Developments in Ultrasonic Guided Wave Inspection Wireless Structural Health Monitoring Technology for Heat Exchanger Shells using Magnetostrictive Sensor Technology N. Muthu, EPRI, USA; G. Light, Southwest Research Institute, USA INTRODUCTION There are a large number of heat exchanger units throughout nuclear power plants. The heat exchanger shell is a pressure vessel that surrounds and encloses a heat exchanger unit. The shell is susceptible to corrosion and wear, especially around the nozzle area where steam is often deflected inside the shell toward the shell wall. The nominal dimensions of these shells are 30 to 50 feet long with diameters ranging from 30 to 60 inches and wall thickness of ranging from approximately 19mm (0.75 inch) up to 32mm (1.25inches). The heat exchanger shell is covered with insulation. At the present time, periodic inspection of the shell wall is required to monitor wall loss and establish trending data. This process usually consists of collecting ultrasonic wall thickness measurements in specific grid locations. This is a laborious process that often provides data that is difficult to correlate and to use because data is not always collected from the grid locations. A method that provides consistent monitoring of the entire shell, requires minimal labor, and provides data that is easy to correlate to damage is needed. To meet this need, EPRI NDE Center funded SwRI to evaluate a long range, magnetostrictive guided wave technology for monitoring the shell wall. The magnetostricitve sensors (MsS) are low profile (approximately 0.25mm (0.01 ) in height), produce a guided wave that can propagate in the steel up to distances of 10m or more, and are relatively inexpensive. This means that the MsS can be installed with the insulation placed over the sensors and left permanently installed. The effort consisted of initial evaluation, fabrication of the monitoring system, and testing of the system in the laboratory. The system functions are conceptually illustrated in Figure 1 consisting of segmented exciter magnetostrictive pulse-echo sensors bonded to the heat exchanger shell, 20 channel multiplexer, MsSR3030 instrument, a remote station, a wireless dualband 802.11 n communication system, and a desk top computer. The magnetostrictive (MsS) sensors are thin ferromagnetic strips bonded to the surface of the heat exchanger shell that are excited to generate ultrasonic guided waves in the shell wall using a thin excitation coil. The MsS sensors generate ultrasonic guided wave beam in the shell wall as illustrated in Figure 2.
Figure 1 - Illustration of MsS Monitoring System for the heat exchanger shell + direction Directionally controlled plate monitoring beams, one in the + direction and then one in the - direction - direction ~20-30ft 10-20ft Figure 2 - Illustration of the plate monitoring beams generated by the MsS sensors on a heat exchanger shell. The defect detection sensitivity is governed generally by the ratio of the defect cross section to the beam cross section at the location of the defect The purpose of this paper is to describe the system components and discuss testing that characterized the performance of the system components and the overall system.
DISCUSSION OF SYSTEM COMPONENTS AND RESULTS OBTAINED FROM CHARACTERIZING THE COMPONENTS Even though various versions of each of the components had been fabricated before or were off the shelf, the characteristics of each component had to evaluated and optimized for use in the Heat Exchanger Shell Monitoring System. The key components of the system including the probes, the multiplexer, the MsSR3030, the wireless system, and the data acquisition and analysis system. The following sections discuss the experimental work conducted to evaluate, characterize, and optimize the components for the MsS Heat Exchanger Monitoring System. MsS Probe Coverage and Defect Detection Sensitivity The MsS probes consists of the FeCo magnetostrictive material bonded to the outside shell wall and the excitation coil. To understand the coverage of the sensors, it was important to characterize the beam patterns produced by the sensors. To provide optimized frequency and directionality control, the coil design had to be optimized. Finally, the epoxy bonding material had to operate at temperatures as high as 300oF. The FeCo material used is a 0.127mm (0.005 inch) thick strip that is 50.8mm (2 inches) wide. The strip is bonded to the shell using DP125 epoxy. The sensors size and frequency was evaluated to determine coverage and guided wave mode generation. The test plate (illustrated in Figure 3) was used to evaluate the beam radiation pattern and coverage as a function of frequency and sensor length. The MsS was used to generate the guided wave and a small MsS sensor was moved along the FeCo material at each of the three locations shown to receive the transmitted wave. One of the radiation patterns obtained is shown in Figure 4. Figure 3 - Illustration of the test plate used to evaluate the sensor radiation pattern and coverage as a function of frequency and sensor length Figure 4 - Radiation pattern obtained from a 64KHz MsS sensor The theoretical expression for the radiation pattern at the -6dB from maximum is given by θ 1/2 = 0.44 sin -1 (λ/d)
