TABLE OF CONTENTS

Signal Origins from the Near Inside Surface Extrapolation to the Near Outside Surface Conclusion Acknowledgement

Signal Origins from the Near Inside Surface

During recent calibrations on Mechanised UT on several pipeline projects there has been some operators were having difficulty obtaining clear signal responses from the LCP (Land for Cross Penetration) zone in the modified J-bevel and the Root zone in a standard J-bevel. These areas are traditionally inspected using a high angle transverse mode, typically 68-74 degrees refracted angles.

Over the past decade or so it has commonly been accepted that the signal off this zonal target is the "direct path" signal. The LCP and Root Land areas are vertically oriented so the centre of the beam makes a relatively low angle of incidence with it (16-22 degrees). The targets for these zones have been 2mm diameter flat bottom holes centred 2-3mm above the inside surface of the calibration block. The assumption has been that the signal received by positioning the probe to hit the target directly is simply the direct return path along the beam centreline. Another option, in some configurations, is to pull the probe back to cause the beam to skip off the inside surface and then strike the target. This option is considered less desirable because it relies on the skip path and it also results in the tightly focused beam interacting with the target in a region of the beam that was not at the optimum focus.

Fig 1: LCP Direct path

Fig 2: LCP Reflection of Main Beam

Figure 1 shows what has been idealised as the preferred probe positioning and expected sound path. The figure shows the centre of beam path for a nominal 72 degree refracted beam directed at the centre of the vertical LCP zone in a 15.8mm wall pipe section.

Upon closer examination one can see that the interaction is not so simplistic. To expect the main signal to return along the same path as the incident path when a 20 degree incidence on a specular reflector has occurred is not reasonable. In an unbounded condition a signal could be detected along the same line of travel but it could not be considered the reflected signal. To return along the incident path a "tip diffraction" effect must occur. Reflectivity shows a strong signal should occur at a reflected angle equal to the incident angle (as predicted by Snell). The tip diffracted signal would, by comparison, be very weak.

Simple travel time calculations would indicate that we should be able to easily separate these two signals (the reflected and tip diffracted signals) since the reflected path must traverse about 6mm more steel path than the direct diffracted signal prior to returning to the Wedge/Steel interface. (See Figure 2)

The direct path in Figure 1 shows the transmit time from the centre of the element then through steel to the target as 22.463 us. Doubled for the return we get 44.926 us. The steel path distance would be 91.312mm
The skip path taken by the centre of beam on the reflected transverse wave is 102.26mm in steel. It would SEEM that a reasonable time exists to gate out the reflected wave and concentrate on the diffracted wave as there would appear to be 11mm of steel path difference.

However, when we look at the signal on the scope no such signals occur having nearly 3.4 us of separation. (See Figure 5). This is accounted for by the change in path length through the wedge material. After reflecting off the target and the ID surface the beam centre is shifted back 2-3mm and when it re-enters the wedge it takes a shorter path to the piezo-element. The arrival time of the reflected wave is therefore approximately the same as the diffracted wave (45.2us compared to 44.9us). Even if the diffracted wave does manage to arrive SLIGHTLY earlier its amplitude is so much lower (15-20dB) that it is not reasonable to use it to set sensitivity of the instrument.

Fig 3: Focal Distance optimised for reflection in first half skip (left), in the second half skip (right)

This is not to imply that we should not use the high angle direct transverse wave to inspect the LCP or vertical Root land. However, we should position the beam to optimise on the focal characteristics of the beam and accept that we are using the Reflected not the Diffracted signal. Figure 3 shows how the tight focus used by this probe minimises effects of nearby geometries but this is true only when the smallest beam profile interacts at the zone of concern. A technique using a skip off the inside surface to the target should be just as effective as the direct path provided that the characteristics of the beam provides for a slightly longer focal distance. Therefore the probe beam focal length must be matched to the technique used to minimise other geometric effects. Simply pulling a probe back to skip off the ID surface will provide a poor signal if the focusing characteristics of the beam were designed to provide the smallest beam profile at the target when the shorter "direct path" is used.

Extrapolation to the Near Outside Surface

The above discussion points out that the originally perceived signal for the near Inside Diameter (ID) zone is not the direct path it was originally thought to be. It also points out that for more than a decade acceptable results have been obtained in detecting incomplete penetration using a "quasi-tandem" effect of the reflected transverse mode.

Fig 4: Fill Zones

Fig 5: LCP Target Signals from the LCP 2mm diameter FBH target in 9.8mm Wall pipe

Fig 6: Fill 2 Target

A corollary can be extrapolated from this situation. For over a decade now, a small faction in the pipeline inspection industry has insisted that a high angle (75-85 degrees) compression mode be use to generate a "so-called" creeping wave for inspecting the upper 3mm of welds in wall thicknesses over about 15mm. The rationale for this has always been that the "quasi-tandem" effect does not ensure detection of near surface flaws, especially due to the interfering effects of weld cap geometry. Clearly this rationale must now be reconsidered. The quasi-tandem effect has unwittingly been used as the ONLY method for LCP and vertical Root land inspections. These lower zones are closer to the inside surface than the upper fill zone is for the outside surface (centred 2mm from the ID surface versus 1.2mm for the OD surface targets). The quasi-tandem effect has consistently been effective on the lowest zone (as shown above) and is as subject to the same potential interactions with the excess metal in the root reinforcement geometry as would be any beams directed at the upper zone near the weld cap. The concern that the quasi-tandem effect be unusable for the upper fill due to geometric interference is no longer a valid argument. It is obvious, therefore, that the demands made to resort to the noise-prone "same-side creeping wave" are therefore unreasonable.

For all applications to date, the Fill Zones have had targets that are Flat Bottom Holes (2 or 3mm diameter) inclined at the bevel angle. The upper most Fill target is typically centred 1.2mm to 1.6mm below the OD surface. This is shown in Figure 4 for a simple 2 Fill Zone arrangement. Even on wall thicknesses over 15mm the space between the target and nearest surface is smaller than the gap for the LCP target near the Inside Diameter surface.

The predicted diffracted wave signals for the lower (LCP) and upper (Fill 2) are shown in Figures 5 and 6. These show the target responses from 2mm diameter Flat Bottom Holes in a calibration block for a 9.8mm wall in a pipe with an outside diameter of 762mm. The probe used in Figure 5 is a spherically focused element at a nominal 71 degree refracted angle. The probe used in Figure 6 is a cylindrically focused element at a nominal 65 degree refracted angle.

Conclusion

It has been shown that the previously considered "direct path signal" for near inside surface targets has been incorrectly evaluated and the actual signal being used for calibration is the reflected signal. Extension of this signal analysis to the signals originating from the Outer near surface has shown the existence of a similar pre-cursor signal.

Requirements imposed by some agencies to use the "same-side creeping wave" techniques to inspect the upper 3mm of girth welds are ill founded. Use of high angle compression mode creeping waves can provide detection of targets in this upper Fill zone but signal quality suffers from very poor signal-to-noise ratio. During actual field inspections this poor signal-to-noise ratio is made even worse by any surface irregularities and the effects of water moving over the wedge (items not considered during probe performance evaluations under lab conditions).

Reflected near ID (LCP and Root) and near OD (upper Fill) signals have been the actual signals on which mechanised UT systems used for pipeline girth weld inspections have been calibrated for many years with no deterioration to inspection integrity.

Acknowledgement

The author would like to thank Mr. Dave Stewart of Canspec (Edmonton, Canada) for providing the images of the tip diffracted pre-cursor signals in Figures 5 & 6.