Model Parameter Identification Of Bridge Structure Based On Empirical Wavelet Transform

Modal parameter identification is an important step to implement bridge health monitoring,and it is also the basis of system identification and damage identification.Due to the density modal problems of bridge structure and high noise intensity of test signal,common time domain and frequency domain identification methods are difficult to obtain satisfactory results.In recent years,the new time-frequency domain methods have become a new research direction.However,the most widely applied methods such as wavelet transform and HHT all have insuperable defects.To identify the modal parameters of bridge structure in the time-frequency domain,the latest signal processing method called Empirical Wavelet Transform was introduced.And some improvements were put forward considering the problems encountered in the processing of bridge dynamic test signals,so that it can effectively decompose the signals.On this basis,a selection rule for effective components was proposed,the Hilbert transform and stochastic subspace identification were introduced,and a set of modal parameters identification process for bridge structure was established.To verify the effectiveness of the proposed method based on improved Empirical Wavelet Transform,through processing the dynamic test data of a scale test model for a curved cable-stayed bridge and a practical long-span cable-stayed bridge,the natural frequency,damping ratio and mode shape both vertical and horizontal of main girder were successfully identified.The identification results are consistent with the theoretical value of finite element calculation and show that the method proposed in this paper can identify the bridge modal information more comprehensively compared with the relevant literature research results.The research results show that: whether the vibration response signal was obtained in the form of artificial excitation in the laboratory environment or gathered through environmental excitation in the actual operation state,this method can effectively decompose test signals,extract components containing structural modal information,and identify modal parameters of bridge structures.Therefore,this paper provides a new method for further research in this field.