ASTM E-1065 has addressed methods for ascertaining beam shapes in Section A6 "Measurement of Sound Field Parameters"; however, these measurements are limited to immersion probes (see Note 2 after Paragraph 22.214.171.124 page 455 of the 1995 edition). In fact, the methods described in E-1065 are primarily concerned with the measurement of beam characteristics in water and as such are limited to measurements of the compression mode only . Techniques described in E-1065 include pulse-echo using a ball target and hydrophone receivers. Such setups allow the soundfield of the probe to be assessed for the entire volume infront of the probe.
Measurements of beam shapes in solids, even for immersion probes, must rely on sound interactions with targets placed in the solid. These would include flat bottom or side drilled holes machined in metal blocks or steel balls or wires imbedded in plastics. The advantage of continuous measurement of pressure variation with change in target position afforded by the immersion techniques cannot be duplicated in all planes for characterizations of beams in solids. For angle beam probesmeasurement of beam spread in the vertical plane along the axis relies on placing several targets at different depths and interpolating results by connecting the points having a common pressure drop from the central maximum (see ASTM E-164 Section A 2.9). This method is commonly recognized as the most expeditious way of determining the beam profile in a solid.
In this paper we use ultrasonic beam profiles modelled to simulate the process of obtaining refracted angles and beam spread plots from side drilled holes.
The first of the programmes "Beam" is used to indicate lateral beam dimensions at the focus or near zone and the amplitude distribution along the beam axis. No refracting medium is used and calculations used in this project are based on a typical SV shear wave velocity in steel, 3230 m/s. Lateral beam dimensions were calculated for the transmission of the beam only, i.e. the one-way path. Measurements made in the lab using pulse-echo techniques would use the square of the dB drops, e.g. the 6dB drop of pressure as determined in the lab by pulse-echo is represented by the 3dB drop in the Beam model.
The second programme "RF Beam" allows correction for angle of refraction at a boundary. To approximate contact testing parameters, the first medium is given a velocity of 2760 m/s and the second medium 3230 m/s. These are the velocities commonly used for Lucite compression mode and Steel transverse mode. According to Snell's Law using a 37° angle of incidence provides a refracted beam angle of nearly 45° (actual refracted angle 44.8°) when these velocities are used.
By presenting cross sections of the beam amplitude profile at depths comparable to the side drilled holes in the IOW block (Institute of Welding or British Standards A5 block), data points can be generated similar to those made using manual or encoder assisted measurements. Maximum amplitude at each depth can be determined and the displacement to the points at the desired amplitude drop for the same depth can be recorded to generate a beam spread diagram similar to that described in ASTM E-164 or other codes. This is equivalent to finding the centre, front and back of beam. Accuracy of these measurements is based on the sampling size used to generate the representation. When 120mm x 120mm dimensions are used with 200 samples in each direction a resolution of 0.6mm is obtained. Resolution in the vertical scale for amplitude is 0.391% (1/256).
To illustrate the lateral beam spread the RF Beam programme was used with the same media and probe parameters and the angle of incidence was changed to be zero. The centre of the probe is positioned to provide about the same soundpath in the coupling medium as would occur for the conditions illustrated in the vertical beam spread illustrations.
RF Beam provides two options for the calculations; Quick and Good Algorithms. The Quick Algorithm simplifies the calculations using a Gaussian distribution factor and preserves symmetry of beam shape. This is adequate for normal incidence but, apart from an aesthetic presentation, the angled incidence suffers a departure from the true asymmetry that actually occurs. For a more realistic representation of angled incidence the Good Algorithm is run. In addition to defining acoustic velocities on either side of the boundary, densities are also entered. This corrects for transmission coefficient and the resulting asymmetry is preserved.
RF Beam parameters were arranged to provide an incident point less than 0.1mm from the zero X coordinate. This allows the Arctan function to be applied to X displacements without correction for offset. This simply stated means we were able to read the X and Y coordinate for the peak amplitude directly off the plot and the calculate the refracted angle using the Arctan function. These images were grouped according to element size and nominal frequency.
|The following information is provided to indicate limits normally found in the most common calibration block used for beam profiling, the IOW block. Note that the 22mm depth of the side drilled holes may result in portions of the beam not interacting with the target at some depths for some probes. This would be the case for highly divergent beams and large diameter probes that do not allow the refracting wedge to approach the edge of the block. British Standard BS 4331 recommends this block for profiling but acknowledges that for some probes the 50mm dimension may not be adequate (just as the 22mm depth of the holes may not be adequate in some cases).|
Figure 1 The IOW Block
| Depth below surface to hole centres|
(for the four holes used for beam profiling)
These eight depths are used as references for representation|
of beam amplitudes in the RF Beam profiles.
|Height||75 mm||Top surface||Opposite surface|
|Hole diameters||1.5 mm||25||50|
|Hole depths||22 mm||43||32|
Examples of the D/ effects on refracted angle determination are illustrated using plots from RF Beam and the amplitude profiles at various depths. Charts of the maximum amplitude and relative distance from the exit point are used to generate graphs of the beam spreads and illustrate how apparaent deviation from Snell's Angle result.
Although pressure distributions are incorporated into the software calculations, numerous assumptions and approximations must be made. These would include; isotropic conditions exist in both coupling and test materials, the test material is infinitely thick and only two modes of wave propagation are considered (compressional and transverse).
Possible distortions to the ideal beam shape generated by the transducer result from numerous sources;
Assessing sound field dimensions in solids for field applications relies on amplitude measurements from a target signal with respect to probe positions in some coordinate system. Effects of the target size, shape, orientation with respect to the beam axis and surface quality, all add to the variables to consider. This need for a target for evaluation is especially troublesome in the vertical plane. A series of targets at different depths is required and then interpolations and extrapolations must be made as though to connect the dots to form the isobars.
Use of EMAT sensors to assess beam dimensions reduces the problems with designing targets. A series of semicylindrical blocks can be made with the probe being evaluated placed with the probe exit point aligned on the axis of the flat surface and the beam directed at the radiused surface. This suffers from two significant problems, many blocks must be machined and specialized equipment must be used to address the EMAT sensor which suffers from narrow bandwidth. As well, theta - Y motion control and encoding is needed to present accurate results. Another option is to use Schlieren optics. As with EMAT setups this equipment is also very elaborate. However, visualization by Schlieren methods is far more detailed than with the pulse-echo or EMAT options on metal blocks. To visualize beam characteristics within a solid the Schlieren setup must be used with a transparent solid. Metals do not permit internal evaluation by Schlieren methods; however, Hall  has achieved good results using glass models.
Both Schlieren and EMAT methods are too costly and too elaborate to be considered by the general NDT practitioner.
| Frederick H.C. Hotchkiss
is product manager for Ultrasonic Transducer Products, Nondestructive Testing Division, Panametrics Inc. in Waltham, Massachusettes, USA. He received a B.Sc. from MIT in 1966, M.Phil. and Ph.D. from Yale in 1972 and 1974 respectively. He was a postdoctoral fellow at the Smithsonian Institution in 1975. Hotchkiss is responsible for the scientific and technical development of transducers at Panametrics. Panametrics Inc.
Ed Ginzel |
born 1952/04/04 in Kitchener, Ontario in Canada. Ed Ginzel is an independent consultant with the Materials Research Institute. Materials Research Institute provides training and consulting to the NDT industry. Consulting is provided to the nuclear and petrochemical industry to assist clients developing NDT techniques and to evaluate existing techniques and inspection systems.
|| UTonline ||