ASTM E2661/E2661M-20e1

This international standard was developed in accordance with internationally recognized principles on standardization established in the Decision on Principles for the
Development of International Standards, Guides and Recommendations issued by the World Trade Organization Technical Barriers to Trade (TBT) Committee.
ϵ1
Designation: E2661/E2661M − 20
Standard Practice for
Acoustic Emission Examination of Plate-like and Flat Panel
Composite Structures Used in Aerospace Applications
ThisstandardisissuedunderthefixeddesignationE2661/E2661M;thenumberimmediatelyfollowingthedesignationindicatestheyear
of original adoption or, in the case of revision, the year of last revision. A number in parentheses indicates the year of last reapproval.
A superscript epsilon (´) indicates an editorial change since the last revision or reapproval.
ε NOTE—An editorial change was made to the References section in November 2020.
1. Scope* 1.7 This standard does not purport to address all of the
safety concerns, if any, associated with its use. It is the
1.1 This practice covers acoustic emission (AE) examina-
responsibility of the user of this standard to establish appro-
tion or monitoring of panel and plate-like composite structures
priate safety, health, and environmental practices and deter-
made entirely of fiber/polymer composites.
mine the applicability of regulatory limitations prior to use.
1.2 The AE examination detects emission sources and
1.8 This international standard was developed in accor-
locates the region(s) within the composite structure where the
dance with internationally recognized principles on standard-
emission originated. When properly developed AE-based cri-
ization established in the Decision on Principles for the
teria for the composite item are in place, the AE data can be
Development of International Standards, Guides and Recom-
used for nondestructive examination (NDE), characterization
mendations issued by the World Trade Organization Technical
of proof testing, documentation of quality control, or for
Barriers to Trade (TBT) Committee.
decisionsrelativetostructural-testterminationpriortocomple-
tion of a planned test. Other NDE methods may be used to
2. Referenced Documents
provide additional information about located damage regions.
2.1 ASTM Standards:
For additional information, see X1.1 in Appendix X1.
E543 Specification forAgencies Performing Nondestructive
1.3 This practice can be applied to aerospace composite
Testing
panelsandplate-likeelementsasapartofincominginspection,
E976 GuideforDeterminingtheReproducibilityofAcoustic
during manufacturing, after assembly, continuously (during
Emission Sensor Response
structural health monitoring), and at periodic intervals during
E1067 PracticeforAcousticEmissionExaminationofFiber-
the life of a structure.
glass Reinforced Plastic Resin (FRP) Tanks/Vessels
E1106 Test Method for Primary Calibration of Acoustic
1.4 This practice is meant for fiber orientations that include
Emission Sensors
cross-plies, angle-ply laminates, or two-dimensional woven
E1316 Terminology for Nondestructive Examinations
fabrics. This practice also applies to 3-D reinforcement (for
E1781 Practice for Secondary Calibration ofAcoustic Emis-
example, stitched, z-pinned) when the fiber content in the third
sion Sensors
direction is less than 5 % (based on the whole composite).
E2533 Guide for Nondestructive Testing of Polymer Matrix
1.5 Thispracticeisdirectedtowardcompositematerialsthat
Composites Used in Aerospace Applications
typically contain continuous high modulus greater than 20 GPa
2.2 Other Documents:
[3 Msi] fibers.
ANSI/ASNT CP-189 ASNT Standard for Qualification and
1.6 Units—The values stated in either SI units or inch-
Certification of Nondestructive Testing Personnel
pound units are to be regarded separately as standard. The
ISO 9712 Non-destructive Testing—Qualification and Cer-
values stated in each system are not necessarily exact equiva-
tification of NDT Personnel
lents; therefore, to ensure conformance with the standard, each
system shall be used independently of the other, and values
from the two systems shall not be combined.
For referenced ASTM standards, visit the ASTM website, www.astm.org, or
contact ASTM Customer Service at service@astm.org. For Annual Book of ASTM
This practice is under the jurisdiction of ASTM Committee E07 on Nonde- Standards volume information, refer to the standard’s Document Summary page on
structive Testing and is the direct responsibility of Subcommittee E07.04 on the ASTM website.
Acoustic Emission Method. AvailablefromAmericanSocietyforNondestructiveTesting(ASNT),P.O.Box
CurrenteditionapprovedJune1,2020.PublishedJuly2020.Originallyapproved 28518, 1711 Arlingate Ln., Columbus, OH 43228-0518, http://www.asnt.org.
in 2010. Last previous edition approved in 2015 as E2661/E2661M – 15. Available from International Organization for Standardization (ISO), 1, ch. de
DOI:10.1520/E2661_E2661M-20E01. la Voie-Creuse, CP 56, CH-1211 Geneva 20, Switzerland, http://www.iso.org.
