Operational Modal Analysis of Historical Buildings and Finite Element Model Updating Using α Laser Scanning Vibrometer

Abstract

Without affecting the integrity or stability of the heritage monuments, vibration-based techniques provide useful solutions for acquiring global information about them. By studying the dynamic response to suitable excitation sources, it is feasible to define the mechanical characteristics of structures and identify and locate defects in their global behaviour. Laser Doppler vibrometry (LDV), which enables non-contact measurements of the vibration velocity of moving surfaces using a focused laser beam, is a highly desirable technique for qualitative dynamic characterisation and damage assessment. LDV is a simple and non-intrusive approach. It permits remote measurements and has a high degree of sensitivity and frequency adaptation. In addition, the system is entirely computer controlled, providing simple data storage, processing, and analysis. LDV has been originally researched and developed for structural and modal shape analysis of physical prototypes, in-service devices (e.g., machinery components), medical imaging applications, and damage detection and analysis relevant to small-scale non-destructive testing (NDT), and evaluation of micro to meso-targets (e.g., fracture detection and mapping in composites, modal shape and vibration analysis of objects, etc.). In spite of several successful applications in the case of bridges and thin structures, ambient vibration testing in an integrated form that includes dynamic identification, sensitivity analysis, and numerical modelling update employing modern sensor non-contact technologies is still uncommon. In this paper, the authors intend to explore further the possibility of combining ambient vibrations and OMA in combination with the non-contact LDV sensing technique in order to remotely acquire mechanical waves travelling in historical structures, track the actual behaviour of such structures, and calibrate their finite element numerical models.