Uncertainty quantification of damage detection features derived from frequency response models

System identification in the frequency domain plays a fundamental role in many aspects of structural health monitoring (SHM). Frequency domain approaches all typically involve estimation of a transfer function, the most common of which is the usual frequency response function (FRF). The FRF serves as a fundamental basis from which many features are derived that are used for detection, localization, and even quantification of damage in structures. Fundamentally, SHM must involve hypothesis testing, since inevitably estimated FRFs involve uncertainty or noise at all stages of the measurement and estimation process. Regardless of source, these errors propagate to the final decision-making process, resulting in either classification rate performance reduction or false positive detection increases. This paper builds an uncertainty model for FRF magnitude and phase estimation and uses receiver operating characteristic curves to evaluate the model’s performance on data from a plate-like structure.