Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
CA 02312101 2000-06-22
FIELD OF THE INVENTION
The present invention relates, in general, to the
detection and characterization of defects in metal structures
and, in particular, to a method for determining the structural
integrity of multi-layer structures such as an aircraft
fuselage lap splice.
BACKGROUND TO THE INVENTION
Corrosion is a major problem that can compromise the
structural integrity of equipment in many diverse industries
such as in pipelines (gas or oil) or in the aircraft industry.
In the aircraft industry, this situation is a primary concern
to the engineering authority responsible for aircraft
airworthiness. One structure closely scrutinized in the
aircraft industry is the fuselage lap splice. This metallic
multi-layer structure has a design such that a crevice exists
where conditions are favourable for corrosion. Corrosion in a
lap splice could, ultimately, lead to a structural failure of
the fuselage.
Visual inspection is one method for detecting
corrosion in multi-layer structures such as in aircraft
fuselage lap splices. This technique is based on the
principle that, when corrosion takes place between the layers,
the metal lost to corrosion results in a product that forces
the plates apart and causes surface distortion. The visual
inspection, however, does not provide a fool proof indication
that the deformation is actually due to corrosion. Such
distortion may exist, for example, as a result of poor quality
- 1 -
CA 02312101 2000-06-22
control during manufacturing or from a previous repair.
Ultrasonic techniques have also been used to detect corrosion
in metal pipes such as the techniques described by David Wang
in Canadian Patent Application 2,258,439. The detection of
defects, other than those in the first layer, using ultrasonic
techniques requires a mechanical bond between plates. The
absence of a bond will preclude detection in second, third or
fourth layer defects.
Another method for detecting corrosion in metal
structures is the use of a low-frequency eddy current
inspection method. A low-frequency eddy current inspection
technique uses a coil to induce eddy currents in a test
object. The induced eddy currents produce a time-varying
magnetic field which can be measured by magnetic flux sensors
to yield information about the condition of that test object
and determine whether a loss of material due to corrosion has
occurred. A low-frequency eddy current inspection method can
detect a loss of material in a metallic structure but is not
always reliable. It often requires the use of dual frequency
methods and signal mixing to detect corrosion.
Canadian Patent 2,102,647 by John H. Flora et al is
directed to detecting defects in a metal component using a low
frequency eddy current technique. John H. Flora et al uses an
excition coil wound on a yoke and a pair of magnetic flux
sensors differentially connected with respect to each other in
an area under the yoke. The differential connection will
result in the cancellation of common signals detected by the
sensors, those which would be generated by the coil, but allow
- 2 -
CA 02312101 2000-06-22
the detection of other signals generated by eddy currents in
the metal component. The yoke is then placed near the metal
component and a low frequency alternating current applied to
the coil to generate eddy currents in the metallic component,
which currents are detected by the sensors. The yoke is moved
along the surface to scan for defects by changes in the
generated eddy currents at defect locations.
US Patents 4,843,319 by Pedro F. Lars and 4,843,320
by Brian R. Spies are directed to corrosion detection where a
transmitting antenna coil is placed next to a metal container,
in this case a pipe with layer of insulation on it, and
applying a train of pulses to that transmitting coil. The
pulses are shaped so the coil is energized for a sufficient
period of time to stabilize the magnitude of the field, with
no eddy current then being generated, and then de-energizes
abruptly to generate eddy currents in the metal which are
detected by a receiving coil. Those eddy currents decay and
are gradually dissipated within the metal with the rate of
diffusion being dependent on the conductivity and thickness of
the metal. The decay of those eddy currents is detected by a
receiving coil and used to determine if defects in the metal
exist such as caused by corrosion and a resulting change in
thickness of the metal. However, errors in responses will
occur due to variations in distance between the antenna and
the metal wall of the container. Pedro F. Lara discusses some
methods for correcting those errors in responses. The pulses
used in these US Patents operate in the time domain rather
than in a frequency domain manner as used in Canadian Patent
- 3 -
CA 02312101 2000-06-22
2,102,647. In the time domain, the information needed to
probe a conductor wall for reasonably accurate detection can
be obtained with one transmitted pulse. Each pulse contains
an infinite number of frequencies. In frequency domain
methods, however, only a few frequencies are used to probe a
conductor wall which results in a limited amount of
information from which the wall thickness is to be determined.
