MAGNETIC PARTICLE TESTING I 283
Three-phase full-wave direct current has all of the FIGURE 41 . Three-phase full-wave direct current
advantages of single-phase full-wave direct current plus waveform
some additional benefits. The current draw on the power
line is spread over three phases, reducing the demand by TIME
nearly half. The demand on the line is also balanced, with
each leg providing a portion of the current (single-phase
pulls all of the current from one leg, resulting in an unbal-
anced line). Many power companies charge a higher rate to
customers with unbalanced, high current requirements.
The three-phase design also permits incorporating a quick
break circuit that improves the formation of indications at
the ends of longitudinallymagnetized test objects.
284 I NONDESTRUCTIVE TESTING OVERVIEW
PART 9
DEMAGNETIZATION PROCEDURES
Objects that have been magnetic particle tested retain 5. Demagnetization is required when objects are to be
some magnetism. The amount of residual magnetism electric arc welded. A residual magnetic field can
depends on the material and its condition. Low carbon steel cause the arc to be deflected or to wander. Arc deflec-
in the annealed condition retains little or no magnetism tion (called arc b.Jow) is a particular problem in auto-
while hardened alloy steels retain strong magnetic fields for mated welding systems that do not compensate for a
shift in arc.
long periods of time.
Reducing or removing residual magnetism is not a direct 6. Demagnetization may be required when remagnetiz-
ing in another direction, if the second magnetizing
function of the intensity of the retained magnetic field but is field intensity is less than the original. If the second
a direct function of the coercive force of the material. There magnetic field strength does not equal or exceed the
are materials with high retentivity and low coercive force. initial field strength, the initial magnetic field remains
The ease of demagnetization depends on the magnetic dominant.
properties or hysteresis curve of the material.
Justification for Demagnetizing Reasons for Not Demagnetizing
There are several ways that an object can be magnetized: Although demagnetization is generally required, there
induced magnetization from earth fields; use of a magnetic are occasions when it is not necessary. Demagnetization is
chuck or plate during machining; mechanically induced not required when the test objects have very low retentivity
magnetization; and magnetic particle testing. Demagnetiza- (such materials are demagnetized when the magnetic field
tion is required for the following reasons, despite the source strength is removed). On some occasions, the residual mag-
of the magnetization. netic field is such that it does not affect the function of the
object nor its service life. Occasionally, the test object is
1. A magnetized object can affect the accuracy and magnetic particle tested a second time, with equal or
function of some instruments and meters. A common greater magnetic field strength in another direction.
occurrence in aircraft is the induced magnetization
resulting from traversing through the Earth's mag- Demagnetization is not necessary when test objects are
netic field. subjected to external magnetic fields such as clamping with a
magnetic chuck during machining or hoisting with an electro-
2. Removal of the magnetic particle media following magnetic crane. Finally, there is no need for demagnetization
testing is necessary because residual particles can if the test object is exposed to a subsequent heating above the
cause problems during subsequent operations such as Curie point (the temperature where magnetic domains
machining and surface coating. Retained particles become random and the material becomes unmagnetized).
can also cause excessive wear on moving components
in assemblies. Demagnetization is necessary because Methods of Demagnetization
flux leakage can retain particles despite a typical
cleaning process. Curie Point Heating
3. Machining of magnetized objects is objectionable All ferromagnetic materials containing magnetic flux may
because chips and shavings may adhere to the sur- be demagnetized by heating to a specific temperature and
face, disrupting the surface finish and dulling the cut- allowing the material to cool in the absence of an external
ting tool. magnetic flux. The temperature at which the material
changes from ferromagnetic to paramagnetic is called the
4. Magnetized objects attract and retain metallic debris Curie point. This temperature varies widely depending on
during handling and cleaning before application of
surface coatings. The entrapped metal particles create
serious imperfections in painted or plated surfaces.
MAGNETIC PARTICLE TESTING I 285
alloy composition. For example, the Curie point for nickel In Fig. 42, the bottom curve illustrates the magnetic
containing 1 percent silicon is 320 °C (608 °F) while the field strength used to generate the flux intensity curve at
Curie point for nickel containing 5 percent silicon is 45 °C right. As the current diminishes in value with each reversal,
( 113 °F). The Curie point for ferrous alloys ranges from the hysteresis curve traces an increasingly smaller path. The
about 650 to 870 °C (1,200 to 1,600 °F). top right curve illustrates the decreasing residual flux inten-
sity in the object, indicated by the shrinking hysteresis
The transition from ferromagnetic to paramagnetic at the loops. The magnetizing current and flux intensity curves are
Curie point reverses on cooling and the material becomes plotted against time. When the current reaches zero, the
ferromagnetic in an unmagnetized condition. Some X-ray residual magnetism approaches zero.
diffraction studies show that this transition is not a c:rystalline
structure transformation but a rearrangement of magnetic Successful demagnetization depends on several require-
domains. Demagnetization by heating through the Curie ments. First, the magnetic field strength at the start of the
point is the most thorough demagnetization possible but demagnetizing cycle must be high enough to overcome the
because of its expense it is not commonly used. coercive force and to reverse the direction of the residual
field. This is accomplished by demagnetizing at a slightly
Electromagnetic Demagnetization higher current than that used in the magnetizing cycle. The
second requirement is that, in each successive cycle, the
There are several techniques for demagnetizing an reduction of magnetic field strength must be small enough
object using electromagnetic energy. All of these techniques that the reverse magnetic field strength exceeds the coer-
subject a magnetized object to a magnetic force that is con- cive force and reverses the flux direction from the previous
tinually reversing its direction and gradually decreasing in reversal. This requires a number of cycles, depending on
intensity. the permeability of the material. Ten to thirty reversals are
often required.
FIGURE42. Demagnetizationhysteresisloopswith currentand flux intensitycurves
B+ +
I FLUX CURVE
I
µ - ----
~' -,;I--I - - -
I
I
H+
CURRENT CURVE
j
B-
286 I NONDESTRUCTIVE TESTING OVERVIEW
Alternating Current Demagnetization placing them in a well separated single layer with their long
dimensions parallel to the axis of the coil.
A common method of demagnetizing small objects is by
passing them through a coil carrying alternating current Direct Current Demagnetization
(Fig. 43). The objects are moved into the coil while the cur-
rent is flowing for exposure to the maximum magnetic flux. The principle of demagnetizing with direct current is
The objects are then slowlyand axially withdrawn some dis- identical to that of alternating current demagnetization. The
tance from the coil. Because flux intensity decreases with magnetic field strength or current must be sequentially
distance from its source, this procedure serves to reduce the reversed and gradually reduced. One of the advantages of
magnetic field strength. To ensure that the flux is reduced to reversing direct current demagnetization is the deep pene-
a minimum, the objects should be withdrawn to a distance tration that is possible.
at least twice the coil diameter.
Because reversing the direction of direct current is done
Alternating current demagnetization can also be accom- through electrical circuits, it is possible to control the rate of
plished by placing the object in the coil and gradually reduc- reversal. The most commonly used reversal rate is one cycle
ing the current to zero. Some coils and some magnetic per second or a frequency of 1 Hz. This produces the opti-
particle system designs have built-in circuitry for current mum depth of penetration, permitting the demagnetization
reduction. When decaying alternating current is available on of large test objects. Direct current demagnetization often
wet horizontal units, the current can be applied directly to the reduces the residual field to a value lower than is possible
object through the head stock and tail stock instead of passing with alternating current. In practice, the test object is
the object through the coil. This is more effective than the placed within the coil where it remains until the demagneti-
coil technique for long, circularly magnetized objects. zation cycle is complete.
There are some limitations to alternating current demag- Yoke Demagnetization
netization. Most important is the fact that alternating cur-
rent concentrates the magnetic flux at the object surface. Yokes or probes are often used for demagnetization
Large test objects are not effectively demagnetized by the when portability is required. Either alternating current or
alternating current method because of its skin effect. reversing direct current can be used, depending on the
available power supply. Pulsating half-wave direct current
This lack of penetration also prohibits demagnetization of found in many self-contained power yokes cannot be used
a number of small objects piled in a basket (the alternating unless the unit also contains a current reversing circuit.
current skin effect demagnetizes only the outside surface of Demagnetization is accomplished by passing objects
objects on the outer layer). Quantities of small objects can be through the poles of the yoke and withdrawing them while
demagnetized with alternating current techniques only by the current is flowing.
FIGURE 43. Alternating current coil Yokes can also be used to demagnetize local areas on
demagnetizing unit with rail assembly large objects. The poles are placed on the surface to be
demagnetized, moved in a circular pattern and then slowly
withdrawn while the yoke is energized. When demagnetiz-
ing small areas on a large object, care must be exercised to
avoid magnetizing adjacent areas.
DemagnetizationPractices
There are practical limits to the demagnetization pro-
cess. These limits are controlled by the equipment, by the
size and material of the object and by the Earth's magnetic
field. Generally, the practical limit of demagnetization
occurs at a point where a residual field remains but at a level
that does not interfere or complicate the intended function
of the object in service.
Longitudinal, residual magnetic fields are usually mea-
sured with a field meter. Some meters read relative units
and are useful for comparison purposes only; other meters
MAGNETIC PARTICLE TESTING I 287
read directly in tesla or gauss. The greatest flux leakage in a a circularly magnetized object. A common practice is to per-
form longitudinal magnetization as the last step in a two-
longitudinally magnetized object is at the ends or comers of step operation or to remagnetize in a longitudinal direction
before demagnetization. This procedure allows the use of a
the object. These are the best places to check for the effec- field meter to check the effectiveness of demagnetization.
tiveness of demagnetization. Note that when the readings
are in relative units there may be differences between the The Earth's magnetic field is in a north-south direction
readings of different manufacturer's field meters. and can cause problems when demagnetizing objects with a
high length-to-diameter ratio. When low residual fields are
The magnetic field of a circularly magnetized object is required, these problems can be reduced by placing the
completely contained within the object and there are no demagnetizing unit's coil axis in an east-west direction.
flux leakage points except at discontinuities. Therefore,
field strength meters cannot indicate residual magnetism of
288 I NONDESTRUCTIVE TESTING OVERVIEW
PART 10
MEDIA AND PROCESSES IN MAGNETIC
PARTICLE TESTING
The formation of reliably visible discontinuity indica- from discontinuities. As the particles become magnetized,
tions is essential to the magnetic particle testing method. they then attract additional particles to bridge and outline
An important factor in the formation and visibility of indi- the discontinuity, thus forming a visible indication.
cations is the use of the proper magnetic particles to obtain
the best indication from a particular discontinuity under Magnetic permeability alone does not produce a highly
the given conditions. Selection of the wrong particles can sensitive particle material. For example, iron based dry
result in (1) failure to form indications; (2) the formation of powders have a higher permeability than the oxides used in
indications too faint for detection; or (3) a distorted pattern wet method suspensions. Yet a typical dry powder does not
over the discontinuity and the resulting misinterpretations. produce indications of extremely fine surface fatigue cracks
that are easily detected with wet method suspensions. High
In magnetic particle tests, there are two classes of media permeability is desirable but is no more important than
that define the method: dry and wet. Dry method particles size, shape or the other critical properties. All of these char-
are applied without the addition of a carrier vehicle. Wet acteristics are interrelated and must occur in appropriate
method particles are used as a suspension in a liquid vehi- ranges in order for high permeability to be of value.
cle. The liquid vehicle may be water or a light petroleum
distillate similar to kerosene. Coercive Force
Magnetic particles are also categorized by the type of pig- Materials used in dry method powders and wet method
ment bonded to them to improve visibility. Visible particles suspensions should have a low coercive force and low reten-
are colored to produce a good contrast with the test surface tivity. If these properties are high in dry powders, the parti-
under white or visible light. Fluorescent particles are coated cle can become magnetized during manufacture or during
with pigments that fluoresce when exposed to ultraviolet their first use, making them small, strong permanent mag-
light. A third pigment category includes particles coated with nets. Such particles have an increased tendency to magneti-
a material that is both color contrasting under visible light cally adhere to the test surface where they first touch it. This
and fluorescent when exposed to ultraviolet light. reduces their mobility and produces a high background,
reducing contrast and masking relevant discontinuity indi-
Magnetic Particle Properties cations.
The media used in magnetic particle testing consist of When wet method particles have a high coercive force,
finely divided iron powder ferromagnetic oxides. The parti- they are also easily magnetized, producing the same high
cles can be irregularly shaped, spheroidal, flakes or rod level of background. Magnetized particles are attracted to
shaped (elongated). The properties of different materials, any ferromagnetic material in the testing system (bath tank,
shapes and types vary widely and some are discussed below. plumbing system or rails) and this causes an extensive loss of
particles from the suspension. Particle depletion creates
The level of particles in suspension should be main- process control problems and requires frequent additions of
tained consistently, preferably about 0.4 mL in a 100 mL new particles to the bath.
settling test. For consistent results, the suspension vehicle
must be changed frequently because of foreign material Another disadvantage of magnetically retentive wet
contamination. method particles is their tendency to clump, quickly form-
ing large clusters on the test object surface. These clumps
Magnetic Permeability have low mobility and do not migrate to leakage fields. This
causes distorted or obscured indications in heavy, coarse
Magnetic particles should have the highest possible per- grained backgrounds.
meability and the lowest possible retentivity. This allows
their attraction only to low level leakage fields emanating Strongly magnetized particles form clusters and adhere
to the test object surface as soon as bath agitation stops.
Particles with low magnetic field strength cluster more
slowly while indications are forming. The leakage field at
MAGNETIC PARTICLE TESTING I 289
the discontinuity draws the particles toward it and the clus- object. Stranded particles often line up in drainage lines that
ters are constantly enlarging due to agglomeration. At the could be confused with discontinuity indications.
same time, the clusters sweep up nearby fine particles as
they move toward the discontinuity. Wet Method Fluorescent Particles
Effects of Particle Size Particles treated with a fluorescent pigment or some of
the visible pigments differ in size and behavior from black
The size and shape of magnetic particles play an impor- or red (uncoated) visible particles. Fluorescent particles
tant role in how they behave when subjected to a weak mag- must be compounded and structured to prevent separation
netic field such as that from a discontinuity. Large, heavy of the pigment and magnetic material during use. A mixture
particles are not likely to be attracted and held by a weak of loose pigment and unpigmented magnetic material pro-
leakage field as they move over the object surface. However, duces a dense background and dim indications. In addition,
very small particles may adhere by friction to the surface the unpigmented magnetic particles may be attracted and
where there is no leakage field and thus form an objection- held at leakage fields bu:t their lack of contrasting color
able background. makes them difficult to see.
Dry Powder Particles Producing fluorescent magnetic particles involves bond-
ing pigment around each magnetic particle. The bonding
Within limits, sensitivity to very fine discontinuities typi- must resist the solvent action of petroleum vehicles and sur-
cally increases as particle size decreases. Extremely small factants and the abrasive action occurring in pumping and
particles, on the order of a few micrometers, behave like agitation systems. Some manufacturers encapsulate the
dust. They settle and adhere to the object surface even bonded dye particle in a layer of resin. Particles built up syn-
though it may be very smooth. Extremely fine particles are thetically are larger than the minimum sized visible parti-
very sensitive to low level leakage fields but are not desir- cles; such powders have much fewer very fine or small
able for production tests because of intense backgrounds particles. As a result of their processing, fluorescent parti-
that obscure or mask relevant indications. cles have a definite size range that is maintained throughout
the suspension's service cycle.
Large particles are not as sensitive to fine discontinuities.
However, in applications where it is desired to detect large Effect of Particle Shape
discontinuities, powders containing only large particles may
be used. Magnetic particles are available in a variety of shapes:
spheres, elongated needles (or rods) and flakes. The shape
Most commercial dry powders are a carefully controlled of the particles affects how they form indications. When
mixture of particles containing a range of sizes. The smaller exposed to an external magnetic field, all particles tend to
sized particles provide sensitivity and mobility while larger align along the flux lines. This tendency is much stronger
sized particles serve two purposes. They assist in building up with elongated particles such as the needle or rod shapes.
indications at larger discontinuities and help reduce back- Elongated shapes develop internal north-south poles more
ground by a sort of sweeping action, brushing finer particles reliably than spheroid or globe shaped particles, because
from the test object surface. A balanced mixture containing they have a smaller internal demagnetization field.
a range of sizes provides sensitivity for both fine and large
discontinuities, without disruptive backgrounds. Because of the attraction of opposite poles; the north
and south poles of these small magnets arrange the particles
Wet Method Visible Particles into strings. The result is the formation of stronger patterns
in weaker flux leakage fields, as these magnetically formed
Particles used in a liquid suspension are usually much strings of particles bridge the discontinuity. The superior
smaller than those used in dry powders. Typically,the parti- effectiveness of elongated shapes over globular shapes is
cles for wet method testing range from 1 to 25 µm (4 x 10-5 particularly noticeable in the detection of wide, shallow dis-
to 1 x 10-3 in.) while dry powder ranges from under 10 to continuities and subsurface discontinuities. The leakage
250 µm (0.0004 to 0.01 in.). The upper limit of particle size fields at such discontinuities are weaker and more diffuse.
in most wet method visible materials is 20 to 25 µm (0.0008 The formation of particle strings based on internal poles
to 0.001 in.). makes stronger indications in such cases.
Larger particles are difficult to hold in suspension. Even The disadvantage of elongated particles is their tendency
to mat and form clusters that mask indications. Particle
20 um (0.0008 in.) particles tend to settle out of suspension mobility is greatly enhanced by a spheroid or globular shape.
rapidly and are stranded as the suspension drains off the test
290 I NONDESTRUCTIVETESTING OVERVIEW
Dry Powder Shapes are black or red. Manufacturers bond pigments to the parti-
cles to produce a wide selection of other colors: white,
The superiority of elongated particles in diffuse mag- black, red, blue and yellow, all with comparable magnetic
netic fields holds true for dry powder testing. However, properties.
there is another effect that must be considered. Dry pow-
ders are often applied to object surfaces by releasing them The white or yellow colors provide good contrast against
out of mechanical or manual blowers. It is essential that the mill surface objects. They are not effective against the silver
particles be dispersed as a uniform cloud that settles evenly gray of grit blasted or chemically etched surfaces or against
over the object surface. Magnetic powder containing only bright, polished machine ground surfaces. For those appli-
elongated particles tends to become mechanically linked in cations, black, red or blue is used. The choice of color
its container and is then expelled in uneven clumps. depends on the surface colors of the test objects and on the
prevailing test site lighting.
When a powder behaves in this manner, testing becomes
very slow and it is difficult to obtain a smooth application The ability to bond fluorescent dyes to magnetic pow-
over the test object surface. Powders made of spherical par- ders has produced a particle material that provides the best
ticles flow evenly and smoothly under the same conditions. possible visibility and contrast under proper lighting condi-
A dry powder must have free flowing properties for ease of tions. When test objects are examined in ultraviolet light, it
application, yet must also have an optimum shape for the is difficult not to see the light emitted by a few particles col-
greatest sensitivity in forming strong indications over weak lected at a discontinuity.
leakage fields. These two conflicting requirements can be
met by selectively blending particles of different shapes. Fluorescent particles are magnetically less sensitive than
visible particles but the reduction in magnetic sensitivity is
A specific proportion of elongated particles must be pre- more than offset by the increase in visibility and contrast.
sent for sensitivity and enough globular particles must be
added to permit smooth and uniform application. Visibilityand 'contrast of fluorescent particles are directly
related to the darkness of the testing site. In a totally dark-
Wet Method Particle Shapes ened area, even a small amount of ultraviolet energy acti-
vates fluorescent dye to emit a noticeable amount of visible
The performance of particles suspended in a liquid vehi- light. When the test site is partially darkened, the amount of
cle is not as shape dependent as that of dry particles. The required ultraviolet energy increases dramatically yet the
suspending liquid is much denser and more viscous than air; emitted visible light is only barely noticeable, especially in
the movement of particles through the liquid is slowed so the conventional yellow-green range.
that they accumulate more reliably at discontinuities.
Most military and commercial specifications require the
Because of this slower movement, wet method particles test site to be darkened to 20 lux (Ix) or 2 footcandles (ftc) or
form minute elongated aggregates. Even unfavorable less, with a minimum ultraviolet intensity of 1,000µW·cm-2 at
shapes align magnetically into elongated aggregates under the test object surface.
the influence of local, low level leakage fields. In suspen-
sion, the particles are kept dispersed by mechanical agita- Particle Mobility
tion until they flow over the surface of the magnetized
object. There is no need to add certain shapes to improve When magnetic particles are applied to the surface of a
the dispersion of the particles. magnetized object, the particles must move and collect at
the leakage field of a discontinuity in order to form a visible
Visibilityand Contrast indication. Any interference with this movement has an
effect on the sensitivity of the test. Conditions promoting or
Visibility and contrast are properties that must be con- interfering with particle mobility are different for dry and
sidered when selecting a magnetic particle material for a wet method particles.
specific testing application. Magnetic properties, size and
shape may all be favorable for producing the best indication, Dry Powder Mobility
but if an indication is formed and the inspector cannot see
it, then the test procedure has failed. Dry particles should be applied in a way that permits
them to reach the magnetized object surface in a uniform
Visibility and contrast are enhanced by choosing a parti- cloud with minimum motion. When this is properly done,
cle color that is easy to see against the test object surface. the particles come under the influence of leakage fields
The natural color of metallic powders is silver gray. The col- while suspended in air and are then said to possess three-
ors of iron oxides commonly used in wet method powders dimensional mobility. This condition can be approximated
on surfaces that are vertical or overhead.
MAGNETIC PARTICLE TESTING I 291
When particles are applied to horizontal surfaces, they When the requirement is to find extremely fine surface
settle directly onto the surface and do not have mobility in cracks, the wet method is superior, regardless of the magne-
three dimensions. Some extension of mobility can be tizing current in use. In some cases, direct current is consid-
achieved by tapping or vibrating the test object, agitating the ered advantageous because it also provides some indications
particles and allowing them to move toward leakage fields. of subsurface discontinuities. The wet method also offers
Alternating current and half-wave rectified alternating cur- the advantage of complete coverage of the object surface
rent (pulsed direct current) can give particles excellent and good coverage of test objects with irregular shapes.
mobility when compared to direct current magnetization.
Visible or Fluorescent Particles
Wet Method Particle Mobility
The decision between visible particles and fluorescent
The suspension of particles in a liquid vehicle allows particles depends on convenience and equipment. Testing
mobility for the particles in two dimensions when the sus- with visible particles can be accomplished under common
pension flows over the test object surface and in three shop lighting while fluorescent particles require a darkened
dimensions when the test object is immersed in a magnetic area and an ultraviolet light source.
particle bath.
Both wet method visible and wet method fluorescent
Wet method particles have a tendency to settle out of the tests have about the same sensitivity,but under proper light-
suspension either in the tank of the test system or on the test ing conditions fluorescent indications are much easier to see.
object surface short of the discontinuity. To be effective, wet
method particles must move with the vehicle and must Magnetic Particle Testing Processes
reach every surface that the vehicle contacts. The settling
rate of particles is directly proportional: (1) to their dimen- A test object may be magnetized first and particles
sions; and (2) to the difference between their density and applied after the magnetizing current has been stopped
the lower density of the liquid vehicle. Their settling rate is (called the residual method) or the object may be covered
inversely proportional to the liquid's viscosity. As a result, with particles while the magnetizing current is present
the mobility of wet method particles is never ideal and must (known as the continuous method). With test objects that
be balanced against the other factors important to wet have high magnetic retentivity, a combination of the resid-
method test results. ual and continuous methods is sometimes used.
Media Selection Residual Test Method
The choice between dry method and wet method tech- In the residual method, the test object is magnetized, the
niques is influenced principally by the following considera- magnetizing current is stopped and then the magnetic parti-
tions: cles are applied. This method can only be used on materials
having sufficient magnetic remanence. The residual mag-
1. Type of discontinuity (surface or subsurface): for sub- netic field must be strong enough to produce discontinuity
surface discontinuities, dry powder is usually more leakage fields sufficient for producing visible test indica-
sensitive. tions. As a rule, the residual method is most reliable for
detection of surface discontinuities.
2. Size of surface discontinuity: wet method particles are
usually best for fine or broad, shallow discontinuities. Hard materials with high remanence are usually low in
permeability, so higher than usual magnetizing currents may
3. Convenience: dry powder with portable half-wave be necessary to obtain an adequate level of residual mag-
equipment is easy to use for tests on site or in the netism. This difference between hard steels and soft steels is
field. Wet method particles packaged in aerosol spray usually not critical if only surface discontinuities are to be
cans are also effective for field spot tests. detected.
The dry powder technique is superior for locating sub- Either dry or wet method particle application can be
surface discontinuities, mainly because of the high perme- used in the residual method. With the wet method, the
ability and favorable elongated shape of the particles. magnetized test object may be immersed in an agitated bath
Alternating current with dry powder is excellent for surface of suspended magnetic particles or may be flooded with
cracks that are not too fine but this combination is of little particle suspension in a curtain spray.
value for cracks lying wholly beneath the surface.
292 I NONDESTRUCTIVETESTING OVERVIEW
In the immersion technique, the strength of discontinu- The wet continuous method requires more operator
ity indications is directly affected by the object's dwell time attention than the residual method. If bath application con-
in the bath. By leaving the object in the bath for extended tinues, even momentarily, after the current is stopped, parti-
periods, leakage fields have time enough to attract and hold cles held by a discontinuity leakage field can be washed
the maximum number of particles, even at fine discontinu- away. If there is a pause between stopping the bath applica-
ities. If the test object has high retentivity, longer dwell time tion and applying the magnetizing current, the suspension
increases the sensitivity over that of the wet continuous can drain off the test object leaving insufficient particles for
method. Note that the location of the discontinuity on the producing discontinuity indications. Careless handling of
object during immersion affects the accumulation of parti- the bath and current sequence can seriously hinder the pro-
cles. Indications are strongest on upper horizontal surfaces duction of reliable test results.
and weaker on vertical or lower horizontal surfaces.
The highest possible sensitivity for very fine discontinu-
Care must be exercised when removing the test object ities is typically achieved by the following sequence:
from the bath or particle spray. Rapid movement can literally (1) immerse the test object in the bath; (2) pass magnetizing
wash off indications held by weak discontinuity leakage fields. current through the object for a short time during immer-
sion; (3) maintain the current during removal from the bath;
Continuous Test Method (4) maintain the current during drainage of the suspension
from the test object; and (5) stop the magnetizing current.
When a magnetizing current is applied to a ferromagnetic
test object, the magnetic field rises to a maximum. Its value Conclusion
is derived from the magnetic field strength and the magnetic
permeability of the test object. When the magnetizing cur- Magnetic particle tests are effective nondestructive pro-
rent is removed, the residual magnetic field in the object is cedures for locating material discontinuities in ferromag-
always less than the field produced while the magnetizing netic objects of all sizes and configurations. It is a flexible
current was applied. The amount of difference depends on technique that can be performed under a variety of condi-
the B-H curve of the material. For these reasons, the contin- tions, using a broad range of supplementary components.
uous method, for any specific value of magnetizing current,
is always more sensitive than the residual method. Application of the magnetic particle method is decep-
tively simple - good test results can sometimes be produced
Continuous magnetization is the only method possible with little more than practical experience. In fact, the devel-
for use on low carbon steels or iron having little retentivity. opment of the technique has been almost entirely empirical
It is frequently used with alternating current on these mate- rather than theoretical.
rials because of the excellent mobility produced by alternat-
ing current. However, the method is founded on the complex princi-
ples of electromagnetics and the magnetic interactions of at
With the wet method, the surface of the test object is least three materials simultaneously. In addition, there is the
flooded with particle suspension. The bath application and critical consideration of the operator's ability to qualitatively
the magnetizing current are simultaneously stopped. The and quantitatively evaluate the results of the inspection.
magnetic field strength continues to affect particles in the
bath as it drains. In some continuous method procedures,
the magnetizing current remains on during interpretations.
MAGNETIC PARTICLE TESTING I 293
REFERENCES
1. Nondestructive Testing Handbook, second edition: 2. Hausman, E. and E. Slack. Physics, second edition.
Vol. 6, Magnetic Particle Testing. Columbus, OH: Princeton, NJ: Van Nostrand Publishing Company
American Society for Nondestructive Testing (1989). (1939).
294 I NONDESTRUCTIVE TESTING OVERVIEW
BIBLIOGRAPHY
1. ASME Boiler and Pressure Vessel Code. Section V, 13. Bray, D.E. and R.K. Stanley. Nondestructive Evalua
Article 7. New York, NY: American Society of
Mechanical Engineers. tion A Tool for Design, Manufacturing and Ser
vice. New York, NY: McGraw-Hill Publishing
2. "Failure Analysis and Prevention." Metals Handbook, Company (1989).
ninth edition. Vol. 11. Metals Park, OH: American
Society for Metals (1986): p 120-127, 245-248, 309- 14. Chedister, WC. and J. Long. "Issues of Magnetic Par-
ticle Weld Inspection." Materials Evaluation. Vol. 51,
338. No. 9. Columbus, OH: American Society for Nonde-
3. Magnetic Particle Inspection Symposium Reference
structive Testing (September 1993): p 976-977.
Book. Columbus, OH: American Society for Nonde- 15. Cheston, W.B. Elementary Theory of Electric and
structive Testing (1993).
4. Nondestructive Testing Handbook, second edition: Magnetic Fields. New York, NY: John Wiley and Sons
Volume 6, Magnetic Particle Testing. Columbus, OH: (1964).
American Society for Nondestructive Testing (1989).
5. Nondestructive Testing: Magnetic Particle, second 16. Edwards, C. A Study of Magnetic Particle Inspection.
edition. Classroom Training Handbook, CT-6-3. Fort Ph.D. dissertation. Hull, United Kingdom: Univer-
Worth, TX: General Dynamics, Convair Division FsiltayhoefrtHy,ulJl. (October 1986).
(1977). 17. "Automation and NDT." Manufacturing
6. Nondestructive Testing: Magnetic Particle, fourth Systems. Chicago, IL: Hitchcock Publishing Com-
edition. Programmed Instruction Handbook, pany (August 1985).
PI-4-3. Fort Worth, TX: General Dynamics, Convair
Division ( 1977). 18. Flaherty, J. "History of Magnetic Particle Testing."
Materials Evaluation. Vol. 48, No. 8. Columbus,
7. Ashton, C. The Fundamental Approach to Magnetic
Particle Calculations and Testing Methods in MKS OH: American Society for Nondestructive Testing
Units Using D.C. Current, second edition. Brisbane, (August 1990): p 1,010-1,022.
Australia: Copyright Publishing Company (1905).
19. Forster, F. "On the Way from the 'Know How' to the
8. Astroem, T. "Use and Misuse of Settling Measure-
ments of Magnetic Particle Suspensions." Non 'Know Why' in the Magnetic Leakage Field Method
Destructive Testing: Proceedings of the 4th European of Nondestructive Testing" (Part 1). Materials Eval
Conference [London, United Kingdom, September
1987]. Vol. 4. J.M. Farley and R.W. Nichols, eds. uation. Vol. 43, No. 10. Columbus, OH: American
London, United Kingdom: Pergamon Press (1988):
p 2,748-2,753. Society for Nondestructive Testing (September
1985): p 1,154, 1,155, 1,158-1,162. See also Part 2,
9. Bayliss, M., D. Short and M. Bax, eds. Underuiater
Inspection. London, United Kingdom: Plymouth Volume 43, No. 11 (October 1985): p 1,398, 1,400-
Ocean Projects Limited; E. and F.N. Span (1988). 1,402, 1,404.
10. Betz, C.E. Principles of Magnetic Particle Inspection, 20. Forshaw, M.E. and P.J. Mudge. "Optimisation of
first edition. Chicago, IL: Magnaflux Corporation Magnetic Particle Inspection." NonDestructive Test
(1967).
ing: Proceedings of the 4th European Conference
11. Blitz, J. "Magnetic Methods." Electrical and Mag
netic Methods of Nondestructive Testing. New York, [London, United Kingdom, September 1987]. Vol. 4.
NY: Adam Hilger (1991): p 44-88. J.M. Farley and R.W. Nichols, eds. London, United
12. Borucki, J.S. "Utilization and Future Developments Kingdom: Pergamon Press (1988): p 2,729-2,740.
of Automated Magnetic Particle Inspection Systems
in the USA." NonDestructive Testing: Proceedings of 21. Goldberg, L. Inspection of Fixed Offshore Structures:
the 4th European Conference [London, United King- Inspector/Diver Training. Columbus, OH: American
dom, September 1987]. Vol. 4. J.M. Farley and R.W.
Nichols, eds. London, United Kingdom: Pergamon Society for Nondestructive Testing (1994).
Press (1988): p 2,754-2,761.
22. Jiles, D. Introduction to Magnetism and Magnetic
Materials. London, United Kingdom: Chapman &
Hall (1991).
