Acoustic Emission Testing of

Composite Materials
Larry A. Duesing 0 Lockheed Aeronautical Systems Company D Burbank

Key Words: Failure detection/isolation, Failure reporting, Reliability growth, Testing method.

ABSTR ACT high modulus, and low density, and in understanding their
behavior under certain types of loading. Glass-fiber reinforced
Acoustic Emission (AE) has developed rapidly over the last
•

plastics (GFRP) are now widely used in boat construction,

two decades as a nondestructive evaluation technique and as
for vessels made to hold hazardous environmental effect
a tool for materials research. AE is a passive monitoring
chemicals, and in the manufacture of pressure vessels designed
technique which, rather than requiring 100070 scanning (such
to hold several hundreds of atmospheres. Carbon-fiber com­
as Radiography, Ultrasonics, or Eddy Currents) can inspect
posites in which the matrix material is either polymeric or
the entire structure volumetrically using remote sensors elec­
carbon are now used in transport applications such as military
tronically tied to a computer for analysis and real-time loca­
aircraft. As with all new materials, studies that show the
tion. It, therefore, can be more efficient with respect to pre­
materials' uses and advantages also show properties that are
paration and inspection time, considerably reducing cost. In
denigrating. Therefore, testing and inspection of composites
many applications it is also more sensitive than other non­
has taken on new impetus. In order to do this, it is important
desfructive testiiig(NDT)iecliruques: It is otten this increased
to know the characteristics of the composites now used in
sensitivity which is exploited by using it to determine areas
commercial and military applications.
of a structure requiring more detailed characterization by
Composite materials can be divided into three broad
other NUT methods.
categories4, dispersion strengthened, particle reinforced, and
With the advent of glass/epoxy and graphite/epoxy com­
posite use in aircraft and aerospace applications, acoustic fiber reinforced. In each category, the composite is made of
a matrix material and of a second-phase material, which is
emission inspections will play a key role in the field of Reliabil­
distributed throughout the matrix. Dispersion-strengthened
ity and Maintainability. Furthermore, acoustic emission could
composite materials are distinguished by a hard and small
provide valuable insights into material behavior and as a
(0.01- to O.l-um-diameter) second phase (with volume con­
design aid for composite components.
centration less than 15007 ) dispersed throughout the matrix
INTRODUCTION material. These materials are distinct from precipitation alloy
Acoustic Emission) (AE) is the accepted term describing a systems in that they are normally made by techniques such
class of phenomenon whereby transient elastic waves are as powder-metallurgy and that the second phase does not go
generated by the rapid release of energy from localized sources into solution when the material is heated near the melting
within a material. With the application of stress or temper­ temperature of the matrix.
ature fields to a material, it results in elastic energy being Particle-reinforced composite materials are distinguished
stored in the material. The sudden local release of such energy from dispersion-strengthened composites by larger size of the

is the basic cause of the AE from the material. Several other second-phase particles (> 1 urn-diameter) and greater second­
terIns used fo describe Uie§e phenomena have included stress­ phase volume concentration (> 25070). An additional differ­
wave emission, microseismic activity and micro seismic ence is that particle-reinforced composites are strengthened by
emission. the relative hardness (compared to the matrix) of the second
The first documented observation of AE in 1953 by J. phase and by second-phase constraint of matrix deformation.
Kaiser2, was the modern day starting point for its active use. These strengthening mechanisms are distinct from the disper­
In fact, Kaiser characterized a basic irreversibility phenomenon sion-strengthened composites, where restriction of the motion
which bears his name. In the Kaiser Effect, when a material is of dislocations by the second phase provides the strength
stressed to a given level and then removed, upon reapplication enhancement.
of stress there is no detectable emission at a fixed sensitivity The distinct microstructural difference of fiber-reinforced
level until previously applied stress levels have been exceeded. composites is that the second phase (the fiber) has one dimen­
In subsequent years, McGonnagle3 has suggested that new sion that is larger than the other t wo dimensions. This
physical phenomena be evaluated for their potential as characteristic leads to anisotropic composite properties rather
nondestructive testing tools and that resolution power and sen­ than the macroscopically isotropic properties of the first two
sitivity-of all-methods be increased and nev-; instrumentation categories. Isotropic fibers have equal physical properties
be developed. along all axes, whereas anisotropic fibers have unequal pro­
perties along different axes.
BACKGROUND

