|NDT.net February 2004 Vol. 9 No.02|
| 2nd MENDT Proceedings
Health Monitoring of Piping and Plate Using the Magnetostrictive Sensor (MsS) Guided-Wave TechnologyGlenn M. Light, Ph.D., Director
Hegeon Kwun, Ph.D., Staff Scientist
Sang Y. Kim, Ph.D., Senior Research Scientist
Robert L. Spinks, Jr., Research Technologist
Southwest Research Institute®
San Antonio, TX 78228
Corresponding Author Contact:
SYNOPSISSouthwest Research Institute has developed a guided-wave sensor that can be installed permanently on piping or plate for long-term structural health monitoring. The sensor is based on the thin-nickel-strip magnetostrictive sensor (MsS) and is operated in the torsional wave mode for piping and the shear horizontal wave mode in plate. With the nickel-strip/MsS-coil fixed to the surface of a structure being monitored, waveform data are collected periodically and carefully compared to detect any structural condition changes occurring over time, such as formation and growth of cracks and corrosion wall loss. MsS sensor technology is easy to apply in the field.
INTRODUCTIONGuided waves refer to mechanical (or elastic) waves in ultrasonic and sonic frequencies that propagate in a bounded medium (such as pipe, plate, rod, etc.) parallel to the plane of its boundary. The wave is termed "guided" because it travels along the medium guided by the geometric boundaries of the medium.
Since the wave is guided by the geometric boundaries of the medium, the geometry has a strong influence on the behavior of the wave [1,2]. In contrast to ultrasonic waves used in conventional ultrasonic inspections that propagate with a constant velocity, the velocity of the guided waves varies significantly with wave frequency and geometry of the medium. In addition, at a given wave frequency, the guided waves can propagate in different wave modes and orders.
Although the properties of guided waves are complex, with judicious selection and proper control of wave mode and frequency, the guided waves can be used to achieve 100-percent volumetric inspection of a large area of a structure from a single sensor location [3,4].
The MsS, developed and patented by SwRI® , generates and detects guided waves electromagnetically in the material being tested. For wave generation, it relies on the magnetostrictive (or Joule) effect: the manifestation of a small change in the physical dimensions of ferromagnetic materials-on the order of several parts per million in carbon steel-caused by an externally applied magnetic field. For wave detection, it relies on the inverse-magnetostrictive (or Villari) effect: the change in the magnetic induction of ferromagnetic material caused by mechanical stress (or strain). Since the probe relies on the magnetostrictive effects, it is called a "magnetostrictive sensor (MsS)."
A schematic diagram of the MsS and associated instruments [Model MsSR 2020 (6)] for generation and detection of guided waves is illustrated in Figure 1. The sensor is configured to apply a time-varying magnetic field to the material under testing and to pick up magnetic induction changes in the material caused by the guided wave. For cylindrical objects (such as rod, tube, or pipe), the MsS is ring-shaped and utilizes a coil that encircles the object. For plate-like objects, the MsS is rectangular-shaped and utilizes either a coil wound on a U-shaped core or a flat coil.
Single MsS generates and detects the guided waves propagating in both directions. In practical inspection applications, the guided wave generation and detection are controlled to work primarily in one direction so that the area of the structure on either side of the sensor can be separately inspected. The wave direction control is achieved by employing two sensors, as illustrated in Figure 1, and the phased-array principle in the MsS instrument.
For operation, the MsS requires that the ferromagnetic material under testing be in a magnetized state. This is achieved by applying a DC bias magnetic field to the material using either a permanent magnet, electromagnet, or residual magnetization induced in the material. The DC bias magnetization is necessary to enhance the transduction efficiency of the sensor (from electrical to mechanical and vice versa) and to make the frequencies of the electrical signals and guided waves the same.
The operating wave mode of the MsS is controlled by the relative alignment between the DC bias magnetic field and the time-varying magnetic field produced by the MsS. For L wave modes in cylindrical objects and Lamb wave modes in plates, a parallel alignment is used. For T wave modes in cylindrical objects and SH wave modes in plates, a perpendicular alignment is used. The guided waves propagate in the direction parallel to the direction of the time-varying magnetic field produced by the MsS.
