 |
Description  |
|
|
TECHNICAL FIELD
In eddy current testing the trend is towards an increasingly more
widespread use of several carrier frequencies in order to overcome
different types of problems, for example the suppression of the lift-off
dependence and the like. One considerable, and limiting, factor in crack
detection--now as well as previously --is, however, that harmless surface
defects, i.e. surface defects in a test object which are not detrimental
to any subsequent process step to which the test object will be subjected,
give rise to false fault signals since they cannot be distinguished from
the surface defects (e.g. cracks) which will be detrimental .
DISCUSSION OF PRIOR ART
In crack detection on hot continuously cast billets, so-called oscillation
marks, for example, cause considerable problems in conventional testing.
Swedish patent application Nos. 7507857-6, 8302738-3, 8400698-0
(corresponding to U.S. Pat. No. 4,661,777 - filed on Feb. 8th 1985 in the
name of Tornblom),. and 8400861-4 (corresponding to U.S. Pat. No.
4,703,265 - filed on Feb. 15th 1985 in the name of Tornblom) describe
methods and devices which may be regarded as part solutions to this
problem. With the present invention in combination with the part solutions
mentioned, a possibility is provided of efficiently suppressing the
influence of oscillation marks on the crack detection process. The
invention operates efficiently also as a separate invention, and together
with, for example the disclosure set out in the afore-mentioned U.S.
patent applications, it provides an almost complete solution.
The reason for the previously experienced difficulties in pluri-frequency
eddy current testing is primarily to be found in the unwanted effects
caused by the different depth of current penetration of the different
carrier frequencies used adjacent to an oscillation mark and the like
harmless surface defect, because of the variation in inductive coupling
arising in that connection, as a function of the position of the
transducer, between the transducer and the test object.
The existing specialist literature--as far as is known--does not describe
any method, nor mention any means, which determines the causes of the
problems discussed, or how they can be overcome and solved.
It should be pointed out in this connection that one reason why the problem
has now become open to analysis and explanation is the introduction of the
imaginary sum currents which are described in the above-mentioned U.S.
Pat. applications and which are also employed as the basis of the
explanation of the present invention given herein.
Since all conventional eddy current transducers provided with a center hole
in the winding normally suffer from the deficiencies mentioned here, the
present invention should therefore result in a marked improvement of the
theoretical limit to the minimum size of cracks that are detectable in
relation to the level of occurrence of oscillation marks and the like,
compared with current technique.
The above-mentioned U.S. patents, incorporated herein by reference,
describe .methods and means by which a vector transformation or the like
can be optimised over the lift-off (LO) operating range by carrying out
the transformation as a function of the LO distance, and also describe how
to compensate for the various depths of the sum currents by a special
design of the transducer. Common to the inventions disclosed in these
earlier U.S. patent applications is that they are primarily effective in
suppressing undesired effects caused by a varying lift-off in combination
with the varying depths of penetration of the currents induced by the
different carrier frequencies. In other words, variations in height of the
transducer (measured perpendicular to the surface of the test object) can
be suppressed. However, the inductive coupling between the transducer and
the surface of the test object is dependent not only on the distance (LO)
but also on the length of the surface current path covered by the
transducer coil in the transverse direction (along the surface).
Therefore, the present invention, which also permits harmless surface
defects with a longitudinal/transverse extension to be ignored, is to be
considered an important complement to the above-mentioned U.S. patent
applications, by means of which a more three-dimensional possibility of
suppression is obtained. Taken together, these inventions then permit, for
example, an efficient suppression of the unwanted influence on, for
example, a crack detection operation, of the presence of oscillation marks
and the like harmless surface defects.
SUMMARY OF THE INVENTION
According to the invention a device utilizing eddy current techniques for
inspecting a test object for surface defects, for example surface cracks,
comprising at least one transducer/sensor, which is made to scan the
surface of the test object, is characterized in that the effect of
harmless surface blemishes on the transducer/sensor is suppressed,
completely or partially, by the device being compensated for the
sensitivity characteristic, in the scan direction with respect to the
surface blemish.
