APPLICATION OF PHASED ARRAY ULTRASONIC TEST EQUIPMENT TO THE QUALIFICATION OF RAILWAY COMPONENTS K C Arcus J Cookson P J Mutton SUMMARY Phased array ultrasonic testing is becoming common in a wide range of industries. For some time it has been used for the qualification at manufacture of components such as wheels; now the advent of robust portable equipment has made its use in all areas of the maintenance workshop and in the field possible. Items such as wheels, and draft gear components, have successfully been tested in situ. Many of the benefits of the new equipment are derived from the computing and data recording capabilities of the equipment as much as from the advances in ultrasonics. Electronic scanning, distance encoding, recording of all A scans and post-test processing of data are all features that can be used to increase the probability of detection for a given test situation. Test data can be displayed in a number of formats including A scan, Sectorial scan, B scan, and C scan. The two dimensional nature of the Sectorial scan can make defect discrimination much easier leading to very low rates of false positives when compared to traditional A scan flaw detection. The analysis of a combination of 3 of the views can give a detailed 3 dimensional analysis of discontinuity locations. This is advantageous for failure analysis or understanding of manufacturing processes. Phased array ultrasonic testing is a useful tool that can provide information not usually available from conventional ultrasonic testing. The results presented in this paper demonstrate the successful application of phased array testing to wheels and draft gear components while in situ. 1 INTRODUCTION Ultrasonic phased array testing has been available to industry for more than 10 years. It is however only in the last 4 or so years that reductions in the size and cost of the computers required to generate and interpret the signals has allowed for genuine portable applications of the technology. Ultrasonic phased array units that at less than 5 Kg in weight are genuinely portable and robust enough for workshop conditions has enabled use of the test method to a whole new range of applications. Now that we can take a phased array unit out into the workshop, and under a train, the question still arrises why? What are the benefits to be gained from using a newer and more costly piece of equipment to do a particular job. Two areas of testing where substantial improvements can be made are in the probability of detection and in defect discrimination. The two applications described in this paper each demonstrate one of these points. Under train auditing of wheels has been successfully conducted. For this application, fully recorded distance encoded phased array ultrasonics was chosen for the high probability of detection it provides. Under train qualification of draft yokes was undertaken with a time encoded phased array ultrasonic sectorial scan to provide greater defect discrimination and eliminate false positive results. Figure 1: Ultrasonic Phased Array Portable Equipment. 549
2 UNDER TRAIN WHEEL RIM TESTING This method has been used for testing both drive and trailing wheels under train, wheel sets in bogies, free wheel sets, and individual wheels. The equipment is shown in Figure 1. A linear phased array probe was used to perform radially aligned electronic scans of the wheel rims from the flange face. Probe alignment was by rollers on the flange tip. The spacing between virtual probe apertures (VPA) was set at 1 mm. The VPAs are activated sequentially, they send and receive in the same fashion as a standard ultrasonic probe and an A scan display is available for each one. As they advance along the phased array probe, this electronic scan advances across the test piece providing full ultrasonic coverage of that probe location. A schematic of the alignment of the probe and the functioning of the electronic scan is shown in Figure 2. An A scan from a single VPA is shown in Figure 3. As can be seen in Figure 2 the radial coverage of the wheel rim, with the exception of the flange region, is complete. All of the A scans from the VPAs of the electronic scan are displayed as a B scan which is an ultrasonic cross section of the rim. A B scan from a wheel rim with the outline of a wheel rim superimposed to provide location reference is shown in Figure 4. Line of A scan from figure 3 Active Virtual Probe Phased Array Probe Electronic Scan Figure 4: B Scan Representation of A Scans From Electronic Scan Figure 2: Phased Array Probe and Wheel Rim Showing Action of Electronic Scan Circumferential probe movement was manually provided and was monitored by a trailing wheel distance encoder. A snap shot of the electronic scans was recorded for every 1.0 mm of circumferential travel. The B scans then are stacked into a C Scan which provides a consolidated view of the data for the wheel rim being scanned. Figure 5 shows a C scan from a wheel rim with a number of internal discontinuities. The location of the B Scan in Figure 4 is shown. B Scan from Figure 4 Figure 5: C Scan of The Full Circumference Of a Wheel Rim. Figure 3: A Scan From a Single VPA. These scan settings have the effect of dividing the wheel rim into a 1 mm square grid. The ultrasonic 550