MsS is directly operable on structures made of ferrous materials such as carbon steel or alloyed steel. MsS is also operable on structures made of nonferrous materials, such as aluminum, by bonding a thin layer of ferromagnetic material (typically nickel) to the structure being tested or inspected and placing the MsS over the layer. In the latter case, the guided waves are generated in the ferromagnetic layer and bonded to the nonferrous structure. Detection is achieved through the reverse process.
In the long-range guided wave inspection, a short pulse of guided waves in relatively low frequencies (up to a few hundred kHz) is launched along the structure under inspection, and signals reflected from geometric irregularities in the structure such as welds and defects are detected in the pulse-echo mode, as schematically described in Figure 2. From the occurrence time of the defect signal and the signal amplitude, the axial location and severity of the defect are determined.
Because of low wave attenuation (at 100 kHz, typically no more than approximately 0.33 dB/m in bare plate and approximately 0.1 dB/m in bare pipe; plate has a higher wave attenuation because of the beam spreading that is absent in pipe), guided waves afford inspection of a long length of structure from a single sensor location. The typically achievable inspection range is more than 30 m in bare pipe and more than 10 m in bare plate. Within the inspection range, the cross-sectional area of detectable defect size in pipes by using MsS is typically 2 to 3 percent of the total pipe-wall cross section or larger. In plates, it is typically 5 percent of the guided wave beam size or larger.
Because of the long-inspection range and good sensitivity to defects, guided-wave inspection technology such as MsS is very useful for quickly surveying a large area of structure for defects, including areas that are difficult to access from a remotely accessible location.
Because of the convenience and better performance, MsS technology now operates primarily in the T wave mode for piping inspection and in the SH wave mode for plate inspection by using the thin ferromagnetic layer approach for generation and detection discussed in the previous section.
The thin ferromagnetic layer approach of the MsS technology has also shown a high potential for application to long-term structural health monitoring (SHM) [7,8]. In this application, the MsS is permanently fixed to the structure. Guided wave data are then periodically obtained from the structure and compared with the initial data taken at the time of sensor installation. From changes in the data, structural degradation such as defect formation and growth and its location in the structure are determined for assessment of structural condition and determination of suitable maintenance measures. As the changes in the data are more readily identified, the sensitivity of defect detection is significantly improved (by a factor of 5 to 10) in the monitoring mode over the inspection mode. Also, because the sensor is already in place on the structure, the time and cost for periodically acquiring data is minimal. Since the guided waves can inspect and monitor large areas of structure and the MsS is rugged and inexpensive, the MsS offers an ideal sensing approach for long-term SHM applications.
EXAMPLE OF PIPELINE APPLICATIONTo show the MsS system capability including the system software for data analysis, 32-kHz T-wave data obtained from a water filled pipeline sample are given in Figure 3, together with a schematic diagram of the pipeline configuration. The sample was 168 mm in outside diameter with a 7.1-mm wall and was approximately 44 m long with a 90-degree elbow. One end of the pipeline was flanged. The sample contained several simulated corrosion defects placed at various locations along the pipe. The data were acquired with the MsS positioned at approximately 11.3 m from the flanged end. The upper data were obtained by launching the guided wave to the positive side of the sensor (i.e., toward the elbow) and the lower data to the negative side of the sensor (i.e., toward the flange), respectively.
Figures 4 and 5 show the processed data and inspection report generated by the system software after the data analysis. The computer reads in the acquired rf data files from both sides, calculates the wave attenuation and velocity, corrects the attenuation effects, converts the data to video data, detects signals that exceed the preset threshold, identifies and characterizes the detected signals, and generates a preliminary inspection report for the inspector's final review and approval. All of the above are automatically performed by the system computer within a few minutes. The inspector then reviews the computer analysis results, confirms and corrects, if necessary, and finalizes and approves the results for reporting.
All the defects placed on the sample were simulated corrosion defects with rounded contours of varying sizes and with a maximum depth of approximately 50 percent of the pipe wall Specifics of the defects are presented in Table 3. In the table, the cross-sectional area refers to the maximum cross section of the defect relative to the total pipe-wall cross section.