U.S. Patent 6,037,768 by John C. Moulder et al
describes another pulsed eddy current (PEC) apparatus to
detect corrosion in metal structures such as aircraft
lapjoints. John C. Moulder et al describes a calibration of
the PEC instrument before the inspection with a reference
structure that the user knows to be flaw-free. The PEC probe,
once calibrated, scans in serpentine fashion a selected
fashion area under computer and motor driven control. John C.
Moulder et al indicates in lines 38 to 44 of column 4 that an
air gap between the probe and lapjoint is known as "lift-off"
and that ideally, lift-off remains constant at 0.007 of an
inch during a scan since the probe has a constant built-in
wear surface. However, possible irregularities in a lapjoint
surface may result in greater lift-off with a possibility of
obtaining anomalous inspecting result. The user, during a
scan is, however, able to filter from the display known
conditions such as the existence of fasteners and airgaps and
excessive probe lift-off.
Prior art methods of detecting corrosion in aircraft
lap splices multi-layer metallic structures have proven
inadequate. The detection of corrosion by either ultrasonic
- 4 -
CA 02312101 2000-06-22
or eddy current techniques is not inherently difficult, but,
there are problems with the identification and
characterization of that corrosion due to the complexity of
multi-layer structures. To quantify the thinning in multi-
layer structures, it is required to determine in which layer
corrosion has occurred. Ultrasounds, for instance, will not
easily penetrate beyond the first layer. Eddy current
techniques, on the other hand, have the ability to perform
multi-layer inspections without requiring a mechanical bond.
Notwithstanding these limitations, most operators have elected
to conduct visual inspections followed by low-frequency eddy
current inspections to detect corrosion in aircraft lap
splices. This approach reduces the number of false
indications but it is not capable of isolating corrosion below
10% thinning in the first layer. Further, second and third
layer corrosion may also progress to much greater amounts of
thinning before they are finally detected by this approach.
SUMMARY OF THE INVENTION
It is an object of the present invention to provide a
pulsed eddy current method for detecting defects in a metallic
structure and allow a quantitative evaluation of any defects
detected.
A method for detecting defects in a metallic structure,
according to one embodiment of the present invention,
comprises locating a transducer at a first distance from said
metallic structure at one area that lacks any significant
defects in the structure, activating said transducer with a
square wave voltage controlled excitation to generate eddy
- 5 -
CA 02312101 2000-06-22
currents in the structure and then sensing, with at least one
magnetic flux sensor, time-varying magnetic fields generated
by the transducer and said eddy currents, signals obtained
from said at least one sensor being recorded, this process
being repeated to obtain at least one other recorded signal
that is obtained with the transducer being locating in the
same location but at a different lift-off distance from said
one area, determining where the recorded signals cross to
establish a Lift-off Point of Intersection at a point in time,
placing said transducer at other areas of said structure which
are to be tested for defects, activating said transducer with
similar voltage-controlled excitation as applied at said one
area, then obtaining and recording signals sensed from the
time-varying magnetic fields generated by the transducer and
eddy currents in a similar manner as at said one area,
comparing the recorded signals amplitudes which are obtained
at said other areas at said point in time with those of
signals obtained at said one area with differences in signal
amplitude providing indications of any defects present at
areas being tested.
