IMPROVEMENT OF TYPICAL DAMAGE IDENTIFICATION TECHNIQUES USING DYNAMIC STRAIN DISTRIBUTION FROM LONG-GAGE FBGSENSORS

摘要

Accurate measurement and monitoring of vibration characteristics is critical to proper detection of damage location and reliable estimation of severity of damage. In recent years, major advances have been realized in our lab on the basis of advanced quasidistributed strain measurements using long-gage fiber Bragg grating (FBG) sensors. This paper presents a thorough and systematic comparative assessment of the performance of typical model free curvature-, strain energy-and flexibility-based damage identification methods based on displacement (or acceleration) and quasi-distributed strain measurements. Through numerical simulation and laboratory experimental verification, the performance of these techniques were studied with due consideration to the sensitivity to mild-to-moderate damage under dense and sparse or limited measurements, few measurable modes and influence of measurement noise. Modified damage indices based on the concept of normalized modal macro-strain are also proposed to improve the quality of damage localization when only limited measurements are available. Finally, modal macrostrain flexibility methods are proposed and validated through numerical and experimental investigations as more reliable non-baseline indices for damage localization in civil structures. It may therefore be concluded that the performance of some algorithms might be improved for practical application to civil infrastructure by using distributed strain fiber optic sensing measurement techniques.