The % defect in the inspection report in Figure 5 was determined by assuming that (1) the weld signal is equivalent to 10% defect and (2) the signal amplitude is linearly proportional to the cross-sectional area of the defect. The above assumptions are not exactly accurate because weld varies from pipe to pipe and from location to location and the defect signal amplitude and waveform vary with the actual
contour and shape of the defect. As a first-order approximation, however, the assumptions used in the data analysis are reasonable, as can be seen in the similarity between the values of the actual cross-sectional area in Table 3 and the estimated % defect in the inspection report in Figure 5.
EXAMPLE OF PLATE APPLICATIONA wide variety of large structures are composed of plates joined together. Examples include ships, above or underground storage tanks, containment liners in nuclear power plants, steel bridges, and steel columns in high risers. The principal defect developed in these structures in service is loss of material wall thickness by corrosion and cracking. Corrosion may be localized or present over large areas. Cracking may occur at high stress concentration areas and at welded joint.
Figure 6 illustrates the configuration of test plate and setup used in the experimental investigation. The test plate sample was 6.45 mm thick and approximately 1.22 m x 12.19 m in overall size. It was made by full-penetration welding two carbon steel plates. Two 20-cm-long, thin-strip sensors were adhesively bonded to the test sample approximately 3.66 m from end 1, as illustrated in Figure 6.
Using the sensors and MsS instrument (Model 2020) , 128-kHz shear horizontal guided waves were launched toward the end 2 of the test plate, and the resulting signals reflected back from the geometric were detected.
To simulate SHM, baseline data were first acquired from the test sample using both sensors. Then the data were acquired periodically after placing a simulated defect in the central area of the guided wave beam of each sensor: for sensor 1, a notch in the weld simulating weld cracking, and for sensor 2, a 6.35-mm-diameter, 50%-thickness-deep drilled hole 10.67 m from end 1 of the test plate simulating corrosion pits. The periodic data acquisition was repeated while incrementing the defect size: for the notch, up to 50 mm long and 50% thickness deep, and for the corrosion, up to five drilled holes, as shown in Figure 7.
The periodically acquired data were then compared with the baseline data using a signal differential algorithm developed by SwRI.
Figure 8 shows a series of periodically acquired data (in 1 volt/ division scale) with sensor 1 before and after placing a notch in the weld and incrementing the notch size up to 50 mm long and 50% thickness deep. The corresponding differential data in 0.2-volt/division scale are shown in Figure 9 that were obtained by subtracting the baseline data from the periodically acquired data. In these figures, signals reflected from the weld and end 2 are indicated as W and E2, respectively. Signals coming from the side opposite to the inspection side of the sensor due to imperfect wave direction control are indicated in parenthesis; for example, (E1) and (WE1) where WE1 indicates the weld signal reflected back from end 1.
The data in Figure 8 showed that the amplitude of the weld signal decreased with increasing notch size. The signal reflected from the notch should have opposite phase to the signal reflected from the weld and, therefore, destructively interfere. The observed decrease in the weld signal was caused by this destructive interference and increased notch signal amplitude with increasing notch size. The differential data in Figure 9 showed the increased changes in the weld signal with increasing notch size, indicating the feasibility of detecting weld cracking during SHM.
Figure 10 shows a series of periodically acquired data with sensor 2 before and after placing a 6.35-mm-diameter, 50%-thickness-deep drilled hole and increasing their numbers up to five. The corresponding differential data in 0.05-volt/division scale are shown in Figure 11.
As can be seen in Figures 10 and 11, the drilled-hole signal was readily detectable when the number of drilled hole was three or more despite of the facts that the defects were 7 m away from the sensor location and small in size. The data in these figures clearly demonstrated the ability of the active guided-wave sensor to inspect and monitor a large area of plate structure for detection of defects and their growth.
The differential data in Figure 11 also showed that the signals from geometric features in the structure such as welds and ends could not be completely subtracted out and their amplitude varied. This would induce error in detecting the presence of defects and their growth in or very near these geometric features; for example, weld cracking. More work is therefore needed on development of suitable calibration procedures and better signal processing techniques to further improve SHM capability.
The drilled holes were not detectable in the data obtained with sensor 1, indicating that the width of monitoring area of the sensor was relatively narrow even at the 7-m distance from the sensor. Further theoretical and experimental studies are recommended to establish the relationship between the sensor length, wave frequency, and the monitoring area.
CONCLUSIONSConclusions that can be drawn from the test results described in this paper include: