You are on page 1of 8

International Conference HR9800024

Nuclear Option in Countries with Small and Mec


Opatija, Croatia, 1996

DETECTING OF CRACKS BY USING +POINT EDDY CURRENT


PROBE

RENATO GRACIN, dipl.ing.


DARKO BARILAR, dipl. ing.

INETEC
ZAGREB, Koturaska 51

Abstract : The development of eddy current "+Point" probing technique makes possible to
meet demands of finding small cracks in difficult parts of steam generator tubing, such as top
of tube sheet and sleeved part of tube. Another application of "+Point" probe is surface
examination of weld or other rough surfaces. This article will present way of detecting
direction of cracks as well as on site inspection experience.

The appearance of circumferential cracking in nuclear power plant steam generator tubes has
been increasing in recent years. Even, many of the cracks were discovered as result of actual
tube leaks. These tube leaks required unscheduled plant shutdowns to be fixed. So, improving
of the minimum level of detection was reason for new development emphasis to be placed.
Initially, multiple-coil surface riding array coils were used for detection of circumferential
cracking in most difficult parts of steam generator tubing (where bobbin coil is not giving
satisfactory results).
The following drawing (Figure 1.) presents standard bobbin coil and multiple coil array probe.

Figure 1.

Next, the most significant step in development of such probe was design of rotating probes.

185
Figure 2. illustrates rotating probe inserted into tube and cross section view of rotating probe.

Figure 2.

During helical scan of tube, presented on Figure 3. this probe gives better results.
Figure 3.

360

tube

crick
axial step

Figure 4.

For this purpose, a new algorithm in analysis


• • • • 1m a ^^ software was developed, i.e. line scans were
joined by accomplishing probe rotation, so two
dimensional presentation (C-Scan) of tube
circumference was can be obtained on analysis
screen (Figure 4.).

Further, to standard pancake coil, on probe body


were additionally mounted two linearly wounded
coils. On was more sensible for axial, and one
Kl.U
for circumferential cracking, as it is shown on
following drawing.

Figure 5.
However, even with all these improvements,
undetected cracking still occurred. The main
problem that remained in many cases, is to detect
and discern cracks clearly from other larger
signals generated by expansion transitions, top of
tubesheet interfaces, magnetic and conductive
deposits. These conditions was reason for
detectable cracks to remain unreported what
resulted in unplanned plant shutdowns.

186
It was clear that additional improvements were required. The entire eddy current system had
to be considered, but for this kind of development focus was on probe design.
The main goal was to increase minimum detection and minimize interference caused by
expansion transitions and top of tubesheet. Therefore, +Point rotating probe was designed.
Why?
Figure 6.
Software for rotating probe already provided advanced analysis tool.
With standard 3-coil rotating probe, this software was not sufficient
because interfering signals were.much larger that signals which
indicates cracks. Therefore, idea was to improve quality of data input
30 mew t0
analysis software, meaning to provide data less influenced by lift-
off, support structures, geometry changes and deposit signals. Solution
found in +Point coil configuration is illustrated on Figure 6..
The design consists of two coils differentially paired electrically, and
physically oriented to generate two eddy current flow paths at 90° to
one another. With this coils configuration, lift-off and magnetic effects
due to geometry changes were significantly reduced, while the
interruption of each of the two eddy current paths made
\ circumferential and axial crack orientations clearly distinguishable
from each other.
What is difference between eddy currents flow with +Point coil if it is compared to bobbin
and pancake coil ?
As eddy currents flow in the same plane in which the coil generating them is wound
(conventional bobbin coil case), it would generate eddy currents parallel to the direction of
circumferentialy orie-nted crack.
Figure 7.
Recommended probe diameter. Because of little distortion
(nominal tube 10 -0.045) of eddy currents caused by
Nominal tube ID+0.015 — circumferential crack, there
is no detection.
Pancake coil is wounded to
generate eddy currents in
circular path that can be
T distorted by all axis of
Rotational cracks - what is desirable
direction
effect. However, conven-
0.115 Primary pancake test coll
tional absolute technique is
+ Point coil highly sensitive to geometry
O.OSO High-frequency shielded coil Pancake test cofl
changes, lift-off and
No. 3 coil No. 2 coll deposits, causing cracks to
be less detectable.
The +Point coil(s) generates axial and circumferential eddy currents to provide multi axis
crack detection possible. But, more important is suppression of unwanted influence of noise
signals what is achieved by operating them in a differential configuration.
Example of +Point probe head is given in Figure 7.. This probe consists of two pancake coils
(one conventional one shielded) and +Point coil. Presented probe is designed for top of tube
sheet examination.

187
Although +Point probe is big improvement, some influences from transition signals and
deposits still remained in eddy current data, but not nearly as badly as with absolute pancake
coil.
Nevertheless, these signals still have potential to detect smaller crack indication signals.
Advanced algorithm available in eddy current data analysis software is one of solutions. With
the raw data containing interfering signals available software algorithm became much more
effective.
In the USA during actual field application of +Point probe, several tubes were pulled out of
various steam generators, and analyzed destructively.
Results showed that rotating +Point coil technique is better than all previously used eddy
current techniques, especially in critical steam generator regions as expansion transitions, top
of tubesheet, sleeve attachment welds and U-bends where large signal interference occurs.

