ASNT NDT Handbook Volume 7 ultrasonic Testing

42. Knab, L.J., G.V. Blessing and 52. Öztürk, T., J. Rappaport, J.S. Popovics
J.R. Clifton. “Laboratory Evaluation of and S.P. Shah. “Monitoring the Setting
Ultrasonics for Crack Detection in and Hardening of Cement-Based
Concrete.” ACI Journal. Vol. 80. Materials with Ultrasound.” Concrete
Farmington Hills, MI: American Science and Engineering. Vol. 1, No. 2.
Concrete Institute Bagneux, France: RILEM Publications
(January-February 1983): p 17-27. (1999): p 83-91.

43. Rebic, M.P. “The Distribution of 53. Valic, M.I. “Hydration of Cementitious
Critical and Rupture Loads and Materials by Pulse Echo USWR
Determination of the Factor of Method, Apparatus and Application
Crackability.” ACI SP 82, In Examples.” Cement and Concrete
Situ/Nondestructive Testing of Concrete. Research. Vol. 30. Amsterdam,
Farmington Hills, MI: American Netherlands: Elsevier (2000):
Concrete Institute (1984): p 721-730. p 1633-1640.

44. Rhazi, J., Y. Kharrat, G. Ballivy and 54. Voigt, T., Y. Akkaya and S.P. Shah.
M. Rivest. ACI SP 168, Application of “Determination of Early-Age Mortar
Acoustical Imaging to the Evaluation of and Concrete Strength by Ultrasonic
Concrete in Operating Structures. Wave Reflections.” ASCE Journal of
Farmington Hills, MI: American Materials in Civil Engineering. Vol. 15.
Concrete Institute (1997): p 221. Reston, VA: American Society of Civil
Engineers (2003): p 247-254.
45. Gomm, T.J. and J.A. Mauseth. “State of
the Technology: Ultrasonic Bibliography
Tomography.” Materials Evaluation.
Vol. 57, No. 7. Columbus, OH: Achenbach, J.D., I.N. Komsky and
American Society for Nondestructive P.J. Stolarski. “A Self-Compensating
Testing (July 1999): p 747-752. Ultrasonic System for Flaw
Characterization in Steel Bridge
46. Martin, J., K.J. Broughton, Structures.” Structural Materials
A. Giannopolous, M.S.A. Hardy and Technology: An NDT Conference
M.C. Forde. “Ultrasonic Tomography [Atlantic City, NJ, February 1994].
of Grouted Duct Post-Tensioned Lancaster PA: Technomic Publishing
Reinforced Concrete Bridge Beams.” (1994): p 26-30.
NDT&E International. Vol. 34, No. 2.
Amsterdam, Netherlands: Elsevier Finch, K.J., M.P. Freeman and S. Snyder.
(March 2001): p 107-113. “NDT (UT) Inspection and Data
Management for Bridge Pins and
47. Schickert, M., M. Krause and Trunnions.” Structural Materials
W. Muller. “Ultrasonic Imaging of Technology: An NDT Conference
Concrete Elements Using [Atlantic City, NJ, February 1994].
Reconstruction by Synthetic Aperture Lancaster PA: Technomic Publishing
Focusing Technique.” ASCE Journal of (1994): p 257-260.
Materials in Civil Engineering. Vol. 15.
Reston, VA, American Society of Civil Gessel, R.D. and R.A. Walther. “Ultrasonic
Engineers (2003): p 235-246. Inspection of Bridge Pin and Hanger
Assemblies.” Structural Materials
48. Krause, M., F. Mielentz, B. Milman, Technology: An NDT Conference
W. Muller, V. Schmitz and [San Diego, CA, February 1996].
H. Wiggenhauser. “Ultrasonic Imaging Lancaster PA: Technomic Publishing
of Concrete Members Using an Array (1996): p 28-33.
System.” NDT&E International. Vol. 34.
Amsterdam, Netherlands: Elsevier Harland, J.W., R.L. Purvis, D.R. Graber,
(2001): p 403-408. P. Albrecht and T.S. Flournoy. Inspection
of Fracture Critical Bridge Members.
49. Popovics, J.S. “NDE Techniques for McLean, VA: Federal Highway
Concrete and Masonry Structures.” Administration (September 1986).
Progress in Structural Engineering and
Materials. Vol. 5. New York, NY: Wiley Harm, E.E. and G.A. Washer. Problem
(2003): p 49-59. Statement 95-D-33, “Improving the
Correlation of Ultrasonic Testing
50. Boumiz, A., C. Vernet and F. Cohen Results to the In Place Condition of
Tenoudji. “Mechanical Properties of Bridge Suspension Pins.” Springfield,
Cement Pastes and Mortars at Early IL: Illinois Department of
Ages.” Journal of Advanced Transportation (1993).
Cement-Based Materials. Vol. 3, No. 3-4.
New York, NY: Elsevier Science (1996):
p 94-106.

51. Arnaud, L. “Rheological
Characterization of Heterogeneous
Materials with Evolving Properties.”
ASCE Journal of Materials in Civil
Engineering. Vol. 15, No. 3. Reston, VA,
American Society of Civil Engineers
(June 2003): p 255-265.

490 Ultrasonic Testing

Hosseini, Z., M. Momayez and F. Hassani. Structural Materials Technology II: An NDT
“Application of SAFT for Inspection of Conference [San Diego, CA,
Cracks in Concrete.” NDE Conference February 1996]. Lancaster PA:
on Civil Engineering: A Joint Conference Technomic Publishing (1996).
of the 7th Structural Materials
Technology: NDE/NDT for Highways and Structural Materials Technology III: An NDT
Bridges and the 6th International Conference [San Antonio, TX,
Symposium on NDT in Civil Engineering September 1998]. Bellingham,
[Saint Louis, Missouri, August 2006]. Washington: Society of Photographic
Columbus, OH: American Society for Instrumentation Engineers
Nondestructive Testing (2006): p 438. [International Society for Optical
Engineering] (1998).
Komsky, I.N. and J.D. Achenbach. “A
Computerized Imaging System for Structural Materials Technology IV: An NDT
Ultrasonic Inspection of Steel Bridge Conference [Atlantic City, NJ,
Structures.” Structural Materials February-March 2000]. Lancaster PA:
Technology: An NDT Conference Technomic Publishing (2000).
[San Diego, CA, February 1996].
Lancaster PA: Technomic Publishing Structural Materials Technology V: An NDT
(February 1996): p 40-45. Conference [Cincinnati, OH,
September 2002]. Columbus, OH:
Miller, W.J. and M.K. Chaney. “NDT of American Society for Nondestructive
Bridge Pins on PennDOT Structures.” Testing (2002).
Structural Materials Technology: An NDT
Conference [Atlantic City, NJ, Structural Materials Technology VI: An NDT
February 1994]. Lancaster PA: Conference [Buffalo, NY,
Technomic Publishing (1994): September 2004]. Columbus, OH:
p 252-256. American Society for Nondestructive
Testing (2004).
Prine, D.W. et al. FHWA/RD-83-006,
Improved Fabrication and Inspection of NDE Conference on Civil Engineering: A Joint
Welded Connections in Bridge Structures. Conference of the 7th Structural Materials
Springfield, VA: National Technical Technology: NDE/NDT for Highways and
Information Service for the United Bridges and the 6th International
States Department of Transportation, Symposium on NDT in Civil Engineering
Federal Highway Administration [Saint Louis, Missouri, August 2006].
(1984). Columbus, OH: American Society for
Nondestructive Testing (2006).
Thomas, G., S. Benson, P. Durbin,
N. Del Grande, J. Haskins, A. Brown
and D. Schneberk. “Nondestructive
Evaluation Techniques for Enhanced
Bridge Inspection.” Review of Progress
in Quantitative Nondestructive
Evaluation [Brunswick, ME, August
1993]. Vol. 13. New York, NY: Plenum
(1994): p 2083-2090.

Woodward, C., G. Reyes and B. Stone.
“Guided Wave Evaluation of Steel
Bridge Beams.” Topics on Nondestructive
Evaluation [TONE]: Vol. 2,
Nondestructive Testing and Evaluation of
Infrastructure. Columbus, OH:
American Society for Nondestructive
Testing (1998): p 25-36.

Structural Materials Technology
Proceedings

Proceedings: Nondestructive Evaluation of
Civil Structures and Materials [Boulder,
CO, May 1992]. Boulder, CO:
Atkinson-Noland and Associates
(1992).

Structural Materials Technology: An NDT
Conference [Atlantic City, NJ,
February 1994]. Lancaster PA:
Technomic Publishing (1994).

Infrastructure Applications of Ultrasonic Testing 491



14

CHAPTER

Aerospace Applications
of Ultrasonic Testing

Richard H. Bossi, Boeing Aerospace, Seattle,
Washington (Parts 1 to 3)
Laura M. Harmon Cosgriff, National Aeronautics and
Space Administration, Langley Research Center,
Hampton, Virginia (Part 4)
Andrew L. Gyekenyesi, National Aeronautics and Space
Administration, Glenn Research Center, Cleveland,
Ohio (Part 4)
Donald J. Hagemaier, Huntington Beach, California
(Parts 1 to 3)
Eric I. Madaras, National Aeronautics and Space
Administration, Langley Research Center, Hampton,
Virginia (Part 4)
Richard E. Martin, National Aeronautics and Space
Administration, Glenn Research Center, Cleveland,
Ohio; Cleveland State University, Cleveland, Ohio
(Part 4)
William H. Prosser, National Aeronautics and Space
Administration, Langley Research Center, Hampton,
Virginia (Part 4)
Gary L. Workman, University of Alabama, Huntsville,
Alabama (Part 4)

Part 4 is a work of the U.S. government and not subject to copyright by ASNT.

PART 1. Overview of Aerospace Applications of
Ultrasonic Testing

Ultrasonics plays a critical role in the the advantage of being sensitive to
production and inservice testing of changes in material properties because the
aerospace structures. It is applied to sound passes through the sample twice. It
metallic and composite parts using a wide is also very useful for parts that have
range of techniques, frequencies and surfaces slightly nonparallel by several
waveform types. The applications can be dozen millirad (a few degrees) because the
routine or unique. The ultrasonic reflector plate can be aligned normal to
techniques check for discontinuities such the beam.
as cracks, delaminations, porosity and
inclusions. Ultrasonics may also be used The through-thickness or pulse echo
to measure dimensions and material technique with longitudinal waves is used
properties. Table 1 lists aerospace at frequencies appropriate for the
materials that are ultrasonically tested. thickness and attenuation characteristics
Eddy current, radiographic, liquid of the materials. The sensitivity to fine
penetrant, magnetic particle and other detail is a function of wavelength and
nondestructive test methods may be more active beam size. At long wavelengths and
appropriate for particular types of quality large beams, details are lost but
issues. Table 2 lists the advantages and attenuation is less, so penetration is
limitations of ultrasound as a function of greater. Short wavelengths and small
test issues. For the most part, ultrasound beams have greater detail resolution on
is best when inspecting for planar thinner structures. Table 3 lists several
discontinuities lying parallel to the test aerospace materials with their wave
surface. If the back surface of the object is speeds and the wavelengths as a function
also parallel, it simplifies the test. of frequency. For composite materials
with layers typically around 0.2 mm
Figure 1 shows basic configurations (0.008 in.) thickness, the ability to count
used in aerospace ultrasonic testing. individual plies is lost below 5 MHz
Through-thickness ultrasonic testing uses frequency or about 0.6 mm wavelength.
two probes placed on each size of the
object. The alignment for Sizing of discontinuities is a function
through-thickness testing may be by of orientation and beam size. When the
robotics, magnetic coupling or simply beam is normal to the discontinuity and
hand held. For hand held the beam is smaller than the
through-thickness testing, one transducer discontinuity, the size can be estimated
is usually held at a fixed position while on the basis of signal loss as the beam is
the other is moved to obtain the peak scanned over the discontinuity zone.
signal. The pulse echo technique is simple When the beam is larger than the
to use and is the most common hand discontinuity, the sizing is based on the
held test. amplitude of the signal. The amplitude
based sizing must be calibrated against a
The reflector plate pulse echo standard. When using angle beams and
technique can also be very useful. It has corner trap detection for discontinuity
detection, sizing is based on the signal
FIGURE 1. Basic ultrasonic test techniques. amplitude compared to a standard.
Discontinuities whose reflection intensity
Through-transmission Pulse echo is different from the standard may not be
with reflector plate sized precisely. However, accept/reject
Pulse echo criteria for discontinuities are normally
based on the reflection amplitude in the
Metallic reflector plate standard, usually a notch machined and
set to be a conservative signal. Cracks in
thick metal structure may be sized by
crack tip diffraction. Crack tip diffraction
can be more accurate than amplitude
sizing for uneven crack discontinuities but
is rarely used on aerospace structures.

The ability of ultrasound to detect
features during testing is a function of the
changes in acoustic impedance at
interfaces. The transmission and reflection

494 Ultrasonic Testing

TABLE 1. Aerospace material test issues.

Material Inspection Issues Ultrasonic Testing

All types curved surfaces, radii, discontinuity Ultrasonic testing requires precise beam alignment, either normal to test
orientation, noodles surface or at precise angles. Normality is found by peaking signal response
for entry surface, simulated discontinuity or alignment of through-
Fiber reinforced consolidation, porosity, inclusions, transmission transducers. Curved surfaces require surface following so
polymer composite fiber-to-resin ratio, delaminations, beam can be oriented normal to surface. Radii require special orientation of
laminate/glass epoxy wrinkles, surface layers, curved probes to remain normal in radius.
surfaces, radii, noodles
Through-transmission or pulse echo testing are main techniques for
acceptance of composite laminates, using automated scanning with water
coupled piezoelectric transducers. Test variations depend on material,
configuration, thickness and sensitivity requirements and include handheld,
resonance, laser coupled, air coupled, lamb wave, spectroscopic and roller
probe techniques. All testing requires correlation with acceptance
standards.

Generally, consolidation and porosity are monitored by acoustic
attenuation using through-transmission or pulse echo testing with reflector
plate. Delaminations, voids and some inclusions are detectable with
through-transmission testing but pulse echo reflection may be more
sensitive to inclusions in some materials. Wrinkles are detectable with pulse
echo B-scanning at high frequency (>3 MHz).

Surface coatings are monitored with high frequency (such as 20 MHz)
pulse echo or resonance testing. In some cases, pitch catch configuration
m__I sahyabpeedussetrducatcurores.s radii to check for quality of noodles in root of T and

Foam core composite cracking, voids, density, bonding to Through-transmission technique is common. Water squirter systems at
skin, inclusions, fluid ingress, skin 1 MHz look for wide range of discontinuities. Air coupling may be
Honeycomb core quality, skin porosity acceptable.
structure
bonding of core to skin, Through-transmission technique is common with standard water squirter
crushed/damaged core, filled core, systems at 1 MHz, looking for delaminations and porosity. Air coupling may
inclusions, skin quality, skin porosity be acceptable.

Carbon-to-carbon consolidation, dry ply, porosity, Ultrasonic testing detects delaminations and porosity. Concerns exist with
delamination, inclusions, wrinkles means of coupling to carbon-to-carbon surface.

Castings cracks, voids/porosity, inclusions, Pulse echo angle beams, normal beams and phased arrays are used to
Forgings shrinkage, weld repairs, dimensional detect and locate discontinuities. Grain size noise affects sensitivity.
Machined parts tolerances
cracks, inclusions, grain size, residual Pulse echo testing of billets for inclusions, porosity and voids. Angle beams
Fastened structure stress look for cracks.

Welded joints cracks, residual stress, dimensional Pulse echo angle beams and phased arrays are used to detect cracks.
tolerance, repairs Internal dimensional checks can sometimes be performed at high
Bonded joint frequency.
assemblies cracks, corrosion, alloy type
Pulse echo ultrasonic testing with angle beams is used for cracks around
Coatings voids, porosity, lack of fusion, lack of fasteners. Normal beam ultrasonic testing is used for corrosion detection,
penetration, undercut, shrinkage, loss of material in top layer.
Subsystems cracks, slag, inclusions, residual
Inservice or damaged stress Normal or angle beam ultrasonic testing is used for cracks, voids,
inclusions, lack of fusion and lack of penetration in welds. Phased arrays
structure surface wetting, bond strength, can be used for beam steering in both pulse echo and pitch catch modes
voids, disbonds, degradation along welds.

Ultrasonic testing normal to bond interface detects disbonds and voids.
Spectroscopy can be sensitive to interface that correlates to bond quality.
Under special conditions, changes in acoustic attenuation or wave speed
can indicate adhesive degradation.

paint thickness, conductive layers, High frequency pulse echo technique can gage paint thickness. Thermal
thermal coatings, insulation, low coatings, insulation and low observables may need low frequency in
observable coatings through-transmission or resonance mode.

cracks, residual stress, surface Pulse echo angle and normal ultrasonic testing are useful for crack
condition detection.

impact damage, heat damage, Ultrasonic testing is useful for composite impact damage, disbonds and
moisture ingress, fatigue cracks, delaminations with normal beams. Moisture ingress can be detected
corrosion, lightning strike, disbonds, through changes in wave speeds and attenuation. Fatigue cracks in top
delaminations layers are detectable with normal or angle beams. Thickness changes in
accessible layers indicate corrosion.