*A Summary of Changes section appears at the end of this standard
Copyright © ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959. United States
ϵ1
E2661/E2661M − 20
NAS-410 NAS Certification and Qualification of Nonde- damage, growing flaws, and friction-based sources. For addi-
structive Personnel (Quality Assurance Committee) tional information, see X1.3.
SNT-TC-1A Recommended for Personnel Qualification and
4.2 This practice provides an approach to determine the
Certification of Nondestructive Testing Personnel
localregionsoforiginoftheAEsourcesandanypotentiallocal
regions of large accumulation(s) of AE sources.
3. Terminology
4.3 This practice can provide an approach to use AE-based
3.1 Definitions—See Terminology E1316 for general termi-
criteria to determine the significance of flaws.
nology applicable to this practice.
3.2 Definitions of Terms Specific to This Standard:
5. Significance and Use
3.2.1 characteristic damage state, n—transverse matrix
5.1 This AE examination is useful to detect micro-damage
cracking during the virgin loading of a composite; often
generation, accumulation, and growth of new or existing flaws.
resulting in reaching a limit of the crack density prior to
The examination is also used to detect significant existing
reaching failure.
damage from friction-based AE generated during loading or
3.2.1.1 Discussion—Resultsinareductionofstiffnessofthe
unloading of these regions. The damage mechanisms that can
composite. For additional information, see X1.2.
be detected include matrix cracking, fiber splitting, fiber
3.2.2 flat panel composite, n—any fiber reinforced compos-
breakage, fiber pull-out, debonding, and delamination. During
ite lay-up consisting of laminas (plies) with one or more
loading, unloading, and load holding, damage that does not
orientations with respect to some reference direction that result
emit AE energy will not be detected.
in a two-dimensionally flat article of finite thickness (typically
5.2 When the detected signals from AE sources are suffi-
relatively thin).
ciently spaced in time so as not to be classified as continuous
3.2.3 plate-like composite, n—any fiber-reinforced compos-
AE, this practice is useful to locate the region(s) of the 2-D test
ite lay-up consisting of laminas (plies), which is not strictly
sample where these sources originated and the accumulation of
flat, but for purposes of theAE examination, can be considered
these sources with changing load or time, or both.
as a two-dimensional (2-D) structural plate for wave propaga-
tion and for location of the region of AE source origin. 5.3 The probability of detection of the potentialAE sources
3.2.3.1 Discussion—Applies for a minimum radius of cur- depends on the nature of the damage mechanisms, flaw
vature of greater than about2m[6 ft], so curvature does not characteristics, and other aspects. For additional information,
change group velocities. see X1.4.
3.2.4 quasi-isotropic lay-up, n—a plate where the group
5.4 Concentrated damage in fiber/polymer composites can
velocities of both the fundamental modes have been shown to
lead to premature failure of the composite item. Hence, the use
be independent of propagation direction; for example: [+45/-
of AE to detect and locate such damage is particularly
45/0/90] (1).
s important.
3.2.5 wideband AE sensors, n—wideband (broadband) AE
5.5 AE-detected flaws or damage concentrated in a certain
sensors, when calibrated according to Test Method E1106 or
region may be further characterized by other NDE techniques
Practice E1781, exhibit displacement or velocity response over
(for example, visual, ultrasonic, etc.) and may be repaired as
several hundred kHz with a coefficient of variation of the
appropriate. Repair procedure recommendations and the sub-
response in dBs that does not exceed 10 %.
sequent examination of the repair are outside the scope of this
3.2.6 wideband-based (modal) AE techniques, n—AE tech-
practice. For additional information, see X1.5.
niques with wideband AE sensors that subject waveforms of
5.6 This practice does not address sandwich core, foam
the signals to combined time and frequency analysis to obtain
core, or honeycomb core plate-like composites due to the fact
mode-based arrival times (for source location calculations) and
that currently there is little in the way of published work on the
modal amplitudes for potential source identification.
subject resulting in a lack of a sufficient knowledge base.
3.2.6.1 Discussion—Note that mode-based arrival times can
also be obtained with resonant sensors, but only at certain 5.7 Refer to Guide E2533 for additional information about
types of defects detected by AE, general overview of AE as
experimentally determined frequencies.
applied to polymer matrix composites, discussion of the
4. Summary of Practice Felicity ratio (FR) and Kaiser effect, advantages and
limitations, AE of composite parts other than flat panels, and
4.1 This practice consists of subjecting flat composite pan-
safety hazards.
els or plate-like composite structures to loading or stressing
while monitoring with sensors that are sensitive to AE (tran-
6. Basis of Application—Personnel Qualification—
sient displacement waves) caused by the creation of micro-
Contractual Agreement
6.1 The following items are subject to contractual agree-
ment between the parties using or referencing this practice.
Available from Aerospace Industries Association (AIA), 1000 Wilson Blvd.,
Suite 1700, Arlington, VA 22209, http://www.aia-aerospace.org.