23. Lovejoy, D. Magnetic Particle Inspection: A Practical
Guide. London, United Kingdom: Chapman and
Hall (1993).
24. Hagemaier, D. "A Critical Commentary on Magnetic
Particle Inspection." Materials Evaluation. Vol. 41,
No. 9. Columbus, OH: American Society for Nonde-
structive Testing (August 1983): p 1,063-1,068.
MAGNETIC PARTICLE TESTING I 295
25. Hagemaier, D.J. "Quality Indicators for Magnetic 31. Level III Study Guide: Magnetic Particle Method.
Columbus, OH: American Society for Nondestruc-
Particle Inspection." Materials Evaluation. Vol. 49, tive Testing (1980).
No. 3. Columbus, OH: American Society for Nonde-
structive Testing (March 1991): p 346-350. 32. Moake, G. and R. Stanley. "Inspecting OCTG Using
26. Hopwood, T., VG. Oka, R.C. Deen. Reliability Capacitive Discharge Systems." Materials Evalua
Assessment of HighRisk Steel Bridges by Nonde tion. Vol. 41, No. 7. Columbus, OH: American Soci-
structive Test Methods. PB 89-117717/GAR. Lexing- ety for Nondestructive Testing (1983): p 779-782.
ton, KY: Kentucky Transportation Research Program
(October 1987). 33. Stanley, R. K. "Magnetic Particle Inspection: Some
27. Kountouris, I.S. and M.J. Baker. Defect Assessment: Unsolved Problems." Materials Evaluation. Vol. 52,
Analysis of the Dimensions of Defects Detected by No. 8. Columbus, OH: American Society for Nonde-
Magnetic Particle Inspection in an Offshore Structure. structive Testing (August 1994): p 917-920.
CESLIC Report OR6. London, United Kingdom:
Engineering Structures Laboratories, Department of 34. Stanley, R.K. "Magnetic Field Measurement: The
Civil Engineering, Imperial College (February 1988). Gauss Meter in Magnetic Particle Testing." Materials
28. Lovejoy, D .J. "A Practical Study of the Behaviour of Evaluation. Vol. 46, No. 12. Columbus, OH: Ameri-
Magnetic Particles in Magnetic Leakage Fields." can Society for Nondestructive Testing (November
NonDestructive Testing: Proceedings of the 4th 1988): p 1,5091,512.
European Conference [London, United Kingdom,
September 1987]. Vol. 4. J.M. Farley and R.W. 35. Sturges, D. and R. Wagner. "Factors Affecting Inten-
Nichols, eds. London, United Kingdom: Pergamon sity of Black Lights Used in FPI and MPI." Proceed
Press (1988): p 2,808-2,813. ings of the Fall Conference on Nondestructive
29. Lovejoy,D.J. "Investigation of the 'Skin Effect' When Testing. Columbus, OH: American Society for Non-
Using Alternating Current Magnetisation for Mag- destructive Testing (1980).
netic Particle Inspection (MPI)." Proceedings of the
12th World Conference on NonDestructive Testing 36. Swartzendruber, L.J. "Quantitative Problems in Mag-
netic Particle Inspection." Proceedings of the 15th
[Amsterdam, Netherlands, April 1989]. Vol. 1. J. Annual Review of Progress in Quantitative Nonde
structive Evaluation [La Jolla, CA, 1988]. Vol. 8B.
Boogaard and G.M. van Dijk, eds. Amsterdam, D.O. Thompson and D.E. Chimenti, eds. New York,
Netherlands: Elsevier (1989): p 348-351. NY: Plenum Press (1989): p 2,133-2,140.
30. Lovejoy, D.J. "Standardisation of Light Levels for
Magnetic and Penetrant Inspection." British Journal 37. Wong, L.C. "Test Parameters of Wet Magnetic Parti-
of Nondestructive Testing. Vol. 36, No. 1. Northamp- cle Inspection of Steel Billets." Materials Evaluation.
ton, United Kingdom: British Institute of Nonde- Vol. 46, No. 12. Columbus, OH: American Society
structive Testing (January 1994): p 8-9. for Nondestructive Testing (November 1988):
p 1,548-1,557
9SECTION
ACOUSTIC EMISSION TESTING
Ronnie K. Miller, Physical Acoustics Corporation, Princeton, New Jersey
Thomas F. Drouillard, Golden, Colorado (Bibliography)
298 I NONDESTRUCTIVETESTINGOVERVIEW
PART 1
FUNDAMENTALS OF ACOUSTIC EMISSION
TESTING
The Acoustic Emission Acoustic Emission Nondestructive
Phenomenon Testing
Acoustic emission is the elastic energy that is sponta- Acousticemissionexaminationis a rapidly maturing non-
neously released by materials when they undergo deforma- destructive testing method with demonstrated capabilities
tion. In the early 1960s, a new nondestructive testing for monitoring structural integrity, detecting leaks and
technology was born when it was recognized that growing incipient failures in mechanical equipment, and for charac-
cracks and discontinuities in pressure vessels could be terizing materials behavior. The first documented applica-
detected by monitoring their acoustic emission signals. tion of acoustic emission to an engineering structure was
Acousticemission is the most widelyused term for this phe- published in 19642 and all of the availableindustrial applica-
nomenon; it has also been calledstress wave emission, stress tion experience has been accumulated in the comparatively
waves, microseism, microseismic activity and rock noise. short time since then. Table l lists organizationsthat publish
standards for acoustic emission testing.
Formally defined, acoustic emission is "the class of phe-
nomena where transient elastic waves are generated by the Comparison with Other Methods
rapid release of energy from localized sources within a
material, or the transient elastic waves so generated:' 1 This Acoustic emission differs from most other nondestruc-
is a definition embracing both the process of wave genera- tive methods in two significant respects. First, the energy
tion and the wave itself. that is detected is released from within the test object rather
than being supplied by the nondestructive method, as in
Source Mechanisms ultrasonics or radiography. Second, the acoustic emission
method is capable of detecting the dynamic processes asso-
Sources of acoustic emission include many different ciated with the degradation of structural integrity. Crack
mechanisms of deformation and fracture. Earthquakes and growth and plastic deformation are major sources of acous-
rockbursts in mines are the largest naturally occurring emis- tic emission. Latent discontinuities that enlarge under load
sion sources. Sources that have been identified in metals and are active sources of acoustic emissionby virtue of their
include crack growth, moving dislocations, slip, twinning, size,location or orientation are alsothe most likely to be sig-
grain boundary sliding and the fracture and decohesion of nificant in terms of structural integrity.
inclusions. In composite materials, sources include matrix
cracking and the debonding and fracture of fibers. These Usually, certain areas within a structural system will
mechanisms typify the classical response of materials to develop local instabilities long before the structure fails.
applied load. These instabilities result in minute dynamic movements
such as plastic deformation, slip or crack initiation and prop-
Other mechanisms fall within the definition and are agation. Although the stresses in a metal part may be well
detectable with acoustic emission equipment. These below the elastic design limit, the region near a crack tip
include leaks and cavitation; friction (as in rotating bear- may undergo plastic deformation as a result of high local
ings); the realignment or growth of magnetic domains stresses. In this situation, the propagating discontinuity acts
(Barkhausen effect); liquefaction and solidification; and as a source of stress waves and becomes an active acoustic
solid-solidphase transformations. Sometimes these sources emission source.
are called secondary sources or pseudo sources to distin-
guish them from the classic acoustic emission due to Acoustic emission examination is nondirectional. Many
mechanical deformation of stressed materials. A unified acoustic emission sources appear to function as point
explanation of the sources of acoustic emission does not yet source emitters that radiate energy in spherical wavefronts.
exist. Neither does a complete analytical description of the Often, a sensor located anywhere in the vicinity of an
stress wave energy in the vicinity of an acoustic emission acoustic emission source can detect the resulting acoustic
source. However, encouraging progress has been made in emission. This is in contrast to other methods of nonde-
these two key research areas. structive testing, which depend on prior knowledge of the
ACOUSTIC EMISSION TESTING I 299
TABLE 1 . Organizations and their acoustic emission standards and practices
Abbreviation Issuing Organization Representative Standards
(Prefix) and Related Documents
MR Association of American Railroads Procedure for Acoustic Emission Evaluation of Tank
Cars and IM-10 l Tanks
API American Petroleum Institute
ASME American Society of Mechanical Engineers SPE.C 1 6A, Specification for Drill Through Equipment
ASNT American Society for Nondestructive Testing ASME Boiler and Pressure Vessel Code; RTP-1 ,
Appendix M-1 0
ASTM American Society for Testing and Materials
ANSI/ASNT CP-189, ASNT Standard for Qualification
CEN European Working Group on Acoustic Emission (EWGAE) and Certification of Nondestructive Testing
CGA Compressed Gas Association Personnel,' ASNT Recommended Practice
No. SNT-TC-1 A
DGZfP German Society for Nondestructive Testing
E 403, E 569, E 650, E 749, E 750, E 751, E 976,
DOT United States Department of Transportation E 1002, E 1067, E 1106, E 1118, E 1139, E 1211,
E 1316, E 1419, E 1495, F 914
HS United Kingdom Health and Safety Executive
NDIS Japanese Society for Non-Destructive Inspection Codes for Acoustic Emission Examination
NRC United States Nuclear Regulatory Commission C-1 8, Procedure for Acoustic Emission Retesting of
SAE Society of Automotive Engineers Jumbo Tube Trailers
SPI Society of the Plastics Industry
SE 1-89, SE 2-91, SE 3-91
8944 Code Variance (Jumbo Trailer Testing); 1 0589
Exemption (Railroad Tank Cars) from 49 CFR
173.31 (c)
G 30, Storage of Anhydrous Ammonia under Pressure
in the United Kingdom
2412, Acoustic Emission Testing of Spherical Pressure
Vessels Made of High Tensile Strength Steel and
Classification of Test Results
Regulatory Guide 1 . 1 33
J 1 232, Acoustic Emission Test Methods
Recommended Practice for Acoustic Emission Testing
of Fiberglass TanksNessels; Recommended Practice
for Acoustic Emission Testing of Fiberglass Reinforced
Plastic Piping Systems
probable location and orientation of a discontinuity in order 3. Since only limited access is required, discontinuities
to direct a beam of energy through the structure on a path may be detected that are inaccessible to the more tra-
that will properly intersect the area of interest. ditional nondestructive methods.
Advantages of Acoustic Emission Tests 4. Vessels and other pressure systems can often be
requalified during an inservice inspection that
The acoustic emission method offers the following requires little or no downtime.
advantages over other nondestructive testing methods:
5. The acoustic emission method may be used to pre-
1. Acoustic emission is a dynamic inspection method in vent catastrophic failure of systems with unknown
that it provides a response to discontinuities under sig- discontinuities, and to limit the maximum pressure
nificant imposed structural stress; discontinuities not during containment system tests.
under stress will not generate acoustic emission signals.
Acoustic emission is a wave phenomenon and acoustic
2. Acoustic emission makes it possible for the technician emission testing uses the attributes of particular waves to
to detect and evaluate the significance of discontinu- help characterize the material in which the waves are travel-
ities throughout an entire structure during a single test. ing. Frequency and amplitude are examples of the wave-
form parameters that are regularly monitored in acoustic
300 I NONDESTRUCTIVETESTING OVERVIEW
emission tests. Table 2 gives an overview of the manner by TABLE 2. Factors that affect relative amplitude of
which various material properties and testing conditions acoustic emission response3
influence acoustic emission response amplitudes.3 The fac-
tors should generally be considered as indicative rather than Factors That Tend to Factors That Tend to
as absolute.
Increase AcousticEmission DecreaseAcousticEmission
Application of Acoustic Emission
Tests ResponseAmplitude ResponseAmplitude
A classification of the functional categories of acoustic High strength low strength
emission applications is given below:
High strain rate low strain rate
1. mechanical property testing and characterization;
2. preservice proof testing, Low temperature high temperature
3. inservice (requalification) testing;
4. online monitoring; Anisotropy isotropy
5. in-process weld monitoring;
6. mechanical signature analysis; Nonhomogeneity homogeneity
7. leak detection and location; and
8. geological applications. Thick sections thin sections
By definition, online monitoring may be continuous or Brittle failure (cleavage) ductile failure (shear)
intermittent and may involve the entire structure or a lim-
ited zone only. Although leak detection and acoustic signa- Material containing material without
ture analysis do not involve acoustic emission in the strictest
sense of the term, acoustic emission techniques and equip- discontinuities discontinuities
ment are used for these applications.
Martensitic phase diffusion controlled phase
Structures and Materials
transformations transformations
A wide variety of structures and materials (metals, non-
metals and various combinations of these) can be monitored Crackpropagation plastic deformation
by acoustic emission techniques during the application of an
external load. The primary acoustic emission mechanism Cast materials wrought materials
varies with different materials and should be characterized
before applying acoustic emission techniques to a new type Largegrain size smallgrain size
of material. Once the characteristic acoustic emission
response has been defined, acoustic emission tests can be Mechanicallyinduced twinning thermally induced twinning
used to evaluate the structural integrity of a component.
and locating areas of matrix damage, fiber breakage, delami-
Testing of Composites nations and other types of structural degradation.
Acoustic emission monitoring of fiber reinforced com- Pressure System Tests
posite materials has proven quite effective when compared
with other nondestructive testing methods. However, atten- Pressure systems are stressed using hydrostatic or some
uation of the acoustic emission signals in fiber reinforced other pressure test. Bending loads can be introduced to
materials presents unique problems. Effective acoustic emis- beamed structures. Torsional loads can be generated in
sion monitoring of fiber reinforced components requires rotary shafts. Thermal loading may be created locally. Ten-
much closer sensor spacings than would be the case with a sion and bending loads should either be unilateral or cyclic
metal component of similar size and configuration. With the to best simulate service induced loads.
proper number and location of sensors, monitoring of com-
posite structures has proven highly effective for detecting Successful Applications
Examples of proven applications for the acoustic emis-
sion method include the following:
1. periodic or continuous monitoring of pressure vessels
and other pressure containment systems to detect
and locate active discontinuities;
2. detection of incipient fatigue failures in aerospace
and other engineering structures;
3. monitoring materials behavior tests to characterize
various failure mechanisms;
4. monitoring fusion or resistance weldments during
welding or during the cooling period; and
5. monitoring acoustic emission response during stress
corrosion cracking and hydrogen embrittlement sus-
ceptibility tests.
ACOUSTIC EMISSION TESTING I 301
Acoustic Emission Testing be used to isolate the sensor from the environment. This is a
Equipment convenient alternative to the use of high temperature sen-
sors. Waveguides have also been used to precondition the
Equipment for processing. acoustic emission signals is acoustic emission signals as an interpretation aid.
available in a variety of forms ranging from small portable
instruments to large multichannel systems. Components Issues such as wave type and directionality are difficult to
common to all systems are sensors, preamplifiers, filters and handle in this technology because the naturally occurring
amplifiers to make the signal measurable. Methods used for acoustic emission contains a complex mixture of wave modes.
measurement, display and storage vary more widely accord-
ing to demands of the application. Figure 1 shows a block Preamplifiers and Frequency Selection
diagram of a generic four-channel acoustic emission system.
The preamplifier must be close to the sensor. Often it is
Acoustic Emission Sensors actually incorporated into the sensor housing. The pream-
plifier provides required filtering, gain (most commonly
When an acoustic emission wavefront impinges on the 40 dB) and cable drive capability. Filtering in the preampli-
surface of a test object, very minute displacements of the fier (together with sensor selection) is the primary means of
surface molecules occur. A sensor's function is to detect this defining the monitoring frequency for the acoustic emission
mechanical displacement and convert it into a specific, test. This may be supplemented by additional filtering at the
usable electric signal. The sensors used for acoustic emis- mainframe.
sion testing often resemble an ultrasonic search unit in con-
figuration and generally use a piezoelectric transducer as Choosing the monitoring frequency is an operator func-
the electromechanical conversion device. tion because the acoustic emission source is essentially wide
band. Reported frequencies range from audible clicks and
The sensors may be resonant or high fidelity. The main squeaks up to 50 MHz.
considerations in sensor selection are (1) operating fre-
quency, (2) sensitivity and (3) environmental and physical Although not always fully appreciated by operators, the
characteristics. For high temperature tests, waveguides may observed frequency spectrum of acoustic emission signals is
significantly influenced by the resonance and transmission
characteristics of both the specimen (geometry as well as
acoustic properties) and the sensor. In practice, the lower
FIGURE 1 . Schematic diagramof basic four-channelacoustic emission testingsystem
PREAMPLIFIER$ MAIN AMPLIFIERS MEASUREMENT
WITH WITH FILTERS CIRCUITRY
FILTERS
SENSORS DISK STORAGE
SCREEN
DISPLAY
DATA MICROCOMPUTER
BUFFERS
OPERATOR
KEYBOARD PRINTER
AND
GRAPHICS
COPY
302 I NONDESTRUCTIVETESTING OVERVIEW
frequency limit is governed by background noise; it is distribution functions, crossplots of one signal descriptor
unusual to go below 10 kHz except in microseismic work against another and source location plots. Installed systems
The upper frequency limit is governed by wave attenuation of this type range in size from 4 to 128 channels.
that restricts the useful detection range; it is unusual to go
above 1 MHz. The most common frequency range for acous- Operator Training and System Uses
tic emission testing is 100 to 200 kHz.
