Improvement of forced bending-torsion coupling vibration method for identifying flutter derivatives of bridge sections based on the same frequency

To expand the application capability of the forced vibration method in the field of bridge section flutter derivative identification, the bending-torsion coupled simple harmonic motion mode of the section oscillated twice and the same frequency with different phases was adopted, and the bridge section flutter derivative identification formula under arbitrary non-zero phase difference was derived with theoretical analysis. The Theodorsen solution of an ideal flat plate was used as the aerodynamic data source, and the identification capability of the method was numerically tested. Additionally, the impact of various factors such as the phase difference of the bending-torsion coupling motion, Gaussian white noise, and integer octave colored noise on the identification results of flat plate flutter derivatives was investigated. The method was combined with the computational fluid dynamics (CFD) technique to identify the flutter derivatives of the girder section of the Great Belt Bridge in Denmark based on Fluent® software. Research results show that when the flat plate aerodynamics do not contain noise, the flutter derivatives are in full agreement according to this method with the theoretical values for any non-zero phase difference. When the aerodynamic forces contain Gaussian white noise, the noise conditions investigated under the three low-level (i. e. 5%, 10%, and 15%), the identification error of flutter derivatives is positively correlated with the noise level and the reduced wind speed, but the sensitivity to Gaussian white noise varies for different flutter derivatives. When the aerodynamic forces on the flat plate contain integer octave colored noise, the identification accuracy of flutter derivatives is significantly affected by the third-order and fourth-order harmonic noises, while the influence of other low frequency and ultra-high frequency noises is relatively small. The flutter derivatives H* 2, H* 4, A*2, and A*4 are found to be more sensitive to octave colored noise in the identification results of 8 flutter derivatives. The flutter derivatives of the main girder of the Great Belt Bridge by this method combined with CFD numerical simulation and are in good agreement with the literature results, and the calculated results of flutter critical wind speed are also relatively close. This research results can achieve the engineering calculation requirements for similar bridge flutter designs. © 2024, Central South University Press. All rights reserved.