STRUCTURAL DAMAGE LOCALIZATION UNDER VARYING ENVIRONMENTAL CONDITIONS

Structural condition assessment by means of structural health monitoring has in recent years evolved into an actionable practice. For diagnosing structural health, a number of damage detection methods have been proposed, relying on vibration response data, for extraction of features that are characteristic of the intact or unsound structure. In this context, environmental variation comprises a severe challenge, since it induces deviations in the measured structural vibration characteristics, often masking the changes induced by damage. This work offers a remedy to this issue by adoption a Principal Component Analysis (PCA) based approach to account for variations induced from environmental condition variation and separate these from contributions corresponding to damage. Beyond mere detection, the proposed framework offers the possibility for localizing damage.