Study On Modal Parameter Identification Of Aeroelastic Model For Long-span Bridges

Wind tunnel test of full-bridge aeroelastic model is an important means for investigating wind-induced performance of long-span bridges.In order to ensure the accuracy of the test results,the model needs to simulate both the shape and the dynamic characteristics of the real bridge.Therefore,it is very important to accurately identify the modal parameters of the model.However,the existing model parameter identification method of wind tunnel test has a great limitation,by which the high order modal parameters of the model can't be identified.To solve the problem,the methods of modal identification under ambient excitation,including stochastic subspace identification(SSI),wavelet transform(WT)and Hilbert-Huang transform(HHT),were applied in wind tunnel tests in the dissertation.Meanwhile,the improved algorithms were carried out and implemented with MATLAB.The main work of the dissertation is summarized as follows.(1)The importance of modal tests in the wind tunnel tests of bridges was elaborated,and the limitations of existing modal parameter identification methods in wind tunnel tests were analyzed.Besides,three methods,including stochastic subspace identification,wavelet transform and Hilbert-Huang transform,were selected from commonly used modal parameter identification methods to identify the modal parameters of aeroelastic models in wind tunnel tests.(2)The basic theory of stochastic subspace algorithm for modal parameter identification and the stability diagram method for system order determination were introduced.In view of the existing damping ratio stability judgment process in stability diagram method may cause some modes be overlooked,an optimization method based on Grubbs' test was proposed.Finally,the accuracy of the improved algorithm was validated by the numerical simulation case of the Oujiang Bridge.(3)The dissertation introduces the basic theory of continuous wavelet transform applied in model parameter identification and its existing problems,such as the selection of optimal wavelet basis,the edge effects of the continuous wavelet transform and the extraction of wavelet ridges.Aiming at these problems,the improved methods were put forward.And based on the traditional Crazy Climber algorithm,a supplement algorithm is proposed to improve extraction accuracy of the wavelet ridges.Finally,using the same case as SSI,the accuracy of the improved algorithm of wavelet transform was validated.(4)The basic theory of Hilbert-Huang transform for modal parameter identification was introduced.To restrain the end effects of empirical mode decomposition(EMD),the endpoint data continuation approach base on SVM(Support Vector Machines)was presented.Meanwhile,a bandpass filter was designed to eliminate the mode mixing problem.Finally,using the same case as SSI,the accuracy of the improved algorithm was validated.(5)Based on the fluctuating wind dynamic responses of Jinsha River bridge aeroelastic model,the SSI,WT and HHT methods were applied to identify the modal parameters of the model in wind tunnel tests.The result,such as frequency,damping ratio and mode shape of the model were obtained and compared with the theoretical value,which validates the reliability of improved algorithms for identifying modal parameters of aeroelastic model under fluctuating wind excitation.