A comparison of operational modal parameter identification methods for a multi-span concrete motorway bridge

Operational modal analysis deals with the estimation of modal parameters using vibration data obtained from unknown or ambient excitation sources in operational rather than laboratory conditions. The paper compares three operational modal parameter identification approaches for identifying modal parameters of a multi-span concrete motorway bridge when subjected to either ambient or broad-band linear chirp excitation. The algorithms examined include the frequency-domains methods, peak picking and frequency domain decomposition, as well as the time-domains method stochastic subspace identification (SSI). The results show that natural frequencies can be identified from either the frequency domain methods or the time domain methods with high accuracy. On the other hand, the damping identification results given by SSI have large scatter. In terms of mode shapes, SSI technique is more robust for dealing with noisy testing data from real-world bridges since it produces smoother vibration mode shape curves.