• Federal Highway Administration >
• Publications >
• Research Publications >
• 04124 >
• 03.Cfm >
• Lab & Field Testing of AUT Systems for Steel Highway Bridges

Lab & Field Testing of AUT Systems for Steel Highway Bridges

3. AUT EQUIPMENT DESCRIPTION

AUT systems have been introduced for many different applications--from the production of engineering materials to the inspection of rails, tubes, and piping.(11) The use of AUT systems in the nuclear field has been widespread for many years,(12) and many different types of AUT systems have been developed for specific applications. One particular application involves the inspection of pipeline girth welds. The American Society for Testing and Materials (ASTM) standard test method, Standard Practice for Mechanized Ultrasonic Examination of Girth Welds Using Zonal Discrimination With Focused Search Units,(13) is one of the first general standards to specifically allow the use of AUT systems for weld inspection. The procedures described in this standard are more complex than typically are applied to welds in highway bridges. AUT systems have also been developed for civil engineering applications (e.g., pin inspection).(14) Recently, both manual and more automated scanning systems have been used to examine bridge components such as hanger pins.^(15,16) These systems typically generate C-scans to characterize any observed indications.

AUT systems typically consist of ultrasonic pulse-echo instrumentation and a computerized data acquisition (DAQ) system, coupled with a spatial control system that tracks the position of the ultrasonic transducer during testing. The spatial control system may be active (where a robotic scanner moves the transducer across a specified scanning pattern) or passive (where the position of the transducer is recorded as the transducer is controlled manually). The position of the transducer and the associated ultrasonic waveforms are recorded by the DAQ system and can be presented in a variety of formats for interpretation by qualified personnel. Typical formats for data presentation include A-scans, B-scans, and C-scans. A-scan displays are commonly used in UT and consist of a two-dimensional image of the ultrasonic waveform with time along the abscissa (x-axis) and signal amplitude along the ordinate (y-axis). A B-scan display consists of time plotted along the ordinate and scan position along the abscissa. The amplitude of the ultrasonic waveform is color-coded and displayed along with its associated time data. A C-scan display is a planar image with three dimensions (the x- and y-positions of an indication plotted along the abscissa and the ordinate, respectively, and the third dimension being a particular characteristic of the A-scan data plotted by color). The amplitude of the ultrasonic signal is the most commonly plotted characteristic; however, a time or frequency characteristic may also be displayed.

SYSTEM SELECTION

To select an appropriate AUT system for application to highway bridges, a series of commercially available systems were considered. These systems were compared with the existing code requirements to determine if the systems were capable of conducting an inspection that substantially conformed to the existing requirements. There are several reasons to use the existing code as a guideline to AUT system selection: First, the code is widely accepted as being effective by bridge owners and, as such, an AUT system meeting the current requirements could be adopted easily for use. The code currently allows the use of UT for compression members, all tee and corner CJP welds, and uses UT to enhance RT for the inspection of fracture critical tension welds (see section 12.16).⁽¹⁾ Second, revisions to portions of the code that define rejectable and acceptable defect criteria may need to be adjusted for AUT systems that cannot perform the scanning described in the code. Specifically, AUT systems that are not capable of articulating a transducer ±10 degrees during scanning will be unable to maximize the reflected signal. As a result, the rating for a particular indication may be different when an AUT system without articulation is employed. Adjustments to the decibel rating scale would be required to provide a level of sensitivity equal to manual UT. Modifications to this particular portion of the code would be difficult and, at a minimum, would require significant additional research. Third, inspectors are knowledgeable in the existing code, so changing from UT to AUT would require minimal retraining.

P-SCAN SYSTEM CHARACTERISTICS

The projection image scanning (P-scan) system in figure 2 was selected for this project. The P-scan system is an integrated computer-based, digital ultrasonic testing and analysis system.