BRIEF DESCRIPTION OF THE DRAWINGS
The following detailed description of the invention
will be more readily understood when considered in conjunction
with the accompanying drawings in which:
Figure 1 is a composite graph of signals showing the
effects of lift-off (the distance between the
transducer and test object) on balanced transient
responses for a single coil transducer,
- 6 -
CA 02312101 2000-06-22
Figure 2 is an expanded view of Figure 1 near the
Lift-off Point of Intersection,
Figure 3 is a composite graph of signals at the Lift-
off Point of Intersection and illustrates the ability
of the present invention to quantify material loss
independently of lift-off variations using a single
coil transducer,
Figure 4 is a composite graph of signals that
illustrates the effect of lift-off on balanced
transient response for a driver-pickup transducer and
the separation of signals at an area with defects and
one without defects,
Figure 5 is an expanded view of Figure 4 near the
Lift-off Point of Intersection, and
Figure 6 is a composite graph of signals near the
Lift-off Point of Intersection to illustrate the
ability to quantify material loss independently of
lift-off variations for a driver pickup transducer.
CA 02312101 2000-06-22
DESCRIPTION OF THE PREFERRED EMBODIMENTS
Corrosion is a major problem that can compromise the
structural integrity of equipment in diverse industries, such
as pipelines or in the aircraft industry. In the aircraft
industry, for example, corrosion is a primary concern for the
engineering authority responsible for airworthiness with
fuselage lap splices, in particular, requiring close
scrutinization. Fuselage lap splices have a multi-layer
metallic structure such that a crevice exists where conditions
are favourable for corrosion and this corrosion, if it exists,
could eventually lead to a structural failure of the fuselage.
One method for detecting corrosion in aircraft
fuselage lap splices is by visual inspection. This technique
is based on the principle that, when corrosion takes place,
the metal lost to the corrosion forms a product that forces
the multi-layer plates apart and causes surface distortion.
Visual inspection, however, does not provide a fool proof
indication that the deformation is actually due to corrosion.
Distortion could also exist as a result, for example, of poor
quality control during manufacturing or from a previous
repair.
A low-frequency eddy current inspection method is
another technique to detect corrosion in metal structures. In
this technique, a transmitting coil is used to induce eddy
currents in a metal structure. Those induced eddy currents
produce a time-varying magnetic field that can be measured by
magnetic flux sensors to yield information about the condition
of the metal structure and provides an indication as to if any
_ g _
CA 02312101 2000-06-22
corrosion has occurred. This technique, however, is not
always reliable and often requires the use of dual frequency
methods and signal mixing to detect corrosion.
Another technique to detect corrosion is to apply a
train of pulses to a transmitting antenna coil which is
located near a metal structure and to measure the magnitude of
fields generated by eddy currents produced in the metal
structure by those pulses. US Patents 4,843,319 by
Pedro F. Lars and 4,843,320 by Brian R. Spies describe such
methods for detecting corrosion using pulses to induce eddy
currents in pipelines for transporting oil or gas. The pulses
applied to the transmitting coil in these US Patents are
shaped so the coil is energized for a sufficient period of
time to stabilize the magnitude of the field, with no eddy
current then being generated, and then de-energizing the coil
abruptly to generate eddy currents in the pipeline which are
detected by a receiving coil. The generated eddy currents
decay and are gradually dissipated within the metal with that
rate of decay being dependent on the conductivity and
thickness of the metal in the pipeline at the area where those
eddy currents were generated. The fields generated by the
decaying eddy currents are detected by a receiving coil and
that information is used to determine if defects in the metal
exists such as those caused by corrosion and a resulting
change in the thickness of the metal. Errors, however, will
occur due to variations in distance between the antenna and
metal wall of the pipeline. Pedro F. Lava discusses some
methods for correcting those errors.