6.2 Personnel Qualification—Unless contractually agreed
The boldface numbers in parentheses refer to the list of references at the end of
this standard. otherwise, personnel performing examinations to this practice
ϵ1
E2661/E2661M − 20
shall be qualified in accordance with a nationally or interna- waves, see X1.6. For a scientific method to select sensors
tionally recognized NDT personnel qualification practice or whose best frequency response corresponds to the frequency
standardsuchasANSI/ASNT-CP-189,SNT-TC-1A,NAS-410, range of the highest amplitudes of the AE waves, see X1.7.
ISO 9712, or a similar document.They shall be certified by the
7.2.1.1 Wideband sensors can be used along with waveform
employer or certifying agency, as applicable. The practice or
recording to enhance AE data analysis by the application of
standard used and its applicable revision shall be identified in
wideband-based AE techniques. A wideband sensor should be
the contractual agreement between the using parties.
chosen with relatively flat response (Test Method E1106 or
Practice E1781) from about 50 kHz to 400 kHz. For additional
6.3 Qualification of Nondestructive Agencies—Unless con-
information,seeX1.7forplateslessthan2 mmthickandX1.8.
tractually agreed otherwise, NDT agencies shall be qualified
7.2.1.2 If resonant sensors are used, the best choice is a
and evaluated as described in Specification E543. The appli-
sensor with its primary resonance in the lower portion of a
cable edition of Specification E543 shall be specified in the
50 kHz to 400 kHz frequency band. Sensors with a lower
contractual agreement.
frequency resonance of about 25 kHz to 50 kHz can be used to
6.4 Procedure and Techniques—The procedures and tech-
increasesensorspacing(forexamplewhenalimitednumberof
niques to be utilized shall be as specified in the contractual
AE channels are available [see Practice E1067]) in AE testing
agreement.Inparticular,thecontractualagreementshouldstate
of composites, but such sensors increase the likelihood that
whether full monitoring of the test sample is required or if only
unwanted extraneous noise will be recorded. To minimize the
partial monitoring of certain expected critical areas is required.
effects of airborne noise the lower resonant-frequency sensors
6.5 Timing of Examination—The timing of examination
can be wrapped with sound absorbing material.
shall be in accordance with 1.3, unless otherwise specified.
7.2.2 Sensors should be shielded against electromagnetic
interference (EMI) through proper design practice or differen-
6.6 Reporting Criteria—Reporting criteria for the examina-
tial (anti-coincidence) element design, or both.
tion results shall be in accordance with Section 12, unless
otherwise specified. 7.2.3 Sensors should have omni-directional response, with
directional variations not exceeding 4 dB from the average
7. Apparatus
peak response of the set of sensors.
7.1 Refer to Fig. 1 for a typical AE system block diagram
7.3 Sensor Couplant:
showing key components.
7.3.1 The sensors must be acoustically coupled (to remove
7.2 AE Sensors: air from between the sensor face and the composite surface)
7.2.1 The selection of a wideband or resonant sensor is directly to the test sample. Commercially available couplants
describedhere.ForinformationonthefrequencycontentofAE for ultrasonic flaw detection may be used. Silicone-based
FIG. 1 AE System Block Diagram
ϵ1
E2661/E2661M − 20
high-vacuum grease has been found to be particularly suitable, to conclude whether the loss is acceptable. Note that integral
but it may not be desirable for all test locations and all test preamplifier sensors eliminate issues resulting from sensor
samples. Adhesives may also be used. Note: the sensor cables, but in some approaches such sensors increase the size
attachment procedure as well as the couplant or adhesive may and mass of the sensors and impact the mounting of this type
require approval prior to sensor installation due to special of sensor.
requirements for materials placed in contact with composite 7.5.2 Power-Signal Cable—The cable and connectors that
structures (compatibility or contamination control, or both).
provide power to the preamplifiers and conduct the preamp-
7.3.2 Couplant selection should be made to minimize amplified signals to the main processor shall be shielded
changes (for example, drying out of the couplant or movement
against electromagnetic interference. The typical standard
of the couplant due to gravity over the range of test tempera- coaxial cabling used in AE testing is RG-58 at about 50 Ω
tures and test time duration) in coupling sensitivity during a
impedance. Dual and quad shielded cable is also available to
complete examination. improve noise rejection in particularly noisy (electromagnetic)
environments. When RG-58 is used, the maximum recom-
7.4 Sensor Attachment Apparatus:
mended length is 330 m [1000 ft] to avoid excessive signal
7.4.1 Adhesives—Various adhesives can be used to attach
attenuation. Smaller diameter cables, RG-174 (50 Ω
sensors and provide acoustic coupling. The bond line created
impedance), can be used if the cable diameter is a concern for
by the adhesive must be much thinner than the shortest
a bundle of cables, but the effects of possible EMI from
wavelengths of interest. Adhesives such as two-part epoxies,
external sources and cross talk between cables must be
silicone adhesives, and cyanoacrylates have been successfully
accounted for. Some systems may use 75 Ω coax cables such
used for attaching sensors. Sensors attached with some adhe-
as RG-59. In all cases, the operator should follow the AE
sivescanbedifficulttoremovewithoutdamagingthesensoror
equipment manufacturer’s recommendations.
