10 1 1 553 9349 PDF
10 1 1 553 9349 PDF
10 1 1 553 9349 PDF
00222461
C 2005 Springer Science + Business Media, Inc. 561
damage can be detected and identified. By measuring
the resistance distribution, the damage distribution can
be determined.
Prior work in the use of DC electrical resistance mea-
surement to assess damage in carbon fiber polymer-
matrix composites has shown the effectiveness of this Figure 1 Composite specimen testing configuration (top view). Con-
method in sensing damage inflicted by tension [8, 9, tacts A1 , A2 , A3 , A4 , A5 and A6 are on only the top side of the specimen.
12, 13, 1519], flexure [8] and impact [7]. The resis- Contacts B1 , B2 , B3 , B4 , B5 and B6 (not shown) are on the bottom side,
tance measured has included the volume resistance and such that B1 is directly opposite A1 , B2 is directly opposite A2 , B3 is
directly opposite A3 , B4 is directly opposite A4 , etc. The point of impact
the surface resistance. The surface resistance is ob- is at the center of the specimen along its 200-mm length. All dimensions
tained with electrical contacts only on one side (e.g., are in mm.
the tension side or the compression side of a com-
posite under flexure), whereas the volume resistance
is obtained with electrical contacts that are not only on
one side of the composite. The volume resistance can though the entire surface was sanded in this work, only
be measured in the longitudinal, oblique and through- the portions beneath the electrical contacts needed to
thickness directions. In case of the longitudinal vol- be sanded.
ume resistance, the electrical contacts are around the Five types of laminate were studied, namely an
whole perimeter of the composite in planes perpen- eight-lamina unidirectional [0]8 laminate (thickness =
dicular to the longitudinal direction. In case of the 1.0 mm), an eight-lamina crossply [0/90]2s lami-
through-thickness resistance, the electrical contacts are nate (thickness = 1.0 mm), a quasi-isotropic [0/45/
directly opposite one another on the two opposite sides 90/ 45]s laminate (thickness = 1.0 mm), a 16-
of the composite. In case of the oblique volume re- lamina quasi-isotropic [0/45/90/ 45]2s composite
sistance, the electrical contacts are on the two oppo- (thickness = 2.1 mm), and a 24-lamina quasi-isotropic
site sides, such that they are not directly opposite one [0/45/90/ 45]3s laminate (thickness = 3.2 mm).
another. For each composite, six electrical contacts were ap-
Composites that are used in practice differ in thick- plied on each of the two sides. Each contact was in the
ness and lay-up configuration. In spite of the sub- form of a line along the 10-mm width of the specimen,
stantial prior work in this area, no prior work has as shown in Fig. 1. The point of impact was at the center
been done in studying the effect of the thickness along the specimen length.
of the composite or that of the lay-up configura- DC electrical resistance measurement was conducted
tion on the self-sensing behavior. The thickness re- using the four-probe method. A Keithley 2002 multi-
lates to the number of lamina. The closest prior work meter was used. The surface resistance of the top side
[7] studied an eight-lamina unidirectional composite, (referred to as the top resistance) was measured by us-
an eight-lamina quasi-isotropic composite and a 24- ing A1 and A6 as current contacts and A2 and A5 as
lamina quasi-isotropic composite. In contrast, this pa- voltage contacts; the surface resistance of the bottom
per provides a systematic study of the effects of thick- side (referred to as the bottom resistance) was measured
ness and lay-up configuration by investigation of (i) by using B1 and B6 as current contacts and B2 and B5 as
an eight-lamina unidirectional composite, (ii) an eight- voltage contacts; the oblique resistance was measured
lamina crossply composite, (iii) an eight-lamina quasi- using A1 and B6 as current contacts and A2 and B5 as
isotropic composite, (iv) a 16-lamina quasi-isotropic voltage contacts; the through-thickness resistance was
composite, and (v) a 24-lamina quasi-isotropic compos- measured using A3 and B3 as current contacts and A4
ite. Comparison of (i), (ii) and (iii) allows investigation and B4 as voltage contacts (Fig. 1).