Microcomputer based systems are usually very versatile,
System Mainframe allowing data filtering (to remove noise) and extensive post-
test display capability (to analyze and interpret results). This
The first elements in the mainframe are the main ampli- versatility is a great advantage in new or difficult applica-
fiers and thresholds, which are adjusted to determine the tions but places high demands on the knowledge and tech-
test sensitivity. Main amplifier gains in the range of 20 to nical training of the operator. Other kinds of equipment
60 dB are most commonly used. Thereafter, the available have been developed for routine industrial application in
processing depends on the size and cost of the system. In a the hands of less highly trained personnel.
small portable instrument, acoustic emission events or
threshold crossings may simply be counted and the count Examples are the systems used for bucket truck testing
then converted to an analog voltage for plotting on a chart (providing preprogrammed data reports in accordance with
recorder. In more advanced hardware systems, provisions recommended practices) and systems for resistance weld
may be made for energy or amplitude measurement, spatial process control (these are inserted into the current control
filtering, time gating, transient recorders and spectrum ana- system and terminate the welding process automatically as
lyzers, and automatic alarms. soon as expulsion is detected).
Acoustic Emission System Accessories Acoustic emission equipment was among the first non-
destructive testing equipment to make use of computers in
Accessory items often used in acoustic emission work the late 1960s. Performance, in terms of acquisition speed
include oscilloscopes, magnetic tape recorders, root-mean- and real time analysis capability, has been much aided by
square voltmeters, special calibration instruments and spe- advances in microcomputer technology. Trends expected in
cial devices for simulating acoustic emission. A widely the future include advanced kinds of waveform analysis,
accepted simulator is the Hsu-Nielsen source, a modified more standardized data interpretation procedures and more
draftsman's pencil that provides a remarkably reproducible dedicated industrial products.
simulated acoustic emission signal when the lead is broken
against the test structure. Characteristics of Acoustic Emission
Techniques
Microcomputers in Acoustic
Emission Test Systems The acoustic emission test is a passive method that moni-
tors the dynamic redistribution of stress/strain levels at or
Signal Processing and Displays immediately adjacent · to latent discontinuities or other
anomalies within a material or component. Therefore, acous-
Nearly all modern acoustic emission systems use micro- tic emission monitoring is only effective while the material or
computers in various configurations, as determined by the structure is subjected to an applied load. Examples of these
system size and performance requirements. In typical stresses include pressure testing of vessels or piping, and ten-
implementations, each acoustic emission signal is measured sion loading or bend loading of structural components.
by hardware circuits and the measured parameters are
passed through the central microcomputer to a disk file of Irreversibility and the Kaiser Effect
signal descriptions. The customary signaldescription includes
the threshold crossing counts, amplitude, duration, rise time An important feature affecting acoustic emission appli-
and often the energy of the signal, along with its time of cations is the generally irreversible response from most
occurrence and the values of slowlychanging variables such metals. In practice, it is often found that once a given load
as load and background noise level. has been applied and the acoustic emission from accommo-
dating that load has ceased, additional acoustic emission
During or after data recording, the system extracts data will not occur until that load level is exceeded, even if the
for graphic displays and hardcopy reports. Common displays load is completely removed and then reapplied. This often
include history plots of acoustic emission versus time or load, useful (and sometimes troublesome) behavior has been
ACOUSTIC EMISSION TESTING I 303
named the Kaiser effect in honor of the researcher who first For most engineering structures, sensor selection and
reported it. placement must be carefully chosen based on a detailed
knowledge of the acoustic properties of the material and the
The degree to which the Kaiser effect is present varies geometric conditions that will be encountered. For exam-
between metals and may even disappear completely after ple, the areas adjacent to attachments, nozzles and penetra-
several hours (or days) for alloys that exhibit appreciable tions or areas where the section thickness changes usually
room temperature annealing (recovery) characteristics or require additional sensors to achieve adequate coverage.
metals fatigued by repeated loading. Some alloysand materi- Discontinuities in such locations often cause high localized
als may not exhibit any measurable Kaiser effect. stress and these are the areas where maximum coverage is
needed.
Acoustic Emission Test Sensitivity
Interpretation of Test Data
Although the acoustic emission method is quite sensi-
tive, compared with other nondestructive methods such as Proper interpretation of the acoustic emission response
ultrasonic testing or radiographic testing, the sensitivity obtained during monitoring of pressurized systems and
decreases with increasing distances between the acoustic other structures usually requires considerable technical
emission source and the sensors. The same factors that knowledge and experience with the acoustic emission
affect the propagation of ultrasonic waves also affect the method. Close coordination is required between the acous-
propagation of the acoustic (stress) waves used in acoustic tic emission system operators, the data interpretation per-
emission techniques. sonnel and those controlling the process of loading the
structure.
Wave mode conversions at the surfaces of the test object
and other acoustic interfaces, combined with the fact that Because most computerized, multichannel acoustic
different wave modes propagate at different velocities, are emission systems handle response data in a pseudo batch
factors that complicate analysis of acoustic emission procedure, an intrinsic dead time occurs during data trans-
response signals and produce uncertainties in calculating fer. This is usually not a problem but can occasionally result
acoustic emission source locations with triangulation or in analysis errors when the quantity of acoustic emission sig-
other source locating techniques. nals is sufficient to overload the data handling capabilities of
the acoustic emission system.
Background Noise and Material Properties
Compensating for Background Noise
In principle, overall acoustic emission system sensitivity
depends on the sensors as well as the characteristics of the When acoustic emission monitoring is used during
specific instrumentation system. In practice, however, the hydrostatic testing of a vessel or other pressure system, the
sensitivity of the acoustic emission method is often primarily acoustic emission system will often provide the first indica-
limited by ambient background noise considerations for tion of leakage. Pump noise and other vibrations, or leakage
engineering materials with good acoustic transmission char- in the pressurizing system, can also generate background
acteristics. noise that limits the overall system sensitivity and hampers
accurate interpretation.
When monitoring structures made of materials that
exhibit high acoustic attenuation from scattering or absorp- Special precautions and fixturing may be necessary to
tion, the acoustic properties of the material usually limit the reduce such background noise to tolerable levels. Acoustic
ultimate test sensitivity and will certainly impose limits on emission monitoring of production processes in a manufac-
the maximum sensor spacings that can be used. turing environment involves special problems related to
high ambient noise (both electrical and acoustical). Preven-
Effects of System Sensors tive measures may be necessary to provide sufficient electri-
cal or acoustical isolation to achieve effective acoustic
Sensor coupling and reproducibility of response are emission monitoring.
important factors that must be considered when applying
multiple acoustic emission sensors. Careful calibration Various procedures have been used to reduce the effects
checks should be performed before, after and sometimes of background noise sources. Included among these are
during the acoustic emission monitoring process to ensure mechanical and acoustic isolation; electrical isolation; elec-
that all channels of the instrumentation are operating prop- tronic filtering within the acoustic emission system; modifi-
erly at the correct sensitivities. cations to the mechanical or hydraulic loading process;
special sensor configurations to control electronic gates for
304 I NONDESTRUCTIVETESTINGOVERVIEW
noise blocking; and statistically based electronic counter- the event of discontinuity growth during a working period,
measures including autocorrelation and cross correlation. subsequent proof loading would subject the material at the
discontinuity to higher stresses than before and the disconti-
The Kaiser Effect nuity would emit. Emission during the proof loading is
therefore a measure of damage experienced during the pre-
Josef Kaiser-is credited as the founder of modern acous- ceding working period.
tic emission technology. Albert Dietz noted the phe-
nomenon in 1941 and related it to fiber breakage and This so-called Dunegan corollary became a standard
change in stress-strain properties in timber.4 But it was diagnostic approach in practical field testing. Field opera-
Kaiser's pioneering work in Germany in the 1950s that trig- tors learned to pay particular attention to emission between
gered a connected, continuous flow of subsequent develop- the working pressure and the proof pressure, and thereby
ment. He made two major discoveries. The first was the made many effective diagnoses. A superficial review of the
near universality of the acoustic emission phenomenon. He Kaiser effect might lead to the conclusion that practical
observed emission in all the materials he studied. The sec- application of acoustic emission techniques requires a series
ond was the effect that bears his name. of ever increasing loadings. However, effective engineering
diagnoses can be made by repeated applications of the same
Intranslation of his own words: "Tests on various materi- proof pressure.
als (metals, woods or minerals) have shown that low level
emissions begin even at the lowest stress levels (less than The Felicity Effect
1 MPa or 100 lbrin.-2). They are detectable all the way
through to the failure load, but only if the material has expe- The second major application of the Kaiser effect arose
rienced no previous loading. This phenomenon lends a spe- from the study of cases where it did not occur. Specifically in
cial significance to acoustic emission investigations, because fiber reinforced plastic components, emission is often
by the measurement of emission during loading a clear con- observed at loads lower than the previous maximum, espe-
clusion can be drawn about the magnitude of the maximum cially when the material is in poor condition or close to fail-
loading experienced before· the test by the material under ure. This breakdown of the Kaiser effect was successfully
investigation. Inthis, the magnitude and duration of the ear- used to predict failure loads in composite pressure vessels7
lier loading and the time between the earlier loading and and bucket truck booms.8
the test loading are of no importance."5
The term felicity effect was introduced to describe the
This effect has attracted the attention of acoustic emission breakdown of the Kaiser effect and the felicity ratio was
workers ever since. In fact, all the years of acoustic emission devised as the associated quantitative measure. The felicity
research have yielded no other generalization of comparable ratio is the stress at onset of emission divided by the previous
power. As time went by, both practical applications and con- maximum stress and serves as an indication of the severity of
troversial exceptions to the rules were identified. a discontinuity. The felicity ratio has proved to be a valuable
diagnostic tool in one of the most successful of all acoustic
The Dunegan Corollary emission applications, the testing of fiberglass vessels and
storage tanks.9 Infact, the Kaiser effect may be regarded as a
The first major application of the Kaiser effect was a test special case of the felicity effect (a felicity ratio of 1).
strategy for diagnosing damage in pressure vessels and other
engineering structures.6 The strategy included a clarifica- The discovery of cases where the Kaiser effect breaks
tion of the behavior expected of a pressure vessel subjected· down was at first quite confusing and controversial but
to a series of loadings (to a proof pressure with intervening eventually some further insights emerged. The Kaiser effect
periods at a lower working pressure). fails most noticeably in situations where time dependent
mechanisms control the deformation. The rheological flow
Should the vessel suffer no damage during a particular or relaxation of the matrix in highly stressed composites is a
working period, the Kaiser effect dictates that no emission prime example. Flow of the matrix at loads below the previ-
will be observed during the subsequent proof loading. In ous maximum can transfer stress to the fibers, causing them
to break and emit. Other cases where the Kaiser effect will
fail are corrosion processes, low cycle fatigue and hydrogen
embrittlement, which are also time dependent.
ACOUSTIC EMISSION TESTING I 305
PART 2
BUCKET TRUCK AND LIFT INSPECTION
A complete list of acoustic emission testing applications FIGURE 2. Components of extensible aerial
would be too extensive to present in this overview. In an device monitored with acoustic emission 15
attempt to cover some of this spectrum, several examples
are given. The majority fall into the category of proof testing
although one example of process monitoring is also given.
If an aerial equipment or bucket truck fails mechanically
or structurally it could result in injury or death to workers
using the equipment. Many utilities have used regular test
and inspection programs to ensure the safety of their line
personnel and to minimize equipment repair costs consis-
tent with safe operation.
Accidents, overloads and fatigue can occur during opera-
tion. Discontinuities in original materials also can cause
problems. A thorough, regularly scheduled inspection can
identify potential problems before they cause injuries or
downtime. Furthermore, discontinuities discovered early
are less expensive to repair than if left to develop.!"
Insulated aerial devices can be divided into several cate-
gories: (1) extensible boom aerial personnel devices;
(2) articulating boom aerial personnel devices; and (3) a
combination of these two types. Figures 2 and 3 show the
first two types of aerial devices or bucket trucks. Table 3
indicates which of their components may be inspected with
acoustic emission methods.
Acoustic Emission Inspection 7
Development
LEGEND BEARING
When acoustic emission techniques were being devel- I. PLATFORM
oped for metal and composite materials, several utilities, 2. PLATFORM ATTACHMENT
bucket truck manufacturers and independent testing com- 3. UPPER BOOM
panies began designing acoustic emission test methods. 4. INTERMEDIATE BOOM
Acoustic emission from metals such as steel is a result of 5. LOWER BOOM
crack formation and propagation. On the other hand, acous- 6. TURNTABLE AND ROTATION
tic emission from fiber reinforced plastics, used for plat- 7. PEDESTAL
forms of bucket trucks, indicates matrix cracking or fiber
breakage. Work on acoustic emission inspection of bucket FROM F 914-85, AMERICAN SOCIETYFOR TESTINGAND MATERIALS.
trucks can be found in the literature.l'U? REPRINTEDWITH PERMISSION.
The American Society for Testing and Materials (ASTM)
issued standard F 914, Standard Test Method for Acoustic
Emission for Insulated AerialPersonnel Devices in 1985.15
The ASTM standard method and most past applications use
the acoustic emission technique for zone location of discon-
tinuities. Verification of emission sources may require use of
other nondestructive test methods such as radiographic,
306 I NONDESTRUCTIVE TESTING OVERVIEW
TABLE3. Aerial device componentsmonitored According to ASTM F 914, the amplitude of an acoustic
with acousticemission techniques emission signal at the output of a sensor is expressed in
terms of decibels (dB):
Articulated Aerial Device Extensible Aerial Device
Components Components
Pedestal pedestal signal peak amplitude (Eq. l)
Platform platform
platform attachment Where:
Platform attachment upper boom
Upper boom turntable and rotation bearing A0 one microvolt as the reference voltage; and
Turntableand rotation intermediate boom A 1 peak voltage of the measured acoustic emission
lower boom
bearing signal at the sensor output.
Lower boom insulator
Upper and lower boom To distinguish significant events, an instrument usually
has at least two threshold levels for inspections. A high
attachment of elbow
Upper and lower section
of lower boom
FIGURE 3. Componentsof articulatingaerial
device monitoredwith acoustic emission15
ultrasonic, magnetic particle, liquid penetrant, leak and
visual testing.
Described below are the acoustic emission methods most
commonly used for bucket truck inspection. Some refined
techniques for other applications may be valid as well.
Instrumentationfor Bucket Truck LEGEND
Inspection PLATFORM
PLATFORM ATTACHMENT
Because sound attenuation for a given frequency in UPPER BOOM
fiberglass composite material is higher than that in metal,
the acoustic emission sensor frequency for the fiber rein- 4. UPPER AND LOWER BOOM ATTACHMENT OF ELBOW
forced plastic component of bucket trucks should be lower 5. UPPER AND LOWER SECTION OF LOWER BOOM
than that for use in metal. The transducer frequency for 6. LOWER BOOM INSULATOR
these components is usually in the range of 20 to 75 kHz for 7. TURNTABLE AND ROTATION BEARING
acoustic emission waves that travel sufficiently long dis- 8. PEDESTAL
tances in the components. The frequency range from 100 to
400 kHz is usually suitable for detection of acoustic emis- FROM F 914-85, AMERICAN SOCIETYFOR TESTINGAND MATERIALS.
sion in metal components for bucket truck inspections. REPRINTEDWITH PERMISSION.
During proof load testing of aerial devices, a number of
channels are needed to monitor individual components
simultaneously. Typicalacoustic emission monitoring systems
have eight to sixteen independent channels. A typical sixteen
channel system will have ten to twelve high frequency chan-
nels for the metal components and four to six low frequency
channels for the fiber reinforced plastic components.
A plot or display of acoustic emission events versus mon-
itoring channels indicates which sensor is closest to the
emission source. Most commercial instruments also have a
first-hit circuit which registers the first channel when sev-
eral sensors detect the same event.
