Aircraft Technologies
Damage Assessment in Smart Composite Structures: the DAMASCOS Programme S.G. PIERCE, F. DONG, K. ATHERTON, B. CULSHAW, K. WORDEN, G. MANSON, T. MONNIER, P. GUY, J-C. BABOUX, J. ASSAD, E. MOULIN, S. GRONDEL, C. DELEBARRE, V. AGOSTINI, P-P. DELSANTO, I. GENESIO, E. MINO, C. BOLLER
The DAMASCOS (Damage Assessment in Smart Composite Structures) project is a European Union funded programme of work bringing together a number of academic and industrial partners throughout Europe. The aim of DAMASCOS is to apply new ultrasonic detection and generation techniques integrated within the structure, together with advanced signal processing to realise damage assessment and ageing characterisation in composite structures. Thispaper describes the background, experimental findings and future applications of the technology as the project moves into its final phase.
he fundamentals behind the DAMASCOS project are best illustrated in fi,k~un' 1. The system formed an ultrasound based interrogation system for damage assessment in advanced glass and carbon reinforced plastic materials (GRP and CFRP). Acoustic Lamb waxes could be launched into the sample from piezoelectric sources (PZT), and detected using either similar PZT transducers or optical fibre receivers. Since the sampie materials were typically of thin plate construction, we have concentrated exclusi\'ely on the propagation of ultrasonic Lamb waves [1] within the sampies. Changes in the condition of the sample under test, affected parameters of the l~amb wave propagation characteristics, and in this faslnion, by the application of suitable signal processing procedures, it was possible to infer the presence and position of structural damage within the sample plates. The experimental programme was complimented by modelling of Lamb wave propagation within the samples, and the interaction with defects. Typical acoustic sources and detectors used in DAMASCOS are shown in fi,k'tm' 2 together with some of the final target inspection structures in ti~~' ~,~,,.~,,
o 05
o it} Oe*O0 le 04 2e 04 3e 04 4e 04 5e 04 TbTe (secor ds}
~j
impact 4
aa le~03
i .....:-V. 10+o2 •
1 ¢~* 01 0
•
1{' 20 q0 40 T~'St ~:atle h u u m b e r
0 04 5{'
0 15
0 05
Figure 12. Impact results from GRP plate, SO mode, (a) row data, (t9) novelty index, (c) Sammon mapping.
0 05
I 015
code from 2 to 3 dimensions ~xas necessao; due to the 3-dimensinnal nature of a Lamb wave propagating thorough a thin plate flawed bv a hole. Tile intrinsic geometry required for [~amb wa\'e propagation in the plate required that its thickness be much smaller than the other two sizes. We avoided the problem o f using the same discretisation step in all three directions (for which the number of nodes becomes prohibitively large) by using a discretisation step t'; along the thickness smaller t h a n the steps in the other two dimensions. ~,~'e tried to optimise our solutions bv varying the ratios ex/e~ and Sv/S~. Froma numerical point of view, inorder to generate Lamb modes one could directh' input the analytical Lamb wave displacement profiles. However, in order to reproduce the physical mechanisms underlying the Lamb modes generation, we chose to perform an 'experimentallike' injection, in particular, we reproduced the strain field created by the transducer using as a reference the 1 cm diameter disk-shaped PZT sensors ot INSA. These transducers allow to select either the So or the A0 Lamb mode operating at the frequency 300 kHz and 100 kHz, respectively. In the first case the specimen was excited mainly through vibrations parallel to the length of the specimen, while in the latter perpendicular vibrations prevailed. Preliminary simulations were performed to investigate Lamb wave interaction with a fully penetrating hole. In fi\~ure 15 we present snapshots at a fixed time (t = 126 Ius) of the out-oF plane displacement component w fl~r a Lamb wave propagating in the plate with a hole of diameter 1 mm (a), 4 mm
