Notches

The austenitic and, hence, anisotropic V-butt weld is embedded in isotropic steel it has a width of 10 mm at its baseline and a height of 30 mm. If a notch is present, it has a width of 1 mm and a height of 15 mm it is located at the right-hand side of the V-butt weld. The simulated transducer is a commercial 45°-shear wave probe (MWB45-2E). The parameters varied during the simulations are ... 

Simulations of that kind result in a wide variety of A-scans and wavefront snapshots. The first screening of this material reveals, that the simulations in which the transducer is coupling partly to the V-butt weld and partly to the steel exhibit quite a number of pulses in the A-scans because the coupling at the interface of the weld results — due to the anisotropic behavior of the weld — in a complicated splitting of the transmitted wavefront. The different parts of the splitted wavefront are reflected and diffracted by the backwall, the interface, and — if present — by the notch and, therefore, many small signals are received by the transducer, which can only be separated and interpreted with great difficultie.s. 

Only the simulations in which the transducer is coupling either to the V-butt weld or to the surrounding steel can be analyzed in a simple and intuitive way, which means that the different pulses in the A-scan signals can be related uniquely to the reflection or diffraction of the wavefront at the weld, the backwall, and/or the notch. 

Apart from the well-known notch base corner reflection for isotropic welds, the anisotropic case results in a second corner reflection for transducer positions well above the weld. 

Signals from the notch tip can only be identified for the herringbone grain structure and not for the perpendicular one. 

Second corner reflection The first corner reflection appears as usual when the transducer is coupled to the probe at a certain distance from the V-butt weld. The second corner reflection appears if the transducer is positioned well above the V-hutt weld. If the weld is made of isotropic material the wavefront will miss (pass) the notch without causing any reflection or diffraction (see Fig. 3(a)) for this particular transducer position. In the anisotropic case, the direction of the phase velocity vector will differ from the 45° direction in the isotropic case. Moreover, the direction of the group velocity vector will no longer be the same as the direction of the phase velocity vector (see Fig. 3(b), 3(c)). This can be explained by comparing the corresponding slowness and group velocity diagrams. 

Figure 7 V-butt weld with perpendicular grain orientation and an inclination of the interface of 15° left without, right with backwall breaking notch...

Notch tip With perpendicular grain orientation no notch tips are detected. The snapshots (see Fig. 7 second and third snapshot from top on the right) show that in the direction to the receiver there is a gap in the reflected / diffracted quasi shear wavefront. In contrast, the notch tips can be detected within the herringbone structure. 

Becker et al. (11) have performed extensive experiments on surface-breaking cracks, tilting both the cracks or the back-side. The cracks are like a half ellipse, but could presumably be reasonably approximated by a strip-like crack. Figure 3 shows a comparison between the experiments and UTDefect for a 2.54 mm crack with varying tilt. The thickness of the plate with the crack is 15.24 ram. The probe is a circular 45 SV probe with frequency 2.25 MHz and diameter 12.7 mm. The experiments are calibrated with a notch but this is presently not... 

Secondly, the linearized inverse problem is, as well as known, ill-posed because it involves the solution of a Fredholm integral equation of the first kind. The solution must be regularized to yield a stable and physically plausible solution. In this apphcation, the classical smoothness constraint on the solution [8], does not allow to recover the discontinuities of the original object function. In our case, we have considered notches at the smface of the half-space conductive media. So, notche shapes involve abrupt contours. This strong local correlation between pixels in each layer of the half conductive media suggests to represent the contrast function (the object function) by a piecewise continuous function. According to previous works that we have aheady presented [14], we 2584... 

Now, we can make the comparison beween the real defect signal and the simulated one which have been computed by solving the linearized direct problem. The measurements were made at 300,150,50 kHz. The flaw is a notch of 8mm length, 1mm width, and 1mm depth. Representative data (300 kHz) for the notch-shaped flaw are shown in Fig. 3. 

Before starting the realisation of silicium coils, we construct a simpler probe with a 27 mm long excitation coil and 16 wired coil sensors, with 1 mm diameter and 1 mm long. We cannot hope any reconstruction with such a probe, but this allows to validate the whole approach. We tried it with test tubes with longitudinal or circonferential notches, external or internal, 100 pm wide, 100%, 60%, 40%, 20% and 10% deep. Our attention has been especially focused on circonferential notches, which are difficult to detect with usual probes. For example, the measurement signals at 240 kHz standard frequency are shown figure 6 7. 

Figure 11 Single detector signal provided by the inspection of inner circonferential notch tube sample, and corresponding scalogram. Time axis (in s) and frequency axis (in Hz) have been scaled according to the speed of evolution of the detector in ihe tube (500 mm/s).

Easily interpretable patterns are observable the first one is an horizontal pattern, located around 15 Hz, and is due to Pilgrim noise. Four regularly spaced vertical patterns are also observable, and shows the detection efficiency. They are due to the presence of the 4 inner circonferential notchs of successive 60%, 40% 20% and 10% deepness notchs. 

Scalogram applied to single detector signal allows notch localization along the main axis of the tube. However, no circonferential localization is possible so far. Of course, this objection can be bypassed by computing the scalogram simultaneously for the 16 detectors. But then the difficulty lies in the representation process, because of the need of real 3 dimensional representation. 

The comparison between the detection cartography (fig. 12) and the signal cartography given by the sensors (figure 7) for the same tube sample, shows the efficiency of our detection technique. Indeed, in the detection cartography, every inner notch present in the tube sample, even the 10% deep notch, is detected... 

Figure 8 Sensitivity of the new photothermal camera to small depth defects. lmage of an EDM notch of 1mm long, 100 pm width and 200 pm depth on ANSI 304 stainless steel with a bad surface condition (ground surface, "Vi 2 -6 ).

COLOURED PRODUCTS BOX MANIFOLD NOTCH SPECIMEN 1 NOTCH SPECIMEN 2... 

Shear Horizontal (SH) waves generated by Electromagnetic Acoustic Transducer (EMAT) have been used for sizing fatigue cracks and machined notches in steels by Time-of-Flight Diffraction (TOED) method. The used EMATs have been Phased Array-Probes and have been operated by State-of-the-art PC based phased array systems. Test and system parameters have been optimised to maximise defect detection and signal processing methods have been applied to improve accuracy in the transit time measurements. 

Due to the limitation posed by the initial electrical leakage signal and by the chosen angle of incidence of 52 deg. diffracted signals from 8 mm deep machined notch were hidden. Defects with depth exceeding 12 mm could be detected and sized. The same difficulty was observed when the thickness of the sample was less than 30 mm. 

Hwang G S and Glapis K P 1998 The Influence of surface currents on pattern-dependent charging and notching J. Appl. Phys. 84 154... 

Rice J.R. (1968) A path-independent integral and the approximate analysis of strain concentration by notches and cracks. J. Appl. Mech. 35, 379-386. 

---