A Data Driven Solution to Synchronicity Deviations in Operational Modal Analysis

A scheme to minimize errors from time misalignments in the measured output of linear systems subjected to stationary excitations is presented. The objective is to provide estimates of synchronicity faults (if any) so the signals can be realigned and the identification carried out subsequently. The approach announces that no corrections are needed when a characteristic phase of the fundamental eigenvector is bellow a specified threshold and otherwise computes the time shifts as those required to move the fundamental mode phases to zero. Simulations that include noise, non-classical damping and input spectral densities not diagonalized by the eigenvectors suggest the approach can provide sufficient accuracy to be of practical value. (c) 2017 The Authors. Published by Elsevier Ltd.