ACOUSTIC EMISSION TESTING I 307
threshold level for 70 dB peak signals at the sensor output 1. The method of load application and attachment shall
indicates serious damage, such as fiber fracture and metal evenly distribute the load and shall not permanently
cracking. The ASTM standard also uses a 40 dB threshold deform the platform. The load system is attached to
level for the acceptance criteria in terms of acoustic emis- the bucket or platform so that the centerline of load
sion counts. application passes through the centerline of the
bucket. On units with two platforms, the load must be
Test Procedure for Bucket Truck distributed to both platforms. On units with platform
Inspection rotators, the platform shall be rotated full around
toward the end of the boom.
The test configurations for the proper load of articulated
and extensible aerial devices are shown in Figs. 4 and 5. 2. The load connection technique should simulate in-
Most aerial devices have two test configurations for com- field use as closely as possible.
plete monitoring of all potential damage.
3. The test load employed shall be two times the rated
A load measuring device such as an electronic load cell platform capacity.
must be attached to the load application system, which in
tum is attached to an adequate dead weight or anchor. The loading sequence is illustrated in Fig. 6. Before
ASTM F 914 recommends the following method to provide applying the load, possible background noise picked up by
sufficient load on the aerial lift. each acoustic emission sensor should be observed for two
minutes. Potential background noise for bucket truck test-
ing includes electromagnetic interference, radio frequency
interference, improper grounding, rubbing interfaces and
FIGURE 4. Test positions for articulated aerial FIGURE 5. Test positions for extensible aerial
devices: (a) over center model (upper boom devices15
travels past vertical); and (b) non-over center
model (upper boom does not travel past LJfa)
vertical) 15 •
fa) ... fb) LOAD
LOAD
...
LOAD
fb) ...
LOAD
FROM F 914-85, AMERICAN SOCIETYFOR TESTINGAND MATERIALS. FROM F 914-85, AMERICAN SOCIETYFOR TESTINGAND MATERIALS.
REPRINTEDWITH PERMISSION. REPRINTEDWITH PERMISSION.
308 I NONDESTRUCTIVETESTING OVERVIEW
impact. If the background noise is excessive and cannot be The high amplitude events indicate serious damage such as
eliminated, the test may be rescheduled or relocated. fiber fracture or metal cracking. The low amplitude events
may result from other discontinuities such as insufficient
Loading starts with a uniform rate, between 15 and lubricating and nonpropagating cracks. These low to
45 N-s-1 (3 and 10 lb-s'), until the maximum test load is medium amplitude events (40 to 70 dB) are detected during
attained. The load is held at maximum for four minutes, the loading periods but are quiet at a constant load.
then removed at a constant rate between 15 and 90 N-s-1
(3 and 20 lbrs-1). After another four-minute hold period at The events shown in Fig. 7a indicate that high amplitude
zero load, the load cycle is repeated. emissions were detected at channels 2, 4, 6 and 7. Because
there were no high amplitude events detected by the sen-
During the loading, unloading and holding at maximum sors of channels 3 and 5, the sensors of channels 2, 4 and 6
load and zero load, events, counts, amplitude distributions must pick up those events independently. The events shown
and location data for all metal and fiberglass channels are on channel 7 might be the acoustic emission pulses propa-
recorded versus time and load. If any acoustic emission test gating from the component with the channel 6 sensor to the
results indicate that damage could be occurring to the aerial channel 7 sensor component through acoustic coupling of
device, the test should be stopped. the two components. This interpretation can be confirmed
through the first hit registration records.
Typical Test Data
After subtracting the high threshold events from the
Figure 7 shows typical sets of data that represent accu- events in Fig. 7b, channels 1, 3, 5, 8, 9, 10, 12, 14 and 15
mulations of events in each channel.14 The chart in Fig. 7a is detected low amplitude emissions. Only channels 9, 12 and
obtained for those events that exceed the high amplitude 15 definitely picked up low amplitude emissions indepen-
threshold at 70 dB. Figure 7b is for the low threshold at dently, because channel 9 had more events than adjacent
40 dB. The total events in each channel in Fig. 7b should be channels 8 and 10; channel 12 had a significant amount of
subtracted by the high amplitude events in the correspond- events and no event in either channels 11 and 13; and chan-
ing channel in Fig. 7a to obtain the low amplitude events. nel 15 had more events than channel 14. However, channels
1 and 3, 5 and 7 detected low amplitude emissions possibly
coming from the same emission sources in the components
with sensor 2 and 6, respectively.
FIGURE 6. Aerial device loading sequence15
FIRST LOADING SECOND LOADING
TEST LOAD
0d t3 t4 ts TIME
.....J ~ TWO FOUR MINUTES MINIMUM TWO TWO
TWO MINUTES MINUTES MINUTES
oz MINUTES
~~un::: START
OF
TEST
t
r, t2
L_,,)
TWO MINUTES
FROM F 914-85, AMERICAN SOCIETYFOR TESTINGAND MATERIALS. REPRINTEDWITH PERMISSION.
ACOUSTIC EMISSION TESTING I 309
There were two possible reasons for this: (1) the source Acceptance Criteria
that emitted high amplitude acoustic emission also emitted The ASTM Standard F 914-85 provides the following
guidelines as the acceptance criteria for the acoustic emis-
low amplitude emissions; and (2) the high amplitude emis- sion testing of insulated aerial personnel devices:
sion pulses became weaker through the joints between two 1. less than 150 total events first hit, with amplitude
greater than 70 dB on fiber reinforced plastic and
components. Therefore, the exact number of either high or metal channels combined; and
low amplitude events generated in one component should 2. less than 15,000 total counts first hit, with amplitude
greater than 40 dB on fiber reinforced plastic and
be obtained in cooperation with the first hit registrations. metal channels combined.
Figure 8 shows two amplitude distributions obtained The above guidelines are only applied to an entire device
in accordance with either total events or total counts in all
from a metal component with a high frequency sensor and test channels. There are no guidelines for a single compo-
nent of either metal or fiber reinforced plastic.
from a fiber reinforced plastic component with a low fre-
quency sensor.14 The histograms represent the number of
events in every 5 dB interval, started at the 40 dB low
threshold level. The total number of events in one his-
togram should equal the number of events in the corre-
sponding channel in the events versus channels chart for the
low threshold level.
Count histograms of all channels for either high or low
threshold are similar to those of event histograms as shown
in Fig. 8.
FIGURE7. Histogramsof events versus FIGURE8, Histogramsof events versus
channel:(aJ events versuschannelat 70 dB amplitude:(aJ at high frequency;(b) at low
high threshold;(bJ events versuschannelat frequency14
40 dB low threshold14
(a) z~ 100
fa)
~
zLI.J
0
vi
~V1
LI.J
u
i=
:V:1:,
0
10 12 14 u<( o~~~~~~----..-A----_.._...__....._.__..__
46 16 20 40 60 80 I 00
CHANNEL AMPLITUDE
(b) (b)
6 8 10 12 14 16 z~ 100 40 60 80 100
~ AMPLITUDE
CHANNEL zLI.J
0
vi
~V1
LI.J
u
i=
V::1:,
0
~0
20
31 0 I NONDESTRUCTIVE TESTING OVERVIEW
PART 3
ACOUSTIC EMISSION TESTS OF FIBER
REINFORCED PLASTIC VESSELS
Testing Procedures for Pressure, the easiest procedure to reference is probably that code's
Storage and Vacuum Vessels Article 11, Section V, which describes the requirements for
acoustic emission examination of new and inservice fiber
Three organizations have accepted a procedure (with reinforced plastic (FRP) vessels under pressure, vacuum or
minor variations) for the examination of fiber reinforced other applied stress.
plastic vessels. The originator of the procedure was the Soci-
ety of the Plastics Industry (SPI), which developed proce- Figures 9 through 13 exemplify the stressing sequence
dures for glass reinforced vessels. Both the American Society and test algorithm flow chart. Table 4 lists evaluation criteria.
for Testing and Materials (ASTM) and the American Society
of Mechanical Engineers (ASME) adopted similar examina- Applications in the Chemical
tion procedures. In addition to fiberglass, the ASME proce- Industries
dure applies to vessels reinforced with other fibers.
Tests of Composite Vessels
The ASME Boiler and Pressure Vessel Code typically Use of acoustic emission for testing fiber reinforced plas-
has state-passed legislative standing instead of serving as
voluntarily applied consensus standards. Because of this, tic vessels has been extremely successful and the technique
TABLE 4. Evaluation criteria for atmospheric (liquid head) and above-atmospheric superimposed
pressures in fiber rei~forced plastic vessels
Criteria First Loading Subsequent Loading Comments
Emissionsduring hold less than EH events beyond less than EH events beyond measure of continuing permanent
Felicity ratio darnaqe?
time TH, none having time TH
measure of severity of previous
amplitude greater than AM value" induced damage
greater than felicityratio FA greater than felicity ratio FA measure of overall damage
during a load cycle
not excessive" less than Ne total counts
measure of delamination, adhesive
no events with duration no events with duration bond failure and major crack
greater than M greater than M growth
Number of events greater less than EA events less than EA events measure of high energy
microstructure failures (this
than reference criterion is often associated with
amplitude threshold fiber breakage)
~' EA- EH, FA, Ne AND M ARE ACCEPTANCE CRITERIAVALUES SPECIFIED BY THE REFERENCING CODE; TH IS SPECIFIED HOLD TIME.
a. SEE APPENDIX 11-1140 FOR DEFINITION OF ~.
b. PERMANENT DAMAGE CAN INCLUDE MICROCRACKING, DEBONDING AND FIBER PULL OUT.
c. VARIES WITH INSTRUMENTATION MANUFACTURER. NOTE THAT COUNTS CRITERION Ne MAY BE DIFFERENT FOR FIRSTAND SUBSEQUENT FILLINGS.
d. EXCESSIVE COUNTS ARE DEFINED AS A SIGNIFICANT INCREASE IN THE RATE OF EMISSIONS AS A FUNCTION OF LOAD. ON A PLOT OF COUNTS
AGAINST LOAD, EXCESSIVE COUNTS WILL SHOW AS A DEPARTURE FROM LINEARITY.
e. IF USED, VARIES WITH INSTRUMENTATION MANUFACTURER.
ACOUSTIC EMISSION TESTING I 311
has become a standard test for fiber reinforced plastic equipment fabricators, acoustic emission instrument manu-
facturers, fiber reinforced plastic designers and vessel own-
equipment and piping. ers was a major factor in the rapid development and ready
acceptance of these test procedures. Initially, the research
In the chemical process industry, performance of fiber was performed to meet the needs of the chemical process
industry. However, the technology has been adopted by
reinforced plastic pipe and vessels has been poor and many other industries including aerospace, automotive, utility
failures have been reported.16 Apart from acoustic emission, (lifts) and sports equipment manufacturers. Acoustic emis-
there is no satisfactory test for determining structural ade- sion testing is accepted as the primary technique for evalu-
quacy of fiber reinforced plastic equipment. Accordingly, ating most composites.
the technique has filled an important need so that develop-
ment and application of the technology has been rapid. Figure 14 shows catastrophic fiber reinforced plastic
tank failures of one major company during a twelve-year
Cooperative research and development programs orga- period.19 The rate of more than two failures per year is
nized by the Society of the Plastics Industry have resulted in
practical field test methods for tanks and piping.17.18 Broad
support from resin and glasssuppliers, fiber reinforced plastic
FIGURE 9. Atmospheric vessels stressing sequence; Dn represents data recording points
00 00 lJJ
I Iz
100.0 - ~ ~! I0 z~02 Cuc.o5, z1::-Jo_,
::::i ~~
2
0 _,
f0 I0
Mcz c<z(
87.5 g cs ~I
___
80.0 s ~i :I l jo aI,1
75.0 00 0
I
d0 60.0 Iz f
[2 lJJ
-.l <(
Q::'. co !/ I~ ~ Iz1:::-:-i
r5/
Oz- 0 v
~ fD - I
I
~ ~ 50.0
I
VUlJ-Q_ I
Q::'.
n, 40.0
f-
VfU-lJ
20.0 .- .B. A10i~T~OC3N0o1 /II
~~i~E II
0 ..D_ETERMINATIO..N..;.1
D1 D1 TIME
312 I NONDESTRUCTIVE TESTING OVERVIEW
clearly unacceptable. These major failures were accompa- of the nature of composites have also contributed to the
nied by many minor failures including leaks, cracks at noz- reduction in failures.
zles, breakage of holddown lugs and attachments, internal
surface cracking and blistering. As a result, many tanks were The failure in 1982 in Fig. 14 occurred in a tank that had
been tested with acoustic emission. The test showed the
unsuitable for their intended function. tank to be unsafe for continued operation and a replace-
Reasons for this poor performance are varied but they ment tank was ordered. In the interim, the tank was oper-
ated at reduced levels and appropriate precautions were
can be grouped into the following general categories: inade- taken to protect against the possible failure. Failure pre-
quate design; variability of fabrication quality; transporta- dicted by the acoustic emission tests occurred during the
tion, handling, storage and installation damage; inservice period following the acoustic emission test and before
abuse; and corrosion attack Compounding these problems, installation of the replacement vessel.
the lack of an adequate test method permitted discontinu-
ities to go undetected until they reached serious propor- Composite Pipe Testing
Applications
tions.
Some companies test all of their fiber reinforced plastic In addition to fiber reinforced plastic vessel tests, several
hundred tests have been conducted on fiber reinforced plas-
tanks as part of the initial purchase specification. The tanks tic pipe sections. For both tanks and piping, discontinuities
are retested after installation and at intervals not exceed-
ing two years. The dramatic reduction in catastrophic fail-
ures shown in Figure 14 corresponds closely with the
implementation of acoustic emission testing in 1980.
Other factors such as improved design methods, better
control of fabrication quality and a greater understanding
FIGURE 1 o. Vacuum vesselsstressingsequence; Dn represents data recordingpoints
0
JO TO 30 ~
i.BACMKINGURTOEUN]D' 00
NOISE \ I
- BASELINE .
::;,;;
DETERMINATION \ z~9 :::;J,;;
\ 00
20 \ 0
I
~ ~z zuiui
d0 \\~o~I t:-J-
:t-J-
~__J -40 -e- 9
00
U0-J;:;c- \0 0
:~:) 2(JJ ,1 l);I I I
~ \0w \wv (Il 9 ::;,;; 9
QV) (JJ -60 :J 00
0 0
V) I z~::;,;; I I
w ::;,;;
~UJ - w :J ~
~zt:-J-
0.... ~z:t-J- ~z::;,;; u:::
I- 9 ::;,;;'
w :i
V) 0
I ~zt:-J- z~::;,;;
UJ w 0
I- w
~z:t-J- z- 0,;- .....J
-80 y -I 0 ~zt:-i-
6 O
0 0
TIME I clw5o M
I
\6 I
~z:tu-J-i
\1
IAE<EH
JOO
ACOUSTIC EMISSION TESTING I 31 3
FIGURE 11 . Flow chart for acoustic emission testing of atmospheric vessels
RECORD RECORD
AE
DATA CONTINUE ETC. AE
DURING STRESSING DATA
HOLD DURING FINAL
4 MINUTES 100 PERCENT
STRESS LEVEL
30 MINUTE
RECORD HOLD
AE
CONTINUE
STRESSING DATA
DURING
STRESSING
NO
NO
USE
RECORDAE
DATA DURING
STRESSING
EVALUATE
DISCONTINUITIES
REJECT
REPAIR RETEST
REDUCE BACKGROUND TO
ACCEPTABLE LEVEL
314 I NONDESTRUCTIVE TESTING OVERVIEW
identified by acoustic emission have been confirmed in all Information of the type shown in Fig. 15 and research
stated cases by subsequent visual or ultrasonic inspection. In work25•26 suggest that acoustic emission has an important
addition, no structurally significant discontinuities escaped role in establishing basic design criteria for fiber reinforced
detection by acoustic emission. plastic structural components. Acoustic emission testing
results in safer, better designed fiber reinforced plastic
Technical background for recommended practices in equipment. This in tum leads to reduced design factors and
pipe testing is available in the literature.20-24 A number of has produced valuable economies of material.
important features of the SPI practices are relevant as well
for metal vessel testing.