the examination sample. Also, due to the larger design defor-
mations of composite materials (relative to metals designed to 7.6 Preamplifier:
operate in their elastic range), adhesively bonded sensors may
7.6.1 The preamplifier converts the high impedance signal
debond during test sample stressing or during thermal cycling
from the sensor to a low impedance signal and amplifies the
of the test sample.
signalstoacceptablelevelstoallowthesignaltobetransmitted
7.4.2 Tape—Elastic adhesive tapes have been successfully
over longer distances of cable. The preamplifier also reduces
used for mounting transducers (for example, taping the sensors
the sensitivity to extraneous electromagnetic signals in the
to one side of a large composite panel).
power-signal cable.
7.4.3 Elastic Bands—An elastic band (for example, rubber
7.6.2 Integral preamplifiers (within the sensor case) reduce
bands) can be placed over the sensor and anchored to the test
the sensitivity to extraneous electrical noise, and they perform
sample to hold sensors in place.
as stated in 7.6.1.
7.4.4 Spring Loaded—Sensorsmaybespringloadedagainst
7.6.3 The preamplifier should include a filter with a band-
the test sample by fixturing (that does not generate extraneous
width that includes the useable frequency range of the sensors
noise during testing). Such mounting must be able to accom-
being used. Typically, a filter bandwidth of 50 kHz to 400 kHz
modate the deformation of the test sample without losing
or high-pass filters with a low cut-off frequency in this range
acoustic coupling.
may be used (the low cut-off frequency is altered if a low
7.4.5 It is generally unacceptable to modify a composite by
frequency resonant sensor is used). If extraneous mechanical
machining a “flat” to mount a sensor (creates potential dam-
noise is present, then the lower frequency may need to be
age). Thus, with surfaces that are rough, or have curvature, or
increased (but ideally it should remain at least 15 to 20 kHz
both, it is typical that the sensors will have less sensitivity than
below the resonant frequency of the sensor).
when they are mounted on flat and smooth surfaces.
7.6.4 Preamplifier gain should vary not more than 61dB
7.4.6 This practice does not address the use of waveguides
within the actual frequency and temperature ranges.
for fiber/polymer composites.
7.6.5 The input capacitance of the preamplifier should be
low (typically less than 25 pf) to limit the loss of sensor
7.5 System Cabling:
sensitivity.
7.5.1 Sensor Cable—AE systems typically use a standard
7.6.6 The preamplifier output should have a noise level not
low noise shielded coaxial cable that is not susceptible to
greater than 5 µV RMS (root-mean-square) (referred to a
triboelectric noise (from mechanical movement of the cable)
shorted input or a 50 ohm terminator at the input) within the
for this connection, due to its ability to shield the low level
actual frequency range.
signal out of the sensor from electromagnetic interference.The
7.6.7 The output impedance of the preamplifier should
cable should be kept short, 1-2 m [3-6 ft], to reduce attenuation
match the input impedance of the signal processing unit
of the signal, to reduce the length of cable possibly exposed to
(typically 50 ohms).
electromagnetic interference, and to create the best signal-to-
7.6.8 Preamplifiersshouldbeshieldedfromelectromagnetic
noise ratios. If it is absolutely necessary to use a longer length
during testing, the effect of the longer length on the attenuation interference.
of signal amplitudes should be evaluated (for example, by 7.6.9 Due to possible high amplitude AE signals in some
PLBs with a short cable length versus the longer length). If the composites, care should be exercised to eliminate voltage
loss is greater than 6 dB, the measured loss should be saturation of the signals in the preamplifier. For example, if
compared to the signal amplitudes obtained during pre-testing peak signal amplitudes are ≥97 dB , then a preamplifier that
AE
ϵ1
E2661/E2661M − 20
has 40 dB of gain and a maximum output of 20 Vpp into 50 Ω similar way as waves from real AE sources. Performance
may experience saturation. In such a case, a preamplifier with verifications verify performance of the entire system including
a lower amplification gain should be used, for example 20 dB. the couplant.
