Download LAMB WAVE PROPAGATION IN THERMALLY DAMAGED

LAMB WAVE PROPAGATION IN THERMALLY DAMAGED COMPOSITES
Michael D. Seale* and Barry T. Smith**
*Department of Physics
**Department of Applied Science
The College of William and Mary
Williamsburg, VA 23187
INTRODUCTION
The use of composites in primary and secondary structures of aerospace vehicles is
important for increased performance with little weight penalty. Determining the response to
thermal damage is necessary for a complete understanding of the total use environment of
these materials. The objective of the research presented here is to provide a method of quantifying the amount of thermal damage in composite materials. Components which have nonvisible damage, but have degraded performance on the order of several percent, are of interest. At this level of damage the safety margin designed into the structure may be compromised.
Thermal damage of these structures occurs in three different scenarios. The first is
damage caused during repair operations of an aircraft component. A thermal cure cycle is
used to effect repairs using thermoset or thermoplastic polymers, and the heater blanket may
undergo an undocumented or unintentional over temperature excursion leading to a loss of
integrity. Second is damage that is incurred in a composite in a heat affected zone of an aircraft, such as near engine exhausts or leading edges. A third is a fire either on board the aircraft or in the immediate vicinity of a parked aircraft.
Verification of the structural integrity of composite components is needed to insure
safe operation of these aerospace vehicles. Traditional NDE techniques used on metals are
not easily adaptable to composites, due to the inhomogeneous and anisotropic nature of
these materials. Finding an effective means of nondestructively monitoring thermal damage
is extremely important to the safety and reliability of such structures.
Lamb waves offer a promising method of evaluating damage in composite materials
[1-2]. As a material is thermally damaged, the material properties are altered. Since the
Lamb wave velocity is directly related to these parameters, an effective tool can be developed to monitor damage in composites by measuring the velocity of these waves. Lamb
waves can propagate several tens of centimeters for most thin planar structures and thus,
provide a large gauge length measurement. If the measurement is made as a raster scan in
two orthogonal directions over the area of interest, a rapid detection technique is possible.
McDonnell Douglas Corporation provided several samples which simulate typical
thermal damage on carbon epoxy aircraft structures. The experiment and results will be
shown, along with a model calculation of the dispersion curves for this material. First a
review of some related measurements will be presented.
PREVIOUS REPORTS ON FIRE DAMAGE IN COMPOSITES
Since the advent of polymer matrix/graphite fiber composite components in aircraft,
the fire damage and subsequent property degradation has been important to designers. Sited
below are several studies which have examined thermal damage in composite structures.
Leach, Briggs, and Carlie [3] investigated a range of material properties for a thermoplastic PEEK matrix with carbon fibers. The temperature range used was -50OC to 240OC. It
was found that above the glass transition temperature, strength decreased by 50% and modulus fell on the order of 25%. Below the glass transition temperature the modulus only
decreased slightly while the strength fell by 20%. Flame resistance of this particular material was noted to be better than that of aluminum.
The results for heating in air of AS4/3501-6 composites was examined by Street, Russell, and Bonsang [4]. In this study, the temperature was extended to 350OC and the damage
assessed by measuring interlaminar fracture toughness, interlaminar shear strength, and
hardness. The data suggested that for overheating at 350OC, less than 5 minutes is sufficient
to seriously degrade all the parameters measured. A combined time temperature profile of
300OC for 15 minutes led to a loss of interlaminar fracture toughness. Up to 30 minutes at
temperatures of 225OC caused insignificant damage, while a temperature of 300OC for the
same period resulted in a 50% loss of interlaminar shear strength. Barcol hardness tests of
the 300OC damage only dropped by 20% for a 30 minute exposure. The sensitivity of the
Barcol hardness test as an indicator of thermal damage seems somewhat low as illustrated
by this study [4].
Operating conditions can cause what are labeled as “thermal spikes”, or temperature
fluctuations in aircraft. Damage generated under these conditions was studied by Stansfield
and Pritchard [5]. For composites that undergo these thermal fluctuations in service where
moisture ingress is possible, the matrix dominated mechanical properties can be severely
limited. It was concluded that matrix cracking and fiber debonding were the main damage
mechanisms.