difficulties encountered in treating 0.2 0.2 sharp or imperfect -=_ o ~ ~ contact interfaces ~" -0.2 by means of the -o.4 (a) -0.4 (b) usual Finite Dif-0.6 -0.6 o 1 2 3 2 ference techniques. time (ms) time (ms) The proposed ap0 . 6 proach shows good ~ 0.4 ~ 0.4 > convergence and ~" 0.2 / ~" 0.2 stability and is g o g o ~ -0.2 ~-0.2 then to be consid'~ -o.4[ (d) ered a very valu"l'III II ~ 0 . 6 ~ -O.SL ...... able tool in the preo 1 2 0 1 2 3 time (ms) time (ms) diction of the system signature and Figure 13. Acoustic emission signals during impacts, (a) 2J impact, (b) 4J impact, (e) 8J impact. identification of the (d) Second 8J impact. optimum signal extraction routines. The modelling apknown in acoustic emission of composproach is easily visualised as a set of ites, these high frequency components interconnected springs (fi,gure 14). can be interpreted as the signature of Changes in the system under consideracrack occurrence. tion (such as the inclusion of defects) can be easily accomplished by either cutting springs, or changing the relative spring constants. Initial propagation Computer simulation techniques prostudies were performed with simple 2D vide an important role as basic tools in sample geometries. The extension of the fields, such as damage assessment of composite structures, in which it is 2 i important to gain a good physical understanding of the propagation mechanisms of ultrasonic waves or pulses through these complex materials. A method, which has been designed for the above purpose and is particularly efficient, especially if applied in conjunction with parallel processing, is the Local Interaction Simulation Approach (LISA) [15-17]. As spin-offs of LISA, the 3 8 Sharp Interface Model (SIM)I18] and the Figure 14. The spring model approach, Spring Model have overcome [19] the 0.61'
0.6
0.4[
0.4
°i
Modelling
4
W
Ij "
%.li
8
[1
4
Aircraft Technologies
STRUCTURES o~
(b) and 10 mm (c). In these plots both o~ the faster Stl and the A o mode are riy[r o sible; the SI/ mode is seen already reflected back from the edge of the "* specimen. We do xlm] :, not report the refo • r'n O2 erence (unfaulted) case, since it is very to the case o 05 " i o 1 similar Ol of a 0 = 1 mm y ira] o hole. In fact, the interference patt) 0215 o 1 tern caused by the hole is clearly visi-o 2 025 ble onlv for tl~e 0 = x [ml b) 4 m m a n d 0 = 10 o2 mm cases. o05 + Subsequent work o, has concentrated Ol on the representay[ml o tion of more realis015 tic structural de-o~ fects. Of particular 0.2 interest was the 025 -o~ case of a delamina0 0.1 02 03 04 × [ml tion. This could be c) modelled as either Figure 15. Map of the out-of-plane c o m p o n e n t a single delaminaw of the displacement at time t= 126 ps. tion (with sharp or o
.
.
.
.
35/xs
51/xs
60 #s
.4
I
74 #s
Figure 16. Time evolution of in-plane c o m p o n e n t of displacement ~hrough the thickness for a conical delaminafion stack, every second iayer damaged. I IIE~
AIR & SPACE
smooth edges); or as a stack of delaminations with a conical profile through the plate thickness to better approximate a typical impact induced delamination stack (fi~\~ure 10). For a conical delamination stack with defects in every second layer, the displacement distribution is shown in.figure 16. Note that displacements are abated inside the cone (high stiffness), but not null; a certain amount of energy was transferred across the cone.
Conclusions This paper has discussed the development of the DAMASCOS project, whose aim is to apply new ultrasonic detection and generation techniques integrated within composite structures, together with advanced signal processing to realise damage assessment and ageing characterisation. The structures were probed with ultrasound that was generated using conventional piezo-electric sources. Detection was accomplished using either similar piezo-electric transducers, or optical fibre sensors. Since the sample materials were typically of thin plate construction, we have concentrated exclusively on the propagation of ultrasonic Lamb waves within the samples. Changes in the condition of the sample under test, affected parameters of the Lamb wave propagation characteristics, and in this fashion, by the application of suitable signal processing procedures, it was possible to infer the presence and position of structural damage within the sample plates. The experimental programme was complimented by modelling of Lamb wave propagation within the samples, and the interaction with defects. A central theme of DAMASCOS was the approach to signal processing and data interpretation which relied on the development of statistical tools, most notably novelty detection, to indicate the presence of damage within the system under test. •
Acknowledgements DAMASCOS is funded by the European Community under the industrial & Materials Technologies Programme EUROPE
• VOL.