System Components

The P-scan system has three main components: (1) the processor unit, (2) the scanner module, and (3) the laptop personal computer (PC) platform. The data acquisition, digitizing, and digital signal processing functions are all carried out in the P-scan processor unit. The scanner module is a position encoding device designed to record the position and rotation of the ultrasonic transducer during scanning. The laptop PC is used for post-processing and graphical presentation of data.

The scanner module, known as the manual weld scanner (MWS-1), includes a magnetic base that is attached to the steel plate adjacent to the weld being inspected. A 300-mm- (12-inch-) long scanner arm is attached to the magnetic base and has three degrees of freedom (DOF) as indicated in figure 3. The arm can rotate about the magnetic base and extend along a linear bearing attached to the base. The rotation (a) and extension (L) of the scanner arm are recorded by encoders in the magnetic base so that the position of a transducer is defined in polar coordinates. A third DOF is provided at the end of the scanner arm where the transducer is mounted. The rotation of the transducer or skew angle (ß) is also recorded by an encoder. The operating range of a is limited to 180 degrees, while L is limited to 300 mm (12 inches) and ß is limited to 300 degrees. Even though the transducer can freely rotate 360 degrees, the rotation output of ß is only valid within the specified range of 300 degrees (as shown in figure 3). The positional data (a, L, and ß) are combined with the time-of-flight information contained in the A scan to determine the reflection location. The length of the scanner arm (300 mm (12 inches)) is user-selectable. Longer scanner arms are available from the manufacturer.

Figure 2. Photo. Photograph of the P-scan system showing the data acquisition system and the scanning arm holding the ultrasonic transducer.

Figure 3. Diagram. Schematic diagram of the MWS-1 scanner, indicating three DOF: (1) scanner arm rotation angle alpha, (2) scanner arm extension L, and (3) transducer skew angle beta.

System Setup

MWS-1 can be positioned on the plate in three different scanning configurations. Figures 4 through 6 show schematic diagrams of the scanner positioning configurations. Depending on the plate width, plate thickness, and the availability of a convenient location adjacent to the weld to place the magnetic base, a suitable configuration can be selected for the inspection. For example, consider the plate shown in figures 4 and 5. The coordinate origin (or datum) is placed at the far left edge of the plate on the centerline so that the positive y-axis coincides with the weld centerline. The x-axis lies along the left edge of the plate with the positive x-axis directed upward and the positive z-axis directed into the plate. For convenience, the section of the plate that falls in the positive x-axis is designated as the top section of the centerline (TSC), and the section of the plate that falls in the negative x-axis is designated as the bottom section of the centerline (BSC). The TSC configuration shown in figure 4 has the scanner's magnetic base placed on the BSC side, with the transducer scanning the weld from the TSC side. Conversely, the BSC configuration shown in figure 5 has the scanner's magnetic base placed on the TSC side, with the transducer scanning the weld from the BSC side. Using the TSC and BSC configurations together, the weld can be inspected from both sides of the weld centerline; however, separate setups and data files should be created for each configuration. By placing the magnetic base at the positions shown in figures 4 and 5, the BSC and TSC configurations allow plates up to 609.6 mm (24 inches) wide and 31.75 mm (1.25 inches) thick to be inspected even though the scanner arm's extension length is only 300 mm (12 inches). For plates greater than 609.6 mm (24 inches) wide and 31.75 mm (1.25 inches) thick, the scanner arm cannot reach the entire weld using the BSC and TSC configurations. Thus, the width of the plate must be segmented and each segment must be scanned separately. The third configuration is designated as the Both configuration (figure 6) and permits the inspection of the weld from both sides of the centerline with a single setup and data file. With the Both configuration, the setup time is reduced by approximately 66 percent, leading to a faster inspection. Plates up to 304.8 mm (12 inches) wide and 12.7 mm (0.5 inch) thick can be inspected using the Both configuration. For plates greater than 304.8 mm (12 inches) wide and 12.7 mm (0.5 inch) thick, the scanner arm does not extend sufficiently to scan the entire weld from both sides. Thus, the width of the plate must be segmented and each segment must be scanned separately. Carefully examining and comparing figures 4 and 5 with figure 6 reveals the reason behind the differences in the width and thickness limitations of each configuration. In the TSC or BSC configuration (figure 4 or 5, respectively), the scanner base is placed on one side of the weld in the middle of the plate. This allows the scanner arm to essentially rotate along the length of the weld. Meanwhile, in the Both configuration (figure 6), the scanner base is positioned in line with the weld but to the side of the plate. This forces the scanner arm to essentially rotate transverse to the weld axis. Thus, the thickness and area of the plate that can be scanned using the TSC or BSC configuration is approximately twice as large as the thickness and area that can be scanned using the Both configuration.

