Keywords: crack detection, plate structure, structural damage detection, non-destructive testing, vibration test, frequency response function, coordinate incompatibility, noise.

One of the most interesting inverse problem in vibration and modal analysis is to determine the spatial model represented by the structural stiffness matrix , the mass matrix and the damping matrix from the response model represented by the receptance matrix , with direct application in system identification and application in structural damage detection. While modern engineering practice has come to terms with the presence of cracks and no longer considers their presence as tragic, it is desirable to be able to detect an invisible (to the naked eyes) crack at an early stage, preferably by non-destructive testing (NDT) methods. The Damage Location Vector method (DLV) method has previously been shown to be able to locate crack and determine extent of damage in structures of simple element topology like discrete systems of mass and springs, space truss, and frame structures [1,2,3]. DLV uses frequency response functions (FRF) obtained from non-destructive vibration tests. It investigates the case of early structural damage in which there is no appreciable change in mass and damping. The resulting change in structural stiffness matrix is reflected in changes of FRF which can be exemplified by the evaluation of damage location vector (DLV), hence the name of the method. DLV requires the dynamic stiffness matrix of the original undamaged structure and the frequency response curves of the currently damaged structure. The former is obtained from a finite element model of the virgin structure and the latter are obtained from impact hammer test.

In this paper it is shown that DLV can be used for more complex plate structures. By using the abundance of FRF data, it can overcome the problem of noise and coordinate incompatibility between the theoretical model and the experimental vibration model that experimental FRF data obtained by vibration tests would inherently possess. It was found that for the cantilever plate the effect of experimental noise on damage detection by DLV was negligible and randomly distributed, even when only one third of the FRFs could be obtained by experiments, indicating that the method of dynamic expansion could be employed to estimate the unmeasured FRFs. It was also found that in the case of a plate structure, the effect of damage in an element is felt primarily by its nodes and to some lesser extent by the neighbouring nodes. Although the effect decays quickly, it complicates the identification of damages in the case of multi-site damage. Further studies are needed to improve the dynamic expansion of unmeasured FRF data and to provide scheme of automatic interpretation of damage and make DLV a practical tool in structural damage detection for more complex hybrid structures.

1

Choudhury, A.R., "Damage detection in structures using measured frequency response function data", Ph.D thesis, Department of Mechanical Engineering, Victoria University of Technology, 1996.

2

Huynh, D., He, J., Tran, D, "Damage Location Plot: A Non-Destructive Structural Damage Detection Technique", Proc of the 9th Int Conf on Civil and Structural Eng Computing, Paper 117, Egmond-aan-Zee, The Netherlands, Civil-Comp Press, Stirling, UK, 2003. doi:10.4203/ccp.77.117

3

Tung, S., Tran, D., "Non-destructive structural damage detection using frequency response functions from vibration test", Proceedings of the 10th Asia-Pacific Vibration Conference, pp 737-742, Gold Coast, Australia, 2003.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £135 +P&P)