USING AN ARRAY OF GUIDED WAVE MAGNETOSTRICTIVE TRANSDUCERS
Abstract
Above ground storage tanks are used to store various fluids and chemicals in many industries. The structural integrity of these tanks must be regularly assessed, but the required inspection operations are hazardous due to the chemicals themselves as well as the requirement to operate within confined spaces. The accepted practice for inspection of these tanks, particularly the tank bottoms, requires removing the tank from service, emptying the tank, and interior entry for direct inspection of the structure.
An inspection from outside the tank would have significant cost and time benefits and would provide a large reduction in the risks faced by the inspection personnel. Segmented Magnetostrictive Transducer guided wave probes are now being developed for this purpose. For the inspection process, the probe is sequentially coupled to the chime (skirt) at multiple locations around the tank bottom. Data is acquired using full matrix capture (FMC) and the acquired waveforms are analyzed using the total focusing method (TFM) to generate an image of the acquired data. This paper presents the detection capability of this novel method. Results shown are from data acquired on both purpose built mock-ups and in-service and out-of-service tanks.
Introduction
Typical storage tanks range in diameter from 2.5 m – 30 m (8.2 feet to 100 feet). The concept of using guided wave methods for testing of the tank floor from the exterior has been investigated by different research groups for at least a decade. However, tank inspection using guided waves is currently still considered to be an emerging method due to multiple challenges, including the following:
- Uneven and sometimes heavily rusted tank extension surface
- Small remaining tank extension (less than 25 mm (1 in.) to mount the probe
- Guided wave energy leaking into the vertical wall of the tank
- High attenuation of guided waves in the presence of generalized corrosion, deposits, or liners 14
- Great variety of tank bottom geometries (lap welded, butt welded or both, patches and penetrations) and the absence of tank geometry configuration
documentation
- Presence of multiple areas with difficult access to the tank bottom extension due to other geometry features
- Presence of a long non-inspectable region (dead zone) between the probe and the beginning of the inspection region
The above factors required the development of a new probe design and a new data collection and analysis process. On the processing and reporting side, this work was focused on utilization of advanced phased array algorithms to provide an intuitive map of anomaly distributions in the tank bottom.
Array of Guided Wave Magnetostrictive Transducers
Magnetostrictive sensor (MsS) technology utilizes magnetostrictive properties of ferromagnetic ribbons to generate ultrasonic vibrations. In the past, SwRI used single element probes with apertures up to 100 mm (4 in.) for data collection around tank bottoms. The probe was manually moved around the tank; time-amplitude waveform data was plotted into an ultrasonic image of the tank floor. The next step in technology evolution utilizes a probe with an array of sensors and new data post-processing algorithms. Figure 1a shows a SwRI 8x8© probe with 8 magnetostrictive sensor segments. Each segment of the array is 76 mm (3 in.) wide and consists of two independent magnetostrictive elements that are used for direction control. The probe has flexible joints between segments and can bend around a large radius. Each segment is individually connected to an SwRI MsSRV5M guided wave instrument with integrated multiplexer, shown in Figure 1b. This pulser-receiver instrument facilitates pitch-catch and pulse-echo data collection methods from every possible combination of transmit/ receive sensor segments. The probe can be installed on the tank bottom extension or tank walls using a magnetic clamping fixture (shown in Figure 1a and 1c).
8x8 MsT Probe Evaluation on Mockups
The system was tested on various mockups representing tank bottoms. The laboratory mockup had a 6.4 mm (0.25 in.) thick aluminum plate representing the tank floor; the plate is attached to a vertical plate representing the tank wall. The bottom plate had a series of anomalies, including 4 mm (0.16 in.) diameter 25% deep drilled holes and gradual wall loss defects, with maximum wall loss of 25% to 90%. The mockup is shown in Figure 2a and 2b.