Figure 4. Diagram. Positioning/setup configuration of MWS-1 scanner on the plate: TSC configuration describes scanning the weld from the TSC side of the centerline.

Figure 5. Diagram. Positioning/setup configuration of MWS-1 scanner on the plate: BSC configuration describes scanning the weld from the BSC side of the centerline.

Figure 6. Diagram. Positioning/setup configuration of MWS-1 scanner on the plate: "Both" configuration describes scanning the weld from both sides of the centerline in a single setup.

System Output

The P-scan system creates and displays the geometry of an indication on three projection planes. To illustrate this, consider laboratory specimen S033 shown schematically in figure 7. This 12.7-mm- (0.5-inch-) thick plate contains two implanted cracks?one at the toe of the weld (figure 8) and one at the root of the weld (figure 9). The appearance of these cracks on a standard radiograph is shown in figure 10. The radiograph shows the entire weld, including the heat-affected zones (HAZ). Figures 11 through 13 show images created by the P-scan system from the inspection of laboratory specimen S033. Figures 11 and 12 display the actual images created by the P-scan system during and after scanning, respectively. Figure 13 illustrates the three projection planes associated with the P-scan images. The geometries of the toe and root cracks are clearly shown in these figures. The region that is inspected by the P-scan and displayed in the P-scan images includes the weld area and the HAZ (i.e., the HAZ is generally taken to be 6.35 mm (0.25 inch) on either side of the weld bevels at the end of the toe). Detailed descriptions of the P-scan images in figures 11 through 13 are as follows:

• C-Scan (Plan View): The C-scan image is created from the projection of the indication geometry in the weld onto the horizontal plane. The dashed lines are the representation of the weld bevels. The C-scan image reveals the x-position, the y-position, and the length of the indications. The C-scan image is analogous to the standard radiographic film.
• B-Scan (Cumulative or Noncumulative End-View Image): The B-scan image is created from the projection of the indication geometry in the weld onto the vertical plane perpendicular to the centerline. The solid lines represent the weld bevels. The B-scan image reveals the orientation of the indications.
• Side View: The side view image is created from the projection of the indication geometry in the weld onto the vertical plane along the weld centerline. The side view image reveals the depth of the indications.
• Amplitude Profile: The bottom two images in figures 11 and 12 display the amplitude of the echo detected by the transducer in decibels. The positive decibel (+dB) readings in figures 11 and 12 indicate that the echoes from the indications have exceeded the calibration reference-level echo. Conversely, the negative decibel (-dB) readings indicate that the echoes from the indications are below the calibration reference-level echo. Generally, large indications are positive, while small indications are negative. The horizontal line marked as the threshold level for imaging is set up solely for imaging purposes by the P-scan system. Any echoes crossing this line will create colored pixels in the three projection images (i.e., C-scan, B-scan, and side view). The threshold level for imaging is generally selected based on the plate thickness and the UT acceptance-rejection criteria in tables 6.3 and 6.4 from the AASHTO/AWS D1.5M/D1.5: 2002 Bridge Welding Code.(1) In this report, the color bars in figures 11 and 12 associate decibel levels with specific colors. Specifically, the bars and corresponding images contain nine colors (i.e., red, orange, light yellow, dark yellow, light green, dark green, light blue, dark blue, and purple), with red being the highest amplitude and purple being the lowest amplitude responses.

Figure 7. Diagram. Laboratory specimen S033: Schematic diagram showing two implanted cracks.

Figure 8. Diagram. Laboratory specimen S033: Schematic diagram of toe crack.

Figure 9. Diagram. Laboratory specimen S033: Schematic diagram of root crack.

Figure 10. Photo. Laboratory specimen S033: Radiographic image shows the two implanted cracks.

Figure 11. Screen capture. P-scan images of laboratory specimen S033: Displayed in logarithmic mode during scanning to ensure full coverage of the weld.

Figure 12. Screen capture. P-scan images of laboratory specimen S033: Displayed in linear mode after scanning (highlighting only the indications in the weld).

Figure 13. Diagram. P-scan images on three projection planes.

As mentioned earlier, figures 11 and 12 show the P-scan images created during and after scanning, respectively. The logarithmic mode for imaging is selected during scanning, and the linear mode for imaging is selected after scanning. The logarithmic mode can display a wide range of echo amplitudes (160 dB), while the linear mode can only display an echo amplitude range of 20 dB. In the logarithmic mode, the threshold level for imaging can be lowered so that smaller amplitude echoes received by the transducer are used to create a carpet of colored pixels over the gray background color (figure 11). The use of the logarithmic mode ensures full inspection coverage of the entire weld with the region not colored being either missed or not yet inspected. This scenario is clearly shown in the right end of the C-scan and side view images in figure 11. These uncolored areas should be scanned to provide full inspection coverage of the weld. It is important to note that the P-scan system saves the highest amplitude echo received even if the same point or region is scanned multiple times. Once scanning of the weld is completed, the collected data can be displayed in the linear mode (figure 12), highlighting indications only. The threshold level for imaging in the linear mode can be determined using the plate thickness and the UT acceptance-rejection criteria in tables 6.3 and 6.4 from the AASHTO/AWS D1.5M/D1.5: 2002 Bridge Welding Code.(1) In the linear mode, only indications exceeding the threshold level for imaging are displayed. Note that the threshold level for imaging in the linear mode is chosen to be the indication level for the class D indication or slightly below. Also, the regions of the weld without indications are displayed using the gray background color.

The indication rating (d) and the geometric dimensions of a defect in the weld can be determined using marker lines in the P-scan images. The indication rating is the decibel reading relative to the zero reference level after having been corrected for sound attenuation (sections 6.19.6.3 through 6.19.6.5 of the AASHTO/AWS D1.5M/D1.5: 2002 Bridge Welding Code).⁽¹⁾ To illustrate this, P scan images of laboratory specimen S033 in figures 12 and 14 are used as an example. The amplitude profile image in figure 12 shows that the peak amplitudes from the root and toe cracks are off the screen. Therefore, manually increasing the amplitude level will bring the peak amplitudes back onto the screen as shown in figure 14. The indication rating is read from the amplitude profile image in figure 14 by placing the solid horizontal marker line on the peak amplitude obtained from the defect. The numerical decibel value for the indication rating, which is designated as "A" in the P-scan system, appears in the P-scan's marker window. The x-position of the defect is measured from the weld centerline, while the y-position is measured with respect to the datum. The y-position and the length of the defect are determined by moving the horizontal dotted marker to the 50-percent drop in the maximum amplitude as shown in figure 14. Then, the solid and dotted vertical markers are moved to the intersection of the horizontal dotted marker and the amplitude profile of the root crack. The numerical value for the y-position and the length of the root crack appear on the marker window as Y and DY, respectively. Similarly, the depth (Z) and the x-position of the root crack are determined using markers in the side view and C-scan, respectively.

