The use of fibre Bragg gratings to detect ultrasound in anisotropic materials Graham Thursby a1 , Brian Culshaw a , Yakov Botsev b , Eyal Arad b , Roman Zeyde b , Moshe Tur b and Iddo Kressel c , a Dept. of Electronic and Electrical Engineering, University of Strathclyde, Glasgow, UK b Tel-Aviv University, Tel-Aviv, Israel 69978 c IAI Engineering Division, Israel 70100 ABSTRACT In previous work we have described the detection and location of damage in isotropic materials using fibre Bragg gratings rosettes to directionally detect Lamb waves. To extend this technique to composite materials it is necessary to understand the propagation characteristics of ultrasound in these materials as a function of their orientation with respect to the ply, and also the directional response of fibre Bragg gratings to them. Finite element modeling of Lamb wave propagation in a 0°, 90° carbon fibre plate is described, as are experiments to detect these waves for various orientations of the source and alignments of the FBG transducers. Results of the experiments are interpreted with respect to predictions from the FE modeling and are shown to give good qualitative agreement. Keywords: Lamb waves, carbon fibre composites, damage detection, fibre Bragg gratings, finite element modeling. 1. INTRODUCTION The concepts of condition monitoring systems using optical fibre Bragg gratings as a strain detection element are many and varied and have been explored for some time. The majority of system concepts centre upon utilising the Bragg grating in quasi static mode and assessing the response of the structure under test to applied loads. In principle, tracking the history of such loading cycles can give some insight into changes of structural dynamics. In practice whilst the measurements can most certainly be made with very acceptable precision and repeatability there remain numerous questions concerning the interpretation of data from such sensor arrays. These uncertainties arise from two principal sources. The more fundamental of these is how to reliably interpret the strain maps which the sensor array may produce. In practice the load histories may be variable in part due to changes in the load itself and in some cases normal ageing processes can be mistakenly interpreted as structural deterioration or indeed vice versa. The second major issue is that of where to put the sensors and how many are required. There is assumption that there are “vulnerable” sectors within a structure and these are the ones which require instrumenting. However we could argue that one criterion for an “ideal” structure is that all sections of it are equally susceptible to wear and damage. Consequently the locations, distribution and total numbers of sensors within a structure require optimisation and whilst there have been numerous attempts to define robust location optimisation algorithms, there is a significant need for empirical verification. Finally before use any structural assessment technology must pass through a rigorous validation process but we are currently some considerable distance from this situation. The situation is somewhat simplified in terms of sensor location and sensor distribution, when active systems are deployed, they usually, but not always, involve ultrasonic interrogation. The ultrasonic excitation array can in principle be fairly easily designed to insonify all the structure. Consequently measuring the propagation characteristics of the ultrasound through a carefully chosen matrix of positions on the structure can provide reasonably accurate damage assessment. This basic principle has traditionally been applied using piezo ceramics as both source and detector 1 g.thursby@eee.strath.ac.uk Smart Sensor Phenomena, Technology, Networks, and Systems 2008, edited by Wolfgang Ecke, Kara J. Peters, Norbert G. Meyendorf, Proc. of SPIE Vol. 6933, 69330C, (2008) · 0277-786X/08/$18 · doi: 10.1117/12.775442 Proc. of SPIE Vol. 6933 69330C-1 2008 SPIE Digital Library -- Subscriber Archive Copy