SAMPLES
The heat damaged samples were provided by McDonnell Douglas Aerospace. These 6
x 6 inch samples were fabricated as 16 ply, uniaxial plates of AS4/977-3. All samples had
received different levels of heat damage, had been C-scanned, and had Barcol hardness readings performed. The exposure temperatures ranged from 550OF to 1000OF for 3 minutes and
included a sample exposed to 600OF for 10 minutes.
CONTACT MEASUREMENT
Lamb wave modes can be generated by placing a longitudinal wave contact type transducer on a plate structure [6]. Mode conversion of a part of the wave to shear waves and subsequent mixing of SV and L waves form Lamb wave modes. In these studies Lamb waves
were generated by using a 5 cycle, 500 kHz sine wave to excite the sending transducer. The
signal was received at various positions and digitized. The transmitting transducer was a
broadband, longitudinal wave transducer with a nominal center frequency of 1 MHz and a
diameter of 0.5 inches. The receiving transducers were of the pinducer type, with active ele-
Function Generator
Power Amplifier
Out
Computer
Sync.
Out
In
Switch
Box
Digitizer
Trig.
Ch. 1
Amp
Out
In
Figure 1. Experimental arrangement for velocity measurement.
ments of 0.1 inches. A coupling gel was used between the transducers and the plate. Shown
in Figure 1 is a schematic of the experimental arrangement.
The time differences were measured by imposing a computational delay on the first
signal to overlap the signals received at greater distances. The leading part of both waves is
identified as the lowest order symmetric Lamb wave, which is not very dispersive. The trailing part of the waves is the dispersive antisymmetric wave. For these measurements, a leastsquares fit from a plot of time and distance was performed to obtain the velocity of the
prompt S0 mode.
RESULTS
Figure 2 shows the phase velocity of the S0 mode for a set of specimens that were
exposed to the plotted temperatures for a duration of 3 minutes. One exception, plotted at
600OF, saw a 10 minute duration. The measurements were done with the propagation direction in the fiber direction. The error bars represent a 5% absolute error, mainly due to uncertainty in the digital channel alignment, and are quite liberal. The decrease in velocity with
what is assumed to be increasing thermal damage is quite dramatic.
The variation in velocity is on the order of 18% for the two extremes of temperature
exposure. This variation is comparable to the modulus and velocity changes measured for
fatigue damage of one million cycles reported in previous research [1]. The correlation with
extended temperatures is excellent and provides a method to assess thermal damage. The
Barcol hardness numbers showed little variation over this temperature range.
A model calculation to determine the dispersion curves for this material system was
computed using material parameters in Table 1, which includes not only the engineering
constants supplied, but the corresponding elastic constants. The engineering values can be
used to calculate most of the elastic constants using various techniques [7]. Since a value for
G23 was not available, the value of c44 (where c44 =(c22 - c23)/2) must be estimated using
common values for polymer matrix/graphite fiber composites [8]. The density of the material was taken to be 1560 kg/m3. The model calculation was the one dimensional model of
Dong and Huang [9] and is shown in Figure 3.
9.5
1810
1806
Velocity (km/s)
9
1804
1809
1805
8.5
1803
1813 (10 minute duration)
1807
8
1802
7.5
7
500
600
700
800
900
1000
O
Exposure Temperature ( F) For 3 Minute Duration
1100
Figure 2. Plot of S0 velocity for a range of samples exposed to different temperatures for
three minutes. Note: one exposure of ten minutes at 600OF.
Table 1. Engineering constants1 and associated elastic stiffness constants.
1
Constant
Value (MSI)
Constant
Value (GPa)
E1
19.3
c11
135.0
E2
1.27
c12
3.20
G12
0.73
c22
9.09
ν12
0.30
c23
1.59
c44 (see text)
3.75 (GPa)
c55
5.03
Provided by McDonnell Douglas
Theoretical dispersion curves were also calculated for AS4/977-3 using reduced stiffness constants. The effect of reducing different cij’s by 20% on the low frequency, non-dispersive region of the lowest order symmetric mode ( 0.0 ≤ fd ≤ 1.0 MHz•mm) was
investigated. The only parameter which altered the velocity of the S0 mode in this region
was c11.
Figure 4 shows a plot of the experimentally determined normalized velocity (velocity
in the damaged sample divided by the velocity in an undamaged control sample) versus
exposure temperature. Also plotted is the normalized velocity of the S0 mode with reduced
values of c11. Here, the normalized velocity is the velocity with c11 reduced divided by the
velocity with the full values of the elastic stiffness constants. Both velocities were taken at a
constant frequency-thickness product of 0.5 MHz•mm on the dispersion curve.