Aerospace Applications of Ultrasonic Testing 495

of ultrasound pressure across interfaces is Because of the large change in acoustic
given by: impedance at an air interface, it is difficult
to couple sound into air. Cracks are an
(1) T = 2 Z2 interface between the material and air
that do not transmit and thus provide a
Z2 + Z1 large amplitude reflection. Inclusions are
interfaces between the base material and
and: the inclusion material and may not have
(2) R = Z2 − Z1 sufficient reflection for detection,
depending on the relative impedances.
Z2 + Z1 Table 4 points out the difficulty of air
coupled ultrasound because of the low
where R is the reflection coefficient, T is transmission of ultrasound from a lead
the transmission coefficient and Z1 and Z2 zirconate titanate transducer into air
are the acoustic impedances of material 1 compared to the transmission from the
and material 2 of an interface. Table 4 lists transducer into water. The same applies
interface transmission and reflection for coupling through material samples. Air
coefficients and their corresponding coupled ultrasound suffers from the
amplitudes for some possible aerospace decibel insertion losses and thus loses
test interfaces. sensitivity relative to immersion or

TABLE 2. Advantages and limitations of ultrasonic testing.

Material Inspection Issues Ultrasonic Testing

Planar discontinuities Ultrasonic testing is well suited to planar discontinuities. Ultrasonic testing does not see behind one
such as discontinuity to detect another. Inspection from both
delaminations sides of object may be required to verify. Ultrasonic
and disbonds testing is applicable to top layer of fastened structure.

Cracks detection Ultrasonic testing is sensitive to very small cracks. Beam must be oriented to intersect crack so that
sound wave is returned to transmitter/receiver or
separate receiving probe.

Crack sizing Amplitude from well oriented cracks can be calibrated for Cracks that do not grow planar and smooth will not
sizing. Crack tip echoes and timing are used for sizing. be accurately sized by ultrasonic testing. Crack tip
Thickness detection requires materials with little internal scatter.
measurement Ultrasound can be very accurate if material is uniform with
known wave speed. Must be able to sense back echo. Wave speeds must
be known.

Porosity Ultrasonic signal is scattered by porosity and can be detected Porosity can be difficult to calibrate. Usually, acoustic
Thin material by signal loss from back of part. attenuation is a criterion, but calibration standards
Thick materials must match test object.
Inclusions Ultrasonic testing can be performed at high frequencies such
Multilayer structure that the wavelength is much shorter than thickness. Near surface detection of features is subject to quality
Complex structure of signal insertion. Delay lines may be used to avoid
Ultrasonic testing can be performed at low frequencies to having ringdown of pulse. Near field transducer
obtain penetration through thick material. effects limit near surface sensitivity.

Ultrasonic testing can detect inclusions by either reflected As frequency is decreased, sensitivity to fine detail is
echo or an attenuation effect on transmitted beam. lost with ultrasonic testing, and some discontinuities
can be missed.
Ultrasonic testing can be transmitted through layered
structures bonded or in intimate contact. It is useful to detect To detect inclusions, difference between material and
delaminations between layers. Resonance testing is useful for inclusion must be enough to generate echo or
in-service testing of multilayer structures such as honeycomb. reduce transmission. Inclusion must be oriented for
Beams can be oriented to inspect zones in complex structures beam detection.
by adjusting beam angles with specially oriented shoes,
computer control immersion systems or phased arrays. Ultrasonic testing is limited to top layer of fastened
multilayer structure.

Highly complex structures may not be inspectable
with ultrasound because beams may not reach critical
zones or orientation of beam relative to discontinuity
may not be sufficiently normal to provide sensitivity.

496 Ultrasonic Testing

TABLE 3. Wavelength in common materials.

Material Acoustic Wave _______________________________W__a_v_e_l_e_n_g__th________________________________
Type
Density Impedance Speed 0.5 MHz 1 MHz 2.25 MHz 3.5 MHz 5 MHz 10 MHz 15 MHz 20 MHz
(g·cm–3) (kg·cm–2·s) (km·s–1) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm)

Air 0.0012 0.04 longitudinal 0.33 0.66 0.33 0.15 0.094 0.066 0.033 0.022 0.017
Water 1.00 148 longitudinal 1.48 2.96 1.48 0.66 0.423 0.296 0.148 0.099 0.074
Acrylic 1.15 310 longitudinal 2.70 5.40 2.70 1.20 0.771 0.540 0.270 0.180 0.135
transverse 1.10 2.20 1.10 0.49 0.314 0.220 0.110 0.073 0.055
Graphite 1.55 465 longitudinal 3.00 6.00 3.00 1.33 0.857 0.600 0.300 0.200 0.150
epoxy 2.71 1710
longitudinal 6.30 12.60 6.30 2.80 1.800 1.260 0.630 0.420 0.315
Aluminum transverse 2.50 5.00 2.50 1.11 0.714 0.500 0.250 0.167 0.125
longitudinal 5.80 5.80 2.58 1.657 1.160 0.580 0.387 0.290
Magnesium 1.72 1000 transverse 2.30 11.60 2.30 1.02 0.657 0.460 0.230 0.153 0.115
longitudinal 6.07 4.60 6.07 2.70 1.734 1.214 0.607 0.405 0.304
Titanium 4.50 2730 transverse 2.40 2.40 1.07 0.686 0.480 0.240 0.160 0.120
longitudinal 5.90 12.14 5.90 2.62 1.686 1.180 0.590 0.393 0.295
Steel, 7.80 4600 transverse 2.30 4.80 2.30 1.02 0.657 0.460 0.230 0.153 0.115
mild 7.83 4540 longitudinal 5.80 5.80 2.58 1.657 1.160 0.580 0.387 0.290
8.88 5000 transverse 2.30 11.80 2.30 1.02 0.657 0.460 0.230 0.153 0.115
Steel longitudinal 5.63 4.60 5.63 2.50 1.609 1.126 0.563 0.375 0.282
stainless transverse 2.96 2.96 1.32 0.846 0.592 0.296 0.197 0.148
longitudinal 5.82 11.60 5.82 2.59 1.663 1.164 0.582 0.388 0.291
Nickel transverse 3.02 4.60 3.02 1.34 0.863 0.604 0.302 0.201 0.151

Steel, nickel 8.59 5000 11.26
chromium 5.92

11.64
6.04

TABLE 4. Interface transmission and reflection coefficients.

Interface Transmission Reflection Transmission Reflection
Coefficient Coefficient Amplitude Amplitude
2Z2·(Z2+Z1)–1 (Z2–Z1)·(Z2+Z1)–1 (dB)
(dB)

Lead zirconate titanate to air 0.00002 –0.99998 –92.69 –0.0002
Air to lead zirconate titanate 1.99998 0.99998 6.02 –0.0002
Lead zirconate titanate to water 0.08227 –0.75
Water to lead zirconate titanate 1.91773 –0.91773 –21.70 –0.75
Graphite epoxy to air 0.00017 0.91773 5.66 –0.0015
Air to graphite epoxy 1.99983 –0.99983 –0.0015
Water to graphite epoxy 1.51713 0.99983 –75.29 –5.73
Graphite epoxy to water 0.48287 0.51713 6.02 –5.73
Acrylic to graphite epoxy 1.20000 –0.51713 3.62 –13.98
Graphite epoxy to acrylic 0.80000 0.20000 –13.98
Aluminum to air 0.00005 –0.20000 –6.32 –0.00041
Air to aluminum 1.99995 –0.99995 1.58 –0.00041
Water to aluminum 1.84046 0.99995 –1.51
Aluminum to water 0.15954 0.84046 –1.94 –1.51
Water to titanium 1.89715 –0.84046 –86.58 –0.94
Titanium to water 0.10285 0.89715 –0.94
Water to steel 1.93766 –0.89715 6.02 –0.56
Steel to water 0.06234 0.93766 5.30 –0.56
–0.93766 –15.94
5.56
–19.76
5.75
–24.10

Aerospace Applications of Ultrasonic Testing 497

FIGURE 2. Simulated 5 MHz signals as function of interface contact ultrasonic testing. Alternative
materials in acrylic sample for water immersion test: transducers with better acoustic matching
(a) setup; (b) A-scan of crack; (c) A-scan away from to air are needed to improve the air
discontinuity; (d) A-scan of steel inclusion. coupled test sensitivity. The equation for
the reflection coefficient can result in a
(a) negative number: the wave form is phase
reversed at the interface. This effect can
Transducer be seen in the acoustic waveform and can
be useful for the interpretation of
Discontinuity (air or interfaces. Figure 2, using simulation
steel insert) software shows the change in waveform
for different interfaces for an acrylic block
scanned in water at 5 MHz. It can be seen
that an insert of air (crack) versus an
inclusion of steel changes the phase of the
reflected waveform.

(b) Amplitude (arbitrary unit) Insert signal
8 10 12 14
0.8
0.6
0.4
0.2
0.0
–0.2
–0.4
–0.6
–0.8

6

Time (µs)

(c)

0.8

Amplitude (arbitrary unit) 0.6

0.4

0.2

0.0

–0.2 Front face of Back face
–0.4 acrylic of acrylic
–0.6

–0.8

6 8 10 12 14

Time (µs)

(d) 0.8 Amplitude (arbitrary unit) Insert signal

0.6 8 10 12 14
0.4 Time (µs)
0.2
0.0
–0.2
–0.4
–0.6
–0.8

6

498 Ultrasonic Testing

PART 2. Aerospace Material Production
Inspection1

Aircraft production uses a wide range of composites for aircraft. Primary structural
materials. The following discussion composites must be 100 percent inspected
expands on Table 1 for some ultrasonic by ultrasound to verify that the laminate
test applications. is properly consolidated and is clear of
porosity and of foreign materials.
Care must be taken to prepare and use
reference standards in production Carbon fiber reinforced plastics are
ultrasonic testing. Table 5 lists types of anisotropic media: acoustic plane waves
ultrasonic reference standards for moving through the material are often
aerospace materials and issues with only quasi longitudinal or quasi
ultrasonic reference standards. transverse (containing both longitudinal
and transverse characteristics). Because of
Fiber Reinforced Polymer the layered structure, the ultrasonic waves
Composite Laminate Glass interact at the ply layer interfaces.
Epoxy Ultrasound can travel along fibers and
therefore certain ultrasonic test
Composite materials are a matrix material techniques can be highly sensitive to the
reinforced with another material, often a fiber orientation in the structure. For the
fiber. Carbon fiber reinforced polymer general testing of carbon fiber reinforced
composite materials are attractive for plastic, however, ultrasonic beams
advanced structural applications because oriented perpendicular to the ply layers
of their excellent strength-to-weight are used. As the ultrasonic beam passes
ratios, high toughness, controlled through the composite material, it will be
anisotropy and ability to be fabricated in attenuated by scattering and absorption.
any desired shape. Composite materials Features in the composite larger than
are widely used in aircraft and spacecraft about 0.1× the wavelength will contribute
and are the largest area of ultrasonic to the scatter. As shown in Table 3,
testing requirements for aircraft. Figure 3 wavelengths between 0.3 and 0.03 mm
shows the growth in the use of (0.012 and 0.0012 in.) represent 1 and
10 MHz respectively. The attenuation due
to scatter and absorption of the sound

TABLE 5. Ultrasonic testing reference standards.

Material Type Comments

General all Construct standards using material like that being tested. Include
discontinuity simulation sizes above and below critical discontinuity sizes
specified by engineering. Include material thickness ranges in standard
that bound materials being tested. Includes steps reasonably close (for
example, within 10 percent acoustic signal) of material under test.

Composite step wedge, porosity, Inserts represent factory foreign materials that could be left in laminate.
laminates inclusions, delaminations Simulated delaminations may be made with nonbonding materials such as
and voids, flat bottom holes fluorocarbon resin, release coated brass shims and release ply materials.
Flat bottom holes may be used to simulate delaminations or voids for one
sided tests but not for through-transmission testing unless backside is
potted. Locations of inserts include near surface, middle and far surface.

Honeycomb step wedges, voids, Step configuration with sufficient range to cover thickness of structure.
and foam flat bottom holes Plug backside drilled holes.

Metallics International Institute of Step wedge for thickness gaging. Custom configuration to match

Welding blocks, step wedges, structure and discontinuity type.

custom configurations

Bonds voids, unbonds Create joint configuration. Create disbond inserts using release ply material
on each interface side of bond.

Aerospace Applications of Ultrasonic Testing 499

pressure of a plane wave can be squirter systems. For bubbler and squirter
represented by: systems, the surface must be wettable for
adequate coupling. Surfactants are often
(3) p = p0 e−α d ⎡⎣Np⎦⎤ added to the water systems to ensure
coupling. In immersion systems, air
10− α d ⎡⎣dB⎦⎤ bubbles or entrapped air on the part can
20 cause significant signal variations and
= p0 must be removed for proper testing.
Table 6 compares the common scanner
where d is the distance traveled (meter), types. Figure 5 shows an automated
p is the end pressure (pascal), p0 is the ultrasonic system with motion control
initial pressure (pascal) and α is the squirters programmed for contour
attenuation coefficient. The attenuation following that can perform
coefficient (decibel per meter) is obtained through-thickness and pulse echo
from by rearranging and taking the log of scanning simultaneously. Laser ultrasonic
the above equation:
FIGURE 4. Graph of sample composite attenuation at 1 and
(4) αd = 20 log p0 5 MHz.
p
12
where 20 log (p0·p–1) is the ratio of the
initial to final pressure in decibels. 10
Figure 4 shows generic plots of
attenuation of a composite material as a 8 5 MHz
function of thickness at 1 and 5 MHz. In
this example, the attenuation coefficient 6
is in range of 0.5 and 1 dB·mm–1 at 1 and
5 MHz respectively. The attenuation can 4
be highly variable among different
composite materials. The higher testing 2 1 MHz
frequency will be more sensitive to
material changes. However, as the sample 0
becomes thicker, the dynamic range of 0 2 4 6 8 10 12 14
the ultrasonic test system limits the (0.08) (0.16) (0.24) (0.32) (0.40) (0.48) (0.56)
testing and needs lower frequencies.
Thickness, mm (in.)
Most composite material tests are
specified to be through-thickness or pulse
echo. The coupling of the ultrasound to
the composite is usually through water in
immersion tanks, bubbler/dribbler or
Percentage (by mass) of composite material per aircraft
FIGURE 3. The use of composites in aircraft has increased with time. Relative attenuation (dB)

60

7E7
50

OV-22

40 A320 F-22 F-35
30 A310 A340
20
10 AV-8B ATR 72 A400M

0 767 F/A-18 E/F
1940 A380
L1011 757
A330-200
DC-10 MD-11
747-400
707 DC-9 737-300

1960 1980 2000 2020

Timeline (year)

500 Ultrasonic Testing

test systems exist but have been relatively approximately 6 mm (0.25 in.) of material
uncommon in the first decade of the at 2.25 MHz and 10 to 20 dB at 5 MHz.
twenty-first century. Air coupled Through-thickness ultrasonic techniques
ultrasound is used for low sensitivity do not determine the depth of detected
applications, typically of noncritical, discontinuities. Depth is determined by
structural aerospace materials. pulse echo time-of-flight (TOF) scanning.

For through-thickness tests, the pulser The horizontal spatial resolution for
can use a tone burst (several cycles of the the data must allow for an adequate
waveform) to increase the acoustic power. number of test points over the smallest
Through-thickness systems can have a discontinuity of interest, normally 2 to
wide dynamic range (>100 dB on some 3 points in each direction. For 6 mm
systems) and easily detect problems in (0.25 in.) resolution, data spacing of 2 to
multiple-layered structures. By comparing 3 mm (0.08 to 0.12 in.) would be used. For
through-thickness signal attenuation pulse echo testing, a spike or square wave
between a standard and the material pulse is desired for a broad bandwidth
under test, porosity, inclusions, unbonds, and depth resolution. In pulse echo
wrinkles or delaminations can all be testing, the sound is reflected from the
detected. For example, porosity in the 1 to front and back surfaces of the composite
2 percent range corresponds to a change material as well as from internal
of 4 to 8 dB in signal level for discontinuities. Figure 6 shows a 5 MHz

TABLE 6. Advantages and limitations of automated scanning system.

Method Advantage Limitation

Water squirter Handling of large parts, fast scanning Complex manipulators. Alignment of
of up to 1 m·s–1 (40 in.·s–1), contour through-transmission mode transducers to each
following, through-transmission or pulse other. Alignment of pulse echo mode to surface.
echo modes.Resolution and power of sound Water splash. Cannot use focused transducers
can be tailored (spike or tone burst). effectively.

Bubbler or Low implementation cost; surface following; Edge or cutout test limitations. Speeds may be
dribbler transducer arrays often used for high rate. limited (< 0.5 m·s–1). Pulse echo mode

(through-transmission testing can be performed

with magnetic coupling).

Immersion Excellent signal quality; good spatial Part size, buoyancy, tank depth, scanning speed
resolution; ability to use focused transducers, limitations.
arrays, angle beams and either
through-transmission or pulse echo.

Laser ultrasonic Easy to scan contoured parts, noncontact. Surface must be suitable to generate ultrasound
testing without damage. Broadband signal, signal quality
may be degraded relative to standard transducers.
Can be large, expensive and require laser safe room.

Air coupled Noncontact, low frequency Lower sensitivity than water coupled. Lower
frequency applications (<1 MHz). Edge effects.
ultrasonic testing penetration for attenuative materials.

FIGURE 5. Example of overhead gantry FIGURE 6. Pulse echo data trace of composite with internal
water squirter through-transmission and feature.
pulse echo system for large composite
objects. Front
surface
Transducer output (relative unit) 60 Internal Back
40 2 feature surface
20 echo

0 4 6 8 10
–20 Time (µs)
–40
–60

0

Aerospace Applications of Ultrasonic Testing 501

test of composite structure. The first pulse Timesignals for internal features, distance
is the front surface echo of the part (the amplitude correction (also called time
interface between the coupling media and corrected gain) should be used to keep the
the part). The third pulse is the back discontinuity sensitivity constant
surface echo of the part. Any laminar throughout the composite thickness.
discontinuities (inclusions or Discontinuities near the back wall of the
delaminations) appear as echoes between composite are electronically enhanced to
the front and back surfaces (second pulse). bring amplitude up to about the same
The amplitude of the echo correlates to level as a front surface discontinuity.
the acoustic impedance mismatch of the
discontinuity to the composite. If the Composite structure can be made in
discontinuity is longer than the beam size many shapes that complicate ultrasonic
and the impedance mismatch is tests. Figure 8 shows a picture of a
significant, then the back surface echo T section and some options for testing.
will be lost or significantly reduced. The Radius testing can be performed with a
depth of the discontinuity is obtained by special shoe for a hand-held transducer,
knowing the wave speed in the material with a special array of transducers or with
and the time of the reflection from the contour following automated scanning
waveform. Carbon-to-epoxy composite systems. Laser ultrasonic testing is suited
materials have a characteristic wave speed for radius testing because it can be easy to
around 3 mm·µs–1, ±5 percent. set up. Figure 9 shows radius testing using
miniature squirters and an automated
Automated pulse echo examinations
generally create both amplitude and FIGURE 7. Pulse echo data on composite test sample with flat
time-of-flight C-scan images for bottom holes: (a) amplitude display; (b) time-of-flight
interpretation. The amplitude images may display; (c) B-scan.
be gated for the full thickness of the part
or gated in regions of the part, such as at (a)
particular zones or just the back wall.
Time-of-flight images may use the time of 100 mm
either the peak signal or the first signal in (4 in.)
the gate. The imaging possibilities allow
important differences in the 0 25 50 75 100 125 150 175 200 225 250
interpretation of discontinuities in (1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
samples or the detection of subtle
features. For example, some inclusions Distance, mm (in.)
may be difficult to detect with a back wall
echo gate but can be detected by an (b)
internal time-of-flight gate in the
material. Wrinkles are generally below the 0 25 50 75 100 125 150 175 200 225 250
threshold commonly applied for porosity (1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
testing but can be found by B-scan data
sets. Figure 7 shows the scan of a Distance, mm (in.)
composite test sample containing flat
bottom holes. The amplitude and (c)
time-of-flight C-scans show the features.
Different colors or gray scales indicate the Front
echo amplitude or, in the case of time of surface
flight, the part thickness or depth. In
Fig. 7, the amplitude of the echoes from Back
the holes are close to the amplitude signal surface
from the back of the part, so the holes are
detected by the presence of the edges. In
the time-of-flight image, the holes have
different gray scale levels due to the time
of the signal. The B-scan of a line trace
across the time-of-flight image shows the
difference in time location of the signal
between the front surface and the back of
the part. Care must be taken in evaluating
pulse echo amplitude images as
discontinuity reflections or multiple
indications of the discontinuity can be in
the amplitude gate. Porosity detection
with pulse echo amplitude is usually more
sensitive than through-thickness testing
because the back wall or reflector plate
technique of Fig. 1 makes the signal pass
through the part twice. For comparison of

502 Ultrasonic Testing

scanner. When angle beams are used in experimental setup where the angle of a
composites for access to particular composite part in a through-thickness
locations, care must be taken as to where system was changed. The plot is a
the beam is transmitted. As noted above, notional extraction of the data from
the multiple layers in a composite result multiple tests showing how the
in multiple mode conversions. The ultrasound signal varies significantly in
layered structure and strong anisotropy amplitude as a function of the
can have significant effects on the beam orientation. This demonstrates that care
performance. Figure 10 shows an must be taken to maintain normality to
the surface of composites and that an
FIGURE 8. Options for ultrasonic testing of angle beam test can be highly variable.
T section: (a) T section; (b) pulse echo test
of radius; (c) pitch catch test of noodle; Honeycomb and Foam
(d) pulse echo test of noodle. Core Laminates

(a) Honeycomb and foam core laminate
structures are inspected primarily with
(b) through-thickness testing for production.
A lower frequency, around 1 MHz, is used
PE UT of radius compared to the higher frequencies used
for solid laminate. Internal delaminations,
nonbonded inclusions and core damage
will be detected. Figure 11 is a photograph

FIGURE 10. Effect of angle on transmission of ultrasound in
composites: (a) scan diagram; (b) results.

(a)

Noodle

(c) Pitch catch Ultrasonic transducer Metal
reflector
of noodle Laminate section
Peak signal (relative scale)
(d) (b)

PE of noodle 0 0.18 0.35 0.52 0.70 0.87 1.05 1.22
(10) (20) (30) (40) (50) (60) (70)
FIGURE 9. Miniature water squirter radius Angle off normal, rad (deg)
inspection.
FIGURE 11. Honeycomb core sample.

50 mm
(2 in.)

Aerospace Applications of Ultrasonic Testing 503

of a sectioned honeycomb sample having between the transducer and the
a thin skin, two sizes of honeycomb and a carbon-to-carbon surface. The transducer
taper. Figure 12 shows a system for near field and ringdown characteristics
inspecting large aerospace structures. need to be carefully selected to allow
Calibration with a standard helps near-surface testing. The back face
establish sensitivity to the features of reflection was used to detect porosity or
interest. The bonding of the skin to the unconsolidated structure. Delaminations,
honeycomb core should have proper dry ply and inclusions are detectable by
adhesive fillets. Lower frequencies and the internal reflection. Wrinkles will have a
signal variations inherent in many core weak change in the back reflected signal
structures prevent the detection of low and may have a weak internal reflection
levels of porosity (less than 4 percent) in but be detectable because of the linear
the laminate skins. In most cases, the indication in scanned image.
application for the honeycomb and foam Carbon-to-carbon standards can be
core structure and the thin (three-ply to difficult to work with because the
ten-ply) skins do not require an ultrasonic repeated application of couplant changes
test for porosity. When porosity testing the response. An alternate material, such
may be desired, a high frequency pulse as phenolic, matched to the appropriate
echo scan may be performed. Over foam attenuation can be used to calibrate the
core, the back echo from the skin is likely transducer and receiver setup.
to be lost at the interface. High
frequencies (above 10 MHz) and high Castings, Forgings and
scanning spatial resolution may be used Machined Parts
to look in the surface layers and detect
porosity. Full waveform data are taken Ultrasonic testing is not the main method
and then analyzed ply by ply. for inspection of aerospace castings,
forgings and machine parts. The shape
Carbon-to-Carbon and locations of the discontinuities are
better inspected in most cases with other
Carbon-to-carbon structures are used in nondestructive test methods. However
aerospace for their strength at high occasions exist, particularly with cracking
temperature and for their light weight. where ultrasound may be used as a
Ultrasound is typically performed at backup test and sometimes as a primary
1 MHz because the material is relatively test.
attenuative. At 1 MHz, thickness up to
around 20 mm can be inspected. Care Castings can be difficult for ultrasonic
must be taken with coupling to the testing because of the surface conditions,
carbon-to-carbon because water should be grain size and complex structures. Crack
avoided. Moisture content in detection at specific locations where the
carbon-to-carbon is unacceptable. beam can be suitably oriented may be
Alcohol, or alcohol in a 50 percent mix performed using standard angle beam
with deionized water, is acceptable. With ultrasonic test techniques. Phase array
carbon-to-carbon structure, the amount of ultrasonic testing can be useful for these
couplant can affect the signal levels and applications and has been applied to
therefore needs to be applied consistently. complex titanium aerospace castings. The
Contact pressure can also be an issue. An phased array allows the beam to be
elastomeric material is recommended scanned for internal coverage without
moving the probe. This is particularly
FIGURE 12. Large aerospace structure useful for locations that may not be
inspection system. accessible by a standard probe.

Aerospace forgings may be inspected in
the billet stage with ultrasonic testing.
Titanium billets are an example for engine
applications. Focused ultrasonic testing is
performed on the billets by using a range
of transducers to focus the beam at
appropriate depths to achieve required
detection sensitivity.

Welded Joints

Aerospace welds are usually thin and
typically are tested radiographically, as are
other industrial welded joints. Calibration
for ultrasonic testing is performed with
standard International Institute of

504 Ultrasonic Testing

Welding blocks. For aerospace structures, 2. Tap testing with a coin or
a particular weld routinely inspected with computerized tap testing is perhaps
ultrasonic testing is the friction stir weld. the most common bond test. The
Figure 13a shows the weld scheme. This signal’s frequency is in the sonic
type of weld is subject to curved rather than the ultrasonic range. It is
discontinuities in the vertical plane of the most useful on thin structures. In the
weld. The weld discontinuity does not case of composites, it should not be
lend itself to common corner trap considered reliable on structures with
ultrasound detection. Phased array top skins of more than three to five
ultrasonic testing has been found to be an plies. The tap test should be performed
excellent technique to test the weld. The in a consistent grid pattern on the
basic phased array technique is shown in structure. The sound should be clear
Fig. 13b. over good structure and change to a
dull or muffled sound over poor
Bonded Joints bonds.

Bonded structures offer significant 3. Specially dedicated acoustic devices
advantages in aircraft design, manufacture commonly called bond testers use low
and performance over traditional fastened frequency vibration schemes, often in
structures. Bond quality is often assessed a pitch catch mode, to excite the
looking for nonbonded regions with structure and sense a difference
ultrasonic and other techniques. between good and bad bonds. Such
bond testing is similar to the
1. Shearography, an optical rather than ultrasonic technique called resonance
an acoustic technique, can be used as mode testing, using a transducer with
an alternative on bonds with thin continuous or wave train frequencies
skin, <2.5 mm (< 0.1 in.), depending of hundreds of kilohertz. These
on the materials and structures. techniques can be scanned over a part
to extract image data from the sensor.
FIGURE 13. Friction stir weld: (a) welding scheme; Typically, sensing is based on a
(b) phased array testing. comparison of the input waveform to
the received waveform.
(a)
4. In pulse echo mode, ultrasonic
Sufficient download force to waveforms are used to measure bond
maintain registered contact line thickness and interface echo
signals. B-scans are often studied to
Advanced side Join assess the reflection intensity and
of weld uniformity of bond quality. Film
Leading edge adhesives will be relatively thin,
Shoulder of rotating tool usually less than 0.25 mm (0.01 in.).
Paste adhesive bonds tend to be
Trailing edge Probe thicker, from 0.25 mm (0.01 in.) to
of rotating tool Retreating side of weld over 2 mm (0.08 in.). The attenuation
characteristics of the adhesive can
(b) Water containment housing indicate the adhesive condition in
Phased array transducer some bonds.
Water or couplant
5. Ultrasonic spectroscopy has been used
Test object on bonded interfaces and is sensitive
to variations in bond quality beyond
the sensitivity of standard ultrasound.
For critical primary structure
applications, the bond quality
assessment procedure may include a
verification of bond strength.
Nondestructive techniques have not
been successful in determining bond
strength because the strength
parameter is measured only at failure.
Proof testing of a joint is required to
ensure adequate strength. It is possible
to perform localized proof testing
using laser generated shock waves.
This technique makes a dynamic
strength assessment as a very localized
proof test of the bond.

Aerospace Applications of Ultrasonic Testing 505

6. Controlled, localized stress waves in
materials and bonded joints offer
opportunities for the characterization
of structures, in particular the strength
monitoring of bonded joints. Stress
waves of intensity sufficient to
evaluate bonding can be generated by
electron beam, mechanical impact or
laser pulse. The laser beam shape
results in controlled, very localized
testing of internal strength. The
technique is sensitive to weak bonds
created by poor adhesive mixing,
improper surface preparation or
contamination.

506 Ultrasonic Testing

PART 3. Inservice Inspection of Aircraft

Aircraft structures may develop cracks Preparation of Reference
during service because of stresses Standards
experienced during flight and landing.
Composite structures are subject to The first step in creating a reference
impact damage from birds, hail, service standard is to select sound material of
vehicles and other sources. Therefore, convenient size to eliminate edge effects
airline operators require that the aircraft in the area of interest. The material
manufacturer provide a nondestructive should be as free from natural
testing manual with information and imperfections as possible and should be
procedures for determining the condition similar to the test object in chemical
of these structures. The manual supplies composition, heat treatment, attenuation,
information about the possible locations velocity and shape (geometry). Next,
of cracks and other service induced artificial discontinuities, representative of
conditions and recommends applicable those to be detected, are created with saw
techniques for detecting them. If the cuts, drilling, fatigue cracks or electric
anticipated crack is on an accessible discharge machining. These manufactured
surface, test methods such as visual, liquid discontinuities are used to generate a
penetrant, magnetic particle or eddy response equivalent to that expected in
current testing may be selected for actual tests, allowing the instrumentation
detection. However, numerous locations to be set at a specific sensitivity level and
throughout the structure are made up of to indicate discontinuity resolution for
multiple layers of detail components various techniques. The proper selection
joined together by rivets or bolt fasteners. and use of reference standards is the key
At these locations, it is possible for cracks to successful ultrasonic testing of aircraft
to be generated in a subsurface member components and permits the use of terms
and go undetected until they propagate to and values that have significant meaning
a surface. Early detection of subsurface for describing test results. Without proper
cracks is possible with ultrasonic and use of reference standards, test results
radiographic testing. Composite have little or possibly no significant value.
assessment is based on visual detection of Usually, the type and orientation of
damage followed by ultrasonic testing. discontinuities are known in advance. By
Composite structures are designed to using reference standards with calibrated
tolerate barely visible impact damage. reflectors similar to the discontinuities of
Damage that does not cause a visible interest, proper testing procedure can be
indication can be tolerated in the established and validated. If the testing
structure. procedure can clearly indicate artificial
discontinuities in a reference standard,
Reference Standards for then there is a high probability that it can
Tests of Aircraft also indicate natural discontinuities in
actual test objects. If the sensitivity level
The purpose of an ultrasonic reference is too low, harmful discontinuities may
standard (or calibration standard) is to not be detected. Conversely, if the level is
provide a test specimen that simulates as too high, natural material characteristics
nearly as possible discontinuities that may may be mistaken for discontinuities.
be encountered in actual tests. Reference
standards help establish instrument Testing Procedure
calibration and are used to ensure that
particular discontinuities are detected For a typical ultrasonic test, the following
with a predetermined sensitivity. steps are performed.
Reference standards are used not only to
facilitate initial adjustment but also to 1. A reference standard is prepared for
check periodically on the reproducibility calibration.
of the measurement.
2. The component is prepared for testing
by removing loose paint and dirt.

3. The ultrasonic test instrument is set
up, and the calibration standard is
used to adjust controls and get a
discontinuity pattern on the A-scan.

Aerospace Applications of Ultrasonic Testing 507

4. An appropriate couplant is selected example of this application is the
and applied to the area of interest. detection of cracks radiating from
attachment holes in the curved attached
5. The test object is scanned according to fittings on the horizontal stabilizer,
detailed instructions specific to the elevator, rudder, flap and aileron in
component. Fig. 16. Several techniques may be used to
establish the proper angle of the
6. All indications of discontinuities are transducer. The mathematics of these
located and identified. techniques are discussed in detail
elsewhere.1
7. After testing, the ultrasonic equipment
is withdrawn and the couplant is One of these techniques allows testing
removed from the test surface. holes of different diameters with the same
apparatus. The incident beam needs to be
Figures 14 and 15 show an example of a perpendicular to the axis of the hole and
landing gear inspection. Figure 14 shows a allow lateral movement of the transducer
standard created for the test. Figure 15 to achieve a refracted longitudinal wave
shows the ultrasonic waveforms at several tangential to the inner curved surface.
locations along the standard and for a Figure 17 shows this configuration. A
crack in the part (Fig. 15). clear plastic shoe must be fabricated to fit
the outer radius of the test object and to
Cracks allow the transducer to move laterally in
the shoe on a plane perpendicular to the
Ultrasonic testing is often used on aircraft center line of the shoe. The lateral motion
structure to detect discontinuities of the transducer d within the plastic shoe
radiating from attachment holes in effectively changes the angle of incidence
fatigue sensitive areas. Anticipated crack and thus the angle of refraction at the
areas can be tested using one or more curved interface. By adjusting the lateral
wave modes. Holes with access limited to position of the transducer in the shoe, the
a curved surface may require only a
refracted longitudinal wave technique. An

FIGURE 14. Crack testing of landing gear: FIGURE 15. Display patterns for ultrasonic
(a) landing gear; (b) enlarged area of testing of landing gear: (a) signal at location
interest; (c) reference standard. 1 of Fig. 14c; (b) signal at location 2 of
Fig. 14c; (c) signal at location 3 of Fig. 14c;
(a) Forward (d) small crack signal in landing gear.

(a)

1

Area of (b)
interest
2
(b)

Crack location

(c) Saw cut from edge of block (c) 3
to hole (simulated crack) (d) 3
25 mm
(1 in.) 3
75 mm
(3 in.)

2

1 13 mm (0.5 in.)
diameter hole
88 mm centered in
(3.5 in.) reference block

508 Ultrasonic Testing

longitudinal wave may be refracted The technique requires only one shoe
tangentially to the circumference of any fabricated to fit the outer radius,
size hole (Fig. 17): regardless of the hole diameter. The test
object, unlike the reference standard, may
(5) d = r0 V1 sin θr have only a semicircular surface on which
Vm scanning can be accomplished. Therefore,
to obtain maximum coverage around the
and: hole, the transducer and shoe are moved
around the entire curved surface in one
(6) sin θr = ri direction and are then rotated 3.14 rad
r0 (180 deg), and the scan is completed in
the opposite direction. Figure 18 shows an
and: example of piston cylinder lug inspection.

(7) d = ri V1 Such techniques use fundamental
Vm ultrasonic principles programmed into
codes for modeling tests with a computer.
where d is the transducer offset (meter) Fig. 19 shows an example using computer
graphics to model the optimum angles.
from the centerline of the tube or shoe, ri Once the geometry of the object is input
is the inner radius (meter) of the tube or into the program, the orientation of the
transducer can be adjusted to peak the
hinge, r0 is the outer radius of the tube or reflected signal. In Fig. 19, the signal is
hinge, V1 is the velocity (meter per peaked for either a longitudinal wave or a
second) of the sound beam in the plastic transverse wave test of the object. The
angles of refraction where the signals are
offset shoe (meter per second), Vm is the peaked are found by changing the
velocity (meter per second) of the sound parameters in the model. The model also
provides estimates of the timing of the
beam in the tube or hinge depending on reflection which are useful to aid in
interpretation. In some cases, echoes may
critical angle, θ1 is the incident angle be obtained by both the longitudinal and
(degree) of the beam in plastic and θr is transverse waves. The model will help
the desired refracted angle of the test determine the source and timing of these
signals.
beam for beam tangency at the inside

diameter surface (degree).

FIGURE 16. Ultrasonic detection of cracks in FIGURE 17. Transducer offset for longitudinal
hinge fittings: (a) test of assembly with ultrasonic tests.
eyebolt in place; (b) test from above;
(c) test from below. Center line

(a) Center line Transducer
d
Transducer

Hinge fitting assembly

V1 Clear plastic shoe

Eyebolt θr
θ1
Refracted sound beam Vm Notch
ro standard

(b) Center line ri

Scan

Tested zone Legend

(c) Tested zone Center line ri = inner radius of tube or hinge (meter)
d = transducer offset from the centerline of tube or
Scan
shoe (meter)

ro = outer radius of tube or hinge (meter)
θ1 = incident angle of beam in plastic (degree)
θr = desired refracted angle of the test beam (transverse

θ3 or longitudinal θ2) for beam tangency at
inside diameter surface (degree)

V1 = velocity of sound beam in plastic offset shoe
(meter per second)

Vm = velocity of sound beam in the tube or hinge (shear
V3 or longitudinal V2) depending on critical angle
(meter per second)

Aerospace Applications of Ultrasonic Testing 509

Tests for Cracked Structures chromium molybdenum steel, with a
reference notch 0.25 mm (0.01 in.) deep.1
An example of ultrasonic testing as
applied to aircraft structures is shown in The D0 and Di dimensions are used to
Fig. 18. An angle beam technique is used make the reference standard. Figures 20
to detect fatigue cracks at the inside and 21 show a transverse wave check of a
diameter of the attachment lug holes wing spar splice. Transducer location,
without removing the bushings or reference standard dimensions and typical
disassembly from the aircraft. Cracks as scope presentations are illustrated. The
small as 0.25 mm (0.01 in.) deep can be reference standard design is obtained from
detected with a transverse wave incident data in the engineering drawings. The
to the inner surface at an angle of 0.8 or
1 rad (45 or 60 deg) rather than at a FIGURE 19. Computer model of ultrasonic testing:
tangent to it. Acrylic (methyl (a) 0.5 rad (30 deg) for longitudinal wave test;
methacrylate) shoes and transducers are (b) longitudinal peak signal; (c) 1 rad (60 deg) for transverse
used. Reference standards are made of wave test; (d) transverse peak signal.
Unified Numbering System H41300
(a)
FIGURE 18. Ultrasonic testing of piston
cylinder lugs: (a) from side and above; 30 degree wedge
(b) looking inboard; (c) from below.

(a) Forward

Cylinder

(b) 0.03Amplitude (relative units)

0.02

0.01

0.00

B-B –0.01

(b) A-A –0.02
Piston
–0.03
Crack
orientation 15 20 25 30 35 40 45

Travel time (µs)

Typical scan (c)

Transducer position 1 60 degree wedge

Transducer position 2 (d) 0.03Amplitude (relative units)

Crack orientation 0.02
View A-A (looking inboard)

(c)

Transducer position 1
Transducer position 2

0.01

0.00

–0.01

–0.02

–0.03

View B-B (looking up) 15 20 25 30 35 40 45

Travel time (µs)

510 Ultrasonic Testing

transverse wave test is used to locate occur, the object must be exposed to the
fatigue cracks in the horizontal and corrosive environment and subjected to a
vertical legs of the lower spar cap. The sustained tensile stress for an
high stresses associated with these undetermined period of time. The time
positions usually generate cracks adjacent required for the object to crack is directly
to the outer two rows of fasteners (as related to the stress intensity and the type
denoted by the darkened fastener and quantity of corrosive products in the
pattern). The cracks are oriented parallel environment. The variability of the time
to the forward and aft direction so the to crack is a major concern because a
transverse wave testing is done normal to repetitive testing interval is difficult to
the crack direction. The cracks in the spar establish. In some cases, components have
cap are covered by the doubler and the been found cracked by high residual
ultrasonic beam is directed under the stresses before installation on the aircraft.
doublers to detect cracks in the caps. Similar cracks have occurred in the
A97075, temper 6, overwing frame
Unified Numbering System A97075 forgings, as illustrated in Fig. 22. Testing
and A97079 wrought aluminum alloys are these forgings requires removal of cabin
highly susceptible to stress corrosion seats and interior side panels. Visual
cracking when in the temper 6 (solution testing may be difficult to perform or may
heat treatment and aged) condition, not be effective for detecting small cracks.
especially in the short transverse grain
direction. Unfortunately, these alloys, FIGURE 21. A scan display from ultrasonic
when in temper 6, are used for many test of center wing front spar splice:
primary structural applications because of (a) reflection from center hole 1; (b) crack
their high yield and tensile strength extending from edge of hole 1; (c) crack
properties. For stress corrosion cracking to extending from edge of hole 3; (d) no
cracks. Dashed lines represent positions of
FIGURE 20. Ultrasonic tests of center wing front spar splice: holes 1, 2 and 3 of Fig. 20.
(a) diagram of component; (b) diagram of wrought
aluminum alloy reference standard. (a)

(a) 1

Up Scan
Forward

(b) 1

Lower 1
spar 3
cap
321
Lower skin

Lower spar cap (c)
(d)
Scan

(b)

Saw cut notches 13 mm (m0.2m55i(nm1.)1min3.()m1 min.()0.5 in.)
(simulated cracks) 25

Scribe line 1
2
3

13 mm diameter holes 63 mm13(m2.m52in(50.).m52mi5n.(m)1m2in5.()m1 imn.)(1 in.)
(0.5 in.)
in.)
6.4 mm (0.25

Aerospace Applications of Ultrasonic Testing 511

When visual testing is impractical, determine their length. A decision to
ultrasonic longitudinal wave (straight either repetitively test panels with cracks
beam) techniques may be used as shown within flyable limits or replace panels
in Figs. 23 and 24. In this example, the with cracks beyond flyable limits can be
transducer is placed at the edge of the reached to maintain the structural
forging so that the ultrasonic beam is integrity of the horizontal stabilizer.
normal to the crack plane for maximum
reflection. Corrosion

Figure 25 shows an example of cracks Although contact between two
in the horizontal stabilizer constant galvanically dissimilar metals such as steel
section integral machined skin panels and aluminum is a known cause of
(planks) made from Unified Numbering galvanic exfoliation corrosion, the design
System 97075 wrought aluminum alloy, of aircraft structures occasionally requires
temper 6. The cracks typically run forward that such metals be joined. When joining
and aft between the bolt holes common them is unavoidable, it is a design
to the operating bulkhead. Extensive requirement that contacting surfaces be
cracks of this type are possible in the electrically insulated with organic paint or
panels and could reduce the strength of that one of the surfaces be coated with a
the horizontal stabilizer. Ultrasonic testing metallic coating galvanically similar to the
of the skin panels can detect cracks and other surface. For example, on many types
of aircraft, cadmium plated steel bolts are
FIGURE 22. Stress corrosion cracking of used. The cadmium plating not only
frame forging of Unified Numbering System protects the steel bolts from corrosion but
A97075 wrought aluminum alloy, temper 6. also provides a surface galvanically similar
to aluminum so that the possibility of
Cracks corrosion is greatly reduced. However, if
the cadmium is depleted or if a crevice
where moisture can collect exists between
the fastener head and the aluminum skin,
pitting and intergranular corrosion may
occur (Fig. 26).

FIGURE 23. Ultrasonic testing of frame FIGURE 24. A scan display of ultrasonic test
forging for stress corrosion cracks: (a) from of frame forging for stress corrosion cracks:
above; (b) cross section. (a) position 1 back reflection; (b) position 2
reflection from fastener hole; (c) position 3
(a) Crack crack reflection, far flange; (d) position 4
crack reflection, near flange.
Position 4 Position 3 (a)

AA (b)

Position 1 100
percent
Position 2
(c)
(b) Crack
(d)
Transducer
positions

Section A-A

512 Ultrasonic Testing

Intergranular corrosion occurs along FIGURE 27. Detection of intergranularAmplitude (arbitrary unit)
aluminum grain boundaries which in corrosion with ultrasonic tests: (a) response
sheet and plate are oriented parallel to the from unpainted or normal paint thickness
surface of the material because of the with no corrosion; (b) response from paint
rolling process. Intergranular corrosion in buildup area with no corrosion; (c) response
its more severe form is called exfoliation from corroded area with loss of back
corrosion — an intergranular delamination reflection.
of thin layers of aluminum parallel to the
surface, with white corrosion products (a)
between the layers. Where fasteners are
involved, the corrosion extends outward 2
from the fastener, either from the entire
circumference of the hole or in one 1
direction from a segment of the hole. In
advanced cases, the surface bulges 0 Time (relative scale)
upward; in milder cases, there may be no
bulging and the corrosion can be detected (b) Amplitude (arbitrary unit)
only by eddy current or ultrasonic testing.
The ultrasonic testing may be performed 2
as shown in Figs. 27 and 28.
1
FIGURE 25. Stress corrosion cracks in
integrally machined skin panels of horizontal
stabilizer constant section: (a) side view;
(b) upper skin panel; (c) lower skin panel.

(a) Transducer

(b) Crack 0Amplitude (arbitrary unit) Time (relative scale)
Time (relative scale)
Crack location Ultrasonic (c)
testing area
(c) 2
Ultrasonic 1
testing area
0

Crack location

FIGURE 26. Galvanic aspects of corrosion in aluminum FIGURE 28. Typical exfoliation corrosion in
cladding. spar cap of extruded Unified Numbering
System A97075 wrought aluminum alloy,
temper 6: (a) cap; (b) cross section.

Contaminants (electrolyte) (a) A
in moat area formed by
fastener and countersink Wing skin

Portion of aluminum cladding Paint film
corroded away
Aluminum skin
Paint film Paint film Exfoliation
A corrosion

Intergranular Anode Anode Extruded spar cap
corrosion + –+
fissures (b) Exfoliation path
Cathode Cadmium-plated fastener
Transducer
Section A-A

Aerospace Applications of Ultrasonic Testing 513

Composite Damage for interply delamination from one side.
Contact probes with soft tip (dry coupled)
Composites are nondestructively tested in and conventional contact with
service to determine the extent of damage appropriate couplants may be used. The
and to evaluate repairs. Damage may most common frequency range is then
occur from impact, lightning, erosion, 2.25 to 5 MHz. Higher frequencies can be
stress or fatigue. Discontinuities include used for thin samples, such as less than
holes, cracks, delaminations, distortion, 2.5 mm (0.1 in.) thick. The transducer
erosion, water entrapment in honeycomb should be highly damped and delay lines
and burned or overheated surfaces. used to resolve discontinuities near the
Ultrasonic testing is used to detect surface.
interply delaminations in laminates and
skin-to-core disbonds in honeycomb. For Figure 29 shows an ultrasonic C-scan of
honeycomb inspection with access to a damaged composite. The time-of-flight
both sides, through-transmission display shows how in composites the
techniques are preferred using a water damage takes on different shapes with
coupled test yoke. Transducer frequency is depth in the laminate. In a color version,
1 to 2.25 MHz. the near surface appears dark blue. Light
blue, green and yellow are deeper and
Pulse echo techniques are preferred for orange is the back surface. It is common
inspecting a laminate composite structure in composite damage for ply damage to
be larger below the surface. The overall
FIGURE 29. Ultrasonic time-of-flight C-scan of damaged shape of the damage is a function of ply
region in composite sample. layup. Damage must be mapped for
structural repair.
25 mm
(1.0 in.) Bond Testing2

Bond testing is performed on both
aluminum bonded and honeycomb
composite structures. The common
techniques are listed in Table 7 with
comments on their application. Bond
testing using continuous waves is
discussed elsewhere in this volume, in
connection with instrumentation.

Bond Testing Instrumentation

An effective tool for bond testing has a
combined phase and amplitude detection
circuit for detecting and quantifying
certain disbond conditions and marginal
adhesive strength.

TABLE 7. Bond testing technique.

Technique Methodology Comments

Tap testing Coin tapping or mechanical tap hammer. Useful on thin structure (three-ply to four-ply
composite). Coin test requires operator to listen for
Ultrasonic Ultrasonic transducer is driven at resonance dull sound. Tap hammer provides numerical output.
resonance frequency. The electrical impedance of
testing transducer is monitored to detect changes Reference null is used to set signal and then look for
as function of changes in test object. changes as transducer is scanned. Self nulling is also
possible. Resonance test can see significant changes,
Pitch catch Generation and detection of plate waves >12 mm (> 0.5 in.) in composite and honeycomb.
affected by poor bonding. Transducer frequency
may be swept or constant. Detection may test For low frequencies, contact may be made without
for time, amplitude or phase effects. couplant by using transducers with tips.

Mechanical Low frequency (audible range) single-contact Proper test frequency is needed for sensitivity. Setup
uses frequency sweep for operator to choose best
impedance probe. Loading, like resonance ultrasonic frequency on test standard. Single-point contact is
good for uneven structure. Sensitive to honeycomb
analysis testing, affects amplitude and phase. far side discontinuities.

514 Ultrasonic Testing

The bond tester operates by exciting Bond Reference Standards
the test material with a series of pulses
from a transmitting transducer. Adhesive The proper preparation and selection of a
bond discontinuities are detected by bond reference standard is a prerequisite
comparing the wave train from the test to obtaining meaningful and consistent
object to the wave train received during results from ultrasonic tests of bond
calibration from a reference standard of joints. A reference standard compares the
known bond integrity. ultrasonic characteristics of test objects
with those of a reference standard with
Pulse transmission is predominantly by known bond integrity. It is essential that
lamb waves, because the test object the reference standard resemble the test
thickness in this application is only a object in material composition and bond
fraction of the wavelength. Transmitted joint geometry. It should also possess a
pulses are detected by the receiving level of bond integrity equal to or slightly
transducer and electronically compared to better than that required in the final
those received from the reference bond. product.
Phase and amplitude change because the
pulse travels along a layer of unbonded A reference standard is required for
material. The unbonded layer is more free each material type and for geometry and
to vibrate at higher amplitude, with less composition corresponding to the test
energy dissipation and lower wave object. Each reference standard must be
velocity than a bonded reference prepared under adhesive bonding
standard. conditions that closely simulate accepted
and expected production practice. The
Bond Testing System Modes reference standards contain no adhesive
discontinuities.
Alarm Mode. To make a valid comparison
of pulses from test objects of unknown
integrity, the bond tester is first calibrated
to a standard. Amplitude and phase are
adjusted to prescribed values while the
reference standard is monitored. Then the
alarm mode is selected, and its level is set.
There are two choices for discontinuity
detection in the alarm mode: (1) sensing
deviations of the phase from its reference
value; (2) sensing the combined
deviations of the phase and amplitude.
Sensing the combined deviations is
recommended. After the amplitude and
phase levels are adjusted, the alarm’s
activation level determines the increase in
amplitude and shift in phase required for
discontinuity indication by alarm. When
testing bond joints in alarm mode, the
activation level, amplitude and phase
adjustments remain set after calibration.
The alarm activation level can be used as
a quantitative indicator of bond joint
integrity under known conditions.

