Monitoring time in operational modal tests with broad and narrow band excitations

Abstract

Advancements in sensing and computer technologies have made vibration-based techniques in structural health monitoring viable methods of structural assessment. An accurate estimation of modal parameters using vibration tests may be achieved only if an optimum monitoring time is chosen. In this paper, the limitations of the currently available criteria are discussed in detail. A new set of criteria is recommended for monitoring time that encapsulates the effect of input excitation other than white noise. The new criteria consider bias and variance error trade-offs for achieving the optimal monitoring time and the number of windows used in the Welch method for spectral analysis. The recommendations in this paper allows one to select an optimum monitoring time for excitations with narrow and broad band spectral densities. A series of numerical parametric studies were conducted to validate the new recommendations. The proposed criteria are also validated using ambient vibration test data collected from a cable stayed bridge in Canada. The experimental study results were also used to demonstrate the trade-off between the spectral analysis parameters on the accuracy of estimations.