PA Ultrasonic Inspection System

The design uses a customised contour represented by a 5th order set of cosine basis functions. Figure 1 shows the probe at different view planes. The probe is an elliptic shape with a long axis of 98 mm and a short axis of 78 mm. The probe was designed to deliver a 2.5 mm diameter beam spot at all inspection depths from just beyond the blind zone (5 mm) to half an inch past the centre of the 10" billet (139 mm from the surface). The optimisation of the surface shape was targeted at reducing the phase variation across any one phased array element. Phase variation is the maximum difference in the time of arrival for an ultrasonic wave received by an individual element, expressed as a phase angle of the ultrasonic wave. With low phase variation, the voltages induced by the wave arriving at different points on the element add up producing a high output signal. As phase variation tends towards 180 degrees, the voltage induced by the arriving wave cancel out and little signal is produced. This probe design was derived from simulation results.

Furthermore, the probe was designed to support an angle up to ± 3° of beam steering in the circumferential direction of the billet. This provides compensation for initial misalignment of the probe when it is positioned above the billet, and dynamic misalignment that occurs as the billet rotates. When the probe is tilted by just 3° in water the refraction at the boundary with the titanium creates an effective beam angle of approximately 12° inside the billet. Without correction, this means that there may be a significant area at the centre of the billet that is never inspected. It has also been shown by simulation that this design is able to steer in the axial direction of the billet too.

Figure 1. Schematic diagram of the phased array probe from the a) X plane and b) Y plane

The contour and probe segmentation was designed to minimise the phase variation on any element, for the full range of inspection locations (the inspection locations being set by the full range of inspection depths and beam steering angles). Probe Designer iteratively re-calculated the shape of the probe face, determined the phase variation across the surface for all inspection locations, and finally assessed how the probe could be segmented to produce the minimum number of elements that were required to ensure the phase variation stayed below a specified limit. This loop was repeated, changing the surface contour, and seeking a minimum number of elements.

With the phase angle limited to 88°, a solution to the surface contour and element segmentation was determined by Probe Designer, which uses only 132 elements However, the segmentation scheme means that many of the elements had a much greater surface area than others; the largest element being 10 times larger than the smallest. An element's surface area affects its electrical impedance, and hence its sensitivity. Therefore, the segmentation scheme was re-designed to minimise element size variations. The probe was re-segmented into 255 elements (Figure 2) of approximately equal size, and the probe design was complete. The largest element resulted only 1.6 times larger than the smallest.

Figure 2. a) Schematic diagram of the 255 element probe design, b) plot of total area of each element.

At the greatest inspection depths all elements are active, but shallower inspections create an increase in the phase variation across elements. The beam spot is defined at the region of inspection where a reflected signal will be within 6dB of peak amplitude. Therefore, in all design and simulation work only the required elements were included in the focal law for each inspection location. The required active elements were selected automatically by Continuum Ultrasonic Modeller's focal law calculator by determining the inspection aperture required to achieve the beam spot size at a given location and selecting only the elements that fall within that aperture. Figure 3 shows the inspection aperture for an inspection point at the maximum inspection depth. Using only the selected elements, the beam spot is approximately constant across all inspection depths. This has been verified by simulation using both Continuum Ultrasonic Modeller and CIVA, the results of which are shown in Figure 4.

Figure 3. Required aperture for an inspection point at 5.5" below the surface of the titanium billet.
The billet diameter is 10"; the water gap is 3".

Figure 4. Simulated results of beam spot size as function of inspection depth. The points marked axial are the beam
widths measured in the axial direction of the billet. The points marked circumferential are the beam widths measured
perpendicular to the axis of the billet.

Figure 5 shows simulated 2D beam profiles in the plane of beam steering. The focal point is set at a depth of 125 mm, with no lateral translation. Tilt angles of 0°, 2° and 3° have been applied to the probe, and the focal laws calculated with beam steering to compensate. Table 1 shows the amplitude of the sidelobe measured in dB with respect to that of the focused main beam. With 3° of tilt in the circumferential direction the sidelobe is at least 7.6 dB, or 2.4 times smaller than the main beam at the focused point.

Figure 5. Beam profile with the phased array probe focused at 125 mm with the probe tiled at 0° , 2° and 3°.
Left column: Beam steering compensating for tilt perpendicular to the longitudinal billet axis of the.
Right column: Beam steering compensating for tilt along to the longitudinal axis of the billet

Table 1. Sidelobe amplitude relative to focused main beam at depth of 125 mm

Figure 6 shows the same beam profiles as those shown in Figure 5 but with the focus point set at a depth of 15 mm, rather than 125 mm depth demonstrated in Figure 5.

Figure 6. Beam profile with the phased array probe focused at 15 mm with the probe tilted at 0° , 2° and 3° .
Left column: Beam steering compensating for tilt perpendicular to the longitudinal billet axis of the.
Right column: Beam steering compensating for tilt along to the longitudinal axis of the billet.

Figure 7 shows beam spot size as a function of beam steering. In this case, the probe remained normal to the billet, so the beam steering resulted in a lateral shift of the focal spot. It can be seen that the beam spot is approximately 2.5 mm in both the axial direction and the circumferential direction at low beam steering angles, but eventually diverges. As the beam steering exceeds 3°, which corresponds to around 29 mm of displacement, the beam spot size starts to expand rapidly.

Figure 7. Beam spot size as function of lateral shift by beam steering. The points marked axial are the beam widths
measured in the axial direction of the billet. The points marked circumferential are the beam widths measured
perpendicular to the axis of the billet.