Comparison of System Identification Methods using Ambient Bridge Test Data

In this paper the performance of four different system identification methods is compared using operational data obtained from an ambient vibration test of the Swiss Z24 highway bridge. The four methods are the frequency domain based peak-picking methods, the polyreference LSCE method, the stochastic subspace method for estimation of state space systems and the prediction error method for estimation of Auto-Regressive Moving Average Vector models. It is not the intention to elect a winner among the four methods, but more to emphasize the different advantages of each of the methods.