Alternatively, lower sensitivity sensors can be used or pream-
8.2.1 (a) The preferred technique for conducting perfor-
plifiers with a larger maximum output voltage can be used.
mance verifications is a pencil lead break (PLB). The lead
should be broken (see Guide E976) on the material surface at
7.7 Power Supply:
a fixed distance of about 10 to 15 cm [4 to 6 in.] from each
7.7.1 A stable, grounded, and reliable power supply that
sensor. When the composite is not of quasi-isotropic
meets the signal processor manufacturer’s specification should
construction, care should be taken so that the signal propaga-
be used.
tion path from the PLB to the sensor encounters the same fiber
7.8 Main AE Signal Processor:
lay-up for each sensor. Guide E976 specifies 2H, 0.3 mm
7.8.1 The main processor and computer with software (with
[0.012 in.] (or 0.5 mm [0.020 in.] providing a larger signal)
sufficient independent channels) should have electronic cir-
diameter lead. The length of the lead should be 3 mm
cuitryandsoftwarethroughwhichsignalsfromthesensorswill
[0.12 in.].Typically,thepeakamplitudeofthesignalfromeach
be processed. The main processor normally adds additional
sensor is recorded for three identical PLBs, and the results for
gain and appropriate frequency filtering. It shall be capable of
each channel should have an average peak amplitude within
processing each AE hit to determine a threshold-based arrival
64 dB from the average for all the channels. If a channel fails
time and the hit’s duration, counts, peak amplitude, and energy
this test, it should be repeated after re-coupling the sensor for
on each independent channel. In addition, it should process the
that channel, since improper coupling is a common problem. If
average signal level (ASL) or the root-mean-square (RMS)
the system still does not meet the performance requirements,
voltage on each channel. In order to record valid AE data, its
the operator must determine the cause of the deficiency and
capability, to process hits and store the processed AE hit data
take corrective action prior to the start of an examination. (b)
must exceed the rate at whichAE hits will be generated in the
In addition, a pencil lead should be broken in contact with the
examination. Finally, it should process and associate real-time
test sample surface at a location(s) such that the pulse
parametric measurement values (for example, time-driven data
generated leads to anAE hit at all the sensors intended for use
such as load, strain, temperature, etc.) with each hit.
in the application. The peak amplitudes from the signals from
7.8.2 Itmayincludehardwarewithsufficientdynamicrange
each sensor for three identical PLBs should be recorded along
(at least 12-bit) and sufficient digitization rates to properly
with the location of the PLB to provide data that provides a
digitize each AE hit. It should provide capability to store the
measure of attenuation of the wave propagation in the sample.
digitized waveform data and provide the ability to review the
When multiple samples are to be tested that are nominally the
waveforms and perform appropriate data analysis. For this
same,thisattenuationdatacanbeusedtoidentifysampleswith
much greater amount of data, its capability to process hits and
better or worse attenuation than the average. Such data may be
store the processed AE digitized data must exceed the rate at
of use in the evaluation of differences in the AE generated in
whichAE hits will be generated in the examination in order to
different test samples that are nominally the same. It also will
record valid AE data.
provideadatabaseforcomparingrelativesignalstrengthsfrom
7.8.3 The electronic circuitry shall be stable within 61dB
a repeat set of PLBs after the AE examination.
in the temperature range 4° to 49 °C [40° to 120 °F] (based on
8.2.2 An additional step may be useful in certain situations
manufacturer specifications).
(for example, when the sensors are covered with insulation
7.8.4 The electronic circuit threshold shall be accurate
after they are installed or when it is not safe to do the
within 61 dB (based on manufacturer specifications).
performanceverificationsduringanAEtest).Thisstepconsists
of first following the description in 8.2.1, then immediately
8. Calibration, System Performance Verification,
conducting a performance verification by the use of an Auto
Verification of Normal Sensor Response, and System
Sensor Test (AST), where a pulse is applied successively to
Electronic Noise Characterization
each sensor (which operates as a transducer or ultrasonic
pulser) and the signals (typically the peak amplitude) from
8.1 Calibration of AE sensors, preamplifiers, acquisition
each of the adjacent sensors are collected. These results then
system, andAE electronic waveform generator (AE simulator,
provide a database that subsequent AST test results can be
used for locally checking the performance of an AE system)
compared to when such tests are done at intermediate times
should be carried out in accordance with the equipment
during theAE examination and after theAE examination. This
manufacturer’s specifications and requirements. For additional
procedurealsoprovidesadatabaseontherepeatabilityofwave
information, see X1.9.
propagation between the sensors for different test samples that
8.2 System performance verification must be conducted
are nominally the same.
immediately before each AE examination. Performance verifi-
cation uses a mechanical device (see 8.2.1 for the preferred 8.3 Post system performance verification (by either or a
technique for composite samples) to induce (with a fast rise combination of the techniques in 8.2.1 (b) and 8.2.2, selected
time and short duration) displacement waves into the material so that a direct comparison can be made with the pretest
underexamination,ataspecifieddistance(sufficientsothatthe results) is also to be completed immediately after the exami-
preamplifier is not saturated by a very large signal) from each nation (when the test sample does not fail during the test) in
sensor. Induced displacement waves stimulate a sensor in a order to verify that there were no significant changes in sensor
ϵ1
E2661/E2661M − 20
coupling or system performance for each sensor and channel 9. Development of an AE Examination Plan
during the examination. However, in composites a variety of
The AE examination plan includes the AE examination
micro-damage or other test-induced damage can affect the post
preparation. The examination plan is called out by the appro-
examination results due to changes in signal propagation
priate structural test plan for the component/structure to be
characteristics. These changes may be observed and character-
examined.
ized by changes in the PLB or AST results.
