFT method and FB-FFT based sampling method were compared and analyzed,and the additional description of the overall uncertainties of non-stationary bridges was provided based on the identification results of FB-FFT based sampling method.