Quantitative Guided Wave NDE
Quantitative Guided Wave NDE
ABSTRACT
Guided waves have demonstrated their value in flaw detection over long
distances from just a single probe position. Inspection under coatings,
fluids, and insulation is often possible. By way of phase velocity and
frequency tuning, defect detection sensitivity and location analysis can be
superb despite distance and environmental constraints. The ability to go
beyond detection and location to classification and sizing, however, is
extremely difficult. The use of the Boundary Element Method for defect
characterization through guided wave mode conversion is considered in
this paper. An outline of the approaches used, and sample results showing
features that might be useful in classification and sizing analysis is
covered. The question of which mode and frequency should be selected to
impinge onto a defect and have the best chance of classifying or sizing that
defect is addressed.
INTRODUCTION
Ultrasonic guided waves are being used extensively for defect and
materials characterization; some theoretical and experimental results that
using BEM are presented in [1-5]. The finite element method (FEM) for
Lamb and SH wave scattering problem has been analyzed in [6-9]. An
electromagnetic-acoustic transducer (EMAT) technique for SH wave
excitation is presented in [10]. The defect within an incident guided wave
path impacts the signature of the scattered fields by exciting all possible
modes for the given frequency value. The study of the scattered field
(reflected and transmitted) provides essential features for defect
characterization. Efficiency of this approach can be improved by selecting
propagating modes sensitive to the different kinds of defects.
Two different cases of an elastic field are considered: the first is for Lamb
wave propagation modes and the second for a shear horizontal (SH) field.
BEM results are presented for a 10 mm steel plate and elliptical wastage
parameters of 2a=6.35 mm and various b values.
Sample phase and group velocity dispersion curve results are shown in
Figure 2. Details of the computational process and explanations of
dispersion curve generation and BEM development can be found in [1]. To
illustrate the process of BEM utility for wave scattering in a wave guide, a
sample problem of S0 mode impingement onto a series of different depth
elliptical defects is illustrated in Figure 3. Even though reflected and
transmitted modes are numerous including A0, S0, A1, S1, A2, S2, etc.,
whatever modes can exist at the incoming mode frequency value,
emphasis is placed on the A1 reflected and transmitted values. The A1 and
other modes can be identified experimentally by group velocity, as an
example, and measurement or adjustment of an angle beam transducer
receiving angle. The most interesting results are shown over specific
frequency values. Note that there is a monotonic increase in the A1 mode
reflection factor amplitude with defect depths around 0.25 MHz and
around 0.5 MHz also. In the transmission factor amplitude, there is a
monotonic decrease with defect depth around 0.3 MHz.
Fig 2: Lamb wave phase (a) and group (b) velocity dispersion curves for a
10 mm thick Steel plate (cL=5.9 km/sec, cT=3.2 km/sec)
Fig 3: Variation of reflection (a) and transmission (b) coefficients of the A1
mode for different cases of the ellipse parameter b, that is equal to 10%, 30%,
50% of the wall thickness for an S0 incident mode. (Areas of monotonic
behavior of R(A1) and T(A1) are marked with a rectangle.)
We'll now consider shear horizontal impingement onto the same elliptical
defect. Dispersion curves for the steel plate are shown in Figure 4. See [1]
for computational details. To illustrate BEM utility for SH wave scattering
in a wave guide, a sample problem of n=1 mode impingement onto a series
of different depth elliptical defects is illustrated in Figure 5. Note in
through transmission the monotonic decrease in amplitude as defect thru
wall size increases for almost the entire frequency range 0.25 to 0.8 MHz.
Transmitted amplitude as a function of frequency is approximately
constant which is also a useful result.
Fig 4: SH phase (a) and group (b) velocity dispersion curves for a 10 mm
thick Steel plate (cL=5.9 km/sec, cT=3.2 km/sec)
Fig 5: Variation of reflection (a) and transmission (b) coefficients of the n=1
mode for different cases of the ellipse parameter b, that is equal to 10%,
30%, 50% of the wall thickness for the n=1 incident mode
CONCLUDING REMARKS
ACNOWLEDGEMENTS
Thanks are given to the Gas Research Institute, Chicago IL for the support
of this work and in particular to Dr's Harvey Haines and Albert Teitsma.
REFERENCES:
https://www.ndt.net/article/wcndt00/papers/idn272/idn272.htm