3 • No
3/4
2001
DAMASCOS (Brite-Euram 11I), project number BE 07 4213. Note This article has been previously published in SP1E Vol.4327, paper ntunber 31, 2001. References (1) Viktorov I.A, Rayleigh a n d Lamb waves-Physical t h e o r y a n d a p p l i c a tions, 1967, Plenum Press, N e w York. (2) Alleyne D.N., C a w l e y P, The interac tion of l a m b w a v e s with defects, IEEE Trans UFFC, Vol 39 (3), 1992, p p 381-397. (3) G u o N., C a w l e y R, L a m b w a v e reflection for the quick non destructive e v a l u a t i o n of large c o m p o s i t e lami nates, Materials Evaluation Vol 52 (3), 1994 p p 404 411. (4) Pierce S,G,, Culshaw B., M a n s o n G., Worden K., Staszewski W.J., The a p p l i c a tion of ultrasonic L a m b w a v e t e c h niques to the e v a l u a t i o n of a d v a n c e d c o m p o s i t e structures, SPIE International Symposium on Smart Structures And Materials 2000: Sensory P h e n o m e n a And M e a s u r e m e n t Instrumentation For Smart Structures A n d Materials, 5 9th M a r c h 2000, N e w p o r t Beach, California, USA, SPIE Vol 3986, p p 93 103, 2000, (5) Alleyne D,N,, C a w l e y 12,Optimisation of l a m b w a v e inspection techniques, NDT&E International , Vol 25 (1), 1992, p p 11-22, (6) Staszewski W.J., Pierce S G , Worden K., Philp W.R., Tomlinson G.R., Culshaw B,, W a v e l e t signal processing for e n h a n c e d Lamb wave defect d e t e c t i o n in c o m p o s i t e plates using o p t i c a l fibre d e t e c t i o n , O p t Eng Vol 36(7), 1997, p p 1877 1888,
(7) Leontaritis I.J., Billings S.A., Input-output p a r a m e t r i c models for non-linear systems, Part I: deterministic nonlinear systems, I n t e r n a t i o n a l Journal of Control, Vol 41, 1985, p p 303-328. (8) Alleyne D.N, Pialucha T,12,C a w l e y P, A signal r e g e n e r a t i o n t e c h n i q u e for l o n g - r a n g e p r o p a g a t i o n of dispersive L a m b waves, Ultrasonics, Vol 31 (3), 1993, p p 201 204, (9) Ing R,K., Fink M., Time reversed L a m b waves, IEEE Trans UFFC, Vol 45(4), 1998, p p 1032 1043. (10) D e r o d e A., Tourin A., Fink M,, Time reversal in multiply scattering m e d i a , Ultrasonics, Vol 36, 1998, p p 443-447,
(16) S c h e c h t e r R.S, Chaskelis H.H.. M i g n o g n a R.B., Delsanto PP, Science 265, 1188 (1994), (17) D e l s a n t o RP, S c h e c h t e r R S M i g n o g n a R,B,, Wave M o t i o n 26, 329 (1997), (18) D e l s a n t o 12R, M i g n o g n a R.B. Scalerandi M., Schechter R.S., in N e w Perspectives on Problems in Classical a n d Q u a n t u m Physics, Delsanto PP Saenz A,W,, (eds.), G o r d o n & Breach 1998, Vol. 2, p.51. (19) Delsanto RP, Scalerandi M , J A ( . Soc. Am. 104, pp, 2584-2591 (1998)
(11) Perrissin-Fabert ~, Jayet, Simulated a n d e x p e r i m e n t a l study of the electrical i m p e d a n c e of a piezoelectric ele m e n t in a viscoelastic m e d i u m , Ultrasonics, Vol. 32, N ~ 2, pp. 107 112, 1994. (12) Moulin E,, Assaad J., D e l e b a r r e C., O s m o n t D., M o d e l i n g of L a m b w a v e s g e n e r a t e d by i n t e g r a t e d transducers in c o m p o s i t e plates using a c o u p l e d finite e l e m e n t normal m o d e s e x p a n sion m e t h o d , J. Acoust, Soc, Am., vo1,107, 2000, p p 87-94. (13) Dewhurst R.J., Shah Q., O p t i c a l r e m o t e m e a s u r e m e n t of ultrasound, Meas. Sci. Tecnol. Vol 10, 1999, ppR139R168. (14) Pierce SG., Philp WR., Culshaw B.,
Gachagan A., M c N a b A., H a y w a r d G., Lecuyer, Surface b o n d e d o p t i c a l fibre sensors for t h e i n s p e c t i o n of CFRP plates using ultrasonic L a m b waves, Smart Materials a n d Structures, Vol 5, 1996, p p 776-787, (15) P P Delsanto, R.S. Schechter, H.H. Chaskelis, R.B M i g n o g n a a n d R. Kline, Wave M o t i o n 20, 295 (1994).
E
About the authors: S.G. Pierce, F. Dong, K. Atherton, B. Culshaw work at the University of Strathclyde in Glasgow; K. Worden, G. Manson work at the University of Sheffield; T. Monnier, R Guy, J.-C. Baboux work at the Institut National des Sciences Appliqu6es de Lyon; J. Assad, E. Moulin, S. Grondel, C. Delebarre work at the Universit6 de Valenciennes; V. Agostini, P-R Delsanto, I. Genesio work at the Politecnico di Torino; E. Mino works at the Centro Ricerche of FIAT in Piedmont, and C. Boiler works for EADS MT2 in Munich
[email protected] (S.G. Pierce)