20
A2
S2
A3
S3
Velocity (km/s)
15
S0
10
A1
S1
5
A0
0
0
0.5
1
1.5
2
2.5
fd (MHz•mm)
3
3.5
4
Figure 3. Theoretical dispersion curve for an AS4/977-3 unidirectional composite.
Normalized Velocity (km/s)
1.05
1
full
parameters
0.95
c11 reduced
by 10 %
0.9
c11 reduced
by 20 %
0.85
c11 reduced
by 30 %
0.8
500
600
700
800
900
1000
1100
Exposure Temperature (OF) For 3 Minute Duration
Figure 4. Plot of the normalized S0 velocity of samples exposed to different temperatures.
Also shown (dashed lines) is the theoretically determined normalized S0 velocity with
reduced values of c11, as described in the text.
DISCUSSION
Presented above are the results of Lamb wave velocity measurements in thermally
damaged composites. The results of the experimental measurement show that the velocity of
the lowest order symmetric Lamb mode dropped significantly for extended thermal damage.
Theoretical calculations with reduced values of c11 show that a 25-30% reduction in this
parameter lowered velocities of the S0 mode by 15%. This value corresponds to the velocity
reduction of the sample which saw an exposure temperature of 1000OF. Barcol hardness
tests did not show a significant change for the samples studied. This result is consistent with
the observations reported by Street, Russell, and Bonsang [4].
The Lamb wave velocity is a quantitative measurement and can be used to verify the
integrity of a composite structure over the service life of the component. This is an important measurement for flight qualified composite hardware that may have undergone some
measure of thermal damage. Since the Lamb wave velocity depends on the material properties, an ideal technique can be developed to monitor composites as damage is incurred.
Finally, because Lamb waves can propagate a long distance, global information about the
integrity of the material can be obtained.
ACKNOLEDGEMENTS
The authors would like to thank Kevin Slattery of McDonnell Douglas Aerospace for
providing the samples studied. This work was supported in part by NASA Grant NAG-11063.
REFEERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
M.D. Seale, B.T. Smith, W.H. Prosser, and J. E. Masters, “Lamb Wave Response of
Fatigued Composite Samples,” Proceedings of 1993 Review of Progress in QNDE,
eds. D.O. Thompson and D.E. Chimenti, Vol. 13B, Plenum Press, New York, 1994,
pp. 1261-1266.
B. Tang and E.G. Henneke, II, “Lamb-Wave Monitoring of Axial Stiffness Reduction of Laminated Composite Plates,” Materials Evaluation, 47, August 1989, pp.
928-934.
D.C.Leach, P.J. Briggs and D.R.Carlile, “Mechanical and Fire Properties of Aromatic Polymer Composites,” Proceedings of the Third European Symposium on
Spacecraft Materials in Space Environment, Noordwijk, The Netherlands, October
1-4, 1985 (ESA SP -232, Nov. 1985), pp. 153-160.
K.N. Street, A.J. Russell and F. Bonsang, “Thermal Damage Effects on Delamination Toughness of a Graphite/Epoxy Composite,” Composites Science and Technology, Vol. 32, (1988), pp. 1-14.
K.E. Stansfield and E. Pritchard, “Damage Generation and Healing During Composite Thermal Spikes,” Proceedings of the International SAMPE Technical Conference,
21st, Atlantic City, NJ, September 25-28, 1989, pp. 120-129.
J.J. Ditri, A. Pilarski, B. Pavlakovic, J.L. Rose, “Generation of Guided Waves in a
Plate By Axisymmetric Normal Surface Loading,” Proceedings of 1993 Review of
Progress in QNDE, eds. D.O. Thompson and D.E. Chimenti, Vol. 13A, Plenum
Press, New York, 1994, pp. 133-140.
Isaac M. Daniel and Ori Ishai, Engineering Mechanics of Composite Materials,
Oxford University Press, Inc., New York, 1994.
William H. Prosser, “Ultrasonic Characterization of the Nonlinear Elastic Properties
of Unidirectional Graphite/Epoxy Composites,” NASA Contractor Report 4100,
NASA Langley Research Center, October, 1987, p. 65.
S.B. Dong and K.H. Huang, “Edge Vibrations in Laminated Composite Plates,”
Journal of Applied Mechanics, Vol. 52, June 1985, pp. 433-438.