The invention will now be described, in somewhat simplified terms as
follows. It should be pointed out, however, that both the description and
the accompanying drawings are to be considered one of many feasible
alternatives or examples of how the invention can be realized. The
drawings are not accurate as far as scales and dimensions are concerned
and are to be regarded as examples illustrating the principle of the
invention.
To make the description more easily comprehensible, mathematical
derivations have been replaced by relevant--in some cases
approximate--reasonings which, despite their simplicity, are well founded
both in theory and in practice. For the same reason, both the description
and the drawings have been based on the use of only two carrier
frequencies. However, the invention does, of course, include the use of
more than two frequencies or complexes of frequencies.
BRIEF DESCRIPTION OF THE DRAWINGS
The invention will now be further described by reference to the
accompanying drawings, in which:
FIG. 1 is a schematic view, from above, of a sensing transducer used for
defect detection supported over a moving test object,
FIG. 2 is a partial sectional view, on an enlarged scale, showing a sensing
transducer and the high and low frequency currents induced in the surface
region of a test object by the transducer,
FIG. 3 is a graph showing the electrical output of the sensing transducer
as a function of X as a harmless surface defect moves past/the transducer,
FIGS. 4 and 5 show, in transverse cross-section two embodiments of sensing
transducer for use in a device according to the invention,
FIG. 6 a simplified block diagram of crack detection equipment in
accordance with this invention.
FULLER DESCRIPTION OF THE INVENTION AND PREFERRED EMBODIMENTS
FIG. 1 shows a test object in the form of a billet 1 exhibiting on its
upper surface an oscillation mark 3 (hereinafter abbreviated as "oscm"). A
conventional circular surface transducer 2 of external diameter D.sub.Y
and internal diameter D.sub.I moves over the surface of the billet 1 at a
constant height above the surface and at a constant speed of v m/s.
The way FIG. 1 has been drawn, the transducer winding overlies only a
length L of the total length of the oscm. This length L will, of course,
vary as the transducer winding moves across the oscm, that is to say, L is
a function of X in FIG. 1. For reasons which will be easily understood,
the inductive coupling between the transducer winding and the surface of
the billet is greatest directly below the winding.
Since in practice, the oscm almost invariably has an extension extending
transversely across the billet 1, and its width W (see FIG. 2) is small in
relation to the external diameter D.sub.Y of the winding 2, the
change/disturbance of the magnetic field caused by the oscm will
approximately be proportional to the length L. This then means that the
impedance of the winding 2 varies in a manner similar to that in which L
varies as the winding 2 passes over the oscm. Since the winding 2 has a
hole or core in the center, which is the case with all crack-detection
transducer windings known to me, a plot of L as a function of X will be
similar to graph 5 in FIG. 3, which is an important statement for an
appreciation of this invention. The dip shown centrally in the peak of
graph 5 in FIG. 3 is therefore to--to anticipate the reasoning somewhat--a
consequence of the internal hole (with the diameter D.sub.I) in the
winding 2.
The extended shape of the oscm, and the simplicity of showing L as a
function of X, facilitates the understanding of the principle behind the
invention. In general terms, however, for all uncompensated coils the rule
applies that they always have sensitivity characteristic functions at
differing carrier frequencies which exhibit the difference D.
FIG. 2 is an enlarged sectional view of a transducer winding 2 supported
over the surface of a test object 1 which contains a surface blemish 3.
The winding is fed with currents at two different frequencies which give
rise to induced currents 10 and 11 at different distances below the
surface of the object 1. The sum currents 10 and 11 shown are the same
imaginary currents as are shown in FIGS. 1 and 2 of U.S. Pat. No.
4,703,265 and in FIG. 1 in U.S. Pat. No. 4,661,777.