where θ 1/2 is the half angle, λ is the wavelength, and D length of the probe. The experimental data obtained are shown in the Table 1 and validated this expression, so that beam coverage could be predicted. θ 1/2, full beam Theory ( Degrees) θ 1/2, @ -6dB Theory ( Degrees) θ 1/2, @ -6dB Measured ( Degrees) Wavelength (λ Probe length in inches) (D in inches) 32 khz 4 8 30.0 12.7 11.3 64 khz 2 8 14.5 6.3 6.2 128 khz 1 8 7.2 3.2 4.8 64 khz 2 12 9.6 4.2 4.3* 128 khz 1 12 4.8 2.1 1.4* Table 1 Testing to determine defect detection sensitivity and signal attenuation was conducted using frequencies of 32, 64, and 128KHz. Initially, test defects that were round bottom holes (6.35mm (1/4 ) in diameter and 6.35mm (1/4 )deep) were used to test the defect detection sensitivity. Each defect represented approximately 0.1% of the beam cross section at the distance from the probe. To increase the defect cross section in the beam, multiple holes were drilled in the plate at a 6.35mm (1/4 ) edge to edge spacing. Since the system is to be used in a monitoring mode, initial waveforms at each frequency were obtained prior to introduction of defects and subsequent waveforms were taken with each frequency after each defect was introduced into the plate. The difference waveform was obtained for each case and the amplitude of the difference waveform was plotted as a function of the cross section of the defect. At 64KHz, the amplitude and signal to noise data obtained as the defect cross sectional area was increased is shown in Figure 5. If the detection S/N threshold of 2 is used, then defects on the order of 0.4% beam cross section can be detected. Figure 5 - Plot of Defect Signal Amplitude and Defect Signal to Noise as a Function of Defect Cross Section Relative to the Beam Cross Section To validate that the sensors could work at 300 o F and the sensor coverage, sensors were bonded on a test plate that could be heated and insulated. This plate is shown in Figure 6. The sensor layout plan is shown in Figure 7. The nozzle region was approximately 3.5m (11 1/2 ) from the sensors placed along the length of the plate. Because there is a dead zone directly in front of the sensors, a number of sensors were placed parallel to the length of the plate and one sensor was also placed parallel to the length of the plate for propagating a beam behind the nozzle cut out.
Figure 6 - Photograph of test plate used to simulate heat exchanger surface in lab (top) with sensors bonded in place, with plate heaters (middle) and finally insulated (bottom) Figure 7 - Sensor layout plan on test plate Discussion of Data Collected Examples of differential waveform data collected from several sensors at various locations (shown in Figure 8) with defects (increasing in cross section) in the locations shown in Figure 7 are provided in Figures 9 through 11. These differential waveforms show that defects as small as 0.4% of the total beam cross section can be detected. The data in Figure 9 is for a defect that was placed (and increased in size in subsequent steps of 0.2% of the cross section) behind the nozzle. The defect is located approximately 1270mm (50inches) from the sensor. The wave propagated by this sensor would generally go around the shell and not see the large signal at 1900mm (75 inches) that is seen in this simulated lab test.
Figure 8 - Illustration showing locations of sensors. Figure 9 - Plot of Difference of MsS data collected from sensor 1 as the defect was increased from 0 defect to a 1.6% defect. The defect was approximately 50 inches from the sensor. The data shown in Figure 10 is from sensor 4 looking toward the nozzle. The defect detected is approximately 2280mm (90 inches) from the sensor and it is clear that as the defect reach even 0.2% of the cross section of the beam it is detectable. This defect is 4.5mm (0.18inch) in diameter by 6.35mm (1/4 inch) in diameter. Because of the radiation pattern of the sensors, this defect is also seen by sensor 5 as shown in Figure 11. For this sensor, the defect detection sensitivity was approximately 0.4% of the beam cross section.
Figure 10 - Differential data from sensor 4 as the defect grow in cross section from 0 to 1.8% in steps of 0.2% cross section. The defect is approximately 90 inches from the sensor. Discussion of the Wireless Communication System and the Multiplexer The wireless system used for the system was the dualband 802.11n communication system as illustrated in Figure 1 and shown in Figure 12. The multiplexer is shown in Figure 13. However, it is important to note that the wireless function is not necessary. If there is concern at the utility about using the wireless technology, the data can be obtained periodically at the MsS instrument.
Figure 11 - Differential data from sensor 5 as the defect grow in cross section from 0 to 1.8% in steps of 0.2% cross section. The defect is approximately 88 inches from the sensor. Figure 12 - Photograph of the dualband 802.11n communication system
Figure 13 - Photograph of the 20 channel multiplexer Conclusions A multichannel MsS heat exchanger monitoring system has been completed. The system is ready to be tested in a field trial at a utility. The other conclusions were as follows (1) a consistent on-line monitoring using multiplexed sensors was developed, (2) the potential for simultaneous monitoring of many heat exchangers through the use of multiple multiplexers has been illustrated, (3) there is no need to remove insulation to collect periodic data, (4) should allow a utility to better achieve ALARA goals, (5) allows inspection repeatability, (6) should provide a means to avoid unplanned shutdowns, (7) should lead to cost savings in repair and replacement activities, and (8) monitoring may be performed at each heat exchanger location or remotely using acceptable wireless technology.