Mechanical Properties of Composites Composite Testing

During the past decade, significant progess has been made AE is used widely throughout the petroleum and chemical
in developing advanced composites possessing high strength, processing industries. Fluids may be pumped down to relieve
0149 144X/89/0000-0128$01.OO © 1989 IEEE
128 1989 PROCEEDINGS Annual RELIABILITY AND MAINTAINABILITY Symposium
 stress on a fiberglass chemical tank and then refilled. The refill ing (S0-70dB). Most significantly, fiber breakage stands out
process is then monitored with AE to ensure that the tank in the data because it emits high amplitude events (> 70dB)
has not developed cracks, debond areas, fiber pullout, etc. and shows a distinct peak on the amplitude distribution plot.
Proof testing, whether during a scheduled shutdown or (See example plot - Figure 1.)
while the unit is onstream, is done to ensure that the integrity -� 01- ..... c-.-_
3000
_
of vessels or systems is assured. Monitoring procedures have
been developed and approved by various governing bodies over 1000 I
the past decade. Along with ASME fmal approval of utilizing De'ilmLaI:,OO
"
�

Acoustic Emission to test metal pressure vessels in lieu of

radiography, Section 5 Article 12, other standards have been
established �SPI, HASNT, ASfM, and CARP (Committee
�
Z 100 -
Resin Cracking
Flbe'�lng
Fiber Pullout
,
Fiber Breakage
on Reinforced Plastics). AE is fmally getting wide recognition
and acceptance through industry.
�
o

!i! 30
� l
Deformation in composites under changes in stess or
temperature are measured in the same manner as metals and
W
>
W 10 \ \
other base materials. , I ,
Because AE is not restricted to metal vessels (i.e. any
material that can be put under stress emits Acoustic Emis­
IJ W '-
ro ro 30 � � � ro � w m
sions), this method can be used throughout the aerospace and AMPLITUDE (dB)
aircraft industries. Each application has its own unique
Fig. I-Amplitude Distribution of Composite Defects
method of test, but they have been generally classed into
Felicity Flfect
metals or composites testing.
When a load is removed and subsequently reapplied signifi­
There are now established test procedures for testing com­
cant Acoustic Emission is frequently monitored at stress le\'els
posites and they are recognized by all the governing bodies
below those previously applied. This phenomenon is known
stated above. Most general rules of procedure will be similar
as the Felicity Effect. Research has shown that the ratio of
for all composites.
stress at onset of emission to the previous maximum stress,
All of the current field tests on composites have been con­
kDow as the Felicity Ratio, is an indication of the severity of
fmed to fiber-reinforced glass I epoxy structures. The newer
damage. Emissions from damaged areas will occur at loads
composites, (i.e. MMCs and Thermoplastics) have been con­
lower than those previously applied and, for this reason, it
fmed to laboratory. governm�t. or manufacturing facilities.
is not necessary to overload components during an Acoustic
. . ··· ThiS lSiiot to say tlWAEwoIi'f 6Ci::Ome a standard practice in
Emission test. Th make maximum use of the Felicity Effect,
aerospace, it's just that so much attention has been placed on
equipment is unloaded and held for a period of time with
development of the materials that inspection is being forced to
little or no load applied. (See Figure 2. S-3A Radome Loading
catch up.
Sequence) This permits viscoelastic flow of the resin and
Experience and laboratory testing has shown the following
redistribution of the internal stresses. It should be noted that
Acoustic Emission parameters are indications of structumI de­
the Kaiser effect is based on the underlying tenet that energy
ficiencies such as damage, underdesign, or improper fabricatioa
release from a material is irreversible. The presence of the
7btal Emission Felicity Effect does not violate this assumption, rather, it is
At. design stress levels, total emission is an indication of the an indication of additional damage occurring due to low C)'cle
sa'erity and extent of damage. 1btaI emission is also an indica­ fatigue. For low cycle fatigue to occur, major damage (i.e.
tion of underdesign. This parameter is particularly useful for crack, y ield, etc.) must be present.
0.0
detecting corrosion damage. -r:30�m�n
' �. "'\
kkground :2
Noise !\ � q
Emission Under Constant Load
A damaged component will continue to emit when held � :\-= i � g�
�
: ,I � .. 0