The P-scan processor unit is equipped with a built-in logarithmic amplifier for the ultrasonic receiver. The amplifier ensures the accuracy of the echo signal. Unlike linear amplifiers that are used in traditional UT systems, logarithmic amplifiers have amplitude-dependant gain with logarithmic transfer characteristics. This means that the weak part of a signal is highly amplified, whereas the strong part is only slightly amplified. As a consequence, the P-scan system has no manual gain. Note that the echo signal displayed in a logarithmic amplifier scale appears to have a different amplitude from the conventional linear echo signal. Therefore, the P-scan system must have an internal conversion routine that displays a linear echo signal on the screen instead of the logarithmic signal.

Figure 14. Screen capture. Flaw-sizing scheme of the P-scan system for the root crack in laboratory specimen S033.

ANALOGY OF INDICATION RATING IN AUT VERSUS TRADITIONAL UT

The ratings of indications in AUT and traditional UT are analogous; however, in practice, they are executed differently. Note that the final indication ratings measured by AUT and traditional UT are the same. The differences between the AUT and traditional UT systems stem from different signal amplification concepts. The AUT systems have built-in logarithmic amplifiers with no manual gain control. The traditional UT systems have manual gain control knobs where the operator may manually increase or decrease energy from the received echo. This difference in hardware design creates similar outputs, but displays the output differently for each system. A positive amplitude rating will appear as a negative rating using the P-scan system, while a negative amplitude rating will appear as a positive rating using the P-scan system. Care must be taken when comparing and interpreting the AUT output versus traditional UT output.

UT Indication Rating Procedures

The indication rating procedures for the traditional UT systems described in section 6.19.6.5 of the AASHTO/AWS D1.5M/D1.5: 2002 Bridge Welding Code⁽¹⁾ are summarized as follows:

• The reference level (b) is measured from the echo off of a 1.5-mm- (0.06-inch-) diameter hole in the International Institute of Welding (IIW) reference block. The echo signal is recorded after maximizing it to an 80 percent full-screen (FS) level using manual gain control.
• The indication level (a) is measured from a feature in the weld. The echo from the feature is maximized to an 80-percent FS level using manual gain control. Obviously, the smaller the feature, the more gain that is required to maximize it to an 80-percent FS level and, conversely, the larger the feature, the smaller the gain that is required to get to an 80 percent FS level. Thus, a > b for smaller features and a < b for larger features.
• The attenuation factor (c) is determined from the sound path.
• The indication rating (d) may be computed using the formula given in section 6.19.6.5, d equals a minus b minus c.⁽¹⁾ It is obvious that the indication rating (d) becomes a positive number when the feature is smaller than the 1.5-mm- (0.06-inch-) diameter reference hole and it becomes negative when the feature is larger than the reference hole.

AUT Indication Rating Procedures

The indication rating procedures for AUT are calculated differently. Unlike traditional UT, which displays the indication level (a) on the scope's screen, AUT systems display the indication rating (d) on all of the displays and plots. The following outline summarizes the indication rating procedures for AUT systems:

• The reference level (b) is measured from the echo off of a 1.5-mm- (0.06-inch-) diameter hole in the IIW reference block. The echo signal is maximized to an 80-percent FS level by varying the resolution of the ordinate in the amplitude profile plot. The reference level (b) is automatically analyzed by the AUT software, causing the 80-percent FS level tick mark to reset itself to zero. Now, the 80-percent FS level tick mark becomes a reference for measuring the indication rating (d). Obviously, the smaller the feature, the less energy that is reflected back, so the signal falls below the reference line. The larger the feature, the more energy that is reflected back, so the signal falls above the reference line. In other words, the indication level (d) is a negative number for smaller features and a positive number for larger features.
• The attenuation factor (c) is automatically accounted for in all of the P-scan data when the attenuation function is turned on in the software during calibration (i.e., 2 dB per 25.4 mm (1 inch) of sound-path distance after the first 25.4 mm (1 inch)).