Metering Mode. In addition to the alarm
mode for which the bond tester was
designed, a test procedure can be verified
by reading the instrument response to the
ultrasonic properties of the local bond
region. This attempt to quantify bond
tester response allows correlation of
responses with bond strength, at least for
certain failure modes. The bond tester
alarm circuit is altered so that the
potential difference between the base of
transducer and the ground could be
measured with a digital volt meter.
Voltage readings are recorded as bond
integrity measurements for each bond
locale.

Aerospace Applications of Ultrasonic Testing 515

PART 4. Ultrasonic Testing for Space Systems and
Aeronautics

Introduction Lamb Wave Assessment of
Fatigue and Thermal
Ultrasonic nondestructive testing plays a Damage in Composites
major role in the manufacturing and
operations of both space systems and Composite materials are being used more
aeronautics. The National Aeronautics and widely today by both aeronautics and
Space Administration has research space systems. In addition other industry
programs at field centers to develop applications are increasing, including
nondestructive test techniques for automotive, sports equipment and many
assessment of critical structures in others. These increased usages are due to
manufacturing and in operation. These the high strength-to-weight ratio and
programs comprise basic and applied versatility in design. Composites also
research in a broad spectrum of aerospace provide weight savings over traditional
related sciences and technologies. Many metals without sacrificing strength.3
of the nondestructive test activities are
also included in programs for the It is important to understand fatigue
National Aeronautics and Space and thermal damage in composites and
Administration to transfer its how that will affect critical applications.3
technological expertise to other users Various techniques have been tested in
— interacting with universities, industry the past, in particular in ultrasonic testing
and other government agencies to and characterization. Ultrasonic lamb
enhance its research. waves provide a convenient means of
evaluating material changes in
Ultrasonic tests and characterizations composites. As a composite material is
of advanced material systems play a large damaged, the elastic parameters of the
role in the agency nondestructive test structure change and in turn change the
programs, with a major emphasis on lamb wave velocity. This characteristic
aerospace propulsion systems and space provides an effective tool to determine
structures. Feasibility of testing is included damage in composites by monitoring the
in every phase of the design and wave velocity changes. Lamb wave
development activities leading to the measurements are better than
deployment of critical space hardware. In through-thickness ultrasonic
addition, the National Aeronautics and measurements because they are sensitive
Space Administration also has a mission to in-plane elastic properties and can
objective in aeronautics to develop an propagate over long distances.3 It is
understanding of aircraft materials and important to recognize also that thermal
structures. Typical material systems degradation as well as mechanical fatigue
include composite structures, monolithic damage may occur under general thermal
ceramics, superalloys and materials for mechanical loading.
operations at extreme temperature.
The correlation between lamb wave
Applications covered below include velocity and stiffness was measured with
(1) assessment of fatigue and thermal strain gages and by correlating lamb wave
damage in composites using lamb waves, velocity and crack density for
(2) assessment of aging wire insulation, mechanically fatigued composite
(3) studying fatigue of complex samples.3 The fatigue damage experiments
composites and flywheel rotors composed were performed using both strain gage
of carbon fiber reinforced polymer matrix and lamb wave velocities. The composite
composites for energy storage on the samples were AS4/35001-6 graphite epoxy
International Space Station, (4) reinforced with a stacking sequence of [0/903]S. Two
carbon-carbon materials for the space 305 × 381 mm (12 × 15 in.) plates were
shuttle and (5) creep damage in fabricated and C-scanned before being cut
superalloys with acoustoultrasonics. into small test samples to check for
abnormalities. The scans revealed a
porosity level of 5 to 7 percent by volume.
The plates were then cut into 280 × 38 mm
(11 × 1.5 in.) coupons with an average
thickness of 1.2 mm (0.05 in.). Two
6.35 mm (0.25 in.) strain gages were

516 Ultrasonic Testing

attached to each sample: one axial and coupon to obtain lamb wave signals.
one transverse. Young’s modulus and Crack density measurements within the
Poisson’s ratio were then obtained from same 25 × 25 mm (1 × 1 in.) area on the
the measurements of the stress, axial coupon were made optically by counting
strain and transverse strain. Loading to through a microscope. As expected, the
ultimate strength was determined by crack density increased with increasing
loading to failure. cycles and the velocity squared decreased
with increasing cycles, as before. The
The coupons were subjected to decrease in modulus was estimated from
tension-tension fatigue in a 380 MPa the crack density measurements using the
(55 000 lbf·in.–2) capacity load frame at expression reported by Caslini.5 For a
10 Hz and at an R (ratio of minimum load [0m/90n]S laminate, the stiffness loss was
to maximum load) value of 0.3. The upper calculated:
load was set to 160 MPa (23 000 lbf·in.–2)
— 33 percent of ultimate strength. Higher (8) E= 1
fatigue cycle values were successively
applied to the specimens. In this region, E0 ⎛ tanh λ ⎞
only the S0 and A0 modes allow for the 2D ⎟
leading part of the wave to be identified 1+ ⎜ n E2 ⎟
as the lowest order nondispersive ⎜ mE1 λ ⎠⎟⎟
symmetric wave. The distances were ⎝⎜⎜ 2D
measured at a precision of 1 mm
(0.04 in.), and a least squares fit of the where D is crack density (1 meter), E is
time and distance gave the velocity of the
S0 mode. The velocity of the lowest order modulus for the damaged state, E0 is
symmetric lamb mode was measured and modulus for undamaged state, E1 is
the modulus was obtained from the strain longitudinal modulus, E2 is transverse
gage measurements both before and after modulus, G12 is shear modulus, h is
each cyclic loading. The samples were lamina thickness (meter), m is number of
then removed from the load frame at
intermediate values of the fatigue cycles 0 deg layers in the half thickness of the
to make the contact measurements.
plate, n is number of 90 deg layers in the
For both fatigued and unfatigued
samples, an immersion measurement was half thickness of the plate and λ is defined
also performed to obtain dispersion curves
to compare with earlier reported results by Eq. 9:
(Fig. 30).3 The immersion measurement
results agreed with previous tests in a (9) λ2 = 3G12 E0 n + m
similar arrangement.4 The results showed h2 E1 E2 n2 m
that the contact results gave roughly a
7 percent difference in velocity for the The results gave a normalized modulus as
symmetric mode when fatigued.3 a function of crack density (Fig. 31).3

In similar experiments, a correlation Thermal experiments were performed
with crack density was obtained when using heat damaged samples of 16-ply,
performed with the loading raised up to uniaxial plates of woven carbon epoxy
45 percent ultimate failure at 200 MPa prepreg with an average thickness of
using pin transducers placed on the 3.05 mm (0.12 in.). As in the fatigue study
above, the lowest order symmetric mode
FIGURE 30. Plot of lower modes of experimental dispersion was used for the measurement. The results
show a dramatic decrease in velocity with
curves for one undamaged specimen and one fatigued for thermal damage along the fiber direction.
1000 h.3
Follow-up studies were performed by
Seale and Madaras6 to measure the lamb

4.5 Unfatigued FIGURE 31. Calculated normalized modulus (solid line) as
4 Fatigued 106 cycles
3.5 function of normalized velocity squared as function of crack
3 density for all samples.3
2.5
Velocity (km·s –1)2 1.05
Velocity modulus1.5
1 1
(normalized)0.5
0.95
0.75
0.9

0.85

0.8 0.85 0.9 0.95 1 1.05 0.8 0.2 0.4 0.6 0.8 1
0 Crack Density (mm–1)

Frequency × distance (MHz·mm)

Aerospace Applications of Ultrasonic Testing 517

wave velocity over a wide frequency range loading cycle lasted for a total of 15 300 s
while subjecting two different composite (255 min).
specimens to thermal mechanical cycling.
The composite materials were graphite The temperature and strain axes are
fiber reinforced amorphous thermoplastic normalized to sustained levels in the
polyimide laminates and high thermal mechanical loading profile
performance graphite fiber reinforced (Fig. 32).6 The graphite fiber reinforced
bismaleimide thermoset composite.6 The amorphous thermoplastic polyimide
1.22 m by 305 mm samples were laminate samples were subjected to high
manufactured with 16 and 32 plies with temperature profiles and both high and
stacking sequences of [45/0/–45/90]2S and low temperature profiles. The 32-ply
[45/0/–45/90]4S. Thermal mechanical samples were exposed to low temperature
cycling of the samples was performed and high strain profiles only.
with either 151 MPa (22 000 lbf·in.–2) or
345 MPa (50 000 lbf·in.–2) capacity load A digital lamb wave imaging device
frames equipped with environmental (Fig. 33)6 was used in this set of
chambers with a thermal control range of measurements. A unique feature of this
–54 °C to 344 °C (–65 °F to 651 °F). The system is bicycle tire patch material as
chamber dimensions were 400 mm couplant between the test sample and the
(16 in.) wide by 686 mm (27 in.) tall by transducers (Fig. 33).
400 mm (16 in.) deep. The upper and
lower sections of the samples remained The elastic bending stiffness and
outside the chamber and only the middle out-of-plane stiffness of the sample
670 mm (27 in.) section was subjected to materials were computed from a
thermal extremes. reconstruction of the flexural plate mode
dispersion curves according to laminated
For all the samples, the load was plate theory.6 For propagation in the 0 rad
applied along the length of the samples in (0 deg) direction, the only stiffness
the 0 rad (0 deg) direction. Both high and constants significantly affecting the curve
low strain profiles as well as high and low are bending stiffness constant D11 and the
temperature profiles were used. The low out-of-plane stiffness constant A55.
strain profiles had strain levels that Consequently, curve fitting of the
ranged from 0 to 2000 microstrain with a experimental data was restricted to these
sustained strain at or above 1040 two parameters. Figure 34 shows the effect
microstrain for 10 800 s (180 min). The of bending and out-of-plane stiffness
high strain profiles had strain levels constants by reducing the lamb wave
ranging from 0 to 3000 microstrain with a velocity.7 The dispersion curve for
sustained strain at or above propagation in the 1.57 rad (90 deg)
1560 microstrain for 10 800 s (180 min). direction has similar trends except that
The temperature extremes for the high D22 and A44 become the controlling
temperature cycling was –18 °C to 177 °C stiffness coefficients for curve fitting of
with a sustained temperature of 177 °C for the experimental data.
10 800 s (180 min) whereas the low
temperature cycling was –18 °C and A reduction in composite stiffness due
135 °C with a sustained temperature of to matrix cracking was observed. Typical
135 °C for 10 800 s mm (180 min). Each results are shown in Fig. 35 and Table 8.
More results are given elsewhere.6,7 A
FIGURE 32. Thermal-versus-mechanical profile.6 much larger difference in stiffness loss for
the samples loaded at high temperatures
Strain (normalized)2.5 1.2 as compared to low temperatures was
Temperature (normalized)2 1 observed, with the graphite fiber
1.5 0.8 reinforced bismaleimide thermoset
1 0.6 composite not performing well at high
0.5 0.4 temperatures. The results also indicated
0 0.2
–0.5 1 FIGURE 33. Image of scanner, showing
–0.2 frame, bridge and head.6
0 4
1 2 3
Time (h)

Legend
= Strain
= Temperature

518 Ultrasonic Testing

that the out-of-plane shear load carrying an ultrasonic test technique for the aging
capabilities are matrix dominated and effects.9-11 Experiments were performed
that matrix cracking due to mechanical with military specification wire samples
fatigue damage in composites leads to a normally used in aircraft: wiring
decrease in elastic moduli. Bending specimens were heat damaged in an oven
stiffness showed few effects of the thermal over a range of heating conditions while
mechanical loading: the matrix was measuring axisymmetric mode phase
insensitive to matrix cracking as well as velocities. Samples include a range of
the large standard deviations in the types and gages.
frequency range used.8

Assessment of Aging of FIGURE 35. Mapping of A55 stiffness
Wiring Insulation coefficient for composite samples when

Environmental aging of wiring insulation unaged and when aged for 10 000 h at
in critical systems because of the onset of
brittleness and cracking in both insulation indicated strain level: (a) 16-ply
and conductor materials has become an
important issue to government agencies thermoplastic; (b) 16-ply thermoset;
such as the Department of Defense, the
National Aeronautics and Space (c) 32-ply thermoset.
Administration and the Federal Aviation
Administration and to the industry that (a) High
manufactures and maintains these critical Unaged
systems. With this broad interest, a strain
number of different test techniques have
been tried, including the development of Normalized

A55
1.00

0.78

FIGURE 34. Flexural dispersion curves comparing velocities (b) Low
with 25 percent reduction from full values for D11 and A55 Unaged
stiffness constants.7 strain
Normalized

A55
1.00

1.2

Velocity (km·s–1) 1

0.8

0.6 0.76

0.4 (c)

0.2 Unaged High
strain
Normalized

00 50 100 150 200 A55
1.00

Frequency (kHz)

Legend 0.70

= full stiffness values
= D11 reduced by 25 percent
= A55 reduced by 25 percent

TABLE 8. Normalized values for time of flight in aged samples. Time measurements are in seconds.

______1_6_-_P_ly__T_h_e__rm__o_p__la_s_t_ic______ _______1_6__P_l_y_-_T_h_e_r_m__o_s_e_t_______ _______3_2__-P_l_y__T_h_e_r_m__o_s_e_t_______

_____U_n__a_g_e_d_____ ___H__ig_h__S_t_r_a_i_n___ _____U_n__a_g_e_d_____ ___H__ig_h__S_t_r_a_i_n___ _____U_n__a_g_e_d_____ ___H__ig_h__S_t_r_a_i_n___

Direction Time Time Time Time Time Time
of Time
of Flight of Standard of Standard of Standard of Standard of Standard of Standard

Flight Deviation Flight Deviation Flight Deviation Flight Deviation Flight Deviation Flight Deviation

0 rad (0 deg) 1.03 0.01 1.07 0.02 1.11 0.01 1.09 0.03 1.08 0.01 1.07 0.02
0.01 1.07 0.02 1.07 0.01 1.08 0.03 1.06 0.01 1.05 0.02
1.57 rad (90 deg) 1.05

Aerospace Applications of Ultrasonic Testing 519

Development of an applicable model insulation. Experimental details are found
was simplified by the fact that a number elsewhere.9-11
of researchers have studied acoustic
guided wave propagation in cylinders. In The angular dependence of the wave
these models, a number of axisymmetric amplitude was examined by holding the
modes propagate in an isotropic cylinder transmitting transducer while rotating the
as a function of material properties, receiving transducer in a fixed location in
geometry, frequency, propagation order increments of 0.18 rad (10 deg).
and circumferential order. The Axisymmetric modes exhibit no angular
axisymmetric mode was investigated to dependence, making identification
determine its ability to detect and straightforward (Fig. 37).10
quantify degradation in electrical wire
insulation because of its As an example, the effect of heat
nondispersiveness at the low frequencies. damage is shown in Figs. 38 and 39 for
The initial experiments measured the MIL-W-22759/34 wire samples of each
dispersion curves on a simple model
consisting of a solid cylinder and a solid FIGURE 38. Images of MIL-W-22759/34
cylinder with a polymer coating. The 16 gage wire as exposed to room
lowest order axisymmetric mode was temperature: (a) virgin wire; (b) 349 °C
sensitive to stiffness changes in the wire (660 °F) for 1 h; (c) 399 °C (750 °F) for
insulation.9 1 h.10

Two ultrasonic transducers are used in (a)
a pitch catch configuration to generate
and receive an ultrasonic guided wave in (b)
the wire (Fig. 36).10 The wave will
propagate in both the wire and the
insulation, with the amplitude and overall
wave speed being affected by the
condition and stiffness of the wire

FIGURE 36. Attachment of transducers to
insulated wire.10

(c)

FIGURE 37. Typical ultrasonic signals from a FIGURE 39. Bar chart showing phase velocity for each gage
polymer coated aluminum rod, showing
axisymmetric wave and first flexural wave.10 family and each heat damage condition in insulated wire

1.0 (sheathed in polymer of tetrafluoroethylene and ethylene)
using MIL-W-22759/34.9
0.5
4000 C B
0.0 3800 B C
3600 A
–0.5 3400 A
3200
–1.0 3000 12
0 25 50 75 100 125
Time (µs)
Amplitude (relative unit) C
Velocity (m·s–1) B

