TOFD made simple
Response to enquiry from John Johnson at INEEL, Idaho Falls re
"TOFD made simple".
I've been working on the development and application of TOFD
methods at AEA Technology for the past 20 years and I hope
the comments below will be helpful.
The experts on TOFD in the US are at AEA Technology Engineering
Inc., contact Everett@aeatech.com . They can give you any
assistance you might need to inspect welds using TOFD.
Alternatively you can contact any of the AEA Technology offices
worldwide by getting references from our website. TOFD was first
invented by our National NDT Centre here in the UK and we have
accumulated extensive expertise in its application.
In the meantime I can offer you the following answers to your
Q1. Why use L-waves?
A1. There are lots of reasons for the use of L waves and against
Sv waves :-
a) As you said. It reduces ambiguity about sources of signals.
b) The diffraction coefficient for L waves varies with angles
subtended, but there is only one narrow range of angle conditions
for which the coefficient is null (sum of angles subtended at Tx
and at Rx is about 76 degrees). Using Sv waves there are lots of
null amplitude conditions so detection is unreliable.
c) Using L waves the signal phase reliably distinguishes between
defect top and bottom edges except that, if the angle sum (see
above) is less than 76 degrees, then the crack bottom signal
phase is inverted. Using Sv the phase pattern is very dependent
on subtended angles, so you lose any benefit from slower velocity
due to uncertain timing references.
d) Using l waves you can use the lateral wave as a reliable
timing reference. Using Sv there is no equivalent to the lateral
wave, only the Rayleigh surface wave which has a slower velocity
so that it obscures part of the region of interest.
For these reasons Sv waves are not used by well informed
inspection agents, for manual or automated methods, apart from
very specific cases.
You can use Sh waves but they are inconvenient.
The arguments about beam divergence and velocity are not relevant
because you can match wavelengths to achieve comparable
divergences using L or Sv waves.
A2. You have to compromise on your probe selection in all
ultrasonic inspection techniques. There is always a trade off
between s/n ratio and resolution because both vary with
wavelength. You ask how to compromise. Use the highest frequency
at which you can penetrate your material and see your defects of
interest with adequate s/n performance. Begin by trying probes
small enough to give you a good beam divergence to cover your
region of interest. If this gives you inadequate resolution, you
may have to use higher frequency probes to improve the resolution
and you may then also have to choose probes with larger crystals,
which give you a narrower beam, less coverage but better s/n.
If grain size is a problem, use of Sv makes the s/n many times
worse at the same frequency and slightly worse at the same
wavelength. If the material is anisotropic (eg austenitic welds),
then Sv is affected far worse and you cannot expect to get any
useful sizing results without knowing all about the velocity
Q3. Interference from the crown
A3. None. The lateral wave results from the quickest possible
path between the probes (Fermat's principle) and is not affected
by convex features in the surface profile. It is hardly affected
even by shallow concave features as the indentation would have to
be large to significantly obstruct or delay the lateral wave.
Q4. Stainless steel.
A4. We at the National NDT Centre, AEA Technology in the UK, and
also in Holland, have had lots of experience of trying to use
TOFD on austenitic castings and welds. Results have been
exceptionally variable and unpredictable. Some inconel welds
cause virtually no problems and most duplex steels seem to be OK.
All austenitic stainless steel welds are anisotropic, so that the
velocities of the lateral wave, back-wall and intermediate depth
features are significantly different and the calculations become
complicated. eg The lateral wave is always delayed by slow
velocity through the top of the weld. The grain size is more
relevant to the stainless parent materials each side of the weld.
Cast stainless is virtually impenetrable for TOFD applications,
even using Sh waves (despite some claims to the contrary). You
are in with a chance if the parent material is well forged or
rolled so that the grain size is very small - but still beware of
the weld's anisotropy.
Q5. Other non-amplitude methods
A5. I'm not the best person to respond to this question.
TOFD remains the best method for sizing sub-surface defects in
ferritic steel and in several round-robin trials has also been
the best, or among the best, also for detection. (DDT trials in
UK, PISC-II (International), NIL (Netherlands), EPRI (US), etc.),
but it does need to be applied properly to achieve good results.
The book by J P Charlesworth and J A G Temple (John Wiley & Sons,
NY, ISBN 0 471 92386 9) is a good reference but is currently out
of print. A second edition is expected later this year.) Also
there is the British Standard BS7706 and a European working
pre-standard pr ENV-583 Part 6.
c/o AEA Technology, UK