9.1 Number of Sensors, Spacing of Sensors, and Locations
8.3.1 If the post examination performance verification or
of Sensors:
any intermediate performance verification result shows that the
9.1.1 When determination of whereAE sources originate is
system performance changed significantly (a loss of peak
the primary goal, the number of sensors and their placement
amplitude of more than 4 dB for any channel during the are determined differently depending on the AE technique
examination), the operator must note this in the report and used. If first-hit analysis is being used with resonant sensors,
the maximum size of the regions to which it is desired to
determineifthesystemoverallperformancewasstilladequate.
localize where the AE sources originated is approximately the
If not, then either the data analysis must be adjusted to account
total sample area being monitored divided by the number of
for the current system performance, or the test repeated with
sensors being used. If wideband sensors and wideband-based
appropriate modifications (for example, if attenuation has
AE technology is being used, then the number of sensors is set
increased significantly, then more sensors may be added to
by the discussion in 9.1.2.
maintain sensitivity) to ensure valid results.
9.1.2 When the primary goal is to effectively use AE to
NOTE 1—It is not possible to repeat the AE generation from a virgin
monitor the whole or a large part of a composite article, the
loading (that generates the characteristic damage state) of a composite
number of sensors and their locations are best determined by
sample during a subsequent retest.
attenuation measurements (with the selected sensor and se-
NOTE 2—A repeated test must go to a higher load at least 10 % above
lected electronic filters) on the composite article or on a test
the first loading.
sample with the same materials, thickness, and fiber lay-up.
8.4 It is important to have a “reference geometry” for use in
The attenuation measurements combined with the expected
quickly verifying the performance of sensors suspected to be amplitudes of the AE sources of interest and the planned
threshold (above the electronic or other background noise
damaged. This can be done using a typical thickness quasi-
levels) determines how far apart the sensors can be located so
isotropic composite plate (say 1 by 1 m [36 by 36 in.], to
that sources of interest do not have signals below the AE
reduce edge reflections) upon which each sensor can be placed
system threshold. Sensor spacing is normally decreased in the
at a fixed location and subjected to the waves from a PLB at a
directions having higher attenuation (for example, in propaga-
fixed location. Comparing the PLB peak signal amplitude (and
tion directions perpendicular to a large percentage of the
possibly the signal shape when waveform recording is being
fibers). When basing the expected signal amplitudes from a
used) with previous data for that sensor (under the same
database from small (25 mm [1 in.] wide) tensile or bending
conditions) may be used to identify faulty or non-performing
laboratory samples, the expected peak amplitudes should be
sensors.
reduced by about 10 to 12 dB to account for reinforcement of
8.5 Characterization of system extraneous electronic noise
signal amplitude from edge reflections in small tab-type
is recommended by the following: (i) in a “quiet” environment
samples. For additional information, see X1.10.
away from significant electromagnetic noise sources (for
9.1.2.1 Thedesiredmethodtocharacterizeattenuationisthe
example welding or operating overhead cranes; but not requir-
useofPLBsonthetestsampleedges.ThesePLBswiththeaxis
ing a Faraday box or room) and mechanical noise sources,
of the pencil parallel to the plate surface should be done both
characterize the RMS (or equivalentASL) noise level for each
onthetestsampleedgenearthetoporbottomsurfaceandvery
sensor/channel when each sensor is wrapped in foam (to near the mid-plane of the edge. The use of these two locations
eliminate any airborne noise) and not coupled to any solid; and generatesthefullrangeofsignalfrequency(modal)dominance
to be expected during testing. For additional information, see
(ii) in the same sensor environment determine the minimum
X1.11. One sensor should be located very close to the PLB
threshold for each sensor/channel before consistent triggering
location (approximately within 6 mm [0.25 in.]) so that the
on background electronic noise occurs. A typical value to
attenuation information includes the near-field geometric at-
define the minimum threshold would be a total of less than 10
tenuation. In this case, because a PLB is very close to the
hits per channel for a 15 min time period. If there is more than
sensor, care must be taken that the AE signal from the nearest
a 3 dB difference in the noise level or the minimum threshold
sensor (relative to the pencil lead break position) does not
from the average of all channels, the faulty channel/sensor
saturate the AE preamplifier. For additional information, see
should be repaired.