A

2800

20 16

Legend Wire gage

A. Baseline.
B. Aged 1 h at 349 °C (660 °F).
C. Aged 1 h at 399 °C (750 °F).

520 Ultrasonic Testing

gage to 270 °C (518 °F) for up to 200 h.9-13 composites showed further applications of
Ultrasonic measurements were taken on lamb waves and provided useful
samples removed from the oven every 3 h information on composite
for 15 h and then every 20 h. The results characteristics.14
are shown in Fig. 39.9 The data for each
gage show a rapidly increasing phase Ultrasonic Resonance
velocity at short oven exposure times and Spectroscopy Applied to
a slower increasing phase velocity at Composite Flywheel Rotors
longer oven exposure times. The
measurements of the condition of the Flywheel energy storage devices
insulation appeared to approach a comprising multilayered composite rotor
limiting phase velocity value. The systems are being studied extensively for
damaged wire insulation also became the International Space Station. A flywheel
darker and more brittle as the tests system includes the components
continued (Fig. 38).10 necessary to store and discharge energy in
a rotating mass. The rotor is the complete
Similar experiments with the use of rotating assembly portion of the flywheel,
lamb waves on composite materials to composed primarily of a metallic hub and
determine fiber volume fractions in a composite rim. The rim may contain
several concentric composite rings
FIGURE 40. Simple rotor is metallic hub with (Fig. 40).15 Advances in current ultrasonic
rim of eight concentric rings.15 spectroscopy research were used to inspect
such composite rings and rims and a flat
Rings coupon manufactured to mimic the
manufacturing of the rings.
Hub
Ultrasonic spectroscopy can be a useful
nondestructive test technique for material
characterization and discontinuity
detection.16,17 Other approaches have
used a wide bandwidth frequency
spectrum created from a narrow ultrasonic
signal and analyzed for amplitude and
frequency changes. An ultrasonic
resonance spectroscopic system18 has been
developed that uses a continuous swept
sine waveform in a pitch catch or
through-transmission mode and performs
a fast fourier transform on the frequency

FIGURE 41. Through-transmission ultrasonic spectroscopy on acrylic sample: (left) digital input waveform in time domain;

(center) digital output waveform in time domain; (right) typical output display of spectrum and spectrum resonance spacing
domains.15

Thickness 50 Amplitude (V·s)
7.92 mm (0.312 in.) 40
30
024 20
10
0
6

Frequency, MHz

0.400 MHz, 1280.7 A

0.203 MHz, 415.5 B

Ratio, 0.3244 2.0

0.0 20.5 41.0 61.5 82.0 102.5 0.0 20.5 41.0 61.5 82.0 102.5 1.5 Amplitude (V·s)2
Signal time (µs) Signal time (µs)

1.0

0.5

0.0
0.31 0.90 1.50

Resonance spacing
(MHz)

Aerospace Applications of Ultrasonic Testing 521

spectrum to create the spectrum variations between the flat composite
resonance spacing domain, or coupon and composite rings were
fundamental resonant frequency. detected as major differences in the
Ultrasonic responses from composite response signals (Fig. 42). The presence of
flywheel components have been analyzed discrete and clustered voids with widths
to assess this nondestructive test greater than 1.7 mm (0.07 in.) was
technique for the quality assurance of detected in thick composite rings as an
flywheel applications. The objectives of amplitude reduction in the spectrum and
this study were to determine the effects of spectrum resonance spacing.
the constituents of single and multilayer
composite material systems and interfacial A unique detection of kissing disbonds
bond properties within those systems on requires further investigation, as their
the frequency responses and the effects of existence in composite rings was not
intentionally seeded and naturally confirmed destructively or corroborated
occurring discontinuities on the resonant with other nondestructive techniques.
frequencies in the various structures to Voids with a width of 1.5 mm (0.06 in.)
assess the technique for certifying or smaller were not detected in the
composite flywheels in the International multiple-ring composite rim. The
Space Station. ultrasonic responses before and after proof
spin testing contained the same
All specimens were measured in the resonances for the four outer rings,
through-transmission, pitch catch suggesting that damage was not
configuration using liquid or gel couplant introduced to the rim. As a result, the
(Fig. 41). Coaxial cables and medium signals from the multiple-ring composite
damped transducers were used to rim are baseline signatures to be
maximize energy and bandwidth. The compared after fatigue testing. On the
transducer frequency depended on the basis of these findings, ultrasonic
upper limit of the frequency sweep for the spectroscopy is a potential nondestructive
5 MHz and 10 MHz transducers. test tool for flight certification of flywheel
rotors for the International Space Station.
Amplitude and frequency changes in
the spectrum and spectrum resonance Impact Damage in
spacing domains were evaluated from the Reinforced Carbon-Carbon
ultrasonic responses of a flat composite on Space Shuttle Thermal
coupon, thin and thick composite rings Protection
and a multiple-ring composite rim.15 Full
thickness resonance was produced in Following the space shuttle Columbia
discontinuity-free composite rings. accident in 2003, an extensive
Foreign materials and delaminations in investigation was conducted to determine
composite rings were detected as the physical cause of the orbiter failure
amplitude reductions in the spectrum and to prevent future mishaps. Analysis of
domain and changes in fundamental debris, flight data and video revealed the
resonant frequency. Manufacturing loss of Columbia and its crew was caused
by a breach of the thermal protection
FIGURE 42. Response from flat composite coupon compared system on the leading edge of the
shuttle’s left wing. The breach was
with response from composite ring: (a) spectrum; initiated by insulating foam shedding off
(b) spectrum resonance spacing.15 of the external tank and striking the
reinforced carbon-carbon leading edge
(a) 80 Amplified (V·s) structure on ascent. Upon orbiter re-entry,
the breach allowed superheated gasses to
60 1 2 34 56 enter the interior of the wing and melt
40 Frequency (MHz) support structures. Based on these
20 findings, the Columbia Accident
Amplified (V·s)2 Investigation Board made
0 recommendations to develop and validate
0 0.25 0.50 0.75 1.00 1.25 1.50 a physics based model to estimate the
Resonance spacing (MHz) severity of damage to the thermal
(b) protection system for any future debris
impact. The investigators further
3000 recommended establishing damage
2000 thresholds that would trigger corrective
1000 action such as on-orbit inspection or
repair. Nondestructive test methods,
00 mainly ultrasonic and thermal, played a
key role in both of these efforts.
Legend
= flat
= composite

522 Ultrasonic Testing

To achieve the goals requested by the were conducted over speeds from 91 to
Columbia Accident Investigation Board, 732 m·s–1 (300 to 2400 ft·s–1), and test
researchers at the National Aeronautics results were used to validate finite
and Space Administration first developed element analysis predictions, to refine the
a dynamic finite element analysis model model and to determine the threshold of
to predict the amount of damage resulting damage. Results from ultrasonic and
from a debris impact with reinforced thermographic tests were used to measure
carbon-carbon. The two main debris types the damage in the panel because visual
considered were external tank foam testing alone was insufficient.
insulation and ice. To validate this
analytical model, a series of ballistic Each reinforced carbon-carbon panel
impact tests were conducted on used in the validation study was inspected
150 × 150 × 6 mm (6 × 6 in. × 0.25 in.) before and after ballistic impact using
thick reinforced carbon-carbon panels both through-transmission ultrasonic
using projectiles made from representative C-scan and pulsed thermography.
debris types as shown in Fig. 43.19 Impacts Ultrasonic tests were conducted using
water immersion and 1 MHz transducers
FIGURE 43. Optical image of reinforced carbon-carbon panels for sending and receiving the ultrasonic
following ballistic impact: (a) impacted with ice projectile at signal. Because of scattering and
91 m·s–1 (300 ft·s–1) with no apparent visible damage; attenuation in the woven reinforced
(b) impacted with foam at 732 m·s–1 (2400 ft·s–1), with carbon-carbon, higher frequencies were
visible damage on both the front and rear of the panel.19 found to be ineffective. Nondestructive
testing before and after impact was
(a) conducted to provide baseline images for
comparison and to identify material
25 mm anomalies in the as received state.
(1 in.) Thermographic testing of the panels used
xenon flash lamp excitation and a focal
(b) plane array infrared camera for image
acquisition. Through-transmission
ultrasonic testing and thermography
turned out to be complementary methods
in terms of the types of discontinuities
they are most sensitive to (perpendicular
to the excitation direction) and were used
together extensively to validate findings.

Figure 44 shows images of the baseline
and postimpact ultrasonic results for a
reinforced carbon-carbon panel subjected
to foam impact at 643 m·s–1 (2109 ft·s–1).19
In addition, postimpact thermography
results are also shown for comparison.
The ultrasonic images shown represent
the peak amplitude of the
through-transmission ultrasonic signal
received during the tests. It should be
noted that the dark, damaged region in
the image cannot be identified visually
but is easily seen by using ultrasound and
thermography together

This project provided the National
Aeronautics and Space Administration
some very useful data for the Columbia
Accident Investigation Board and the
Shuttle Program. Ultrasonic imaging,
coupled with pulsed thermography,
enabled researchers to determine impact
energies above which damage was
noticeable and below which no detectable
change resulted in the nondestructive
testing signals. Measurement of this
threshold velocity and feedback into the
impact model were a vital part of the
effort to return the space shuttle to flight.

25 mm
(1 in.)

Aerospace Applications of Ultrasonic Testing 523

Acoustoultrasonic result, it is desirable to monitor and assess
Assessment of Creep the current condition of these
Damage in Nickel Alloy components, particularly for early damage
detection as well as to prevent failure
Metallic turbine components in operation during operation. Acoustoultrasonics was
for long periods of time may experience used to monitor the state of the material
creep behavior due to elevated at various percentages of creep life in a
temperatures and excessive stresses. As a precipitation hardenable, nickel based
FIGURE 44. Nondestructive test results for reinforced alloy.20 For polycrystalline metals, the
carbon-carbon panel subjected to ballistic foam impact: permanent deformations associated with
(a) baseline through-transmission ultrasound; (b) postimpact creep are due to various stress assisted,
ultrasonic results revealing damage zones; (c) pulsed thermally activated micromechanisms.
thermography to confirm ultrasonic results.19 These include the generation and
(a) mobilizations of dislocations, escape of
dislocations from their glide planes, grain
(b) boundary sliding and diffusion of atoms
and point discontinuities.19 These
(c) mechanisms characterize the majority of
the creep life as defined by the primary
and secondary portions of a typical strain
versus time creep curve. As failure
approaches, an increase in strain rate is
noticed and is identified as the tertiary
portion of the creep curve. The increase in
the strain rate is assumed to occur because
of additional mechanisms such as the
growth and accumulation of cavities
along grain boundaries. The growing
cavities reduce the effective area of the
material with the final result being creep
rupture.21,22

Acoustoultrasonic testing was used as a
nondestructive testing technique for
disclosing distributed material changes
that occur before the local damage
detected by other means of testing. A
dual-element (25 MHz broad band,
50 MHz acquisition rate, 10 µs window),
contact ultrasonic transducer, in a
send/receive arrangement, was used for
investigating a region of interest. The
intent of the technique was to correlate
certain parameters in the detected
waveform to characteristics of the
material being interrogated. In these
experiments, the parameter of interest was
the attenuation due to internal damping.
The parameters used to indirectly quantify
the attenuation were the ultrasonic decay
rate as well as various moments of the
frequency power spectrum.23,24

Creep rupture tests were conducted on
specimens of Unified Numbering System
N07520, a precipitation hardenable,
nickel base alloy. The specimens were
designed with a multiple-step gage region
to simulate areas of different remaining
life, as shown in Fig. 45c.

The overall specimen length was
150 mm (6 in.) — each step, including the
grip region, having a length of 25 mm
(1 in.). The thickness was 4 mm (0.16 in.).
The four gage widths were designed to
correspond to 12.5, 25, 50 and
100 percent of used life with fracture in
the cross section of smallest area. Hence,
the four gage widths ranged from 22.9 to
30 mm (0.9 to 1.2 in.). Additional

524 Ultrasonic Testing

specimens were stressed at various levelsSpectral energy (arbitrary unit) acoustoultrasonic results as a function of
at 732 °C (1350 °F) and at 816 °C position (that is, used up creep life)
(1500 °F). Specimens were also thermallyAttenuation parameter (arbitrary unit) compared to an image of the failed
aged without load to compare specimen.
acoustoultrasonic results of creep tested
versus thermally aged conditioning. The first waveform parameter used was
Figure 45 shows the obtained the diffuse decay rate, which quantifies
the internal damping of the vibration.
FIGURE 45. Acoustoultrasonic parameters as functions of used Damping was measured through
up creep life, concerning specimen B2 tested at 732 °C determination of the volume averaged
(1350 °F): (a) M0 and decay rate; (b) fc, f0 and fp; decay rate as function of frequency and
(c) photograph of fractured specimen. Four gage areas on time. Then the mean square value of the
test object represent (from left) 100, 50, 25 and power spectral density M0 measured the
12.5 percent used up creep life. Scale above specimen shows overall energy of the received waveform.
position: specimen photo is approximately aligned with Lastly, the shape parameters fc, f0 and fp
positions in graphs.20 represented centroid of power spectrum,
(a) 3 f0 the frequency of mean crossings with
positive slopes and fp the frequency of
2 maxima in the time domain.

1 The data (Fig. 45) show that there was
an overall increase in the attenuation.
0 This increase was indicated by the
0 2 4 6 8 10 12 14 16 18 20 increase in the diffuse field decay rate and
Position number the decrease in the ultrasonic wave energy
M0 as the failure region (100 percent of
(b) 1.10 life used) was approached from the right.
Another observed effect was the
1.08 preferential attenuation of the low
frequency signal components, which
1.06 caused the shape parameters fc, f0 and fp
to increase with damage. A comparison of
1.04 the aged specimens (that is, elevated
temperatures but no stress) to the creep
1.02 tested specimens (that is, stress as well as
elevated temperatures) showed that the
1.00 large majority of the changes in the
acoustoultrasonic parameters were driven
0.98 by the creep mechanisms. As a final note,
0 2 4 6 8 10 12 14 16 18 20 the acoustoultrasonic approach was also
Position number applied to characterize the above
discussed flywheel materials and showed
success in monitoring degradation due to
spin testing.

(c) Position 2 4 6 8 10 12 14 16 18

100 50 25 12.5
Used up life (percent)

Legend

= M0
= decay rate
= fc
= f0
= fp

Aerospace Applications of Ultrasonic Testing 525

References

1. Hagemaier, D.J. “Ultrasonic 9. Anastasi, R.F. and E.I. Madaras.
Maintenance Testing of Aircraft “Investigating the Use of Ultrasonic
Structures.” Nondestructive Testing Guided Waves for Aging Wire
Handbook, second edition. Columbus, Insulation.” Fifth Joint NASA/FAA/DoD
OH: American Society for Conference on Aging Aircraft [Orlando,
Nondestructive Testing (1991): FL, September 2001]. NASA Langley,
p 635-648. VA: National Aeronautics and Space
Administration (2001).
2. Chapman, G. “Ultrasonic Testing of
Automotive Composites.” 10. Anastasi, R.F. and E.I. Madaras.
Nondestructive Testing Handbook, “Application of Ultrasonic Guided
second edition. Columbus, OH: Waves for Aging Wire Insulation
American Society for Nondestructive Assessment.” Material Evaluation.
Testing (1991): p 649-667. Vol. 63, No. 2. Columbus, OH:
American Society for Nondestructive
3. Seale, M.D., B.T. Smith and Testing (February 2005): p 143-147.
W.H. Prosser. “Lamb Wave Assessment
of Fatigue and Thermal Damage in 11. Madaras, E.I., T.W. Kohl and
Composites.” Journal of the Acoustical W.P. Rogers. “Measurement and
Society of America. Vol. 103, No. 55. Modeling of Dispersive Pulse
Melville, NY: American Institute of Propagation in Drawn Wire
Physics, for the Acoustical Society of Waveguides.” Journal of the Acoustical
America (May 1998): p 2416-2424. Society of America. Vol. 97, No. 1.
Melville, NY: American Institute of
4. Balasubramaniam, K. and J.L. Rose. Physics, for the Acoustical Society of
“Physically Based Dispersion Curve America (January 1995): p 252-261.
Feature Analysis in the NDE of
Composites.” Research in Nondestructive 12. MIL-W-22759/87, Wire, Electrical,
Evaluation. Vol. 3. Columbus, OH: Polytetrafluoroethylene/Polymide
American Society for Nondestructive Insulated, Normal Weight, Nickel Coated
Testing (1991): p 41-67. Copper Conductor, 200 Degrees,
600-Volts, for Crimp Applications.
5. Caslini, M., C. Zanotti and Arlington, VA: Defense Information
T.K. O’Brien. “Fracture Mechanics of Systems Agency (1994).
Matrix Cracking and Delamination in
Glass/Epoxy Laminates.” Journal of 13. MIL-W-22759/34, Wire, Electrical,
Composites Technology and Research. Fluoropolymer-Insulated, Crosslinked
Vol. 9. Amsterdam, Netherlands: Modified ETFE, Normal Weight,
Elsevier (Winter 1987): p 121-130. Tin-Coated Copper, 150 Deg. C,
600 Volt. Arlington, VA: Defense
6. Seale, M.D. and E.I. Madaras. Information Systems Agency (1994).
NASA/TM-2000-210628, Use of Guided
Acoustic Waves to Assess the Effects of 14. Seale, M.D., B.T. Smith, W.H. Prosser
Thermal Mechanical Cycling on and J.N. Zalameda. “Lamb Wave
Composite Stiffness. Washington, DC: Assessment of Fiber Volume Fraction
National Aeronautics and Space in Composites.” Journal of the
Administration (2000). Acoustical Society of America. Vol. 104,
No. 3. Melville, NY: American Institute
7. Seale, M.D. and E.I. Madaras. “Lamb of Physics, for the Acoustical Society
Wave Stiffness Characterization of of America (1998): p 1399-1403.
Composites Undergoing
Thermal-Mechanical Aging.” NASA 15. Harmon, L.M. and G.Y. Baaklini.
1998 Technical Documentation CD. “Ultrasonic Resonance Spectroscopy of
New Port Richey, FL: Integrated Composite Rings for Flywheel Rotors.”
Publishing (1998). Nondestructive Evaluation of Materials
and Composites V. SPIE Proceedings,
8. Seale, M.D. and E.I. Madaras. “Lamb Vol. 4336. Bellingham, WA: SPIE
Wave Characterization of the Effects of — International Society for Optical
Long Term Thermal-Mechanical Engineering (2001): p 24-35.
Cycling on Composite Stiffness.”
Journal of the Acoustical Society of 16. Fitting, D.W. and L. Adler. Ultrasonic
America. Vol. 106, No. 3. Melville, NY: Spectral Analysis for Nondestructive
American Institute of Physics, for the Evaluation. New York, NY: Plenum
Acoustical Society of America (1981).
(September 1999): p 1346-1352.