The following will be illustration of all mentioned before. All presented figures are plotted
from data acquired by METEC on inservice inspections and showed in C-Scan form.
Figure 8.
Figure 8. presents ASME calibration
standard obtained by conventional 3-
coil rotating probe. For this purpose is
used pancake coil on 400 kHz. ASME
calibration standard (plotted part)
consists of through wall hole, 80%,
60%, 40% and 4x20% flat bottom
holes (example, 60% means that hole
depth is 60% of tube wall thickness).

Figure 9.

Figure 9. gives the same, just obtained


by using +Point probe. It may be
noticed that signal indicated by
damage is going up and down what
indicates that damage has volumetric
nature (what really hole is).

188
Figure 10.
Part of rotating probe calibration
standard with 4 flat bottom holes and
two circumferential notches is
presented on Figure 10. obtained by
pancake coil.

Figure 11.

Figure 11. gives the same data


acquired by +Point probe.

Figure 12.
Real data from steam generator,
collected in region of top of tubesheet
by using probe head with two pancake
coils and +Point coil (schematic is
given in previous text) is the following
presented (Figure 12.). C-Scan is
plotted from pancake coil data. As it is
visible, geometry change signal is
interfering with circumferential crack
(not deep) signal. Crack is detectable,
but it is not clearly visible and might
happen to remain unreported.

189
Figure 13.
But, if the same crack is reviewed by
using +Point data presented on next
figure, it may be said that crack is easy
detectable. Signal is going up, what
indicates that circumferential crack is
present.

Figure 14.
The following two figures illustrate
ac« MM mo. tun
data from the same tube (top of
tubesheet location) acquired by 3-coil
rotating probe, i.e. way of examination
of top of tubesheet as it was till +Point
probe was designed.
Figure 14. gives pancake coil data
(this data should be the same as on
pancake coil on +Point probe head
data showed before).

Figure 15.
Figure 15. gives data collected by coil
wounded to detect circumferential
cracks. Both coils recognize
circumferential crack located on the
top of tubesheet, but detectability can
not be compared with detectability
provided by +Point coil data (where
presence of circumferential crack is
obvious).

190
Figure 16.
As mentioned in previous text,
+Point probe gives also better
results in sleeve inspection. For this
purpose standard +Point probe is
improved to enable covering of
larger geometry changes of tube
(tube diameter variation). This
probe is called GPP (Gimbaled
+Point Probe). On probe head in
this case is mounted only on +Point
coil. Before +Point probe, for sleeve
inspection is used I-coil probe. This
probe did not give satisfactory
results, and indications on parent
tube where sleeve was installed
were not visible.
Figure 17.
Example of I-coil data from Sleeve
calibration standard is given in
Figure 16.. Presented part of
calibration standard contains two
parent tube through wall holes, two
60% and two 20% flat bottom holes.
As visible, I-coil hardly detected
only through wall holes and 60% flat
bottom holes.
If this data is compared with data
acquired by GPP probe presented on
Figure 17., may be said that GPP
probe gives much better results.

Figure 18.
The following two figures gives data
acquired by GPP probe on real
sleeves installed due to indications
initiated on top of tubesheet location.
On Figure 18. is presented sleeve
which covers large parent tube
circumferential crack. It shall be
noticed that signal is going up what
indicates presence of circum-
ferentialy oriented crack.

191
Figure 19.
Figure 19. presents sleeve on
other tube where +Point probe
detected two parent tube cracks.
First one is on the top of tubesheet
and it is circumferential (crack
signal is oriented up). The other
one is located on tube outside
tubesheet. As visible this crack
signal is going down what
initiates that crack is oriented
circumferentialy.
Considering the fact that main
problem of I-coil is lift-off
interference, and that this problem
is satisfactory resolved with GPP
probe the difference between this
two probes is more than obvious.

The other, very important application of +Point probe (not related to steam generators) is
surface examination which requires scanning over a raised weld, where similar signals
interference has to be overcome. Some residual magnetism, existed in the welded region, is
causing large permeability signals, and the weld rough geometry creates lift-off interference.
Requirements to detect and distinguish transverse ad longitudinal cracks leads to +Point probe
which proved to be the best solution.

INETEC apply improved +Point probe for weld examination in penetration tube to sphere
joint in Reactor Vessel Head, where this probe is giving satisfactory results.
When using surface probe (not rotating) eddy current signals are the same as when using
rotating probe inside tube, with difference that C-Scan (two dimensional presentation) can not
be performed.

Conclusion : Every non-destructive testing technique consists of a set of compromises


between variables such as minimum detection, characterization of tube damage, production
rate requirements, available technology and ultimately costs. The emphasis on improvement
on any of them changes according to requirements required by operating experience. In
nuclear steam generator in-service inspections, the +Point technology resulted from
development initiated by particular requirements of industry being served.

192

You might also like