test data for each square of the grid is recorded and available for review with either the testing equipment or via a PC based data analysis software. It is this scan set up and the data recording that gives the test method its high probability of detection. When reviewing data, areas of incomplete scanning or of loss of coupling are clearly differentiated in the data display and full coverage of the test area can therefore be ensured. Figure 6 is a screen shot from the data analysis software which shows the A scan, C scan and B scans in a combined view. The arrows in the Figure are pointing to the same discontinuity in each view. Figure 6: Data Analysis Software Showing a Full Circumference Ultrasonic Phased Array Scan of a Wheel Rim. Arrows indicate the same discontinuity in A scan, B scan and C scan. The advantages of the test described above are best understood when compared to the other portable alternative method of test. This would typically consist of an A scan ultrasonic flaw detector used in conjunction with a 12 mm diameter 5 MHz probe scanned freehand over the test surface. This test system scans a pencil like beam of sound through the wheel rim in what are hoped to be overlapping scans with an instantaneous read out that the operator must monitor continuously if all defects are to be detected. The ultrasonic phased array electronic scanning distance encoded unit can be thought of as a continuous wall of sound sweeping around the wheel rim and recording all that it finds. The ultrasonic phased array system has been used effectively with the reference sensitivity set as high as 1 mm diameter flat bottom hole equivalent reflector size. Time varied gain can be used to obtain correct sizing of discontinuities through the thickness of the rim. Tabulation of the equivalent reflector size and position for individual discontinuities is carried out by reviewing the test data in either the test unit, or the data analysis software. The data may then be plotted in any Figure 7: Plot of Data from Phased Array Ultrasonic Scan of a Wheel Rim. 551
desired format. One example is shown in Figure 7 This shows all ultrasonic indications from scanning of the full circumference of the rim of an individual wheel. The main part of the figure shows all indications detected around the rim (grouped into three size categories) overlaid on the rim crosssection; the upper part of the figure shows the same data set plotted as a function of both the circumferential and axial axes. Verification of these results has been carried out on selected wheels, confirming that the larger indications (for example that highlighted in Figure 7) are large, exogenous non-metallic inclusions. 3 UNDER TRAIN DRAFT YOKE TESTING This method has been used for testing the back corner inside radius of both rotary and fixed yokes. The yokes have been tested in wagon and as individual units. The equipment is shown in Figure 8. Figure 9 shows a section of a defective yoke with the area of interest outlined and some defects highlighted. A linear phased array probe was used to perform longitudinally aligned sectorial scans of the back corners of the yokes. The scans were performed from the outer surface of the yokes to enable the procedure to be used on both free and in wagon yokes. A wedge with a curved contact surface is used to match the curved surface of the test piece. Probe alignment was by hand due to the coarseness of the as-cast surface and the limited access with the yokes in wagon. Figure 10 shows typical probe location and schematically the operation of the sectorial scanning of the phased array unit. The schematic shows ultrasonic beams at 8 different angles. The Figure 10: A Section of a Defective Yoke With a Schematic of A Sectorial Phased Array Scan Superimposed. Figure 8: Ultrasonic Phased Array Portable Equipment Being Used to Test Draft Yokes in Wagon. Figure 9: A Section of a Defective Yoke With The Area of Interest Outlined and Some Defects Highlighted. set up used for testing scanned from 30 º to 70 º in half degree increments which gives 81 virtual probe apertures each with an angled beam of ultrasound to interrogate the test piece. Once again, each of these virtual probe apertures is the equivalent of a standard ultrasonic probe, and has an A scan associated with it. The virtual probe apertures are fired sequentially by the phased array unit, in the same manner as a multiplexer would cycle through a number of probes. The data from the 81 A scans is presented in a Sectorial Scan display which with some training and practice is easily interpreted. Figure 11 shows a sectorial scan from a reference block that contains 4 side drilled holes. The fourth hole is just outside the scanned area, however the other holes and features of the block can be easily related to the scan. This 2 dimensional display, with angular as well as distance information allows for much greater certainty when discriminating between defects and echoes that are inherent in the structure of the test piece. 552