526 Ultrasonic Testing

17. Krautkrämer, J. and H. Krautkrämer. Bossi, R.H., K.R. Housen and
Ultrasonic Testing of Materials, W.B. Shepherd. “Application of Stress
fourth edition. New York, NY: Waves to Bond Inspection.” SAMPE
Springer (1990). 2004 [Long Beach, CA, May 2004].
Covina, CA: Society for the
18. Tucker, J.R. United States Patent Advancement of Materials and Process
5 591 913, Apparatus and Method for Engineering (2004).
Ultrasonic Spectroscopy Testing of
Materials (January 1997). Brasche, L. “The Use of Nondestructive
Inspection in Jet Engine Applications.”
19. Columbia Accident Investigation Board Materials Evaluation. Vol. 60, No. 7.
(CAIB) Mishap Report. Vol. 1. Columbus, OH: American Society for
Washington, DC: United States Nondestructive Testing (July 2002):
Government Printing Office, for the p 853-856.
National Aeronautics and Space
Administration (August 2003). Bratt, M.J. and V.I.E. Wiegand. “Detection
of Flaws in Jet Engine Parts by
20. Gyekenyesi, A.L., H.E. Kautz and Ultrasonics.” Nondestructive Testing.
W. Cao. NASA/TM-2001-210988, Vol. 13, No. 5. Columbus, OH:
Damage Assessment of Creep Tested and American Society for Nondestructive
Thermally Aged Udimet 520 Using Testing (September-October 1955):
Acousto-Ultrasonics. Washington, DC: p 45-47, 59.
National Aeronautics and Space
Administration (2001). Davis [Witmann], R.S. “Nondestructive
Inspection of Jet Engine Assemblies
21. Shames, I.H. and F.A. Cozzarelli. Elastic and Components.” Materials
and Inelastic Stress Analysis, revised Evaluation. Vol. 48, No. 12. Columbus,
printing. Philadelphia, PA: Taylor and OH: American Society for
Francis (1997). Nondestructive Testing (December
1990): p 1485-1486, 1488-1490.
22. Gittus, J.H. Creep, Viscoelasticity, and
Creep Fracture in Solids. London, Ortolano, R.J., B. Kotteakos and
United Kingdom: Applied Science M.L. Luttrell. “Automated Ultrasonic
Publishers (1975). Inspection of Turbine Blade Tenons.”
Materials Evaluation. Vol. 51, No. 5.
23. Tiwari, A. NASA CRT-198374, Real Columbus, OH: American Society for
Time Acousto-Ultrasonics NDE Technique Nondestructive Testing (May 1993):
Monitoring Damage in Ceramic p 568-571.
Composites under Dynamic Loads.
Washington, DC: National Aeronautics Papadakis, E.P. “Justification for Engine
and Space Administration (1995). Parts Testing in Manufacture.”
Materials Evaluation. Columbus, OH:
24. Lot, L.A. and D.C. Kunerth. “NDE of American Society for Nondestructive
Fiber-Matrix Interface Bonds and Testing (December 2002): p 1399-1400.
Material Damage in Ceramic/Ceramic
Composites.” Conference on Vaerman, J. and H. Forsans. AD0196517,
Nondestructive Evaluation of Modern Ultrasonic Testing of Ball Bearings.
Ceramics [Columbus, OH, July 1990]. Fort Belvoir, VA: Defense Technical
Columbus, OH: American Society for Information Center (1971).
Nondestructive Testing (1990):
p 135-139.

Bibliography: Ultrasonic
Testing of Jet Engines

“NDT of Jet Engines: An Industry Survey.”
Materials Evaluation. Columbus, OH:
American Society for Nondestructive
Testing. Part One, Vol. 44, No. 13
(December 1986): p 1477-1478,
1480-1482, 1484-1485. Part Two,
Vol. 45, No. 1 (January-February 1987):
p 26, 28, 30-34.

Bossi, R.H., K.R. Housen and
W.B. Shepherd. “Using Shock Loads to
Measure Bonded Strength.” Materials
Evaluation. Vol. 60, No. 11. Columbus,
OH: American Society for
Nondestructive Testing
(November 2002): p 1333-1338.

Aerospace Applications of Ultrasonic Testing 527



15

CHAPTER

Special Applications of
Ultrasonic Testing

David S. Forsyth, Texas Research Institute, Austin,
Texas (Part 1)
Lawrence W. Kessler, Sonoscan, Elk Grove Village,
Illinois (Part 4)
Francesco Lanza di Scalea, University of California —
San Diego, La Jolla, California (Part 2)
Roger D. Wallace (Part 3)

PART 1. Reliability of Nondestructive Testing

Nondestructive testing has critical roles in curves were constructed by using moving
process control and in inspection of safety averages, or averaging the response of all
critical physical assets such as aircraft, cracks in an interval and manually fitting
pressure vessels, nuclear reactor a curve through these points.1,2 The
components and pipelines, so the probability of detection for each interval
measurement of the performance of on the moving average was used to
nondestructive testing has become calculate confidence intervals.
important. It is no longer sufficient in
many cases simply to assume that an Early work was sponsored by the
inspection is a perfect process of National Aeronautics and Space
unbounded capability; rather, it is Administration, and the United States Air
imperative to know the probability of Force was moving along a similar path. In
finding discontinuities of interest. This is both cases, the motivation was the use of
usually called the probability of detection. damage tolerance philosophies in design
Probability of detection is described below and maintenance: it would be assumed
in terms of cracks but has been applied to that parts contained discontinuities when
other discontinuities such as corrosion they left manufacturing, and these
loss, impact damage or delamination. discontinuities would grow as cracks
Figure 1 shows a representative curve used under the expected operation of the part.
for analysis of probability of detection for The other technology advance that
a given application. enabled this approach was the maturing
of fracture mechanics models for crack
The most commonly used measure of growth in metals.
nondestructive testing performance is
probability of detection. Other measures The National Aeronautics and Space
of nondestructive test performance are Administration would adopt a damage
considered below. The exact definition of tolerance approach for the space shuttle
probability of detection, and statistical program. The United States Air Force
methods used to calculate probability of adopted fracture mechanics and damage
detection, have evolved over time. In the tolerance in response to structural failures
1970s, some probability of detection in jet aircraft that were virtually new.3 The
Air Force released MIL-STD-1530 in 1972

FIGURE 1. Probability of detection curve from transverse ultrasonic testing of welds for
longitudinal cracks in flush grounded, gas tungsten arc welds made of Unified Numbering
System A92219, heat treatable, wrought aluminum alloy. Three operators detected 291
indications in 345 opportunities for a combined probability of detection of 90 percent
at 7.54 mm (0.030 in.).6

Probability of detection (percent) 100
90
80 1.3 2.5 3.8 5
70 (0.05) (0.10) (0.15) (0.20)
60
50
40
30
20
10
0
0

Actual crack depth, mm (in.)

530 Ultrasonic Testing

to revise their Aircraft Structural IntegrityDiscontinuity size (relative scale)1. A threshold was needed for
Program to include damage tolerance.4 In deterministic fracture mechanics.
1974, MIL-A-83444 detailed the damage Damage tolerance approaches of the
tolerance requirements for airplane safety United States Air Force and the
of flight structure.5 The essence of these National Aeronautics and Space
approaches is illustrated in Fig. 1: the time Administration were such that a crack
for a severe initial cracklike discontinuity growth analysis was started from a
a0 (often called the rogue discontinuity) to rogue discontinuity, and the time to
grow a crack until failure is estimated.6 crack instability was calculated from
Inspections are scheduled to have one or this starting point. Inspections are
more opportunities in this time to detect used to detect the crack before
the crack and to repair or replace the part instability. To reduce risk, one would
before failure. prefer 100 percent probability of
detection at crack sizes greater than
The approach to damage tolerance the rogue size. Because 100 percent
illustrated in Fig. 2 shows a single crack probability of detection is not likely at
size aNDT, defining a threshold above any size of crack, a lesser threshold
which all cracks would be found. Given a had to be chosen.
probability of detection curve, the simple
answer for picking this size for 2. Binomial statistics were in use to
nondestructive testing would be at the calculate probability of detection and
point where probability of detection is confidence bounds.
100 percent. However, the binomial
statistics used to estimate probability of 3. Using binomial statistics to
detection and confidence bounds in the demonstrate high probability of
early works required very large numbers detection numbers requires many
of trials to obtain high values of tests: 29 successes out of 29 trials is
probability of detection and high required for 95 percent confidence
statistical confidence. In the mid 1970s, that probability of detection is
the threshold for probability of detection 90 percent. Many more trials are
was chosen to be the 95 percent lower required to increase this confidence.
confidence bound at the probability of
detection of 90 percent. This is often 4. A confidence level of 90/95 percent
referred to as a90/95 or simply the 90/95 was consistent with the
value. The acceptance of the 90/95 USAF MIL-HDBK-5 B basis materials
criterion as a threshold for inspection allowables, and thus familiar to the
capability was due to four factors. people working with the damage
tolerance approaches.7,8
FIGURE 2. Damage tolerance approach to
structural integrity used by United States Air A published standard or guideline on
Force and National Aeronautics and Space how to perform an experiment to
Administration. estimate a probability of detection curve
was not available immediately. Within the
Unstable growth American Society for Nondestructive
until failure Testing, an effort was initiated by
W.H. Lewis to develop a recommended
acrit practice document. Although the final
version was finished in 1976, publication
aNDT Inspection did not occur until 1982.9,10 This
aO interval recommended practice was based on
binomial methods. In the meantime, the
Usage or time (relative scale) ASM Handbook included a section
describing possible binomial techniques
Legend for estimating the probability of
detection. It was acknowledged that the
acrit = discontinuity of critical size ASM document was developed in concert
aNDT = discontinuity at moment of test with the ASNT authors.11

a0 = discontinuity before propagation The most common technique of
estimating probability of detection curves
from inspection data are based on USAF
MIL-HDBK-1823, released in 1992 as a
document of the North Atlantic Treaty
Organization.12,13 The Federal Aviation
Administration has also published ways to
conduct a probability of detection study,
based on the statistical methods described
in MIL-HDBK-1823.14 These statistical
methods were developed by Berens and
Hovey at the University of Dayton in the
1980s.15,16 Two principles underpin the
approach.

Special Applications of Ultrasonic Testing 531

1. Different cracks of the same size willTest signal (relative logarithmic scale)There are a number of useful general
not be detected with the same statements that can be made.
probability of detection.
1.The process of probability of detection
2. In many cases the relationship estimation requires a number of
between the signal from the inspections to be performed.
nondestructive testing system and the
size of the crack are linearly related if 2. The predefined inspection system
plotted on a logarithmic chart. being assessed should be complete,
including representative equipment,
These principles (Fig. 3) give rise to the procedures, inspectors and target parts.
use of the log normal equation to fit the
relationship of crack size and probability 3. Parts used should have discontinuities
of detection.12 that represent the discontinuities of
interest, or a means to assess the
Nondestructive Test difference between the two: for
Measures in Practice example, using machined notches or
flat bottomed holes can provide a
The techniques of probability of detection useful measure of capability but
have been continually refined to reduce should not be assumed to represent
unpredicted variations in reliability.12,17-25 cracks or other natural discontinuities.

Estimation of Probability of 4. The inspection procedure and
Detection environment should be typical of the
deployed environment. Human factors
Both the United States Air Force and the studies have shown that the
Federal Aviation Administration have relationship of factors — such as
published guidelines that describe in environment (such as lighting and
detail the experiments required to temperature), training, experience,
estimate the probability of detection of an motivation and others — to inspection
inspection system. These documents are performance is not simple and often
in the public domain and can be obtained not intuitive.
for free from the government agencies as
well as the Department of Defense’s To estimate the probability of
Advanced Materials, Manufacturing, and detection, it is key to understand the
Testing Information Analysis Center. The physical parameters that may affect the
United States Air Force and National response to a discontinuity. Parts with
Aeronautics and Space Administration discontinuities arising from actual service
have updated key handbooks and are optimal but often not available.
guidelines.12,17-19 Requirements for other Therefore every reasonable effort should
industries appear in their standards, such be made either to replicate the service
as the ASME Boiler and Pressure Vessel discontinuities as closely as possible or to
Code.26 The reference to be used in a carefully define the difference between
particular application should be based on the probability of detection experiment
the regulating body that requires the and inservice conditions.
assessment.
False Call Rates
FIGURE 3. Log linear relationship observed between test
signal magnitude and crack size. The number of false positives or false calls
made during inspections is very important
Probability of detection = shaded area in practical applications because of the
indicated proportion detected at cost implication and the trust in the
that crack size nondestructive test procedure. The
detection threshold is often lowered to
Hit Decision reduce the quantity of dangerous false
Miss threshold negatives. However, lowering the
detection threshold to capture smaller
Discontinuity size (relative scale) discontinuities increases false calls. The
appropriate level will depend on
individual situations. Key factors to
consider include the cost of false calls and
the cost of passed discontinuities, that is,
false calls.

The tradeoff between false calls and the
detection of discontinuities of a single size
can be visualized using the receiver
operating characteristic curve.

Receiver Operating Characteristic

The receiver operating characteristic is a
measure in a system of the separation of
the noise signals from the signals due to

532 Ultrasonic Testing

the item of interest (for example, cracks). Correct detectionsdistinguish between the two. This line is
This measure originated in the evaluation (percent)at a 0.79 rad (45 deg) angle on the receiver
of radar signals in World War Two. It operating characteristic plot (Fig. 5a).
shows how the decision threshold affects
the tradeoff between false calls and It must be noted that the receiver
detection of discontinuities at a single size operating characteristic curve plot applies
(Fig. 4). only to a single discontinuity size and
thus provides less information than the
To plot the receiver operating probability of detection curve.
characteristic curve requires making
measurements of the signal distributions Common Mistakes in
from noise and from discontinuities, as Estimation of
shown in Fig. 4. The receiver operating Nondestructive Test
characteristic curve is then plotted by Reliability
changing the decision threshold and
plotting the fraction of detected The means used to assess nondestructive
discontinuities as a function of the testing reliability have evolved
fraction of false calls. Starting with a significantly since 1970 and the statistical
decision threshold at positive infinity, one basis of assessment is often at a level
would find zero discontinuities and have beyond the statistical knowledge of most
zero false calls. As the decision threshold nondestructive test practitioners. A few
goes lower, the detection rate goes up errors in the estimation of nondestructive
with few false calls. As you continue to testing reliability are common and must
lower the threshold, the detection rate be avoided.
will increase slowly and the false call rate
will increase rapidly. FIGURE 5. Various receiver operating
characteristic curves for tests of different
A set of receiver operating capability: (a) best and worst case;
characteristic curves is shown below. The (b) realistic case.
perfect curve is one where the noise signal
and discontinuity signal distributions do (a) 100
not overlap, as shown in Fig. 4a. This
makes a receiver operating characteristic Heavy line is
curve as shown in Fig. 5a. A more realistic best case
curve is the one shown in Fig. 5b,
corresponding to Fig. 4b. The worst Dashed line is
possible receiver operating characteristic worst case
curve is when the noise and discontinuity
signal distributions are identical and 0
completely overlap: the inspection cannot 0 100

FIGURE 4. Distributions of signals from noise False calls
and discontinuities: (a) ideal case of perfect (percent)
detection, with zero false calls; (b) realistic
distribution, in which a higher detection rate (b) 100
also gives higher rate of false calls. Distance
between distributions is measure of Realistic case
capability of test and can be represented in Correct detections
receiver operating characteristic curve. (percent)

(a) Signal from
Probability
(relative scale) Noise discontinuities of
particular size

Amplitude (relative scale)Probability 0
(b)(relative scale) 0 100

Noise Signal False calls
(percent)

Amplitude (relative scale)

Special Applications of Ultrasonic Testing 533

1. Often the customer for nondestructive application. The most common technique
testing reliability information is the for assessing nondestructive testing
design engineer, who may have very performance is the probability of
little knowledge of nondestructive detection curve. MIL-HDBK-1823 provides
testing. It is crucial to know what the detailed information on the experiments
end use of the information will be, in required to estimate the probability of
order to provide the best estimate of detection for a specific inspection
reliability. Often designers will ask for technique and application.12
the smallest discontinuity that can be
found. The nondestructive test
practitioner is cautioned to avoid
simply replying the smallest
discontinuity ever found. Rather, the
question needs to be clarified so that
the designer understands the statistical
nature of nondestructive testing. This
understanding will help meet the
customer’s needs and provide the
inspector with reasonable
expectations.