Effect of Acoustic Emission Tests of Zone Location in Fiber Reinforced
Fiber Reinforced Plastic Structures Plastics
Figure 15 is a summary of discontinuities found by Discontinuity location in fiber reinforced plastics is
acoustic emission in one owner's installed fiber reinforced accomplished with zone location methods. Zone location
plastic tanks during a six-yearperiod of testing. Not only has approximates the position of an event based on individual
the percentage of discontinuities dropped sharply but the sensor activity. Figure 16 shows a general sensor layout for a
nature of each discontinuity is less serious. Because the typical fiber reinforced plastic vessel, with the zones marked
acoustic emission test detects discontinuities at an early around each sensor. To ensure complete coverage, the zones
stage, repairs are possible before the discontinuity becomes of adjacent sensors overlap and in these areas acoustic emis-
a hazard or propagates beyond repair. sion will be detected more than once. Approximate location
of the discontinuity can be determined by recording which
FIGURE 1 2. Pressurevessel stressingsequence PHASE II
PHASE I
100
80 gs
8 ~ ~O0-'I0 I0
60 DEPRESSURE INCREMENT
lJ.J lIJ-.J IO PERCENT M.A.XIMUM TEST
40 ~I~z STRESS (TYPICAL)
z::)
0z 95 PERCENT PREVIOUS PEAK STRESS HOLD
co N st" AE < EH (TYPICAL)
::)
TIME
i... f05
20
0
ACOUSTIC EMISSION TESTING I 31 5
sensors detect the emission. Conventional source location, A second problem arises because triangulation calcula-
using time of arrival and triangulation analyses, has been tions assume that stress waves travel along the vessel wall at
less than satisfactory for fiber reinforced plastics. constant velocity. Most fiber reinforced plastic constructions
are anisotropic and stress waves will travel much faster
Problems with Time of Arrival Source Location along the glass fibers than perpendicular to them. In theory,
it might be possible to adjust for the differences in wave
There are three main disadvantages to time of arrival velocity. However, most practical constructions have at least
source location. First, there is severe attenuation of stress two fiber orientations and the mathematics become
waves as they travel through fiber reinforced plastic. extremely complicated. Time of arrival location problems
Because of this, most events will not strike a sufficient num- are further compounded when stress waves reach a sensor
ber of transducers to permit triangulation. Placing the trans- by spiraling around the tank along the glass fibers but,
ducers closer together can reduce the effect of this problem because of attenuation, do not reach the same sensor by the
but may be too expensive for some applications. shorter route along the tank wall.
FIGURE 13. Flow chart for acoustic emission testingof pressurevessels
PHASE I PHASE II
O TO 30 PERCENT STRESS .........__ ......._-i)a~ 30 TO I 00 PERCENT STRESS
START RECORDAE NO
DATA DURING
BACK-
GROUND STRESSING
NOISE DISCONTINUITIES
CHECK
REJECT
REDUCE BACKGROUND
TO ACCEPTABLE LEVEL
316 I NONDESTRUCTIVETESTING OVERVIEW
FIGURE 14. Catastrophicfailures of fiber Fiber reinforced plastic is an extremely good emitter and
reinforcedplastictanks, 1972-91 this leads to the third problem. A large number of events
will be generated by a single source and each event will
COMPANY WIDE ACOUSTIC EMISSION occur closely after the previous one. These events have a
LV.l1J TESTING OF ALL FIBER REINFORCED broad range of amplitudes and some will not reach all the
Cl:::: sensors of a measurement array. Frequently, it becomes
::) PLASTIC VESSELS INTRODUCED impossible to track which event is striking which sensor and
the source location results become invalid.
_J
Zone source location is by definition inexact. It does
~ however limit the areas of a tank that are candidates for sub-
L0L sequent nondestructive testing with other methods. Nor-
mally, it is then a relatively simple matter to identify specific
Cl:::: discontinuities within a zone.
Le.nlJ
2
z::)
72 74 76 78 80 82 84 86 88 90 Felicity Effect in Fiber Reinforced
Plastics
YEAR
The felicity eff~ct is the presence of significant acoustic
FIGURE 15. Reasonsfor failureof acoustic emission at stress levels below those previously applied.
emission test by fiber reinforcedplastic tanks This phenomenon is a breakdown of the Kaiser effect. The
felicity effect is an indication of discontinuities in fiber rein-
60 forced plastics. The felicity ratio is the stress at onset of
emission divided by the previous maximum stress and serves
50 as an indication of the severity of a discontinuity.
V1 Acoustic Emission Signatures
LU
Cl:::: Research has shown that specific types of composite dis-
::) continuities (dry glass, star cracking, internal blistering)
have specific acoustic emission signatures. 27 Comparison of
~_J 40 test data with standard discontinuity signatures permits
identification of some types of discontinuity.
L0L
Cl:::: 30 FIGURE 1 6. Sensor layout and zone location
for acoustic emission tests of fiber reinforced
Le.nlJ plastic tanks
2
z::)
20
/ \ r. / \)( 'v' ~<-.. ~ ..... -1- f-'!-
..L_( )"..__ \ ZONE MONITORED BY
I I•I TWO SENSORS
.... ,-\ .....i,', 1--J. ....
/\ I ZONE MONITORED BY
v\ THREE SENSORS
0 4 ,,\»;I >~
ELAPSED YEARS 'i"I\-\-,, ZONE MONITORED BY
- ' - ONE SENSOR
LEGEND II
/:,,. = TOTAL
0 = CORROSION
~ = ABUSE
eC = LEGEND
FABRICATION • = ACOUSTIC EMISSION SENSOR
= STRUCTURAL
ACOUSTIC EMISSION TESTING I 317
Criteria for Evaluation and Calibration truck testing. The American Society of Mechanical Engi-
neers (Boiler and Pressure Vessel Code committee) speci-
The SPI recommended practices define specific evalua- fies acoustic emission as a mandatory requirement for
tion criteria and the most important is the one based on one-of-a-kind fiber reinforced plastic vesselsw and permits
emission during load hold. Other criteria are based on the design to a lower strength ratio for fiber reinforced plastic
felicity ratio, total counts, number of large amplitude events tanks proof tested with acoustic emission.i'"
and the acoustic emission energy parameter sometimes
referred to as signal value. Acoustic emission is accepted as the standard test
method for glass fiber reinforced plastic equipment. New
Recommended calibration methods define 'instrument applications, such as advanced aerospace composites, car-
operating settings and evaluation criteria for different types of bon and graphite fiber reinforced plastics and high pressure
acoustic emission instruments. As specified, the calibration piping, are regularly emerging and many are the subjects of
procedures require the use of a large lead sheet and a steel intensive research and development programs. Acoustic
bar. This is clumsy to apply in practice. However, laboratory emission is an important quality control tool and it is addi-
testing and field experience has shown that the procedures tionally becoming an important design tool for fiber rein-
provide satisfactorycross calibration between instruments.28 forced plastic components.
Acceptance of Acoustic Emission Some interesting research25,26 is directed toward the use
Techniques for Testing of Fiber of acoustic emission for defining design strain and the limit
Reinforced Plastics ofproportional linearity. Because acoustic emission is a sen-
sitive method of detecting onset of laminate microcracking,
The American Society for Testing and Materials (Com- it may be suitable for experimentally determining the design
mittees E-7 and F-18) has developed standards for acoustic strain value.
emission testing of tanks and piping and a standard for lift
Further research is focused on extending the range and
scope of signature definitions using techniques such as pat-
tern recognition and trend analysis. A better understanding
of the actual failure mechanisms of specific discontinuities
has been a valuable secondary benefit of this work
318 I NONDESTRUCTIVE TESTING OVERVIEW
PART 4
INDUSTRIAL GAS TRAILER TUBE
APPLICATIONS
Recertification of Gas Trailer Tubing TABLE5. Physicalpropertiesof 3AAX and 3T
industrialgas tubes (forgedend seamlesspipe)
Part of the distribution network for industrial gases
involves steel tubes, mounted on trailers, for transporting Property 3AAX 3T
gases over highways. Typically, the tubes contain gas at pres-
sures around 18 MPa (2,600 Ib-in."). Retesting of these Chemistry (percentage by mass)
tubes every five years is mandated by the United States
Department of Transportation. Traditionally hydrostatic Carbon 0:25/0.35 0.35/0.50
testing was performed. Since 1983, acoustic emission test- 0.75/1.05
ing has been used by companies on two types of tubes Manganese 0.40/0.90 0.035
(3AAXand 3T tubes) in place of hydrostatic retesting. This 0.04
was allowed by exemptions issued by the Department of Phosphorus (maximum) 0.04 0.15/0.35
Transportation. 0.80/1. I 5
Sulfur (maximum) 0.05 0. I 5/1 .25
Characteristics of 3AAX and 3T materials and typical 565 (I ,050)
tube dimensions are listed in Table 5. Figure 17 shows a Silicon 0.20/0.35
hydrogen service trailer carrying ten 3AAXtubes. Note that 795 to 895
the tubes are mounted with free space between them; they Chromium 0.80/1. IO (I 15 to 130)
do not make contact with one another.
Molybdenum 0.15/1 .25 930 to I ,070
In acoustic emission recertification, the tubes are instru- 538 (1,000) ( 135 to I 55)
mented and then pressurized to llO percent of normal fill- cTempered at a 0 (°FJ
ing pressure. Acoustic emission data provide location 16
distributions (plots of acoustic emission event counts versus Mechanical Properties 620 to 730
axial location) for each tube. Locations that produce a sig- Yield strength (90 to I 06) 560 (22)
nificant number of events are subject to shear wave ultra- MPa 10.4 (34)
sonic inspection. Ultrasonic testing establishes the presence (103 lb,.in.-2) 725 to 895
of a discontinuity, as well as its circumferential location, (105to125) ~10.5(0.415)
length and maximum depth. Ultrasonic testing is performed Ultimate tensile strength
MP a 20
FIGURE 1 7. Hydrogenservice trailer carrying ( I 03 lbrin.-2)
3AAX tubes
Elongation in 50 mm
(2 in.) gage (percent)
Typical dimensions 560 (22)
Diameter; mm (in.) I 0.4 (34)
Length, m (It)
Wall thickness, mm (in.) ~ 13.6 (0.536)
shortly after the acoustic emission test. Access to disconti-
nuities by the ultrasonic beam has been adequate with the
trailer fully assembled. Disassembly of trailers, for the pur-
pose of accomplishing ultrasonic inspection, has not been
necessary.
This acoustic emission test is normally performed at a
production facility, while the trailer is refilled. Excess gas is
then bled into a receiver. Before this test method could be
implemented, it was necessary to establish the maximum
allowable discontinuity size. In these tubes, the potentially
critical discontinuities are longitudinal fatigue cracks. This is
because the highest stress is in the circumferential direc-
tion. It is first necessary to establish what size discontinuity
will grow to critical size before the next recertification five
years hence.
ACOUSTIC EMISSION TESTING I 31 9
Analysis31 indicates that tubes containing discontinuities An acoustic emission system threshold of about 35 dB (rela-
with a maximum depth of 2.5 mm (0.1 in.) or more should tive to 1 µV reference voltage) is used. Trailer pressure is
be removed from service. Figure 18 shows plots of years measured with a transducer connected to the trailer mani-
required to reach critical discontinuity size (depth) versus fold. The pressure transducer output is also recorded by the
initial discontinuity depth for both 3AAXand 3T tubes. The acoustic emission system.
worst case is a 3AAX tube in hydrogen service. Hydrogen
atmospheres cause accelerated fatigue crack growth. The sensitivity of each mounted sensor is checked by
breaking a pencil lead close to it.32 The time required for a
Test Procedure for Trailer Tubing stress wave to traverse the distance between sensors (typi-
Tests cally 10 m or 32 ft) is also measured on each tube by break-
ing pencil leads or by pulsing outboard of one sensor. The
Setup for Acoustic Emission Testing time required for the resulting stress waves to travel from
the near sensor to the far sensor is input to the acoustic
Acoustic emission sensors (140 to 150 kHz resonance) are emission processor for use in location calculations.
attached to each end of each tube on the trailer with silicone
adhesive; silicone grease is used as a coupling medium. The The tubes are monitored to test pressure while the
sensors are connected to a multichannel acoustic emission trailer is filled. This procedure normally takes four to five
system that performs detection and linear location. hours when using gas plant compression equipment.
The sensor outputs are band passed at 100 to 200 kHz. Results of Acoustic Emission Tests of Trailer Tubes
Location distributions for each tube are monitored dur-
ing the test. Figure 19a shows a four tube location distribu-
tion display. If a sufficient number of events occur at an axial
location (for example, ten events in a 200 mm or 8 in. axial
FIGURE 18. Relationship of critical FIGURE 19. Plots of acoustic emission tube
discontinuity size, time and initial discontinuity tests: (a) event count versus axial location for
depth four tubes; (b) attenuation versus axial distance
for 3T tube tested with 1 50 kHz sensor
100 10,000
90 9,000
80 8,000
70 ~ 3T CRITICAL DISCONTINUITY SIZE 7,000 w
w 60 I 0.4 mm (0.41 in.] 6,000 N
N ' 5,000 vi
s~vi 50 4,000 (a)
40 3,000 sz~ V) .. •.
JO JI .
z;z:::: 30 2,000 z;:::: f>zzww- .. - ~ .
=mDmISCO~NTINUITY 10
1,000 0u
\ 3MX CRITICAL' 900 0 IL. •
800 V)
0u 20 DRONONce,~ 700 vi
600 0
-, Sl~6:g~EN GAS 500 V) JO
<, 400 6......J
V) 300 ~w
0 ' r,<, ;::::
6......J ' <, 200 uc2 -20
;:::: JO u;:::: IL .. .
uc2 9 <, <, <(
8 HALF THE 3MX V)
7 CRITICAL uI
<, <, :::i 0 20 40 60 80 JOO
DISCONTINUITY ~ u0
LOCATION
<( 6 PROPOSED RETEST YEARS <, Cl:::: <( (percent of distance between sensors)
5
Iu 0
~4 r
Cl:::: 3 f- (b) z 0 xA= -15.88 0 178 [meters)
O xA= -8.235 0 178 [inches)
0 ---------- wV) 0 - -10 0
f- .uu>...-..-J : 0 0 0 0 Oo O
__J REJECTION UMrT ~ !~:::,~:9 -20 0
V) _1_.69 mm (0.106 in -30
JNrTIAL
Cl:::: DISCONTINUITY ACCEPT REJECT
~>-- SIZE
100 <(
2.75 3.25
0.25 0.75 l .25 175 2.25 0 2.5 5.0 7.5 10.0
millimeters meters
0.01 0.03 0.05 0.07 0.09 0.11 0.13 0 JOO 200 300 400
inches inches
INITIAL DISCONTINUITY SIZE AXIAL DISTANCE (X)
320 I NONDESTRUCTIVE TESTING OVERVIEW
distance), a discontinuity probably exists and ultrasonic (structurally insignificant)will be measured again during sub-
inspection of that location is performed. sequent tests and discontinuity growth will be monitored.
If a sufficient number of events occur at a tube end (say, Interpretation of Trailer Tube Test Results
outside the sensor locations) with sufficient peak amplitude
(for example, ten events at 43 dB referenced to 1 µVat the Trailer tubes have a relatively large length-to-diameter
sensing element), then that entire tube end receives ultra- ratio and the axial location accuracy is generally very good.
sonic inspection. Even though discontinuities in tube ends Figure 2lb shows a location distribution for a long disconti-
cannot be located with this acoustic emission setup, their nuity on the inside surface of a tube in oxygen service. Fig-
presence is readily detected. ure 2la is a depth profile showing six separate ultrasonic
depth measurements.
Measured stress wave attenuation data for a 3T tube are
plotted in Fig. 19b. A power law curve has been fitted Five different sources of acoustic emission may occur at
through the data points. Both 3AAX and 3T tubes exhibit a discontinuity. These have been identified and measured
about 24 dB of attenuation over the distance between sen- during trailer tests and during laboratory tests on compact
sors on 10.3 m (34 ft) jumbo tubes. tension specimens. The five sources are listed in Table 6 in
order of decreasing intensity (peak amplitude at the source).
Figure 20 shows the acoustic emission reverification
methodology. Using this approach, discontinuities are Emission from all or some of the first four sources might
detected and located with the acoustic emission system. Dis- be measured during a given acoustic emission test. Mill
continuities are then circumferentially located and measured scale, the source that exhibits the highest peak amplitudes,
with ultrasonics. Discontinuities that are apparently passive is always present in natural discontinuities. The sources of
mill scale are: (1) oxidation of surface laps during hot form-
FIGURE20. Flow chart of retestingprocedure ing of seamless pipe and (2) residue from lubricants forced
for jumbo trailer tubes into discontinuities during pipe forming operations.