X1.12.Aseries of additional sensors at several distances (up to
8.6 Routine electronic evaluations must be performed any
or beyond the expected sensor spacing) from the source
time there is concern about signal processor performance. An
providedata(typicallypeakamplitude)todeterminethelossof
AE electronic waveform generator or simulator should be used
amplitude with increasing distance of propagation. To provide
in making such evaluations. Each signal processor channel
a true measure of attenuation for large test items, when a test
mustrespondwithpeakamplitudereadingwithin 62dBofthe
sample of the same thickness and fiber layup (rather than the
electronic waveform generator output. actual test article) is used for these wave propagation studies,
ϵ1
E2661/E2661M − 20
the test sample should be of sufficient size sample (transverse 9.2 The test plan should include the planned settings of the
dimensions should be at least two times the maximum propa- preamplifier gains and their filter range.
gation distance to be characterized) so that edge reflections do
9.3 The AE measurement system setup for proper data
not significantly reinforce the direct path signals. Also, if the
acquisition for the specific structural geometry and materials
material is not quasi-isotropic, the propagation directions
must be specified. This information includes parameters such
should include, at a minimum, the directions with the maxi-
as the threshold (and system gain prior to the signal reaching
mum and minimum in-plane and bending stiffness. In the case
the threshold circuitry), filter range, and other parameters that
of a layup with large differences in the number of fibers in
depend on the particularAE system. In addition, the choices of
different directions, the attenuation measurements should also
the test parameters (for example, load, pressure, test
be made at different angles relative to the direction of the PLB
temperature, etc.) to be recorded by the AE system during the
force. In such cases, modeling has shown there are both
test should be specified.
preferred propagation directions with less attenuation and
9.4 A suitable loading profile for AE testing of composites
non-preferred directions with higher attenuation (2).
mayberelativelysimpleoritcanbecomplicated,consistingof
9.1.2.2 Aprobability of detection, PoD [based onAE signal
many load cases in tension, compression, bending, multi-axial
amplitude], approach to sensor spacing in a test specimen has
loading, and it may include environmental effects (not to
been presented (3). The approach uses a “reference amplitude
induce loads) such as high/low test temperatures, vacuum, etc.
distribution” best determined experimentally from different
For additional information, see X1.13. The specific loading
source types in small test samples, where all the signals are
recommendations are described here. For additional
detected [PoD = 100 %]. This reference distribution is com-
information, see X1.14.
bined with an experimental attenuation database to allow an
9.4.1 Since polymer matrices are normally viscoelastic at
estimate of PoD as a function of proposed/selected sensor
typical test temperatures, ramp portions (increasing and de-
locations in the test specimen. An alternative approach to creasing) of the loading/unloading schedule should have con-
attenuation is to use a conservative estimate of the attenuation
trolled rates, and hold times at load as well as rest times at zero
coefficient. That should provide a corresponding conservative or near zero load should be specific and uniform in length or
estimate for the sensor spacing. match each other. The method used to control the loading/
unloading rate should be specified.
9.1.3 When the primary goal is a comparison between the
9.4.2 At the beginning of the examination, there should be
damage accumulations in virgin samples as a function of
an initial low level loading in the range of 5 % to 10 % of the
increasing stress level for different designs (material compo-
targetmaximumload.(NotethatthisisnotshowninFigs.2-5.)
nents or fiber lay-up, or both) of the same item for well
This loading is done to verify the functional performance of
designed composite items (having relatively uniform stress
every part of the entire system including test controls, loading
fields without regions with stress concentrations), a single AE
paths, and instrumentation. This low-level load also helps
sensor typically is sufficient along with RMS (typically both
verifytheinitiationoftheloadpath.Ifthisisthevirginloading
averaging time and time-driven interval of 200 to 300 ms) (or
of the test article, thenAE will typically be generated from the
itsequivalentASL)measurementsoftheAEtocharacterizethe
start of the formation of the characteristic damage state. After
accumulationofthecharacteristicdamagestateasafunctionof
the specified initial low level loading verification, the AE
applied load. The high hit rates may preclude the use of the
examination proceeds by increased stressing of the structure.
measurement of standard hit features. If the hit rate is not too
9.4.3 A common loading profile that is attractive, particu-
high, then the RMS data can be supplemented by the standard
larly for proof testing, is a load-hold-unload-rest-reload test
hit features. If the test sample does not meet the requirement of
cycleasillustratedinFig.2.TheprimaryAEmonitoringinthis
stress uniformity, then the technique for selection of the
case is during the second loading (used to obtain the FR), the
number and placement of sensors should follow 9.1.2. The
second hold, and the second unload. The initial loading-
measurement technique for the generatedAE would remain the
holding-unloadingportionservestonormalizethesampleprior
same for each channel. ThisAE data for different designs may
totheprimaryAEmonitoring.Amodificationofthisprofilefor
demonstrate optimal designs (material components or fiber
quality control testing or testing to optimize materials or
lay-ups, or both) having the least accumulation of damage up
fabrication parameters, or both, is to terminate the test after the
to a given test level (or design load level).
first unloading, and to focus the AE monitoring on the first
9.1.4 For test articles with known stress concentrations, the
loading and first hold.