2. The most common mistake in
nondestructive test reliability is the
use of pre-existing probability of
detection information beyond its
scope of applicability. The cost and
time required to do a study as
described in the USAF MIL-HDBK-1823
may be prohibitive, and the
temptation to apply data from
previous studies strong. However, old
data can be used only with careful
engineering judgment and an
understanding of all the factors that
affect nondestructive test capability in
a particular situation. Residual stresses
are only one example of factors that
are often undocumented but can have
an enormous impact on the
detectability of cracks.

3. Another common mistake is to assume
that repeated inspections will increase
probability of detection. Recognizing
the statistical nature of the inspection
process, many people have then
assumed that inspections are
independent and therefore repeating
the inspections will improve the
probability of detection. It must be
understood that the concept of
statistical independence applies to
random events like a coin toss, and an
inspection is not a random event.

4. Finally, the use of proper technique
documentation, calibration procedures
and training are essential to ensure
that the carefully controlled data
acquired in a probability of detection
study can be reproduced in service.

Summary

The measurement of the performance of
nondestructive testing is important for
maintaining safety, for scheduling
inspections and maintenance and for
technique selection for a particular

534 Ultrasonic Testing

PART 2. Ultrasonic Testing in the Railroad
Industry

Introduction Ultrasonic Testing of Rail
Track
Railroad components such as tracks, axles
and wheels are usually designed for an Motivation
infinite life performance based on the
fatigue endurance limit of the material. Rail track failures are a cause of great
The need to design for infinite life arises concern to railroad operators and owners.
from the fact that these components far The United States Federal Railroad
exceed the 106 to 107 cycles generated in Administration maintains updated safety
a common fatigue test, which measures statistics on train accidents and associated
the stress versus the number of cycles to causes. According to these records, in the
failure. For example, for a typical service decade from 1992 to 2002, track
duty of 400 000 km per year for high discontinuities causing accidents in the
speed railroad systems, the number of United States were responsible for 2600
cycles that axles and wheels experience is derailments and 441 000 000 dollars in
on the order of 2 × 108.27 Unfortunately, reportable damage cost. Unfortunately,
there is no deterministic guarantee that these numbers get worse with aging
the goal of infinite life will be met, for two infrastructure and heavier tonnages.
reasons.
Rail Track Discontinuities
1. Loads experienced in practice differ
from hypothetical loads considered in Discontinuities in rail tracks are classified
the design phases. according to their orientation with respect
to the major geometrical planes of the
2. Discontinuities can be originated in track (Fig. 6). The major distinction is
the manufacturing process or between longitudinal discontinuities
generated by unexpected service preferentially oriented in the horizontal
events, such as debris impact. plane or in the vertical plane of Fig. 6,
and transverse discontinuities (the most
Thus railroad components are maintained severe ones) preferentially oriented in the
with a damage tolerance concept.28 transverse plane of Fig. 6.30 Figure 7
schematizes the most typical track
According to this philosophy, it is discontinuities, including longitudinal
accepted that cracks grow in tracks, axles discontinuities (horizontal split heads,
and wheels provided that these cracks do vertical split heads and vertical split
not reach their critical size during the webs), transverse discontinuities
component’s life. Nondestructive testing
becomes a critical link of the damage FIGURE 6. Geometrical planes of railroad
tolerant philosophy. Nondestructive track.
testing allows engineers to make decisions
on remaining useful life as well as Vertical plane
schedule the frequency of subsequent
inspections. Without nondestructive Transverse
testing, structural failures caused by plane
growing discontinuities may have
catastrophic consequences with potential Horizontal
for loss of human life and large direct cost plane
associated to repairs and indirect cost
associated to traffic disruptions.

Ultrasonic testing is routinely used for
the nondestructive testing of tracks, axles
and wheels.29 Ultrasonic testing is often
used with other nondestructive test
methods, such as magnetic induction
testing in tracks and magnetic particle
testing in axles.

Special Applications of Ultrasonic Testing 535

(transverse fissures, detail fractures and the track head and web, with only limited
compound fractures) and other types of coverage of the track base.
discontinuities internal to the track (weld
discontinuities, head web separation, split The frequency of rail track inspection
webs, bolt hole cracks and broken bases). varies from country to country. In the
Discontinuities at the surface of the track United States, the Federal Railroad
include shelling, head checks, head Administration mandates that inspections
squats, engine burn fractures and for track discontinuities be made at least
corrugated track. once every 40.6 × 109 kg (40 × 106 long
tons) or once a year, whichever interval is
Some of these discontinuities originate shorter, for tracks over which passenger
during manufacturing, such as the trains operate. For tracks over which
transverse fissure from hydrogen nuclei. passenger trains do not operate, the
Others originate in service because of the inspection must be carried out every
wheel-to-rail fatigue. Rolling contact 30.5 × 109 kg (30 × 106 long tons) or once
fatigue discontinuities initiate at the a year, whichever interval is longer.
surface of the rail head as horizontal head
checks or squats. A few millimeters from Tracks testing are tested ultrasonically
the surface, they can turn to a transverse with longitudinal or transverse
crack and develop a detail fracture. transducers in a pulse echo mode or a
Because rolling contact fatigue pitch catch mode. The transducers are in
discontinuities tend to form almost wheels filled with water or a water
continuously in a given track, they are of solution that runs over the surface of the
greater concern than other track track. Sleds, rather than wheels, can also
discontinuities. be used to host several transducers in a
smaller area. Specialized test cars typically
Practice of Rail Track Ultrasonic perform the inspection. Details on
Testing common transducer configurations can be
found in various references.31-33 The most
In addition to electromagnetic tests, common configuration (Fig. 8a) uses
ultrasonic testing has been widely used for transducer orientations to generate
discontinuity detection in rail tracks since ultrasonic beams propagating at normal
the 1960s. Normally, the ultrasonic test is incidence (0 rad, 0 deg) and at 1.22 rad
targeted to detecting discontinuities in (70 deg) from the normal to the rail
surface. The 0 rad (0 deg) probe targets
FIGURE 7. Typical discontinuities in rail: (a) in the vertical horizontal cracks while the 1.22 rad
plane; (b) in the transverse plane. (70 deg) probe targets the transverse cracks
that tend to grow in a 0.35 rad (20 deg)
(a) Vertical Shelling direction from the transverse plane. A
checks 0.65 rad (37 deg) or 0.79 rad (45 deg)
Horizontal split or head probe is also often used in addition to the
split checks previous two orientations to target other
Vertical web Separation Bolt hole discontinuities, including bolt hole cracks
Weld head split head of head cracks and weld discontinuities (Fig. 8a). To
discontinuity from web target vertical discontinuities, complete
search units also host side looking
Broken transducers, generating beams in the
base transverse plane, rather than in the
vertical plane of the track, typically at
Weld Detail Transverse 0.79 rad (45 deg) orientations (Fig. 8b).
fracture fissure
(b) Wheels or sleds are often used in
tandem to provide complete coverage.
Using tandem configurations also allows
adding pitch catch testing capability to
the pulse echo capability of a single
wheel.

The standard transducer of a rail track
ultrasonic test unit operates at 2.25 MHz;
3 MHz transducers can also be used.

The test car inspections of rail tracks
are followed by manual scanning to
confirm the presence of a discontinuity
and to size it. Generally, both normal
beam transducers and angle beam
transducers (with conventional acrylic
wedges) can be used in manual scanning.
As mentioned above, normal beam
transducers target horizontal
discontinuities whereas angle beam
transducers target transverse

536 Ultrasonic Testing

discontinuities. Discontinuities are sized transverse plane to enhance the
by simply scanning the normal beam interaction with certain types of
transducer for horizontal discontinuities discontinuities.
and by using a conventional technique
while scanning the angle beam Manual ultrasonic testing is required in
transducers. In addition to conventional the United States to verify any
normal beam or angle beam discontinuity indication from the test car
configurations, more complex transducer search units — the stop and confirm test
arrangements can be used in manual mode. Europe operates, instead, in a
scanning. Skewed transducer orientations nonstop test mode, whereby indications
can be used, for example, for the are verified by manual ultrasonic testing
detection of detail fractures.31 In skewed only days or weeks after the test car has
arrangements, the ultrasonic wave passed.34 Modern test cars can operate
propagates along planes inclined with fast, over 64 km·h–1 (40 mi·h–1), although
respect to both the vertical and the stopping and confirming in the United
States limits the effective testing speed to
FIGURE 8. Common transducer as slow as 15 km·h–1 (10 mi·h–1).
arrangements for rail track tests: (a) vertical
plane; (b) transverse plane. One serious drawback of current wheel
or sled transducer arrangements is the fact
(a) 0 deg transducer that internal transverse discontinuities in
the track may be missed by the inspection
1.2 rad (70 deg) 0.8 rad (45 deg) in the presence of shallow horizontal
transducer transducer cracks, such as head checks and shelling.
Another drawback is that the high
1.2 rad probing frequency (2.25 MHz to 3 MHz)
(70 deg) of the transducers is sometimes not
effective in penetrating aluminothermic
0.8 rad welds due to their grain structure, coarser
(45 deg) than that of the surrounding steel.34,35

(b) Side looking Other Techniques of Rail Track
Ultrasonic Testing
0.8 rad (45 deg)
transducer To avoid the drawbacks of contact
between the test probes and the test rail,
0.8 rad noncontact techniques for ultrasonic
(45 deg) testing of rail tracks also exist. These
include techniques based on
electromagnetic acoustic transducers,36 air
coupled transducers,37,38 and
combinations of lasers and air coupled
transducers.39-41 These noncontact
approaches are particularly effective when
generating ultrasonic guided waves along
the rail running direction. Ultrasonic
guided waves have been used to inspect
rails at probing frequencies ranging from
a few tens of kilohertz to a few megahertz
(rayleigh waves). Besides providing large
inspection ranges, guided waves are
particularly sensitive to transverse
discontinuities, the most critical in rails.
More importantly, they do not suffer from
the masking effects of horizontal head
checks or shells that instead affect bulk
waves excited by wheel or sled transducer
configurations as mentioned
above.35,38,39,42,43 Discontinuities are
generally detected and sized by measuring
reflection or transmission coefficients of
the guided waves. A rayleigh wave
arrangement (laser or air coupled) that
works on a transmission discontinuity
detection mode is shown in Fig. 9. In this
system, discontinuities in the rail head
between the two sensors are detected and
sized by monitoring changes in the ratio
between the strengths of the two sensor
readings.

Special Applications of Ultrasonic Testing 537

Automatic discontinuity classification 0.5 percent, and yearly axle replacement
capabilities have also been added to rates were up to 10 percent.49
guided wave ultrasonic testing of rail
tracks through pattern recognition Axle Discontinuities
algorithms.41,44
The causes of failures and methods of
Measurement of Residual Stress in inspection in railway components,
Rail Tracks by Ultrasonic Testing including axles, has been summarized in
the literature.28 Figure 10 shows the most
Another successful application of typical discontinuities found in train
ultrasonic testing of rail tracks is the axles. These include radial fatigue cracks
measurement of longitudinal stresses that from fretting at the fitted parts (that is, at
develop from temperature gradients and the wheel seats), at the brake disk seats
can lead to buckling of the track. (for trailing axles) and at the gear seats
Acoustoelastic ultrasonic tests, based on (for driving axles); radial fatigue cracks in
the measurement of ultrasonic velocity, the transition regions between two
can be used for this purpose.45-47 principal diameters; radial fatigue cracks
Acoustoelastic techniques have proven at the journal fillet; radial fatigue cracks
effective to nondestructively measure in the free region of the axle (less
stress levels once the effects of material frequent); and discontinuities caused by
texture and temperature are removed. corrosion, for example, when protective
coatings are removed accidentally (less
Ultrasonic Testing of Axles frequent).

Motivation Fretting remains the most common
cause of fatigue cracks in axles, despite
Train accidents due to axle failures occur measures in various countries to increase
worldwide. In the three years from 2002 fretting fatigue strength.
to 2004, for example, axle failures were
the eighth largest cause of accidents Practice of Axle Ultrasonic Testing
having electrical and mechanical causes,
with almost 16 000 000 dollars in Railroad axles have been tested by
reportable damage. In the same category nondestructive testing since the 1950s.
of primary cause and during the same Ultrasonic testing is also often combined
period, the journal overheating was the with magnetic particle testing to probe
first cause of accidents, associated to specific critical surfaces of the axle. As for
reportable damage of 35 000 000 dollars. rail track inspections, the frequency of
Reported axle failures in the United axle inspections varies from country to
Kingdom amount to one or two per country. In the United Kingdom, axles are
year.48 In Japan, the number of axle subjected to ultrasonic testing every
failures was greatly reduced with the 200 days of service or 240 000 km
introduction of ultrasonic testing in 1957 (150 000 mi).27 Japanese high speed rail
and, subsequently, of angle ultrasonic axles are inspected for cracks every
testing in 1963. After 1963, nevertheless, 30 000 km (19 000 mi) regularly, every
yearly axle failures were still reported at 450 000 km (280 000 mi) in inspection of

FIGURE 9. Noncontact laser and air coupled FIGURE 10. Common discontinuities in train axles.
arrangement for rail track tests in
transmission mode for discontinuity Wheel Gear or brake disk
detection.
Wheel seat Gear or
Air coupled sensors Journal brake disk seat

Pulsed laser
source

Fatigue crack Fatigue cracks
at journal fillet in free region

Fatigue cracks
in diameter
transition
regions

Fatigue cracks in
press fitted regions

538 Ultrasonic Testing

bogies (speed reducing axle components) between 0 rad (0 deg) normal incidence
and every 900 000 km (560 000 mi) in a and 0.28 rad (16 deg) angled incidence.
general inspection.28 The normal incidence is effective for
detecting internal cracks whereas the
Most commonly, ultrasonic testing of angled incidence configurations are used
axles is performed by longitudinal to probe surface cracks in specific regions
transducers positioned at the axle free of the axle. Although manual scanning of
ends and operated in either a pulse echo the ultrasonic transducer is most common
or a pitch catch mode (Fig. 11a). for the free end inspections, more
Generally the inspections are performed sophisticated systems can be used. These
from both ends of the axle and a include either arrays of up to ten
discontinuity echo is only flagged as such transducers, opportunely multiplexed,
if it is measured in the same axial position that can substantially increase the
of the axle when testing from either end. coverage of the axle cross section, or
The transducers operate in the 1 to 3 MHz motorized systems that provide automatic
frequency range and can be arranged to scanning of the transducer while
generate beams in a given direction maintaining it in constant contact with
the axle surface through liquid couplant
FIGURE 11. Transducer arrangements for axle tests: or dry pressure.
(a) longitudinal transducer at axle free end; (b) angle beam
transducers; (c) phased array transducers. One problem of free end tests is the
(a) fact that ultrasonic echoes from the
diameter fillets must be distinguished
Transducer from discontinuity echoes. Ultrasonic
angle probes positioned at critical
(b) Angle beam transducer positions along the axle length, developed
in the 1960s, have improved the coverage
(c) Phased array transducer of press fitted areas in axles. The
conventional ultrasonic angle probes
require manual scanning of the ultrasonic
beam to cover the entire cross section of
the axle (Fig. 11b). Recently, ultrasonic
phased arrays have been proposed to
detect radial cracks in the press fitted
areas of axles without requiring manual
scanning (Fig. 11c). The phased array
technique consists of a set of several
normal ultrasonic transducers driven at
slightly different times, so that they
interfere constructively and destructively
into an angled wave. Typical ultrasonic
operating frequencies for phased array
transducers are 3 MHz. The ultrasonic
beam is swept electronically with typical
steering angles covering a span as large as
0.3 to 1.3 rad (20 to 80 deg).

The phased array technique for axle
inspection can be computerized and
synchronized with the rotation of the axle
to provide full coverage of the axle cross
section. The section is covered by
mounting phased array probes on
manipulators put in contact with the
surface of the axle through a water gap
while the axle is being rotated. Results are
displayed in terms of ultrasonic amplitude
versus axle rotational position in
cascading charts. Ultrasonic indications
(reflections) from variations in diameters,
that are circular and symmetric, appear as
vertical straight lines; ultrasonic
indications from discontinuities, instead,
appear as isolated events because most
discontinuities do not extend to the entire
circumference of the axle. One such
computerized phased array system for axle
inspection has been used by German
Railways.50

Special Applications of Ultrasonic Testing 539