_ g _
CA 02312101 2000-06-22
The detection and characterization of corrosion in
multi-layer structures according to the present invention is
carried out with a pulsed eddy current technique. Generally
pulsed eddy current techniques, up to present, have attempted
to keep the distance between the transmitter/receiver coils
and the metal surface as constant as possible in order to
avoid errors caused by variations in that distance to the data
collected and to simplify the analysis of that data. It has
been found, however, that when transmitter/receiver coils
(transducers) are used in pulsed eddy current techniques, the
transient responses for various lift-off distances (i.e.
distance between the transducer and test object) intersect at
a given point (i.e. Lift-off Point of Intersection) when no
corrosion is present. This Lift-off Point of Intersection can
then be used for the characterization of corrosion and, most
importantly, the results obtained with this method are
independent of lift-off variations inherent to field
inspections.
The essence of a pulsed eddy current test is that
current pulses drive a test coil assembly whose output signals
are analyzed. The system's architecture is, up to a certain
level, dictated by the transient response. Variations in the
transient response due to defects are so small and the
phenomenon occurs so rapidly that digital data processing
generally is the only viable option. An enormous amount of
digital data can be generated for a scan of a given surface.
The analysis of that amount of digital data would require a
- 10 -
CA 02312101 2000-06-22
microcomputer's data handling capability. A microcomputer can
also be used to carry other functions such as:
a. controlling most of the parameter settings for
the instruments;
b. reading and analysing the outputs of the eddy
current instrument; and
c. sending outputs to external equipment.
An architecture that could be selected to carry these
functions is the star structure when the central unit of the
system's architecture is the computer. External equipment
would include the scanning system, the pulse generator and the
inspection probe with each instrument being individually
connected to the central control unit. The computer is the
central control unit of this type of set-up and it is via one
of its applications that the excitation is digitally triggered
at given spatial positions during the scan of the test object.
When a selected position for the test coil is reached, the
computer sends a trigger to the pulse generator to provide the
test coil assembly with the voltage controlled excitation
signals.
The pulsed eddy current technique is based on the
principle of magnetic induction where a transmitter coil
(transducer) provides a magnetic field when excited with a
square wave current (pulse) and this generates eddy currents
in an adjacent metal structure (test object) to produce a
magnetic field which opposes the field generated by the coil.
The square wave produces a time-varying magnetic field and
provides for a wide range of frequency excitation. The
- 11 -
CA 02312101 2000-06-22
induced eddy currents flow at specific depths within the test
object and decay over a period of time after the magnetic
field being generated by the transducer is terminated.
Various sensors can be used to capture the time-domain
variations of the magnetic flux. Some sensors will measure
the magnetic flux density while others will measure the rate
of change of the total magnetic flux. Coils have the
advantage to be the sensor mostly used for in-service
applications.
The captured transient response in the time domain,
also called A-scan, contains a broadband spectrum of
excitation frequencies that, theoretically, can be analyzed to
determine the test object condition, i.e. defect depth, size
and location. The transient response also contains a large
number of separate components (also transients) from different
parts of the structure being investigated with many of those
remaining constant. The major part of the total transient
are, in fact, due to the field in air and the scattered field
due to the specimen. The total transient can be subtracted in
order to enhance the appearance of the small transients by a
process referred to as a balancing process. The balanced
transient is the traditional means used to determine the
presence, the amount, and the location of corrosion. There is
a significant drawback, however, to the balancing process with
standard pulse eddy current techniques. In order to identify
the flaw transient, the background parts of the transient must
remain constant throughout the duration of even the most
extensive of measurements. Any changes to the background such
- 12 -
CA 02312101 2000-06-22
as lift-off, will be interpreted as changes to the flaw
transient signal thus leading to inadequate interpretation of
the responses received. Specifically, lift-off increases the
balanced transient peak amplitude, advances its location and
advances the time to zero crossing (where applicable) to such
an extent that the defect's size and location cannot be
adequately determined by traditional techniques.
The shape of the time-domain balanced transient
responses sensed by magnetic field sensors changes
tremendously with variation of distance between the
transducers/sensors and test object, i.e. the lift-off
distance. One particular feature, referred to as the Lift-off
Point of Intersection, however, does not vary significantly
with variation in lift-off distances and this feature can be
used to provide a qualitative and quantitative evaluation of
the extent of the corrosion in a given test object.