FIG. 3 shows, as explained above, the sensitivity characteristic
(hereinafter abbreviated to "SC") of the transducer/sensor 2 for a certain
type of surface blemish, in this case an oscm. Thus, SC is the sensitivity
curve sensed by the transducer/sensor 2 in passing over an oscm. The graph
is plotted as a function of X, at a certain LO-distance. X=0 has been
selected to represent the position where the transducer winding is
situated exactly above the oscm.
Because currents of different frequencies are influenced somewhat
differently by an oscm, two SC's are shown in FIG. 3, i.e. one for each
respective carrier frequency. Curve 4 shows SC.sub.L (low frequency) and
curve 5 shows SC.sub.H (high frequency). When the oscm 3 passes under the
transducer winding 2, the sum currents of the high 11 and low 10
frequencies, respectively, will be depressed by Z.sub.H and Z.sub.L,
respectively, by the oscm, as is shown simplified in FIG. 2.
As is also clear from FIG. 2, Z.sub.H /LO.sub.H>Z.sub.L /LO.sub.L, which
means that the impedance change caused by the oscm in the transducer
winding 2 is relatively greater for the high frequency current than for
the low frequency current. This is one of the reasons for the dip on the
SC.sub.L -curve being less significant than the dip on the SC.sub.H
-curve. The result of this is that the oscm will give rise to different
SC-curves for the different carrier frequencies which are used. The
consequence of this, in turn, is that a difference D is obtained between
the curves 4 and 5 which is shown shaded in FIG. 3. Depending on the shape
of the winding 2 etc., this difference D may contain primarily second and
third harmonics but in certain cases also harmonics of a higher order.
Both the difference D per se and its harmonic content are greatly
disturbing for all types of signal processing, for example a
transformation, in which at least two carrier frequencies are used, since
crack detection is normally based on some type of difference measurement
between the frequencies. An example of such a difference measurement is
given in the above-mentioned Swedish patent application No. 7507857-6. The
difference D can here be directly or indirectly construed as a crack (i.e.
a nonharmless blemish), which of course is a considerable disadvantage.
By imparting to different parts of the transducer/sensor different
sensitivites in the direction of relative movement of the transducer past
the object 1 as a function of the carrier frequency in question, it is
possible to compensate for the difference D. Such a compensation will be
referred to as a sensitivity characteristic compensation (hereinafter
abbreviated to "SCC"). It should be mentioned that the SC-functions,
including the SCCfunction, may involve more than just the transducer and
its windings. Thus, the SC-function may also, for example,
comprise--completely or partially--the rest of the crack testing or
measuring device.
The invention proposes both a method and a device for limiting, completely
or partially, the effects on the crack detection process caused by the
difference D and/or its harmonics or frequency contents. Both the
SC-functions for harmless surface deformations and the difference between
different SC-functions of different frequency origins, when using at least
two carrier frequencies, are novel both as regards the definition and as
regards the possibilities of understanding and remedying the undesired
effects caused by different depths of current penetration. This is the
reason why the invention can be considered to be characterized by the
device as described in the Summary of the Invention given above.
According to the invention harmless surface deformations, such as
oscillation marks and the like, can be suppressed, completely or
partially, by the sensitivity characteristic (SC) for the surface
deformation in question being largely the same or similar for/at at least
two different carrier frequencies and/or complexes of carrier frequencies.
The SC's, which are largely similar for the carrier frequencies or
complexes of carrier frequencies in question, may be obtained, completely
or partially, by signal processing, for example amplifying, signals from
at least one part of the transducer/sensor or the transducer/sensor
arrangement, as a function of the carrier frequency in question. Thus, an
SC-function can be formed, for example, by attenuating or amplifying
signals of a certain frequency from a limited part of the transducer
winding, which signals are then, for example, added to the other signals
from the transducer winding of the same frequency origin. It is possible
to obtain, completely or partially, the SC-functions, which are largely
the same or similar for the carrier frequencies occurring at a particular
time, by supplying parts of or a part of the transducer winding, having
different surface coverage, completely or partially with different carrier
frequencies or carrier frequency components.
