Optimal Selection of the Sampling Interval for Estimation of Modal Parameters by an ARMA- Model

Optimal selection of the sampling interval for estimation of the modal parameters by an ARMA-model for a white noise loaded structure modelled as a single degree of- freedom linear mechanical system is considered. An analytical solution for an optimal uniform sampling interval, which is optimal in a Fisherian sense, is given. The solution is investigated by a simulation study. It is shown that if the experimental length T1 is fixed it may be useful to sample the record at a high sampling rate, since more measurements from the system are then collected. No optimal sampling interval exists. But if the total number of sample points N is fixed an optimal sampling interval exists. Then it is far worse to use a too large sampling interval than a too small one since the information losses increase rapidly when the sampling interval increases from the optimal value.