of the effect of the lay-up configuration. Comparison During impact at progressively increasing energy,
of (iii), (iv) and (v) allows investigation of the effect of using a steel hemisphere (diameter 19 mm or 0.75
thickness. in) dropped from a controlled height, measurement
As in the closest prior work [7], damage is inflicted of the top, bottom, oblique and through-thickness re-
in this work by drop impact. This is because impact is a sistances was continuously made. The impact energy
commonly encountered cause of damage of structural was calculated from the weight of the hemisphere as-
composites. sembly (0.698 kg) and the initial height of the hemi-
sphere (up to 850 mm). After an impact, the hemi-
sphere bounced back to a height up to 1/3 of the ini-
2. Experimental methods tial height. Hence, the energy absorbed by a specimen
Commercially manufactured composites in the form of due to an impact was smaller than the energy calcu-
continuous carbon fiber epoxy-matrix laminates were lated from the initial height. Impact was directed at the
cut into strips of size 200 10 mm and then sanded by same point of the specimen at progressively increasing
using 600 grit silicon carbide sand paper for the pur- energy.
pose of removing the surface layer (about 20 m thick) The damage resulted in an indentation, the diame-
of epoxy matrix prior to the application of electrical ter of which was measured by using calipers in order
contacts. The contacts were in the form of silver paint to provide a rough indication of the extent of damage.
in conjunction with copper wire. The sanding step is The depth of the indentation was calculated from the
not essential, but it helps the electrical measurement by diameter of the indentation and the diameter of the im-
increasing the accuracy and decreasing the noise. Al- pacting hemisphere. Each indentation was made with
562
a single impact at a selected impact energy, in con- directional, crossply and quasi-isotropic composites.
trast to the multiple impacts made at the same point at The through-thickness resistance is as sensitive as the
successively increasing energies for the electrical resis- oblique resistance, but it suffers from a variation of
tance monitoring. R/Ro with impact energy that is not monotonic.
Multiple specimens of each type were similarly For the eight-lamina crossply and quasi-isotropic
tested by resistance measurement in order to ascertain composites, the trends for the variation of R/Ro with
the reproducibility of the results. impact energy are similar to those for the eight-lamina
unidirectional composite, except that the through-
thickness resistance decreases with increasing impact
3. Results energy much more clearly. The values of R/Ro are
Figs 26 show the fractional changes in resistance much higher for the bottom resistance than the top resis-
(top, bottom, oblique and through-thickness resis- tance for both crossply and quasi-isotropic composites,
tances) during impact at progressively increasing en- but are comparable for the unidirectional composite.
ergy respectively for the unidirectional, crossply and The values of R/Ro for the through-thickness resis-
quasi-isotropic eight-lamina composites and the quasi- tance are much higher for the unidirectional composite
isotropic 16-lamina and 24-lamina composites. than the crossply or quasi-isotropic composite.
Figure 2 Fractional change in resistance (R/Ro ) vs. time during impact at progressively increasing energy for the eight-lamina unidirectional
composite: (a) Top resistance, (b) bottom resistance, (c) oblique resistance, and (d) through-thickness resistance.
563
Figure 3 Fractional change in resistance (R/Ro ) vs. time during impact at progressively increasing energy for the eight-lamina crossply composite:
(a) Top resistance, (b) bottom resistance, (c) oblique resistance, and (d) through-thickness resistance.
Figure 4 Fractional change in resistance (R/Ro ) vs. time during impact at progressively increasing energy for the eight-lamina quasi-isotropic
composite: (a) Top resistance, (b) bottom resistance, (c) oblique resistance, and (d) through-thickness resistance.
564
Figure 5 Fractional change in resistance (R/Ro ) vs. time during impact at progressively increasing energy for the 16-lamina quasi-isotropic
composite: (a) Top resistance, (b) bottom resistance, (c) oblique resistance, and (d) through-thickness resistance.
Figure 6 Fractional change in resistance (R/Ro ) vs. time during impact at progressively increasing energy for the 24-lamina quasi-isotropic
composite: (a) Top resistance, (b) bottom resistance, (c) oblique resistance, and (d) through-thickness resistance.
565
resistances are particularly sensitive to minor dam- posite than the 16-lamina composite at the same im-
age, which cannot be indicated by the top or bottom pact energy. The through-thickness resistance increases
resistance. monotonically with increasing impact energy for the
The values of R/Ro for the top, bottom and oblique 16-lamina composite (Fig. 8), but mainly decreases
resistances are much higher for the 8-lamina quasi- with increasing impact energy for the 8-lamina com-
isotropic composite than the 16-lamina and 24-lamina posite. On the other hand, the top, bottom and oblique
quasi-isotropic composites. However, the R/Ro val- resistances all increase monotonically with increasing
ues for the through-thickness resistance are compara- impact energy for both 8-lamina and 16-lamina com-
ble for the 8-lamina, 16-lamina and 24-lamina quasi- posites, except for minor irregularity for the top resis-
isotropic composites. tance of the 16-lamina composite.