Another characteristic feature may be that at least one transducer/sensor
includes at least two windings (or loops) of different dimensions and/or
shapes. By making, for example, a part of the transducer/winding
adjustable relative to the rest of the arrangement, the optimum
SC-function for the surface deformation in question can be tested in a
simple manner. As an example, a small coil can be screwed out of or into a
larger coil in order thus to optimize the SC-function. Another way of
achieving a suitable SC-function is to form the coil so that it, per se,
exhibits a suitable characteristic, which normally presupposes a
non-square or rectangular shape of the cross section of the winding. By
making the ratio D.sub.Y /D.sub.I large, for example > 5, the dip on the
SC-curve is reduced, which may in certain cases be a sufficient
compensation.
If it is desired to obtain maximum performance, it may be suitable to
minimize the difference D in FIG. 3 further, in addition to what can be
carried out at the transducer winding. This can be done, for example, by
signal processing, for example shaping, signals originating from
completely or partially different carrier frequencies, for example
rectified carrier frequency signals, differently prior to the
transformation and the like, since in that case the transformation is not
disturbed.
In eddy current testing, magnetic fields of a relatively high frequency are
used, which may, for example, be generated by supplying a winding from a
so-called constant current generator with current of different carrier
frequency contents. In this way, one winding may act both as a primary
coil and as a secondary coil or--where desired--as the transducer and as
the sensor at the same time. FIGS. 4 and 5 may therefore, for the sake of
simplicity, be regarded as a transducer/sensor with a common primary and
secondary winding. Thus, the current supply and the sensing can take place
via the same connection, if this is desired. However, the current
generators have been omitted in these Figures so as not to confuse
matters.
FIG. 4 shows a transducer/sensor comprising two coils 6 and 7 of different
sizes and shapes These coils are also shown in FIG. 6. The larger coil 6
is sensed with respect to both the carrier frequencies, i.e. .omega..sub.L
and .omega..sub.H, whereas the smaller coil 7 is only sensed with respect
to the high frequency .omega..sub.H, which is also clear from FIG. 6. The
SC-function from the larger coil 6 may then, for example, have the
appearance as illustrated in FIG. 3. However, by subtracting from curve 5
in FIG. 3 the contribution obtained from the smaller coil 7, placed at the
center of the transducer, the dip in curve 5 can be reduced and hence be
made equivalent to that in curve 4. In this may the difference D is
completely suppressed and resultant optimum SC-functions are obtained
which are insensitive to harmless surface blemishes.
Because of the fact that the smaller coil 7 has smaller dimensions, for
example--as in this case a smaller diameter, its SC-function (curve 55 in
FIG 3) will have a more limited extension than that of the larger coil 6.
It is then readily appreciated that its dimension can be adapted so as to
become optimal with respect to the difference that it is to be compensated
for.
FIG. 5 shows an alternative form of transducer/winding, in which one coil 8
is fed, for example, from a constant current generator, containing the
carrier frequencies .omega..sub.H and .omega..sub.L, at one point and with
only .omega..sub.H at another point. In this way, with an equivalent
reasoning, sensibly the same result can be obtained as with the
transducer/winding shown in FIG. 4.
In principle, it is possible to design transducers/sensors comprising an
unlimited number of windings of, for example, different shapes and
dimensions, and to allow these to cooperate so that SC-functions are
created which are suited for suppression of an arbitrary type of surface
blemish. It will then be possible to combine parallel or series-connected
channels in signal processing equipment in which, for example, the task of
each respective channel is to suppress the effect of a specified type of
surface blemish.
In the same way as the effect of surface blemishes can be suppressed,
SC-functions can be created which emphasize harmful blemishes in the
surface, for example cracks. "Tailoring" a transducer for detection of a
certain type of surface blemish or other defect is then the inverse use of
the invention.