Determining a Lift-off Point of Intersection for a
test object is relatively simple. A representative area of
the structure without any defects is first selected to be
inspected by the pulsed eddy current techniques described
above in order to provide calibration curves of signals
(responses) detected by the transducer or sensors due to the
eddy currents generated. For that given calibration location,
at least two but preferably three transient responses are
recorded where only the transducer (or sensors) distance from
the structure is varied, i.e. where only the lift-off distance
is varied. The time at which the two or three lift-off
balanced transient responses intersect is the Lift-off Point
- 13 -
CA 02312101 2000-06-22
of Intersection. That time will be the same for any Lift-off
Point of Intersection for responses recorded at other areas of
the structure where defects might be present. If no
significant defects are present in an area being tested at the
Lift-off Point of Intersection, then the voltage amplitude of
responses at that particular time will be close to zero for
the referenced subtracted signals. The responses recorded at
that point of time when other areas are being tested can then
be evaluated to determine if any defect is present and the
extent of that defect. Depending on the amount of material
loss, the voltage amplitude of signals recorded at a test area
will vary and this will provide a means to provide a
quantitative evaluation of the material loss at that area.
Effectively, the use of a calibration standard will allow the
determination of material loss. It should be noted that the
Lift-off Point of Intersection also exists for the transient
responses before the reference substraction. Reference
substraction is, therefore, not absolutely necessary.
The signal display is the real link between the test
equipment and its intended purpose, i.e. detection and
identification of corrosion. The transient response, as
previously indicated, is a signal in the time domain and each
point in the surface scan has a particular transient response.
This situation limits the ability to assess the conditions of
the test object. Advanced imaging and image enhancement
software are generally used to provide a reasonable data
interpretation capability. This will make it possible to
represent the test object with a C-scan using only one feature
- 14 -
CA 02312101 2000-06-22
of the transient response, e.g. the Lift-off Point of
Intersection.
One type of transducer used to test a metal plate
with the pulse eddy current technique was a single coil where
the excitation and the sensing is carried out by the same
coil. The single coil used for the first tests had an inside
diameter of 6.6 mm, outside diameter of 13.9 mm a length of
0.685 mm and a wire gauge 41 AWG. Different size coils could
be used and, effectively, a higher sensitivity could
theoretically be achieved by building a larger inducing coil
as this would allow for a higher depth of penetration of the
magnetic fields. If, however, the same large coil is used for
detection purposes, a low resolution and sensitivity would be
achieved. The reason for this is that the coil would respond
to all magnetic flux lines passing through the coil winding,
regardless of the spatial direction and orientation. This,
together with the size of the coil can impact on the
sensitivity of sensor. Using a smaller coil would provide a
better sensitivity but would limit eddy current penetration
within the test object. Separate coils can be used to provide
the excitation and the sensing.
One setting of particular interest is the pulse
width. The selected width must be sufficient to allow the
single coil transient response to reach its maximum and
subsequently decrease to zero. Valuable information about the
test sample may be lost if the width is inadequate to allow
this to occur. There is also a requirement to determine an
adequate sampling rate for a given transient response. A
- 15 -
CA 02312101 2000-06-22
higher sampling rate gives more data points per unit of time
and therefore produces improved accuracy in representing the
original signal. The sampling rate should be at least twice
the highest frequency measured.
The inspection of any test object is best
accomplished with the help of a scanning system. In an
experimental set-up for testing, the test sample was kept
stationary and a scanning system alters the position of the
transducer to cover the surface of the test sample. The
operation of the scanning system is controlled by commands
from the central computer having pre-selected parameters such
as the scanning and index axis, the dimension of the area to
be scanned and the scanning speed.
