Strategies for aeroelastic parameter identification from bridge deck free vibration data

Several techniques for identification of aerodynamic derivatives (ADs) from free vibration test data are compared using simulated data and test data obtained from wind tunnel tests. These identification methods include system identification from one or two degree of freedom response to either transient excitation or to turbulent buffeting. The experimental and analytical difficulties involved in each method are highlighted and suggestions made for the best approach to determination of ADs in both model and full scale studies. Time domain methods using step relaxation provided the best results as long as airflow turbulence does not cause severe signal to noise ratio problems with the free vibration decay. When, as in full scale, the turbulence is the primary forcing function, time domain and frequency domain methods can be used to recover the full set of ADs concerning vertical and torsional response. (C) 2001 Elsevier Science Ltd. All rights reserved.