Figure 7 Fractional change in resistance vs. impact energy for the 8-lamina quasi-isotropic composite: Top, bottom, oblique, and through-
thickness.
Figure 8 Fractional change in resistance vs. impact energy for the 16-lamina quasi-isotropic composite: Top, bottom, oblique, and through-
thickness.
566
T A B L E I Depth of indentation for various composite configurations at various impact energies
Impact Diametera Depthb Diametera Depthb Diametera Depthb Diametera Depthb Diametera Depthb
energy (J) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm)
a Measured.
b Calculated from the measured diameter.
567
composites irrespective of lay-up configuration or 2. M . D E F R E I T A S , A . S I L V A and L . R E I S , Composites Part
thickness (number of laminae), as shown for impact B: Engineering 31(3) (2000) 199.
3. J . M . A . S I L V A , J . A . M . D E R R E I R A and T . C .
damage. However, the lay-up configuration and thick-
D E V E Z A S , Mater. Sci. Tech. 19(6) (2003) 809.
ness affect the self-sensing characteristics and the rec- 4. K . S . C . K U A N G and W . J . C A N T W E L L , Polym. Compos.
ommended resistance measurement configuration. 23(4) (2002) 603.
The oblique resistance is a particularly effective indi- 5. A . M A R T I N , J . H U D D , P . W E L L S , D . T U N N I C L I F F E
cator of damage. Although the through-thickness resis- and D . D A S - G U P T A , Key Eng. Mater. 167/168 (Damage Assess-
ment of Structures) (1999) 102.
tance is as sensitive to damage as the oblique resistance,
6. T . M A R T I N , A . J O N E S , I . R E A D , S . M U R R A Y , D .
its variation with the impact energy tends to increase H A Y N E S , P . L L O Y D , P . F O O T E , R . N O B L E and D .
and then decrease as the impact energy is increased, T U N N I C L I F F E , Key Eng. Mater. 204/205 (Damage Assessment
when the composite has only 8 laminae. In contrast, the of Structures) (2001) 371.
oblique, top and bottom resistances all increase mono- 7. S . W A N G , D . D . L . C H U N G and J . H . C H U N G , Compos-
ites: Part A, in press.
tonically with increasing impact energy, irrespective of
8. Idem., Smart Mater. Struct., in press.
the lay-up configuration or the number of laminae. For 9. S . W A N G and D . D . L . C H U N G , Comp. Interf. 9(1) (2002)
multidirectional composites with 8 or 16 laminae, the 51.
bottom resistance is more sensitive to damage than the 10. S . W A N G , Z . M E I and D . D . L . C H U N G , Int. J. Adh. Adh.
top resistance. For 8-lamina multidirectional compos- 21(ER6) (2001) 465.
11. S . W A N G and D . D . L . C H U N G , Polym. Polym. Comp. 9(2)
ites, the bottom resistance is more sensitive to damage
(2001) 135.
than the top, oblique or through-thickness resistance. 12. X . W A N G and D . D . L . C H U N G , J. Mater. Res. 14(11) (1999)
For the 8-lamina unidirectional composite, the top and 4224.
bottom resistances are comparably sensitive to damage, 13. X . W A N G , S . W A N G and D . D . L . C H U N G , J. Mater. Sci.
due to damage in the form of longitudinal matrix crack- 34(11) (1999) 2703.
14. S . W A N G and D . D . L . C H U N G , Comp. Interf. 6(6) (1999)
ing. For 16-lamina and 24-lamina quasi-isotropic com-
507.
posites, the oblique and through-thickness resistances 15. X . W A N G and D . D . L . C H U N G , Polym. Comp. 18(6) (1997)
are comparably sensitive and both increase monotoni- 692.
cally with increasing impact energy. 16. Idem., Smart Mater. Struct. 6 (1997) 504.
17. D . D . L . C H U N G and S . W A N G , Polym. Polym. Comp. 11(7)
(2003) 515.
Acknowledgment 18. X . W A N G and D . D . L . C H U N G , Composites: Part B 29B(1)
The authors thank U.S. National Science Foundation (1998) 63.
19. Z . M E I , V . H . G U E R R E R O , D . P . K O W A L I K and D . D .
for financial support of a part of this work. L . C H U N G , Polym. Compos. 23(3) (2002) 425.
20. S . W A N G and D . D . L . C H U N G , in preparation.
References
1. A . S J O G R E N , A .K R A S N I K O V S and J . V A R N A , Compos- Received 24 September
ites Part A: Appl. Schience & Manufact. 32(9) (2001) 1237. and accepted 11 October 2004
568