FIG. 6 shows a simplified block diagram of one form of crack detection
equipment. The transducer/sensor comprises two windings 6, 7. The output
signals from these are amplified in amplifiers 12 and 13 and are then
filtered in a band pass filter 14 with respect to the high frequency and
in a band pass filter 15 with respect to the low frequency. Since the
signals from coil 7 and amplifier 12 are only supplied to the filter 14,
the coil 7 only provides a contribution to the high frequency signal,
which means the SCC of that part of, for example, the billet surface which
is covered by the coil 7, or better still, that part of the
transducer/sensor 2 which is represented by the coil 7. The output signals
from the filters 14 and 15 are further amplified in respective amplifiers
16 and 17 and are then rectified in respective phase-sensitive rectifiers
18 and 19. In case the above-mentioned SCC is not sufficiently efficient,
it may be justified to perform an additional SCC of the rectified signals,
which in FIG. 6 is effected electronically in the blocks 20 and 21, which
may comprise summation amplifiers. These can perform in certain cases
sophisticated, signal processing which may include, for example,
filtering, adjustable signal delay, pulse shaping and so on, and results
in the finally SCC-compensated SC-functions SC.sub.H and SC.sub.L,
respectively, which constitute input signals to subsequent transformation
blocks 22, 23 and 24. As will have been clear from the above, the original
SC-functions have been compensated for both by the transducer/sensor
arrangement used and by the signal processing undertaken. The amplitude of
the SC-functions is suitably adapted to, for example, the current
transformation setting. If, for example, the transformation is chosen for
optimum LO-suppression, a suitable SC-amplitude is chosen among the
functions, starting from the current transformation setting. A band pass
filter 25 finally filters out a current fault signal F, representing, for
example, the presence of a crack on the billet surface or other harmful
blemish on the test object.
Since the SCC is carried out prior to vector transformation, the existence
of harmless surface deformations does not, of course, influence the
possibility for reliable crack detection.
Scope of terms used in the Specification
By SENSITIVITY CHARACTERISTIC (SC) is meant the--e.g.
normative--sensitivity function/curve which is obtained upon relative
movement of the transducer/sensor past a harmless surface blemish or
harmful defect. The SC refers to a certain carrier frequency or complex of
carrier frequencies, and to a certain surface blemish, and in some cases
also to a suitable LO-distance.
By SENSITIVITY CHARACTERISTIC. COMPENSATION (SCC) is meant that the
difference (D) and/or the differences between different original
SC-functions is/are completely or partially compensated for or balanced
out.
By CARRIER FREQUENCY is meant the frequency of that current which generates
a magnetic field, in other words, the frequency of the transducer current.
By TRANSFORMATION and VECTOR TRANSFORMATION is meant for example, vector
transformation as described, for example, by Libby in U.S. Pat. No.
4,303,885, in Swedish Patent No. 7507857-6, and, inter alia, in U.S. Pat.
application No. 699594.
By TEST OBJECT is meant, for example, a hot billet, a sheet, a tube or a
section.
By LO-DISTANCE is meant the lift-off, i.e. the distance between the
transducer/sensor and the surface of the test object.
By TRANSDUCER is means any type of transducer and/or sensor operating with
or being otherwise sensitive to a magnetic field. The term transducer can,
for example, include everything from a simple loop to complicated coil
arrangements of a three-dimensional nature.
A surface deformation, for example an oscm, normally differs from a crack
(see C in FIG. 2) by having a greater width (W) in the scan direction v
than the crack.
In the case of rotary symmetrical transducer embodiments it is simple, for
obvious reasons, to perform an SCC on the transducer in all directions of
movement, which means that the transducer may have an arbitrary direction
of movement with a retained SCC.
The invention can be varied in many ways within the scope of the following
claims.
* * * * *
|
|
 |
|
 |
Description |
|
