Assessment of small damage by direct modal strain measurements

Abstract

Vibration based Structural Health Monitoring was and is still a hot topic in research. Much progress has been made both from the theoretical as well as the practical side. Vibration-based SHM traditionally makes use of uniaxial/triaxial accelerometers or velocity meters. Also sometimes inclinometers are installed. Recent trends in SHM are the use of high-rate GPS receivers [1], wave propagation-based piezoelectric ceramic sensing technology and optical fiber sensors for dynamic strain and temperature measurements [2]. A particular challenge for structural health assessment is the discovery of small local damage. It is well known that for small damages, the changes in natural frequencies remain very low. Moreover, they are considerably influenced by environmental conditions (mainly temperature) [3], which influence has to be eliminated on beforehand [4, 5]. Also modal displacements are rather insensitive to small stiffness perturbations. On the contrary, modal strains (or curvatures) are very receptive to small stiffness changes. Additionally, they immediately spot the damage location. Some authors have tried to derive curvatures from modal displacements but this procedure is very prone to even slight measurement and/or identification errors. Therefore, the key issue is the direct measurement of (modal) strains. However, the development of a distributed strain sensor network able to cope with the very low strain intensities during ambient excitation is still a challenge. In this paper by recent experiments on steel and concrete beams the extreme accuracy of dynamic measurements with optical FBG strain sensors is demonstrated. Another possibility to obtain precise modal strains would be the development of a transducer that amplifies the strains. In a recent research project, by using Topology Optimization [6], a transducer was developed that measures differential axial displacements over a sufficient long distance and at the same time is upscaling the strains. Results obtained with this transducer are reported. In this context, a new challenge for Optimal Sensor Placement [7] is to deal with different sensor types, e.g. displacement transducers, accelerometers and strain sensors. For localization and quantification of damage, the most powerful method is FE-model updating based on minimizing differences between measured and calculated modal parameters [8]. The addition of modal strains to the objective function of the minimization problem will improve the damage identification process.