1 4 3 2 Radius End This is especially so when testing an area such as the back corners of yokes where the probe contact surface and the inner surface are curved, but not concentric. This leads to constantly varying beam paths to both the inner surface of the test piece and any discontinuities that may be present. This lack of a fixed distance reference plus the generally rough and lumpy surface available for probe contact make interpretation of standard A scan results less than reliable. Sizing of detected discontinuities was also problematic with an A scan ultrasonic flaw detector and single element angle probes. The inconsistency of the scanning surface also prevented distance encoding of the recorded scans. As the length of any detected discontinuity was required as part of the accept / reject criteria, the C scan was set to a time encoded distance scale to enable recording of the scans. By scanning at a constant speed, it is possible to estimate the length of any detected discontinuities from the C scan. Several examples are given below in Figure 12. 1 2 Radius 3 4 End Figure 11: A Sectorial Scan of a Reference Block. Figure 9: A Section of a Defective Yoke With a Schematic of A Sectorial Phased Array Scan Superimposed. Figure 12A: A Scan, Sectorial Scan and C Scan of The Back Corner of a Yoke. No Discontinuities Detected. The arrows highlight the same indication in the A, B, and C Scan windows. The indication is from the inner surface of the yoke which is present in most yoke scans and serves as a marker which assists with interpretation of the scans. The boxed area is the inner surface indication across the C scan. The dark red area below the box is interface (surface) noise. 553
Figure 12B: A Scan, Sectorial Scan and C Scan of The Back Corner of a Yoke. Crack Like Discontinuities Detected. Depth 2 mm to 4 mm, length 30 mm. The arrows highlight the same indication in the A, B, and C Scan windows. The indication is an unacceptable discontinuity. The boxed area delineates the defect. Figure 12C: A Scan, Sectorial Scan and C Scan of The Back Corner of a Yoke. Crack Like Discontinuities Detected. Depth 2 mm to 4 mm, length intermittent for 75 mm. The arrows highlight the same indication in the A, B, and C Scan windows. The indication is an unacceptable discontinuity. The boxed area delineates the defect. 554
Figure 12D: A Scan, Sectorial Scan and C Scan of The Back Corner of a Yoke. Crack Like Discontinuities Detected. Depth greater than 16 mm, length full width of yoke. The arrows highlight the same indication in the A, B, and C Scan windows. The indication is an unacceptable discontinuity. The boxed area delineates the defect. Figure 12E: Section through midsection of the defect (crack) shown in Figure 12D. Magnetic particle testing has been used to highlight the defect. 555
Sizing for depth of the yoke discontinuities was by comparison with artificial and real discontinuities of known size. Yokes with real discontinuities were sectioned to verify the discontinuity sizes. As can be seen from the 4 screen captures above, the presence or absence of discontinuities within the area of interest in the back corner of the draft yokes is very easily discernable. Sizing the discontinuities for length is easily done from the screen, but accuracy of the length is dependent on the operator maintaining a constant scanning speed during the recording of the scan. Sizing for depth is easily achieved by providing the operator with reference pieces with discontinuities of different known depths. The operators soon become proficient at this comparison sizing, and discontinuities are quickly classified during testing. 4 CONCLUSIONS Ultrasonic phased array equipment is a useful tool for testing of railroad components. Many of the benefits of the equipment come from the ability of the phased array to electronically scan and bend the ultrasound. The various methods of presenting the data on screen, and the ability to record data to serve as verification of a full test, and to be recalled later for analysis are also of great assistance to the ultrasonic operator. The equipment is more complex than the standard ultrasonic A scan flaw detector and additional operator training is required. However, the benefits in increased probability of detection and better defect discrimination will in many cases make the additional cost of the equipment and training well worthwhile. 5 ACKNOWLEDGEMENTS This paper is based on railway research activities undertaken by the authors for the following organisations: Pilbara Iron Pty Ltd (Rail Division) Rail Corporation of NSW Amsted Maxion (Brazil) BHP Billiton Iron Ore Pty Ltd 556