A multi-layer structure was first tested with no
lift-off and at areas where no defects were present to obtain
a reference signal subsequently subtracted from all other
transient responses. Then, the single coil transducer was
positioned at various distances from the plate in order to
determine a Lift-off Point of Intersection as illustrated by
balanced transient response curves 1, 2 and 3 in Figure 2,
Figure 2 being an expanded view of Figure 1 near the Lift-off
Point of Intersection in order to more clearly illustrate the
balanced transient response curves at that area. In this
case, with this particular metal plate structure, the time
where the measured voltage amplitude of the three balanced
transient responses intersect (the Lift-off Point of
Intersection) was about 36 us. That metal plate was then
tested at other areas where known defects with a 14.4
- 16 -
CA 02312101 2000-06-22
material loss were present. The responses obtained were
recorded and are shown in Figure 1 and the expanded view of
Figure 2. The separation between the obtained signals at
areas where defects are present and an area where no defect
was present at the Lift-off Point of Intersection are clearly
illustrated by these curves. In Figures 1 and 2 the plate was
tested at an area where the 14.4% material loss defect was
present at the bottom of the top plate, at an area where the
defect was present at the top of the bottom plate and at an
area where the defect was present at the bottom of the bottom
plate. These are identified in Figure 2 by the various types
of lines representing the different response curves.
The Lift-off Point of Intersection value (about
36 us) is identical for each curve in Figure 2 and the signal
is almost zero when no defect is present as shown by curves 1,
2 and 3 at the Lift-off Point of Intersection. It is possible
to ascertain the presence of corrosion in a multi-layer
structure independently of lift-off variations from the
separation between the curves at the Lift-off Point of
Intersection located near 36 us. The curves in Figure 2 only
highlight the ability to determine the presence of corrosion.
The Lift-off Point of Intersection, however, also provides
the ability to quantify the amount of material loss in a
multi-layer structure. A variation in the material loss will
translate into a variation of the voltage amplitude of
responses at the Lift-off Point of Intersection and this is
illustrated in Figure 3. The lift-off Point of Intersection
was determined to be around 36 us, similar to that in
- 17 -
CA 02312101 2000-06-22
Figures 1 and 2. The middle group of curves at the Lift-off
Point of Intersection were obtained at an area of the metal
plates where a 14.4°s of material loss existed while the lower
group of curves were obtained at an area where a 25.5%
material loss existed. It is possible, as illustrated in
Figure 3, to quantitatively evaluate the amount of corrosion
(or material loss) by using a calibration specimen to
ascertain the voltage amplitude associated with a given loss
of material.
Further tests were made on the structure with another
type of transducer, a driver-pickup transducer. With this
type of transducer, the excitation and sensing are carried out
by two coils having different characteristics. The driver-
pickup transducer configuration used for these tests consisted
of two concentric coils with an excitation coil having the
same dimensions as the single coil transducer. The detection
coil had an inside diameter of 1.59 mm, an outside diameter of
6.35 mm, a length of 0.660 mm and a wire gauge 45 AWG. The
testing and analysis carried out using this transducer
followed the same procedure as previously done with a single
coil. Figure 4 shows a number of balanced transient responses
combining flaw locations and lift-off distances obtained with
the driver-pickup transducer. Figure 5 is an expanded view of
Figure 4 at the Lift-off Point of Intersection area. This
figure clearly shows the separation between the curves at the
Lift-off Point of Intersection at an area where no defects
were present (calibration curves) and at areas where defects
are located.
- 18 -
CA 02312101 2000-06-22
The possibility to quantitatively determine the loss
of material due to corrosion is not readily apparent with
Figure 5. However, the Lift-off Point of Intersection also
provides the ability to quantify the amount of material loss
in a multi-layer structure. The same type of testing was
carried out as previously done with a single coil transducer.
The balanced transient responses obtained are shown in
Figure 6, and demonstrate the ability to quantitatively
evaluate the amount of corrosion (or material loss) by using a
calibration specimen to ascertain the voltage amplitude
associated with a given loss of material.
Various modifications may be made to the preferred
embodiments of the invention without departing from the spirit
and scope of the invention as defined in the appended claims.
- 19 -