at constant load. In the damaged area, high local stresses will

be present in the resin matrix. These stresses cause viscoelastic
3.0
i�
:: 10L §;:,
flow of the matrix, which, in tutn, can result in fiber debond­
�� �:\� h�

.. � q
mg, fiber pullout, and resin cracking. In cases of severe
damage, load redistribution may even lead to fiber breakage.
Amplitude Distribution
• •
: 10
• I
jlt! 1
:: 10:,
01. .. 0
Amplitude distribution is the key facet of signature analysis : �:\:a �:: �:,� .:.
for determining defect characterization in composites. At. early :: :
::\�c<:
\li'E�: : :\�s�
: \i' E E
stages of the load cycle most events tend to be in the low :
: � ":�: : • � N M
(4()...6()dB) amplitude range. Of the five (5) main failure 5.4 : �!: : L.+J
mechanisms, resin cracking, fiber debonding and fiber pullout TIME ---- -,� :

PH 01 02 PH 03 04 PH os PH 06 07 PH D8 09 PH 10
faD in this range. Delamination bas a cbal8cteristic of having
a separate peak and shows up in the mid-range of a distnbut- Fig. 2-S-3A Radome Loading Sequence

1989 PROCEEDINGS Annual REUABIU1Y AND MAlNTAINABIU1Y Symposium 129

 ACOUSTIC EMISSION SYSTEMS ability to distinguish time of arrival between sensors, it was
only logical that space discrimination was feasible. In the late
Acoustic Emission systems for monitoring structural 70's with the advent of microprocessor and computer storage,
defects were designed to detect microscopic defects when they raw data manipulation capabilities enabled multi-channel
are still at the subcritical crack growth stage. In order to hear systems to monitor large structures and gather data in real­
this phenomenon, it is necessary to "listen" to the defects time information. Advances in naval submarine location gave
in the frequency spectrum in which they occur. This range rise to mathematical algorithms which computers could com­
is just above the audible range and encompasses a spectrum pute in near real-time, thus allowing AE systems to perform
from 20 kHz up to 1.5 MHz. There are instances in which location of defects on many different configurations (struc­
a defect in a material will propagate sound in the AE spec­ tures) in a real-time mode. The speed at which the multi­
trum, and also at the lower frequency detectable by the human channel AE systems could now handle data allowed location
ear. Good examples would be (I) the cracking and splintering detection rates above 5OD/sec.
of wood, (2) cracking of rock and some concrete aggregates, On the standard fiber-reinforced materials, conventional
(3) defect growth in metals and ceramics on cooling, and (4) source location was unsatisfactory. There are three main pro­
crack growth in glass. In order to understand how AE is used blems with time of arrival source location. The first is due
as nondestructive testing (NDf) tool, it is necessary to describe to the severe attenuation of a stress wave as it travels through
-atypical syst�m anah()wto�pplY1ts use. (See Figure 3. AE this composite. Because of this, most events will not strike
Data Processing) a sufficient number of transducers to permit triangulation.
Placing the transducers closer together can reduce the severi­
ty of the problem. However, all transducers in a measurement

,
array need to be within t wo (2) feet of an emission source
AE Sensor Signal AmPlification : Signal to ensure location of all significant events. Such an arrange­
I Processing
Low Noise Preamplifier
, ment would be prohibitively expensive for most components.
I

,
, The second problem arises because triangulation calculations
,
� L _______ , assume that stress waves travel along the structure wall at con­
,
-------- -------

CRT MOnitor :
stant velocity. Most of these types of composite constructions
,
I are anisotropic and stress waves will travel much faster along
the glass fibers than perpendicular to them through the
matrix. In theroy, it might be possible to adjust for the dif­
Raw ference in wave velocity. However, most practical construc­
Dala
Storage tions have two (2) or three (3) fiber orientations and the
x Y Plotter (Optional)
. mathematics becomes extremely complicated. Fiberglass rein­
."1-_---' Video (Optional) forced composites are extremely good emitters and this leads
to the third problem. A large number of events will be
Fig. 3 - AE Data Processing generated by a single source and each event will follow
closely upon the previous one. These events have a broad range
In order to detect microseismic activity on any part, it is of amplitudes, and some will not reach all sensors of a
necessary to attach a sensor directly to the surface of the part; measurement array. Frequently, it becomes impossible to track
much like a physician uses a stethoscope to listen. Unlike the which event is striking which sensor first, and the source loca­
stethoscope, which picks up sound in the audible range, AE tion results become invalid. So, source location on fiberglass
sensors rely on local surface deflection caused by a crack epoxy reinforced composites, and initially on all other com­
releasing energy which propagates as a stress wave to a free posites tested, will be accomplished utilizing individual sen­
surface which then oscillates producing a pressure on a sor activity. By spacing the sensors further apart, it ensures
piezoelectric type sensor. These sensors are called transducers that tracking of first sensor hits can be distinguished. To en­
because of their ability to transfer mechanical energy into elec­ sure complete coverage, the zones of interest of adjacent sen­
trical energy for measurement. sors overlap and in these areas the emissions will be detected
Today, there are a selection of AE systems, which, depend­ by more than one sensor. The approximate location of the
ing on the application needs, have from one (1) channel to defect can be determined by recording which sensor, or sen­
multi-channel systems having 128 channels. The earliest sors, detect the emission. Zone location is not exact, however,
systems, now used primarily in lab environments, were simple it provides an indication of which area of the component
digital machines that measured the number of times a sensor should be inspected by follow-up NDf examination. Normall)\
was pulsed (HI T) and could cumulate totals. As the need it is a relatively simple matter to identify the specific defect
increased to monitor structures outside the lab, and to monitor within the zone.
structures much larger than tensile test specimens, more chan­ Thday's modem systems not only are capable of determin­
nels were added. At this time, it was discovered that data rates ing Delta-T's (L::,. T = time of arrival between two or more sen­
increased exponentially as more sensors were placed on a sor), expanded capabilities of computers enable them to look
structure and in order to discriminate (in time) which sensor at each event ( = a hit at each transducer that is strong enough
was being hit, an accurate recording was necessary. With the to rise above a preset threshold which is above the electronic

130 1989 PROCEEDINGS Annual RELIABILITY AND MAINTAINABILITY Symposium

background noise), and determine its size (amplitude), dura­ For more discrete monitoring, there are three basic resonant
tion (length), and other specific details. (center) frequencies used for metal and composite testing.
These are:
Sensors 75 kHz - Used when attenuation is high as in thin walled
Being the key component of any detection system, this metals, composite structures, and aircraft panels.
section is devoted to more in-depth discussion of AE sensors. 150 kHz - Covers majority of AE tests where range is the
primary consideration (broad band) and attenua­
Main Jypes oj Sensors tion is not a problem.
1. Piezoelectric - These are most widely used and pre­ 500 kHz - Used when high background noise is a problem
dominate the field service applications. They can be as in high flow / stream noise, or aircraft inflight
selected for specific applications by choosing a specific fatigue monitoring.
resonant frequency or by selecting one with a broad
Setllant Couplants
frequency band.
Used to fill the air gap between the sensor shoe and the
2. Electromagnetic - These need no direct contact with
structure and, thereby, couple the AE energy from the struc­
the part, but are == 40dB less sensitive than piezoelec­
ture to the sensor.
tric types. These are seldom if ever used for AE because
Couplant should have these characteristics:
they are subject to welding current interference or air­
1. Wet the test surface and the sensor shoes
borne electrical interference (EMI/RFI).
2. Not trap air bubbles
3. Capacitive - These are used almost exclusively in the
3. Not corrode test surface
laboratory. They are noted for their flat frequency
4. Have sufficient viscosity to rise on vertical surface
response and can be utilized in finite spectral analysis.
5. Be easily removed
4. � - These are laser optical path interferometers.
Some common couplants:
They have a small focused beam == IOmm and make
Natural wax
measurements possible in hostile environments. They
Propylene glycol
can be calibrated accurately and make repeatable finite
Resin V 9
measurements. But, the band widths depends on the
Silicone grease
electronics employed and they are cost prohibitive.
Epoxy
There are presently many AE transducers on the market
Contact cement
and they can be utilized for specific applications. For this
discussionl only sensors used for production and in-service Analysis oj AE Data
monitoring will be detailed with emphasis on testing of metals Due to the complexity of the many parameters that are
and composites. To further focus attention, only piezoelec­ recorded and presented to the inspector, and the amount of
tric sensors are used for most, if not all, metal or composite time, education, and experience it takes to understand a
tests. system's capabilities, it would be impossible to list all of the
As background, the piezoelectric effect was first demon­ information available for analysis in the confmes of this
strated by the Currie brothers in 1880. It is defmed as the report. Therefore, key parameters will be discussed to give in­
phenomenon whereby an applied pressure on certain crystals, sight into the data available during an AE inspection.
the unit cells of which have no center of symmetry, will cause The AE event is observed on the system's oscilloscope and
an electric charge or potential to appear across the faces. There is distinguishable above the ever present background or elec­
are two broad categories for this type sensor, the single crystal trical noise. A threshold is set above this noise by cutting off
and the differential (dual) crystal. Prior to development of the system's counting detector when the signal is below a cer­
newer type case materials and crystal coatings, two crystals tain amplitude, typically 4OdB. There are two types of AE
were used (sometimes split crystals) to provide high common signal that are indicative of 'active" AE sources, burst and
•

mode noise rejection and maximum electrical isolation from continuous. Crack nucleation and propagation often emit as
the part being tested . This required dual leads within the (sen­ bursts. Other defects show up as continuous and can be diffi­
sor to system) cabling. Newer single crystal sensors with cases cult to distinguish from pseudo sources when looking at the
that provide EMI (electromagnetic interference) and RFI waveform alone. These pseudo sources are:
(radio frequency interference) protection, enables the use of Liquid and gas leakage
single power I signal coax cables with BNC connectors. There Crack closure
is no appreciable loss in sensitivity and either can be used, Loose particles and loose parts
but the ease of use could dictate the latter. Slag cracking
As with ali sensors, the signai from the sensor is pre­ Frictional rubbing
amplified before driving the signal up the long cable. In the Cavitation
past, the sensor and pre-amp were separate and required Any sudden volume change suC;h as:
special handling. Thday, the pre-amps are built into the sensor boiling
housing, with only a slight increase in height of the sensor, freezing
which makes for easier and quicker set-ups. melting
The score of diverse sensors cover a range of 25-800 kHz. chemical reaction

