Papers by Author: Cecilia Surace

Abstract: Damage to many structural systems (including bridges, offshore platforms, airplanes and aerospace systems) can occur during normal service due to fatigue loading, a corrosive environment, collisions with extraneous objects, etc. For such structures, in order to guarantee safety, periodic inspections and maintenance are essential: Since undetected damage may result in catastrophic structural failure, the realisation of an accurate and trustworthy damage detection technique is fundamental.Vibration-based inspection offers the potential for detecting faults by monitoring the dynamic response of a structure, exploiting the fundamental principle that structural damage affects the stiffness distribution and hence the presence of the fault will change the dynamic properties of the structure under investigation.Many vibration-based inspection techniques have been developed over recent years which require knowledge of the baseline modal responses of the structure in the original undamaged state.However, for the vast majority of existing structures in operation, such data are simply not available. This keynote presentation reviews past and present research studies in which the author has been involved that aim to detect the presence of structural damage and identify its approximate location, using only post-damage vibration measurements. The techniques presented analyse either the mode shapes, operating deflection shapes or principal orthogonal modes and their corresponding derivatives. These features have been found to be good indicators of damage due to the spatial information that can be provided with respect to location of damage. The methodologies proposed are applied to isotropic/orthotropic uni/bi-dimensional structures and corresponding numerical and experimental results are presented.

Abstract: The present paper describes an experimental validation of a new structural damage detection method based on the Polynomial Annihilation Edge Detection (PAED) technique. It is well known that concentrated damage such as a crack, causes a discontinuity in the rotations and consequently in the first derivatives of the mode shapes. On this basis, the PAED, a numerical method for detecting discontinuities in smooth piecewise functions and their derivatives, can be applied to the problem of damage detection and localisation in beam-like structures for which only post-damage mode shapes are available. As described in this paper, in order to verify this approach experimentally (a numerical assessment having already been documented in previous papers), vibration tests on a cantilever steel beam with a saw-cut have been performed and the Operational Deflection Shapes (ODS) determined. As the approach requires a reasonably high spatial resolution of the ODS, a scanning laser vibrometer, capable of acquiring data rapidly at a very large number of observation points, was used.

Abstract: This paper concerns the analysis of how uncertainty propagates through large computational models like finite element models. If a model is expensive to run, a Monte Carlo approach based on sampling over the possible model inputs will not be feasible, because the large number of model runs will be prohibitively expensive. Fortunately, an alternative to Monte Carlo is available in the form of the established Bayesian algorithm discussed here; this algorithm can provide information about uncertainty with many less model runs than Monte Carlo requires. The algorithm also provides information regarding sensitivity to the inputs i.e. the extent to which input uncertainties are responsible for output uncertainty. After describing the basic principles of the Bayesian approach, it is illustrated via two case studies: the first concerns a finite element model of a human heart valve and the second, an airship model incorporating fluid structure interaction.

Abstract: The aim of the work is to develop a procedure allowing the test engineer to determine the probability of finding a crack in a beam structure. The procedure is based on the use of wavelet analysis and the simulation is performed by taking advantage of spectral elements to represent accurately the dynamic behaviour of beam structures in the high frequency range. In this context, numerical analyses are performed with the final scope of simulating a real testing environment: measurement error is considered and spectral elements are used so as to avoid influencing the capacity of the procedure with regard to solving the inverse problem. In this article the relation between the excitation frequency and the probability of locating the fault is shown. In particular, it is demonstrated by simulation that the probability of correctly determining the fault location increases with the excitation frequency.

Abstract: The paper presents an experimental application of the Proper Orthogonal Decomposition (POD) to damage detection in steel beams. A damaged beam has been excited with a sinusoidal force, the acceleration response at points regularly spaced along the structure has been recorded and the relevant Proper Orthogonal Modes calculated. In this way it is possible to locate damage by comparing the measured dominant Proper Orthogonal Mode with a smoothed version of it which does not exhibit apparent peaks in correspondence with the damage. One of the principal advantages of the proposed damage detection technique is that it does not require vibration measurements to be performed on the undamaged structure. Moreover the ‘optimality’ of the proper orthogonal modes only requires the use of a few (one-two) of them which can be computed in real time during lab experiments or while the structure is functioning in the field.

Abstract: The object of this paper is to illustrate the use of novelty detection techniques in Structural Health Monitoring (SHM) by the consideration of a number of case studies of varying complexity, from a simple lumped-mass system to an FE model of an offshore structure to an experimental study of an aircraft wing.

Abstract: The paper describes an on-going research effort aimed at detecting the presence of delamination damage in composite panels based upon their higher-frequency structural response. Two alternative damage indexes are examined that facilitate the identification of the location and extent of delaminations. The damage indexes do not require vibration measurements to be performed on the undamaged structure. Use is made of the bending and twisting curvatures corresponding to the higher-frequency mode shapes that are post-processed via two different smoothing techniques. The modal data are obtained via finite element models based on Mindlin theory and including delaminations. These are introduced using a sub-laminate strategy that permits multiple damages to be modelled through the thickness. Various delamination sizes and locations are examined with a random noise superposed on the data in order to ascertain the degree of sensitivity of the damage index to the noise in the experimental data.