AE test plan should emphasize placement of sensors in those
9.4.4 Another loading profile is shown in Fig. 3. (Note that
regions. In other regions, a sufficient sensor density should be
the magnitude of the steps may be adjusted to fit particular
used to monitor for possible unknown flaws, or stress
cases.) This loading sequence is more time consuming due to
concentrations, or both.
the repeated unloading. This loading sequence is valuable due
9.1.5 Specific information about the model identification of
to: (i) its elimination of most of the AE signals from the
the sensors (model designation, whether resonant, and the
formation of the characteristic damage state during each
resonant frequency or wideband) to be installed on the struc-
reloading up to near the previous maximum load; (ii) its
ture and the sensor installation techniques (coupling and enabling evaluation of the Felicity ratio for each successive
attachment) and the materials used for the sensor installation
loading; (iii) its enabling of evaluation of the load-hold AE at
should be specified. successively higher stresses; and (iv) its providing the test
ϵ1
E2661/E2661M − 20
FIG. 2 Example of Proof Loading with Holds and Reload; All Ramps and Holds of Equal Duration
FIG. 3 Sequence of Loading, Hold, and Complete Unloading; All Ramps at Equal Rates (not to scale) and Holds of Equal Duration
FIG. 4 Example of Loading Sequence in Steps Toward the Target Load; All Ramps at Equal Rates and Holds of Equal Duration
manager withAE-based feedback on the state of the composite 9.4.5 A sequence of steps of load and then a hold as
such that decisions to continue or stop loading can be made illustrated in Fig. 4 provides an alternate loading sequence.
without generation of significant additional damage or cata- (Note that the magnitude of the steps may be adjusted to fit
strophic failure occurring. particular cases.) This sequence may be dictated by other
ϵ1
E2661/E2661M − 20
FIG. 5 Sequence of Loading, Hold, and Partial Unloading; All Ramps at Equal Rates and Holds of Equal Duration
considerations in a composite structural test. In this case, it is 9.5.4 Development of appropriate acceptance/rejection
better to use longer hold times so as to allow more relevantAE criteria, when AE data is used in this role. Typically this AE
monitoring when the characteristic damage state (virgin load- application requires multiple “acceptable” and “unacceptable”
ing of the test article) is not being formed. This profile also test samples to establish the empirical criteria.
provides the test manager with AE-based feedback (from the
9.5.5 An evaluation of test environmental effects such as
hold portions) on the state of the composite such that decisions
temperature, humidity, and background noise, etc. should be
to continue or stop can be made without generation of
specified in the test plan. Some typical sources of background
significant additional damage or catastrophic failure occurring.
extraneous noises are test fixturing, test grips, operating
9.4.6 Another sequence of load steps, holds, and partial
overhead cranes, hydraulic servo-valves, etc.
unloads is shown in Fig. 5. (Note that the magnitude of the
9.5.5.1 Specifically for background noise, the following
steps may be adjusted to fit particular cases.) This sequence,
should be done. In the actual test environment with the sensors
while not as desirable (lacks sufficient unloads, such as shown
coupled to the test item, the same two noise items specified in
in Fig. 1 and Fig. 2, to fully activate the FR), has the advantage
8.5 should be measured for each channel. Comparing this data
of reducing the total test time.
with that from the testing in 8.5 can be used to identify and
then correct any sensor/channel not consistent with average
9.5 Since many factors influence the nature and amount of
values for the actual test environment. For additional
AE in composite materials (fiber materials (includes fiber
information, see X1.15.
weaving and number of “ends”), matrix material, fiber volume,
9.5.5.2 Specifically for humidity, due to the nature of the
general internal macrostructure such as laminate lay-up, time
anticipated testing (not always well controlled laboratories),
between load cycles, loading rate, cure cycle, sample volume,
the control of the moisture content of test samples is not
subsequent loading cycles, porosity or void content, moisture
generally possible. If different test samples are expected to
content, history of load and temperature and test temperature),
have significant changes in moisture content, then a series of
an AE characterization evaluation for a trial sample or prefer-
tests with different moisture content is recommended to deter-
ably several trial samples may be needed to finalize details
mine the effect of this variable on AE behavior.
associated with the examination plan to determine the follow-
ing:
9.6 Test samples should be identified based on sample type,
9.5.1 Characteristics of AE wave propagation from real
sample condition, manufacturing characteristics, loading
sources such as wave velocities (typically the velocity of the
parameters, etc.
first arrival of the signal (hit)), attenuation, and directional
9.7 The AE data to be observed real time during the AE
variations of these properties in the composite material. This
examination should be specified. The following is recom-
information is used with other data to determine the final
mended.
sensor spacing and sensor locations.
9.7.1 Graph of total first hits (or counts from first hits) in
9.5.2 Characteristics of AE signals from the composite
eachchannelversustimeorload,orboth,asappropriateforthe
sample during loadings so as to enable distinguishing extrane-
loading profile.
ous AE from AE sig
...