Based on the generated image shown in Figure 2c, the probe demonstrated outstanding sensitivity to simulated pitting corrosion and gradual wall loss defects, as well as the ability to map the defect locations. This particular test showed the system performance in the shorter 1.2 m (3.9 ft) range. It should be noted that the reason for selection of aluminum plate was its low weight and ease of transportation for demonstrations. In general, aluminum material introduces lower attenuation to guided waves for the selected 1.2 m (3.9 ft) inspection range, but the difference in performance between aluminum and carbon steel is negligible.
The system was also evaluated on defects located at approximately 6 meters (20 feet) distance from the MsT 8x8 probe in a 6.4 mm (0.25in.) thick carbon steel plate. The objective of this test was to demonstrate the ability of the probe and processing algorithm to detect and locate anomalies at a long distance from the sensor. Figure 3 shows that all three EDM notches in the plate, with depths of 1.3 mm (0.05 inch) to 2.5 mm (0.1 in.) could be clearly mapped from more than 6 meters away.
Separation of Indications in the Vertical Wall
As mentioned earlier, guided wave energy leaks into the vertical wall of the tank from the probe positioned on the tank bottom extension, as shown in Figure 4a. As a result, indications generated in the tank wall (produced by wall loss area 1) and tank bottom (produced by wall loss area 2) will be overlapped on the indication plot. As an example, Figure 4b shows the indication plot obtained from the aluminum mockup used for the previous test where indications produced by the upper edge of the vertical plate and bottom indications overlap. A new test procedure is being developed to identify the wall indications separately from the bottom indications. The method uses an acoustically coupled wall attachment that alters the signal in the tank wall without affecting the signal traveling through the tank bottom.
The test procedure includes two steps: acquiring the first set of data without the wall attachment and then acquiring the second set of data with the wall attachment. Then the two data sets are subtracted from each other before being fed into the image reconstruction algorithm. The resulting plot is shown in Figure 4c. As it can be noted, only
indications relevant to the vertical wall can be seen.
Field Testing Results
The system has been evaluated at a limited number of field trials. Figure 5a shows the results obtained from an 18 m (60 ft) diameter storage tank with a lap welded tank bottom. The lap weld was located at an approximate distance of 60 cm (2 ft) from the tank wall. The aperture of the MsT 8x8 probe used for the test was 60 cm (2 ft) wide. The data in this trial was acquired at three frequencies in the range of 16 – 250 kHz. The data presented in Figure 5 was acquired at 64 kHz.
Based on the results presented in Figure 5a, most indications were located in the area before the lap weld and were clearly mapped in reference to the probe position. Some indications were likely from the second lap weld, as noted in the figure. An indication was also produced by the first weld in the vertical wall, also noted in the figure.
This test also demonstrated that the dead zone (the area saturated by the incident pulse and reverberation produced by the vertical wall) was approximately 0.2 meters (8 in.) at 64kHz. The length of the dead zone depends on the operating frequency and condition of the wall-to-bottom fillet weld. Other laboratory tests demonstrated that this dead zone can be as small as 50 mm (2 inches) at a 250 kHz test frequency on 6 mm (0.25 in.) thick aluminum plate.
Figure 5b shows a typical indication report with indications originating from the tank bottom marked with yellow squares and indications originating from the tank wall marked with blue squares. Indication ranking information based on relative amplitude was also prepared.
Conclusion
A SwRI 8x8© system that combines an eight segment magnetostrictive array probe with FMC acquisition and TFM analysis algorithms demonstrated great potential for mapping anomalies in tank bottoms. This technology represents a promising quick screening method that could be performed between traditional inspections involving emptying and decommissioning the tank; the results could be used to provide guidance for the type and timing of more conventional inspections.
Future work is needed to characterize and to qualify the performance over the wide range of tank configurations in use, to develop commercial quality software that combines data from multiple probe positions and supports common inspection reporting requirements, and to develop inspection procedures and a training program, etc.
Acknowledgements
The authors would like to express acknowledgement to BASF for continuous support of this project.
This article was originally published in MTI CONNECT 2024, Issue 1.