1989 PROCEEDINGS Annual REUABIUTY AND MAlNTAINABIUTY Symposium 131

 In order to determine useful information from these 4. RMS signal level. This measures the root-means-square
signals, the system processes these signals by looking at the voltage level of the AE signal coming from a sensor.
following information contained in each wave packet or event: It does not depend on the threshold level. It can pro­
(See Figure 4. Counting AE Signals) vide information as to when yielding of the metal
Event - The envelope of each signal from initial threshold occurs, existence of leaks and on the total system noise.
crossing to the last threshold crossing. It can be used to detect mechanical degradation, such
Duration - The amount of time involved with the above as loose parts.
event. S. Total number of AE source locations versus a test
Ringdown Count - The number of peaks in a given event. parameter, such as pressure. This is important if the test
Amplitude - Is the measure from a reference signal to the has hold periods since increases in the amount of the
top of a peak signal in an event. AE data produce increases in the height of the curve.
Rise Time - Is the time from first threshold crossing to When AE activity continues to increase with no increase
the peak amplitude. in pressure, this usually indicates that the structure is
EYlnt deteriorating and that a crack may be approaching a
Event Duration DuratiOn
I-t critical size. At the critical size, the crack suddenly
speeds up and spreads across the material or structure
in which it is located and causes failure.
Ring down Counts 6. Rate at which AE sources are located versus a test

�Ij �
, parameter, such as pressure. This is useful in the analysis
of AE data when the test or operational parameters are
'OdB Threshold
_ .. - ---- ---- continually increasing. A "rate" display is frequently
a

Background
Noise
1/ I I 1/ A
easier to interpret in tests which there is
number of AE data.
7. Video display of sensor arrays and AE source location
high total

il I
n
III'I
in respect to sensors. Each AE event will produce one

.. __ .. -- unique point of light on the display, positioned as to

_
I- ------- - - --- - -
its location on the test structure.
8. Amplitude distribution derived from the first sensor to
I receive the data for each valid event . The amplitude of
the signal is typically shown from 0-100dB. This display
has, as one of its most useful features, the ability to
quickly show whether or not the threshold levels are pro­
Fig. 4-Counting AE Signals perly set and can provide a means of quickly ranking
Data Display the AE sources as to their activity. It is used to
A data display method should be capable of performing characterize defects in composite materials by recogni­
any of the following functions: tion of zones of peak amplitude.
I. Show the location of the AE source in real time. Distribution analysis refers to the process of sorting out
2. Display any change of test variable with time. AE signals from one or more sources simultaneously. The
3. Enable the operator to assess the severity of an AE sorting is done on the basis of some measurable characteristic
source. of the signal, such as:
4. Be reproducible in hard copy form for convenient 1. Peak amplitude of each event
storage. 2. Counts per event
S. Be able to be updated as new data are acquired. 3. Duration of each event
4. Energy per event
Selection of a display mode depends primarily on the 5. Arrival time
information desired by the AE test. lYPical displays are: 6. Rise time of each event
1. Test parameter versus time. Pressure, total load or
temperature are plotted versus time. This documents the 1Ypical Data Displays
time during which load increases have occurred or A list of typical X-Y plots used:
----periods when the load was held COnSULTIt. • Distribution of events by event duration
2. Cumulative total AE source locations versus time. One • Log of distribution of events by location, test#, region#
of the more familiar ways of presenting AE data. • Distribution of events by amplitude
Establishes the total activity of the structure independent • Average ringdown counts per event vs. load
of the controlling test parameter. • Average rise time per event vs. cycles
3. Rate at which new sources become active versus time. • Log of cumulative distribution of events by amplitude
This type of display is useful for example, when the test • Log of amplitude vs. time
parameters are not constant with time and can provide • Cumulative events vs. time
an early warning of failure.

132 1989 PROCEEDINGS Annual RELIABILITY AND MAINTAINABILITY Symposium

'Jest Procedures sensor or from the sensors in an array. The simulator sends
Although there are many different applications, with out a signal of constant level. A constant output is more im­
almost as many procedures, there are several guideline pro­ portant than the actual level. A simulator can be used to:
cedures established. A brief outline will give an overall view 1. Check the gain of the sensor channels.
of what is entailed. 2. Measure the velocity of rayleigh waves on the wall of
A pre-test noise assessment survey prior to installation of a structure.
an AE monitoring system should be made to: 3. Confrrm the location of AE sensor positions pre­
I. Determine operating noise level/frequency viously determined manually by direct measurement.
2. Provide data so that proper AE system can be selected 4. Determine the minimum discriminator threshold level
3. Determine if AE signals can be detected above which would have been fIrst crossed by rayleigh wave
background noise fronts.
The fIrst problem to be faced in setting up an AE system S. Estimate the attenuation characteristics of the material
at a fIeld site is to make sure that the AE signals will be above through which the AE burst are traveling.
the background noise level. This can be done with a simulator, 6. Verify the operation of the complete AE system.
a sensor, a preamplifIer, the expected cable length, a main The same simulators used in the cahbration of sensors may
amplifier and an RMS voltmeter. A measurement of the fre­ be used to calibrate the complete AE system.
quency spectrum of the noise might be desirable to select an 1b calibrate the sensitivity of an AE system, fmt determine
optimum operating frequency band to avoid noise. If the noise the sensor signal-to-noise ratio. This is the ratio of the out­
is excessive, the "advance" evaluation crew can recommend put voltage of the preamplifier produced by an input from
modifications of the test procedures or equipment prior to the simulator divided by the noise level voltage when the
shipping equipment to the site. This might, for example, in­ simulator is not on. This ratio is a measure of coupling effi­
clude selection of a sensor with a peak frequency response ciency. Maintaining this ratio for all sensors in a system means
lying outside the frequency range of the noise. Possible sources that the coupling efficiency is the same for all sensors.
of noise that Would interfere with the detection of AE signals Next, the gain of each channel should be adjusted so that
and suggested ways to eliminate or reduce them are listed the peak voltage output of all sensors is within ± IdB.
below. The mechanical AE simulator is essentially a precision
Electrical noise is a function of instrument shielding and mechanical pencil (e.g. Pentel 0.3 mm, 2H). To generate
good electrical grounding. Good shielding practice demands simulated AE, just push the pencil tip against the structure
that the voltage drops caused by currents in ground conduc­ slowly until a segment of the lead breaks. The simulated AE
tors be less than signal voltages. The signal levels detached is a point release force pulse of step-function time dependency.
at the sensor are measured in microvolts and, therefore, a very The pulse height can be independently measured with a force
low impedance ground is needed. gage or a load cell.
The principle methods of eliminating electrical noise are:

I. Use "bipolar" (common mode) sensors Applications

2. Shield all instruments and cables AE is a tool that falls under the general categories of NDf.
3. Use good electrical grounding NOr is used to prevent accidents, protect human life, ensure
AE systems are designed to fllter out most ambient product reliability,used as a manufacturing control, improves
mechanical noise, however, certain types, such as caused by productivity, and reduces production cost. The advantages and
impact, cavitating or very turbulent flow, sealing flange stick­ disadvantages of AE are as follows:
slip or vessel mount motion may be detected. It is not possi­ Advantages
ble to completely eliminate these, but their contribution may - Real-time evaluation
be redu �considerably by a few precautions. These include - Passive technique
protection from rain, mounting on wood or padding, using - Continuous monitoring
flexible hose rather than pipe from pump to vessel , removing - High sensitivity
members in contact with both the vessel and a noise produc­ - Whole structure interrogation
ing structUre and securing all vessel connections against - Source location
leakage. - Active defects detected
The amount of spurious noise reaching a sensor is less if - Pattern recognition aids interpretation
there is no continuous metal path between the noise source Disadvantages
and the sensor. Plastics are good sound attenuating materials. - Discrimination between noise and AE signals
The effectiveness of the plastic isolating material is due to - Requires experience
the difference in the acoustical impedances of metal and - Material behavior must be understood
plastic. An acoustic wave is 98070 reflected at the boundary - Quantitative correlation limited
between a piece of steel and a piece of nylon. - Passive defects not detected
Areas which benefit from AE and possible future test areas
Calibration 0/ 'lest Equipment are listed below:
Calibration of an AE test is performed by placing a stan­ 1. Power Plants/Electric Utilities
dard AE source simulator at a specified distance from a single 2. Chemical/Petrochemical Industry

1989 PROCEEDINGS Annual REUABIUTY AND MAINTAINABILITY Symposium 133

 3. Aerospace simple hand held single channel detectors, to full capability
a. Fatigue testing of airframe members sixty-four (64) channel systems with all the peripheral systems
b. In-flight monitoring of structural members and latest software programs.
c. Crack detection in liquid and solid motor cases Selection needs to be made by an engineer familiar with
d. Proof testing and monitoring of fuel tanks all of the systems, capabilities, and have knowledge of its use
e. Testing integrity of composite material structures now and for the future. Systems should cover current needs,
f. Laboratory study of aircraft materials / parts but also be able to expand or be tied into a second system.
g. Engine analysis and inspection Careful attention must be paid to maintainability in the field
4. Metal Fabrication and location of service centers to the operating area where
a. Spot weld production control the system is employed.
b. Resistance and laser spot weld quality control
c. Immersed arc welding quality control CONCWSIONS
d. Inspection of metal bonding
5. Transportation - Railroads / Highways Acoustic Emission applications abound in the civilian and
6. Mining / Underground Caverns military aircraft markets. Problems such as fuel leaks, fatigue
7. Medical Science cracking, and composite repairs are prime candidates for one
of the many AE type tests.
Test Reports AE was chosen for initial application studies of aerospace
The reports of AE tests are evidence of the condition of composites because it has several positive aspects. It is highly
the structure as determined by the test condition at a par­ portable, cost effective, and tolerant of ambient noise. It has
ticular time. They provide the owner of the structure and a the added advantage of not being so technically sophisticated
regulatory agency with documentation as to the existence of that technician training is difficult.
discontinuities. AE holds potential for future applications because a viable
It is a good idea in critical structures to perform a test regime for many aerospace composites, with minimal
"baseline" inspection prior to putting the structure into use. attention to size and shape, can be developed in a timely
Periodic inspection and simultaneous record keeping make fashion at relatively low cost.
it possible to monitor discontinuities that are not critical and
possibly to relate their growth to an operating parameter. REFERENCES
All relevant data should be reported. Ageneral checklist
follows: I. ASTM E07-04 Subcommittee on AE, "Standard Definition of
1. Date Terms Related to AE", 1980.
2. Time of day 2. J. Kaiser, Untersuchungen Uger Das Auftreten Gerauschen Beim
3. Operator Zugversuch Archiv Fir Das Eisenhuttenwesen. Vol. 24, 1953
4. Location - including city and state 3. W.J. McGonnagle, "Non Destructive Testing" in Fracture, edited
5. Structure by H. Liebowitz, Vol. Ill, 1971.
6. Instrumentation - manufacture and serial numbers
7. Sensor types and serial numbers BIOGRAPHY
8. Sensor location
9. Calibration data - sensor and system Larry A. Duesing
10. Results Lockheed Aeronautical Systems Company
II. Signatures of operator and supervisor p.o. Box 551
12. Photographs Burbank. California 91520-6611 USA

13. Hard copies of displays

Larry Duesing is presently the R&M lead engineer for the 5-3 aircraft
14. Any other data that would be necessary to duplicate
at the LASC. Burbank facility. He has been involved in Indepen­
the test dent Research and Development projects and is currently the principle
. $ystern Selection investigator of Non-avionic Design Requirements for the Product
There are currently three (3) to five (5) system manufac­ Support Division at LASe, Burbank. Larry holds a degree in Business
turers in the commercial sector. The cost of the systems are Administration from National University. He has 25 years experience
proportional to their capabilities. These systems range from in aircraft maintenance and maintenance related disciplines.

134 1989 PROCEEDINGS Annual RELIABILITY AND MAINTAINABILITY Symposium