ORIGINAL INSPECTION PRIOR TO SERVICE FIGURE21 . Characterizationof trailer tube
(ASSIGN RATED WORKING PRESSURE)
discontinuity:fa) ultrasonicdepth
FIVE YEARS SERVICE measurements;fb) locationdistributionfrom
same discontinuity
ACOUSTIC EMISSION TEST AT 1 1 0 PERCENT OF
SERVICE PRESSURE (WHILE FILLING TRAILER WITH fa)
PRODUCT GAS) V") 0. 75 f0.03) I
..OucJ: I
ACTIVE INTENSE EMISSION SOURCE I§_
TEN OR MORE VALID EVENTS WITHIN 200 mm f8 in.) AXIAL DISTANCE ON tw I
CYLINDRICAL PORTION OF TUBE; TEN OR MORE EVENTS FROM TUBE t'.2
END, WITH PEAK AMPLITUDE~ 43 dB O QQ).) 0.50 f0.02)
YES ~E 1
ULTRASONIC TEST FOR ro0.25 0111
EMISSION SOURCE LOCATION
0 (0)
~----·· ····· -- -- 150 mm (6 in.)------·-·----·-
DISCONTINUITY WITH DEPTH z(b) Vf-) - 4l
GREATER THAN 2.5 mm f0.1 in.) >wwz 2~
0 ~----
YES .----'------. NO 0 ... ,..JL, .,.llJ1L.JJ_ . T" U._,
175
REMOVE TUBE CONTINUE V) 200 225 250 275 300 325 350
FROM SERVICE SERVICE (7) (8) (9) (10) (11) (12) (13) (14)
V)
AXIAL LOCATION
~ millimeters (inches)
w
u
i=
V)
:::i
0u
<(
ACOUSTIC EMISSION TESTING I 321
TABLE 6. Sources of acoustic emission at Advantages of Acoustic Emission
discontinuities in trailer tubes, fisted in order of Testing of Trailer Tubes
decreasing intensity
An acoustic emission recertification test data base has
Mill scale within discontinuities · been accumulated since 1983. Companies in the gas and
Inclusion fracture and delamination at crack tip chemical industries are able to contribute to the data. In
Mechanical interaction at the surfaces of cracks time, details regarding various systems, the best signal pr~-
Crack growth through parent metal cessor threshold and the relationship . between acoustic
Plastic deformation emission events and discontinuities should become more
firmly established.
Figure 22 shows a section cut from a 200 mm (8 in.) di~-
continuity located on the outside surface of a 3T tube. T_h1s With the standard hydrostatic test,33 the trailers are
section shows that the discontinuity (a fold created dunng removed from service, transported to a test facility and dis-
fabrication) is filled with mill scale. There is no evidence of assembled. The tubes are then inserted into a water jacket,
fatigue growth at the tip of the discontinuity and the mea- pressurized to 5/3 of service pressure and the v?lumetric
sured acoustic emission was from mill scale. expansion of the tube is measured. Tubes excee~mg allow-
able volumetric expansion are removed from service.
Because much of the measured emission from disconti-
nuities is not from crack propagation, emission can occur at The acoustic emission retest, by contrast, can be per-
relativelylowpressure. Figure 23 shows a plot of events ver- formed at a production facility. The trailer does not leave
sus pressure from the discontinuity pictured in Fig. 22. In service, is not disassembled and the inside surfaces of the
this case, slightly over half of the events occurred at pres- tubes do not become wet. The acoustic emission test
sures less than normal fillpressure. method outlined in Fig. 20 (when combined with ultrasonic
techniques) has these advantages: (1) it produces precise
locations and dimensions of discontinuities; (2) it is per-
formed for less cost; and (3) it produces more information
about the structural integrity of the tubes.
FIGURE22. Sectionof discontinuityon outside FIGURE23. Events versuspressurefor
of 3T tube; maximumdepth fradial direction)is discontinuity(see Fig. 22) on outsidediameter
1.4 mm (0.056 in.) of 3T tube
z1V)-- 20 DATAACQUISITION BEGINS
LU
>zLU I 5
0
vi
~V) 10
LU
ur=
:V)J
0u
<(
7,000 14,000 2 I ,000
( I ,000) (2,000) 13,000)
PRESSURE
kilopasca/s (pounds per square inch gage)
322 I NONDESTRUCTIVE TESTING OVERVIEW
PART 5
RESISTANCE SPOT WELD TESTING
Resistance Spot Welding emission signal from a transducer. From this fundamental
information, it is possible to extract from the signal that por-
Direct current or capacitor discharge welding is a joining tion believed to contribute most for determining an acoustic
method that passes a high current, short time pulse across emission threshold value:34 Note that the total detected
an interface to cause welding. Welding parameters perti-
nent to capacitor discharge welding of contacts have been acoustic emission signal is basically composed of three sig-
extensively studied and characterized. It is more difficult to nal subsegments corresponding to the three aforemen-
dynamically assess weld quality. tioned physical effects.
Although acoustic emission has been used to monitor the The premelting period helps determine the minimum
capacitor discharge welding process,34 the relationship of required welder power, whereas the postmelting period
acoustic emission signal to physical changes occurring in the gives the time needed to complete the welding process. As
weld zone needs to be clarified. Information from the melt- far as the welding itself is concerned, the interest here is pri-
ing and postmelting portions of the weld cycle needs to be marily focused on the melting period. From this period
extracted from the total acoustic emission signal, then ana- comes the extent of melting and consequently a prediction
lyzed and related to weld quality. The total welding cycle is of the thermal time constant for welding. The time constant
apportioned into three parts: (1) premelting; (2) melting; and is in turn used to set the test criterion in a nondestructive
(3) postmelting.35 By thermal analysis of a simplified body acoustic emission test for weld quality under ideal condi-
incorporating a change of phase, the melting portion of the tions and with compatible materials.
cycle is analyzed for penetration depth (extent of melting).
Alternating current resistance spot welding can be
Treating these periods separately is necessary to estab- treated as above by adding several half cycles (direct cur-
lish a correlation between welding and the detected acoustic rent) into multiple cycle welding. Note that the energy of
the half cycles (heat) can be electronically controlled and is
denoted as a percent of heat.
FIGURE 24. Typical acoustic emission response signals during resistance spotwelding
lVOLTS NUGGET
INFORfv1ATION
TIME
--+-
CURRENT INITIATION EXPULSIONS SOLIDIFICATION SOLID-SOLID POSTWELD
CRACKING
PHASE
TRANSFORMATION
ACOUSTIC EMISSION TESTING I 323
Principles of Acoustic Emission occurring during setdown and squeeze can can often be
Weld Monitoring related to conditions of the electrodes and the surface of
the parts. The large, often brief, signal at current initiation
The resistance spot welding process consists of several can be related to the initial resistance and the cleanliness of
stages: setdown of the electrodes; squeeze; current flow;forg- the part. For example, burning through certain oxide layers
ing; hold time; and liftoff. Various types of acoustic emission contributes to the acoustic emission response during this
signals are produced during each of these stages. Often, these time.
signals can be identified with their source. Individual signal
elements may be greatly different or totally absent, in various During current flow, plastic deformation, nugget expan-
materials or thicknesses. Figure 24 is a schematic representa- sion, friction, melting and expulsions produce acoustic emis-
tion showing typical signal elements that may be present in sion signals. The signals caused by expulsion (spitting,
the acoustic emission signature from a specific spot weld. flashing or both), generally have large amplitudes and can
be distinguished from the rest of the acoustic emission asso-
Most of the depicted acoustic emission signal features ciated with nugget formation. Figure 25 shows typical
can be related to factors of weld quality. Acoustic emission acoustic emission response signals during current flow for
both direct current and alternating current welding.
FIGURE25. Typicalacousticresponsesignalsduring currentflow: fa) directcurrentwelding;
[b] alternatingcurrentwelding
(a) CAPACITANCE
DISCHARGE
WELDING CURRENT
l(t)
VOITS ~~~~---,.--~,--~~~~~~__;;::::::...__,~~~-+ TIME (t)
V(t)
ACOUSTIC EMISSION TIME (t)
RESPONSE
L SOLIDIFICATION
CURRENT)) t
INITIATION EXPULSIONS
NUGGET
FORI\IIATION
(bJ ALTERNATING CURRENT
WELD
WELDING CURRENT
l(t)
VOLTS
V(t)
ACOUSTIC EMISSION~~.f+Hl###++H#l##!,lll-AA
RESPONSE
CURRENT NUGGET HOT CRACKING OR EXPULSIONS
INITIATION FORI\IIATION BLOW HOLES
324 I NONDESTRUCTIVE TESTING OVERVIEW
Following termination of the welding current, some mate- are clean or thinner material is welded. It occurs later if the
rials exhibit appreciable acoustic emission during solidifica- welding conditions have deteriorated or thicker stock is
tion. This can be related to nugget size and inclusions. As the welded with the same controller setting.
nugget cools during the hold period, acoustic emission can
result from solid-solid phase transformations and cracking. No matter when it occurs, expulsion is not desirable
because it removes material from the weld nugget area. It
During liftoff, separation of the electrode from the part may, however, be automatically kept to a minimum when
produces signals that can be related to the condition of the expulsion signals are used to generate a feedback signal to
electrode as well as to the cosmetic condition of the weld. control the welding process.
Using time and amplitude or energy discrimination, the When a feedback control arrangement is not used,
acoustic emission response corresponding to each stage can expulsion may be knowingly tolerated because of produc-
be separately detected and analyzed. Although the acoustic tion considerations. Inthis case, several test coupons should
emission associated with each stage of the spot welding be welded and the resulting weld strength or other quality
process can be relevant to weld quality, this discussion gives parameter should be determined through destructive test-
detailed consideration only to acoustic emission generated ing. Weld strength may be correlated with a suitable mea-
by nugget formation and expansion, expulsion and cracking. sure of the expulsion acoustic emission. In this way, a
maximum acceptable level of expulsion acoustic emission
Weld Quality Parameters That can be determined and used to segregate unacceptable
Produce Acoustic Emission welds or welded parts.
Acoustic Emission during Nugget Formation Acoustic Emission from Phase Transformation
When the material in the welding zone is heated, the In certain carbon steel alloys, material in and near the
pressure applied by the top electrode plastically deforms the spot weld undergoes martensitic phase transformation as
material and acoustic emission is generated. The amplitude the nugget cools. The total volume of material experiencing
of the acoustic emission signals associated with this process the transformation is related to both the nugget size and the
is affected by both electrode pressure and the cleanliness of area of fusion bonding. Therefore, the acoustic emission
the part. These acoustic emission signals may be gated out, response during phase transformation may be related to the
since they provide no useful information relative to the strength of the weld.
quality of the weld.
A waveform recorder should be used to isolate the phase
Further heating results in melting within the welding transformation stage of the acoustic emission signal. By
zone and growth of the nugget. Nugget formation and selecting a time and amplitude interval, a measure of the
expansion produce acoustic emission signals that can be cor- phase transformation acoustic emission for several samples
related with the strength of the weld. As soon as the welding may be correlated with the weld strength or size, as deter-
current starts to decrease, the nugget begins to solidify and mined from other tests. Inthis way, a minimum acceptable
residual stresses occur in and around the weldment. If these level of acoustic emission corresponding to acceptable weld
residual stresses are severe, hot cracking occurs between quality may be established.
cycles of alternating current welding.
Acoustic Emission from Cracking
By selecting a time and amplitude interval, measure-
ments of the acoustic emission response for several samples Acoustic emission resulting from postweld cracking can
may be correlated with weld strength or nugget size as be monitored during the time interval between current ter-
determined from other tests. Inthis way, a minimum accep- mination and liftoff (or, in some cases, between phase trans-
tance level of acoustic emission corresponding to acceptable formation and liftoff). A waveform recorder or a storage
weld quality may be established. oscilloscope may be used to identify the cracking interval by
monitoring the acoustic emission response. For applications
A storage scope or other device to record acoustic emis- where crack size affects quality evaluation, acoustic emission
sion response should be used to verify the several stages of monitoring should be supplemented with other test methods.
acoustic emission generation and detection shown for direct
current and alternating current spot welding in Fig. 25. Acoustic Emission Instrumentation
for Resistance Spot Welding
Acoustic Emission from Expulsion
The acoustic emission system used for this kind of test-
Expulsion occurs after a sufficient weld is formed. · ing usually consists of a high frequency acoustic emission
Within the weld period, it occurs sooner if the electrodes
ACOUSTIC EMISSION TESTING I 325
sensor attached to one of the spot welding electrodes. replenished and calibration of the system response is main-
Matching amplifiers and filters with a center frequency of tained. For sustained monitoring, such as online acoustic
0.5 MHz may be used to minimize the effects of mechani- emission examination or control of each nugget, the sensor
cally induced low frequency signals. Detector and energy should be permanently mounted using an epoxy adhesive or
processor follow, producing a digital acoustic emission a similar material.
energy count proportional to the energy of the detected
acoustic emission signal. A preamplifier is usually positioned near the sensor. How-
ever, when instrumentation is located less than 1 m (3 ft) from
The initiation of each weld power impulse is sensed and the sensor, gain otherwise supplied by the preamplifier may
a window is generated synchronous with the effective length be incorporated into the main amplifier of the instrument.
of the impulse. The window gates the acoustic emission
energy count into the counter, display and digital compara- Preliminary Measurements
tor. For multiple impulse welds, the energy counts are
cumulative. The comparator output can activate a control The acoustic emission signal from a single spot weld
circuit to terminate the weld cycle or if the anticipated pre- should be displayed on a waveform recorder. A wire coil or
set energy count was not realized an alarm circuit can be Hall effect sensor positioned near an electrode can be used as
activated. The energy count is displayed on a four decade a current sensor, thus providing a timing reference and trig-
light emitting diode readout. A simplified block diagram is ger signal for viewing and measuring the acoustic emission.
shown in Fig. 26.
This reference signal can also be obtained through an
Sensor and Preamplifier Attachment appropriate interconnection to the weld controller. Having
established a typical acoustic emission trace, characteristic
The sensor should be mounted to the lower (grounded) stages should be identified and one or more selected as an
electrode or electrode holder. Liquid couplant may be used acoustic emission examination parameter. For example, weld
for periodic sampling of weld quality, provided that it is quality indicators may be obtained from the acoustic emis-
sion response to nugget formation, expulsion or cracking.
FIGURE 26. Block diagram of acoustic emission monitor and controller for spot welding
L>7~ [>---TRANSDUCER /: -- - - -- -- - ··- - - - - - -
I
I BANDPASS ENERGY
1 PROCESSOR
ALTERNATING CURRENJEt+PREAMPLIFIER FILTER DETECTOR
SPOT WELDER X l____J
c:=, SENSOR I
I
l~ ~ / ITTSMMPUFIER
WELDER CONTROL
COUNT AND DISPLAY
COMPARE
CONTROL
---------------------~
326 I NONDESTRUCTIVE TESTING OVERVIEW
Setup for Welding Applications Defective spot welds occur randomly or gradually in the
process. Randomly occurring discontinuities are caused by
Preliminary measurements must be made to determine extraneous changes in welding conditions (for example,
the instrument settings and the conditions for monitoring. changes in the composition, the geometry or the surface
The weld controller settings are determined from normal conditions of the welded parts). Gradual deterioration of
weld considerations. First, the complete acoustic emission spot welds are caused by electrode wear.
response is observed on an oscilloscope. Next, the instru-
mentation gain is set to the maximum value where the The heat applied is proportional to ( 1) the electrical
acoustic emission signals representing the selected examina- resistance and weld time and (2) the square of the weld cur-
tion parameter do not saturate the amplifier. This step will rent. Of these, only the time and current are process vari-
ensure that the measurement is made with the best possible ables. Typical constant heat curves are shown in Fig. 27. For
signal-to-noise ratio. example, if the initial settings of the controller were some-
where on the curve H2, then after a large number of welds,
Next, the detection threshold level is established at a the surface area of the electrodes may increase and cause
value slightly above (or below) the peaks of acoustic emis- the current to decrease. This is equivalent to a shift of curve
sion signals that are excluded from the measurement. The H2 to the position H3 and requires compensation by adjust-
timing control is referenced to the onset of the weld current ing either the weld time or the weld current or both.
and consists of a delay and a time interval. The time interval
is selected to catch relevant signals. Online Monitoring
Finally, the count multiplier is set to a value that allows The purpose of online monitoring a spot welding process
utilization of the maximum number of significant digits in by acoustic emission is to identify faulty welds or parts. This
the readout. The finalized settings of the weld controller information can serve as a basis for acceptance or rejection
and the acoustic emission instrumentation are recorded of parts or for establishing a schedule for repair. Theoreti-
along with a photograph of the total acoustic emission sig- cally any combination of the weld quality parameters can be
nal. The counts obtained from individual welds may be be used for this purpose. Selection of the parameters is done by
kept on file for future reference. determining the most prevalent causes of defective welds in
a particular welding application.
