Statistical Damage Detection of Civil Engineering Structures using ARMAV Models

In this paper a statistically based damage detection of a lattice steel mast is performed. By estimation of the modal parameters and their uncertainties it is possible to detect whether some of the modal parameters have changed with a statistical significance. The estimation of the uncertainties is based on ARMAV calibration using a non-linear Prediction Error Method approach. Besides estimating the parameters of the ARMAV model this approach can also provide an estimate of the covariance matrix of these parameters. On the basis of this covariance matrix it is possible to estimate the uncertainties of the modal parameters