Repeated Applications
FIGURE27. Constantheat curvesfor typical
If acoustic emission monitoring was previously applied to spot welding applications
a particular controlling or monitoring problem, the purpose
of the preliminary measurement is to reestablish the known =H 12RT
and recorded original test conditions, including the heat and
the weld time settings on the weld controller. Gain, threshold WHERE:
level, timing and other settings on the acoustic emission H = HEAT (WAIT SECONDS)
instrumentation may be made identical to those on record. I = WELDING CURRENT (AMPERES)
R = RESISTANCE (OHMS)
A few sample welds are made to verify that the general T = TIME (SECONDS)
appearance of the total acoustic emission signal exhibits the
expected characteristic features and that the average count
falls within the range of past measurements.
Typical Applications of the Acoustic WELDING CURRENT
Emission Method
Acoustic emission surveillance of spot welding can take
two forms: monitoring and control. The purpose of online
rrwnitoring is to identify and segregate unacceptable welds
for quality evaluation. Online control uses the acoustic emis-
sion instrumentation to complete a feedback loop between
the process and the source of welding current by automati-
cally adjusting one or two process variables to compensate
for deteriorating welding conditions.
ACOUSTIC EMISSION TESTING I 327
Excessive postweld cracking can be caused by loose toler- As the electrode conditions gradually deteriorate, the
ances of the dimensions and surface conditions of the welded occurrence of expulsion shifts upward with constant cur-
stock; burning; or excessive expulsion due to overwelding.
Such cracking can be identified as randomly occurring dis- rent. Intermediate conditions can be represented by the
continuities using corresponding acoustic emission responses. curve· H2. A shift requires adjustment of the weld current,
weld time or both. Acoustic emission feedback control auto-
aIf all parameters are monitored simultaneously, an accept-
matically adjusts the required process variables. In practice,
able weld may be defined as spot weld with adequate adjustment of the weld time is most common.
nugget volume that (1) isfree of excessive expulsion and post
weld cracking and (2) satisfies the average strength require The input pulse is the timing reference for a single spot
ments. Limits for a gradual deterioration of weld conditions,
represented by a gradually increasing number of defective weld operation. The feedback controller produces a time
welds, can also be detected by monitoring the expulsion. This delay corresponding to the time between points T and A in
is the only examination parameter for certain applications.
Fig. 27. From time A onward, acoustic emission is moni-
Of the weld quality parameters described above, only
nugget formation and expulsion occur during the welding tored and expulsion is detected when acoustic emission rises
process. Therefore, these parameters may be used for con- above the threshold level. Monitoring continues at least
trolling weld process variables. The weld time and heat set- until the time corresponding to point B. If expulsion occurs
tings of the controllers usually produce an overwelded
condition in industrial practice. This is done in order to earlier, the output of the feedback controller terminates the
cover variations of material, material dimensions and sur- weld current at C, for example, thereby overriding con-
face conditions and to allow some wear of the electrodes troller setting B.
before the next maintenance period. ·
The equipment described previously should have the For both acoustic emission weld control parameters, it is
following inputs: (1) preamplified acoustic emission signal possible that the maximum controller settings can be
and (2) pulse or step function from the weld controller reached without producing an acceptable weld. If this
which marks the onset of current synchronously with zero
crossing point. occurs, the electrodes have deteriorated to failure and
should be changed.
Acoustic emission feedback control automatically adjusts
the required process variables. In practice, adjustment of MonitoringCoated Steel
the weld time is most common. After the acoustic emission AlternatingCurrent Welds
weld parameters have been established as described previ-
ously, the acoustic emission monitor may be interfaced with Spot welding steel sheets coated with the zincrometal
an existing welder controller as shown in Fig. 28. process expels the coating from the weld area in a violent
manner, due to outgassing of the coating. With controlled
Nugget Formation as the Control Parameter
FIGURE 28. Acousticemission feedbackcontrol
For a specific weld strength, the corresponding acoustic of spotwelding process
emission value is set on the acoustic emission controller.
When a weld is made, a pulse signals the initiation of the EXISTING EXISTING
gating section of the acoustic emission controller for accep- WELDING WELDING
tance of acoustic emission only during nugget formation. CONTROLLER
POWER
If the acoustic emission value is reached, the controller SUPPLY
sends a pulse to the welder in order to terminate the weld-
ing heat. The termination signal should be synchronous with ELECTRODES /V WELD NUGGET
the zero crossing of the weld current.
z: ACOUSTIC EMISSION
Expulsion as the Control Parameter36
ACOUSTIC EMISSION SENSOR
In Fig. 27, if the curve H 1 represents the heat settings
necessary to form a good nugget under initial and ideal PREAMPLIFIER ACOUSTIC INTERFACE
welding conditions and if curve H 3 represents the actual
controller settings, then the time difference between two EMISSION
corresponding points on the two curves represents the SPOT WELD
amount of overwelding. Under initial conditions, expulsion CONTROLLER
occurs close to H 1.
328 I NONDESTRUCTIVETESTINGOVERVIEW
welding parameters, the expulsion of the coating occurs energy count from galvanized welds are usually of greater
during the first weld cycle. A one cycle weld envelope from magnitude than for steel, as shown in Fig. 31.
a zincrometal spot weld is shown in Fig. 29. The two initial
bursts of acoustic emission are due to coating expulsion. The second feature is apparent later in the weld interval.
Because this emission has little relation to true welding, it is Since the temperature for welding steel is above 1,470 °C
blanked by preset logic in the acoustic emission weld ana-
lyzer, as indicated by the lack of windows during the first FIGURE30. Acousticemissioncontrolof
weld cycle. The following two weld cycles were processed in zincrometalspot welds
normal fashion.
zf- 5,000
Plotted in Fig. 30 is the weld energy, expressed as counts :J
detected in a series of zincrometal spot welds versus nugget 0u
diameters. The curve would approach a straight line if the l>.9- 4,000
counts were plotted against nugget area. This curve is typi-
cal of the relation between weld energy count and spot weld 0:::
size and is the basis for controlling the welding process with
acoustic emission instrumentation. zLU
zLU
AlternatingCurrent Spot Welding 0
Galvanized Steel vi 3,000
~V1
Because of the low melting point of zinc, acoustic emis- LU
sion signatures from spot welds in galvanized steels have two u
features not encountered in ordinary steel welds. The coat- f'.=
ing is rapidly melted and expelled from the pressure zone V1
beneath the electrodes. This mass flow of liquid zinc causes a 2,000
slight to moderate increase in acoustic emission which sub- :J
sides in a few cycles. The acoustic emission amplitude and u0<
/
/ 0.2
/
0.1 0.15
NUGGET DIAMETER
(units)
FIGURE29. Acousticemissionfrom spot weld FIGURE31 . Acousticemissionenvelope
in zincrocoatedsteel for nugget diameterof ( 1 V-cm-1 J detectedduring twelve-cyclespot
0.15 units;acousticemissionenergycount of weld in galvanizedsteel; acousticemission
1,550; sweepof 0.5 ms-mrrr ': and verticalof energycount 21,200 for nugget diameterof
0.05 V·mm-1 4.8 mm (0.19 in.J
ACOUSTIC EMISSION TESTING I 329
(2,678 °F), far above the boiling point for zinc, each power FIGURE 32. Expanded trace showing last four
impulse is followed by residual zinc boil off from the weld cycles of weld; note acoustic emission from
zone. This boiling is a considerable source of acoustic emis- boiling zinc following each current impulse;
sion. By proper setting of delay and window for each power between windows
pulse, the acoustic emission from boiling can be largely
rejected. If the zinc is not removed, the weld nugget is rich BOILING ZINC
in iron zinc-compounds and would be brittle. An extended
trace showing the last four cycles of Fig. 31 is provided in = = ;;i=~ :: -:: =I -.; I :
Fig. 32, to more clearly illustrate the zinc boil off following 1 J~
each impulse while the window was closed. - ..,~I
• =[= ~ .::: :I
Detecting the Size of Adjacent = ~ ==
AlternatingCurrent Welds :: II:
+. n ~ tll:
Shunting is caused by the presence of a previously made t~+++++a !++ ~~ 11•11
spot weld close to a weld being made and offering a shunt II!
circuit for part of the weld power. The effect of shunting on I =II,,.. c
weld nugget size becomes more pronounced with closer i-.,1 ~ I "":l
spacing between spot welds. To illustrate the effect, a series L&
of spot welds were made in the uncontrolled mode. Welding ;~1- -J
parameters were held constant. The first welds made had ~
some scatter of nugget size depicted by the acoustic emis-
sion instrument. These are referred to as initial welds. ~ -~ -
Shunted spot welds were made from 0.25 units to 1 unit c
center-to-center and all were deficient in nugget size, as ""
plotted in Fig. 33. All welds were made with three cycles.
The welder control was reset to a maximum of five cycles FIGURE 33. Effect on nugget size from
and the acoustic emission spot weld controller was preset to adjacent weld shunting
turn off the welder after 1,100 weld energy counts. The
results for a similar series of shunted welds was a uniform z l1--
series of spot welds with 0.2 unit diameters requiring four to
five cycles of weld power. ::) 5,000
Control of Spot Weld Nugget Size 0u
>-
The principle that the size of the spot weld is propor- 1...'.J
tional to the acoustic emission weld energy count is based Oc::
on several measured parameters of the welding procedure.
Figure 30, illustrating the relation between the two, is valid zUJ 1500 ~
even for the extreme case of coated steels. UJ
In Fig. 33, while the weld power was held constant, 0 -·_.•.. ---·.
nugget size varied widely due to shunting. This effect was _
accurately evaluated by acoustic emission. When the acous- .....J
tic emission energy count was used to turn off the welder, , ·-·-·---·---T·--·-·------·-···T·-··--·--······--·--1·--·----·--
that was a constant energy count. The resulting nugget size IUJ
was also constant in spite of wide variations in power input ~z 1,0001 0.05 0.1 0.15 0.2
needed to compensate for shunting.37 i~V)0
vi
Figure 34 is the acoustic emission record of a spot weld in
steel using expulsion limit control.36 A large spike of acoustic UJ 500 ~
emission due to expulsion occurred during the first half of
u I
l::)i==
V)
0
~
NUGGET DIAMETER
(units)
LEGEND
UNCONTROLLED WELDS
• UNCONTROLLED SHUNTED WELDS
" = CONTROLLED SHUNTED WELDS
330 I NONDESTRUCTIVE TESTING OVERVIEW
the third cycle. The expulsion limit of 100 counts preset into before gross expulsion occurs. Thus, a reasonable control of
the instrument was exceeded and the instrument turned off welding is obtained with some assurance that there will be
the welder after 2.5 cycles. The instrument was working in no stickers or underwelded spots.
expulsion mode. In the expulsion mode, a high level detector
is activated. In this example, the expulsion bias was set at 3 V Conclusions
and only acoustic emission signals above that bias was
detected and counted. The resulting spot weld was normal at Acoustic emission monitoring provides valuable insights
0.19 unit diameter with only a trace of expulsion. into the welding process. Every burst of acoustic emission
has a logical source and meaning.38•39 Using acoustic emis-
The principal advantage of expulsion limit control is that sion for online evaluation of spot welding provides new
expulsion of molten metal must occur after fusion. If the applications for weld process control.
expulsion is detected late in. the weld cycles, acoustic emis-
sion detection is sensitive enough to terminate welding 1. With instantaneous analysis, acoustic emission can fill
in the missing link in a control loop including detec-
FIGURE 34. Expulsion limit control of welding tion, feedback and control.
2. Acoustic emission monitoring can be automatic and
integrated with numerical control used with robotics.
3. Acoustic emission techniques are a unique diagnostic
tool for evaluating weld parameters in real time.
4. Overwelding can be reduced, improving weld effi-
ciency.
Every mechanism occurring during a spot weld cycle
results in acoustic emission that can be related to a specific
source. The source mechanism may be beneficial or detri-
mental, with corresponding effects on weld integrity. Acous-
tic emission monitoring provides a tool for spot weld
evaluation and control that is limited only by the initiative of
the user. Other aspects might limit use, such as knowledge
of the relationship between acoustic emission activity and
the desired weld qualities.
ACOUSTIC EMISSION TESTING I 331
PART 6
ACOUSTIC EMISSION APPLICATIONS IN
UNDERSEA REPEATER MANUFACTURE35
For an undersea electronic repeater, lying on the ocean current power supply in the cable. It is an oil filled paper
floor and boosting transmission signals in an undersea cable, capacitor housed in a cylindrical ceramic casing with metal
maintenance is impossible and replacement is extremely end caps, as seen in Fig. 35. The paper capacitor unit is
expensive. Consequently, during the manufacture of these wound on a ceramic core to allow a coaxial cable to pass
repeaters, the device's circuitry and the waterproof closure through its axis. Both the casing and the core have a metal-
of its cylindrical protective housing are subjected to lized band on each end for solder sealing (see Fig: 35).
extremely rigid quality controls.
During the rotational soldering operation (Fig. 36), a
Application of acoustic emission technology helps ensure poor quality ceramic casing or capacitor core could crack
the structural integrity of the blocking capacitor in an because of microscopic discontinuities or deviations from
undersea repeater's circuitry. optimal processing conditions. While external cracks can be
detected visually, it is not possible to detect internal hairline
High Voltage Capacitor in the cracks in the ceramic using visual or optical methods,
Repeater Circuitry Unit because both ends of the capacitor's casing and core are
enclosed by the soldering operation. Even though the sub-
The high voltage direct current blocking capacitor sepa- sequent life test might identify some defective capacitors by
rates the alternating current voice signals from the direct revealing an oil leakage, this test is costly and extremely
time consuming. Moreover, optical inspection and life test-
ing will not reveal cracks covered by the solder.
FIGURE 35. Capacitors ceramic casing is positioned between two metal covers that fit over ends
(left); covers are then soldered to ceramic casing fright); ceramic core, also susceptible to cracking
during soldering operation, is shown projecting from end cap in assembled capacitor (right)
FROM BELLlABORATORIES. REPRINTEDWITH PERMISSION.
332 I NONDESTRUCTIVE TESTING OVERVIEW
Once a capacitor is operating within the repeater hous- problem was also made to distinguish cracking from noise
ing, a hairline crack in a casing or core could propagate, utilizing a signature recognition scheme.
releasing a seepage of oil. Besides contaminating the
repeater, this seepage could cause dielectric breakdown in Figure 37 is a block diagram of the crack detection sys-
the high voltage section of the cable. A shorted capacitor tem used in conjunction with the rotational soldering opera-
would render the repeater inoperable, requiring costly tion that joins the metal caps to the ends of the ceramic
replacement. Hence, a system that can detect cracks as they capacitor case described above. The incoming acoustic
occur during manufacture would help ensure the quality of emission signal from the preamplifier (Fig. 38) is filtered
the capacitor. and further amplified. This analog signal is then passed to
both a threshold detector and an envelope detector. The
Instrumentation and Analysis output of the threshold detector is a wave train of pulses
corresponding to each threshold crossing of the filtered
Ceramic Capacitor Crack Detector acoustic emission signal. The output of the envelope detec-
tor is a voltage following the peak input excursions of the
Because of the continuous nature of the soldering pro- acoustic emission signal.
cess, a window or gate period could not be defined and
existing crack detectors could not be used effectively for the The acoustic emission envelope is then passed to both a
undersea repeater's capacitor soldering operation. The threshold detection circuit and an envelope strength circuit.
required system would have to monitor cracking during the A window opens when the envelope surpasses a threshold
entire processing cycle. The design used to overcome this and persists as long as the envelope remains above that
threshold. The envelope strength circuit converts the acoustic
emission envelope into pulses whose number is proportional
to the area (amplitude and length) of the envelope curve.
FIGURE36. Transducerto detect acousticemissionoriginatingin ceramicduring solderingis seen at
left, also servingas pivot for rotatingcapacitor
FROM BELLlABORATORIES. REPRINTEDWITH PERMISSION.
The words you are searching are inside this book. To get more targeted content, please make full-text search by clicking here.
ASNT NDT Handbook Volume 10 OVERVIEW
Discover the best professional documents and content resources in AnyFlip Document Base.
Search
ASNT NDT Handbook Volume 10 OVERVIEW
- 1 - 50
- 51 - 100
- 101 - 150
- 151 - 200
- 201 - 250
- 251 - 300
- 301 - 350
- 351 - 400
- 401 - 450
- 451 - 500
- 501 - 550
- 551 - 600
Pages: