Combination Near field Remote field probes, a Breakthrough in Ferromagnetic Tube Testing
Defect detection at Support Plates in Ferromagnetic Tubes
Inspecting Ferromagnetic Heat Exchanger Tubes Detecting Outside Defects with Near Field™ Eddy Current
* Eddy Current Technology, Inc., 201-A Horace Avenue, Virginia Beach, VA 23462; (757) 490-1814; fax (757) 490-2778; email: email@example.com.
The remote field and Near Field™ eddy current techniques for inspecting ferromagnetic heat exchanger tubing are similar in that they both take advantage of low frequencies to penetrate the material without the use of magnetic bias to reduce permeability by saturation or partial saturation. Remote field is a send/receive through the wall technique that uses the send and the receive coil(s) separated by approximately two and a half tube diameters, or there may be two send coils the same distance on both sides of the receive coil(s). Near Field™ has the sensor coils close together and is not a through wall transmission technique.
The signals that result from the remote field test are dependent upon the condition of the tube near the send coil(s), near the receive coil(s), and to some degree, in between the coils. Also, as the remote field travels outside the tube wall, the signals are also very dependent upon support structure, such as support plates and tube sheets that may be between the send and receive coils.
On the other hand, the signals that result from near field eddy current are dependent only upon the condition of the tube close to the coils and as the signal barely penetrates outside the tube wall, is only slightly dependent upon support structures, such as support plates.
1.0 22.2 mm (7/8 inch) Diameter, 1.25 mm (.049 inch) Wall 1020 Carbon Steel Tube
For this part of the study a 1020 carbon steel tube with an outside diameter of 22.2 mm (7/8 inch) and a wall thickness of 1.25 mm (.049) was used. Two calibration tubes were used, one a groove standard and the other a big pit standard.
The groove standard is shown in Figure 1.0-1. On the outside diameter it has three concentric grooves that are 3 mm (1/8 inch) long and three concentric grooves that are 25 mm (1 inch) long. The depth of each set of three grooves are 25, 50, and 75 per cent through the wall. There is also a 20 per cent through the wall groove on the inside which is 3 mm (1/8 inch) long.
The big pit standard is somewhat similar to a non-ferrous ASME calibration tube in that it has a through wall hole and 80, 60, 40, and 20 per cent O.D. pits (with the 20 per cent having four pits located circumferentially around the tube), except that all of these flat bottom pits are 1/4 of an inch in diameter. The big pit standard is shown in Figure 1.0-2.
1.1 22.2 mm (7/8 inch) Diameter, 1.25 mm (.049 inch) Wall 1020 Carbon Steel Tube, Near Field™ Probe
The results of testing these tubes with the Near Field™ probe appears in Figures 1.1-1, 1.1-2, 1.1-3, and 1.1-4.
An ect MAD 8D Eddy Current Instrument together with an ect Carbon Steel Filter Amplifier Rev. 3 was used.
The test was conducted in differential and absolute modes simultaneously using a frequency of 380 Hz. The impedance plane for the differential signal appears in the upper left hand corner, and the impedance plane for the absolute mode appears in the lower left hand corner of Figure 1.1-1. There are two strip charts in the figure. The left most strip chart is the differential vertical signal. The right most strip chart is the absolute vertical signal.
In Figure 1.1-1, you can see six signals resulting from the groove standard in the strip charts. Starting with the defects at the top, the signals from the 25 per cent, 50 per cent, and 75 per cent 3 mm long grooves appear first, followed by the 25, 50, and 75 per cent 25 mm long grooves. The length of the 25 mm grooves is fairly obvious in the absolute channel. Immediately below the signal from the 75 per cent 25 mm long groove is the signal as the probe exists from the tube.
In the impedance plane portions of Figure 1.1-1, the 50 and 25 per cent 3 mm long grooves appear. The 25 per cent groove signal is the almost vertical signal that is partially obscured by the 50 per cent groove signal.
In the differential channel the larger of the two signals, the 50 per cent groove, is marked with the cursor; specifically, the square box at the upper left and the cross at the lower right. In eddy current testing of heat exchanger tubes, it is typical that defect signals will go down first and then up. The point on the signal with the cross is the first part of the signal in time, and the portion of the signal marked with the box is the end of the signal. The angle of the signal is measured clockwise from horizontal assuming the square box is the origin, even though it is the later part of the signal. In this case, the result of measuring the signal is shown in the sixth line of the menu as 67 degrees, 4.77 volts, and 49 per cent through the wall. The per cent through wall is calculated by applying the measured angle to a user defined lookup table.
In Figure 1.1-2, the three 3 mm long grooves are shown in both the impedance plane and the strip chart. The 25 per cent groove is marked with the cursor and is measured as 82 degrees, 1.62 volts, and 25 per cent through the wall.
For Figure 1.1-2, the ect AutoMAD Data Analysis Software was enabled. The three defects were scanned by the ect AutoMAD Data Analysis Software and the results are listed in a table below the word mixer. The ect AutoMAD Software has the capability to analyze an entire length of tube and list the ten deepest defects that it encountered. These are shown in order as the 75, 50, and 25 per cent grooves. The good angle difference between these three signals makes it possible to analyze these signals accurately with phase analysis. In fact, the appearance of the signals and the angle spread is very much like eddy current testing of non-ferrous tubes.
Figure 1.1-3 shows the eddy current signal from the 25 per cent O.D. groove and the 20 per cent I.D. groove. For this figure, the screen sensitivity has been increased by a factor of four to 0.5 volts per division.
The angle of the internal groove is measured at 34 degrees and is listed as 21 per cent through the wall from the I.D. The 25 per cent O.D. groove is almost vertical. There is a clear and obvious difference between I.D. and O.D. defects when using Near Field™ technology to inspect ferromagnetic tubes. Near Field™ technology can distinguish between I.D. and O.D. defects in a manner very similar to regular eddy current testing of non-ferromagnetic tubes.
Big Pit Standard
The signals that result from eddy current testing of the big pit standard are shown in Figure 1.1-4. Looking at the strip charts from the top down, you first see the probe entrance in to the tube, followed by the hole and 80, 60, 40, and 20 per cent O.D. pit defect signals. This is then followed by the probe exiting the tube. The first three defects are quite easy to see. The last two defects are somewhat difficult. The 40 per cent defect has a signal to noise ratio of approximately three, and the 20 per cent defect has a signal to noise ratio of approximately two.
In the impedance plane diagram of the differential signals, the hole and the 80 per cent pit are quite easily visible. The 80 per cent pit is the defect marked with the cursor. The three smaller defect signals are covered up by the larger defect signals.
In the absolute impedance plane, the hole and the 80 per cent defects are also quite easy to pick out, but the other signals are obscured by noise and probe wobble. An absolute channel would not be used for the detection and measurement of small defects like these.
The angles of the various defects as measured in the differential channel warrant further discussion. The good angle difference between the hole and the 80 per cent defect would make it possible to distinguish between these two and to measure the depth; however, the angle for the other three defects seems relatively random, making it impossible to measure the depth of these defects by the angle of the eddy current signal on the screen.
Also of note is that the 80 per cent O.D. pit marked by the cursor is listed in the analysis as being 80 degrees, .48 volts, and 29 per cent through the wall from the outside. Exactly as in eddy current testing of non-ferromagnetic tubes, as a defect of a given depth grows in size, its eddy current signal rotates counterclockwise. For more information about this, see Another Thing Not So Easy About Eddy Current Testing of Tubes. The ect MAD 8D Eddy Current System has multiple calibration tables in order to accommodate this phenomenon. A separate calibration table can be prepared and used for different classes of defects, such as grooves and pits. In this case, a calibration table for accurate analysis of the pits was not made because there was not good correlation between defect depth and angle for the shallow defects.
1.2 22.2 mm (7/8 inch) Diameter, 1.25 mm (.049 inch) Wall 1020 Carbon Steel Tube, Dual Send Remote Field Probe
The results of testing these tubes with the Remote Field probe appears in Figures 1.2-1, 1.2-2, and 1.2-3.
An ect MAD 8D Eddy Current Instrument together with an ect Carbon Steel Filter Amplifier Rev. 3 was used.
The test was conducted using a frequency of 1,200 Hz in differential mode only as a reference probe was not available. The impedance plane for the differential signal appears in the upper left hand corner of Figure 1.2-1. There are two strip charts in the figure. The left most strip chart is the differential horizontal signal. The right most strip chart is the differential vertical signal.
In Figure 1.2-1, you can see three signals resulting from the 3 mm long grooves in the groove standard in the strip charts. Starting with the defects at the top, the signals are the 25 per cent, 50 per cent, and 75 per cent 3 mm long grooves. Immediately above the 25 per cent groove signal, you see the signal from the probe entering the tube. Before the beginning of this scan, the send coil near the cable end of the probe and the differential receive coils were inside the tube. The entry signal seen here results from the second send coil entering the tube. Note that this signal continues right up to the 25 per cent groove, even though the groove is 90 mm (three and one half inches) from the end of the tube. The first 90 mm of this tube (a length of tube approximately equal to four times the diameter of the tube) cannot be inspected. Defects in this location would not be detectable.
Keep in mind that this is a dual send Remote Field probe. If a single send Remote Field probe were utilized, then it would be possible to inspect closer to one of the tube sheets, but not both, unless the tube was inspected twice, once from each end of the tube.
In the impedance plane portions of Figure 1.2-1, the 75, 50, and 25 per cent 3 mm long grooves appear. The 25 per cent groove signal is the almost vertical signal that is partially obscured by the 50 per cent groove signal. The cursor marks this 25 per cent defect and lists it as 91 degrees, 3.54 volts, 26 per cent through the wall from the outside.
The ect AutoMAD Data Analysis Software is enabled and the three grooves are listed below the word mixer as 76, 53, and 26 per cent through the wall from the outside.
As there is very good correlation between the angle of the defect signal and the depth of the defect, it is possible to accurately measure the depth of defects with the Remote Field probe.
Figure 1.2-2 shows the signals from a 20 per cent I.D. and a 25 per cent O.D. groove. In both the strip charts and the impedance plane, the 25 per cent O.D. defect can be distinguished from the 20 per cent I.D. defect by the fact that the signal from the deeper defect is larger. In fact, the amplitude of the 25 per cent O.D. defect is 3.5 volts and the amplitude of the 20 per cent I.D. defect is 2.77 volts, a ratio of 1.26, almost exactly the ratio of the defect depth. The shallower I.D. defect is also rotated counterclockwise from the deeper O.D. defect. Phase analysis can be used to measure the depth of I.D. grooves as well as O.D. grooves; however, the angle range the angles of the signals from the I.D. and the O.D. defects is nearly identical. Therefore, it is impossible to determine whether a defect is on the inside or outside of a tube wall using a Remote Field probe.
Figure 1.2-3 shows the signals from the big pit standard. All five defect signals are clearly visible in both the impedance plane and the strip chart presentations. The fact that all signals are drawn in red shows that the defect detector has detected all five of them. The defect depths are listed in the table of defects found under the word mixer as 99, 78, 60, 40, and 19 per cent O.D. The 100 per cent O.D. defect is marked with the cursor in the impedance plane. The other four defects can be seen rotated sequentially clockwise from the 100 per cent through wall hole with the 20 per cent O.D. defect being rotated 90 degrees clockwise from the through wall hole. This is excellent correlation.
As there is good correlation between defect depth and angle, phase analysis can be used to analyze the depth of pit type defects of these sizes. In this case, the calibration table has been prepared for the big pit standard. You will note that calibration table zero is being used. To analyze the grooves as seen in Figure 1.2-1, calibration table one had been used.
1.3 Effect of Support Plate on Near Field™ Eddy Current Testing, Single Frequency
The purpose of support plates is to protect the tubes, so it is ironic that 90 per cent of the outside defects that are in tubes occur close to or underneath the support plates. Obviously, this is an important section of the tube to test accurately.
To simulate a support plate in this study, a clearance hole for the tube was drilled through a 19 mm (3/4 inch) length of mild steel round rod with an outside diameter of 35 mm (1-3/4 inch). This support ring is the same as would be used to simulate a support plate for eddy current testing of non-ferromagnetic tubes. This is possible because most of the electromagnetic field generated by a Near Field™ probe does not extend beyond the tube wall. On the other hand, in the case of remote field testing, the field must travel several tube diameters from the send coil(s) to the receive coil(s) and is blocked when a support plate is between the send and receive coils. The support ring as described above would not block the signal as much as a large support plate. For this reason, to simulate a support plate for remote field testing of tubes, it is necessary to have a plate approximately six times larger in both dimensions than the diameter of the tube in order to adequately simulate the amount of signal blocked by a real support plate.
Figure 1.3-1 is similar to Figure 1.1-3 above in that it shows the 25 per cent and 50 per cent 3 mm grooves with the 25 per cent groove marked with the cursor. Figure 1.3-1 differs from Figure 1.1-3 in that there is the additional signal from a support ring located midway between the two groove signals. In the differential signal, note that the support ring signal is rotated clockwise from the 25 per cent groove. In the absolute signal, note that in addition to being rotated from the 25 per cent groove, the signal moves in the opposite direction; that is to say to the right and up, as opposed to to the left and down. This is what one would expect because the support ring is farther from the probe than the groove; hence, the rotation, and because it adds material as opposed to removing material, it moves in the opposite direction.
Figure 1.3-2 shows the Near Field™ eddy current signals when the support ring is placed adjacent to the 25 per cent 3 mm O.D. groove. Note that the presence of the support ring has rotated the signal clockwise by 15 degrees to 97 and that the analysis is NO because this angle is outside of the range of angles for tube damage. However, the fact that there is a significant vertical component added to the support ring signal and the fact that the defect detector has picked up on it, it should be obvious to an Inspector that there is some form of damage at this tube support.
Figure 1.3-3 shows the support ring adjacent to the 25 per cent 3 mm O.D. groove. Part b of this diagram shows the support ring positioned over the 25 per cent 25 mm long O.D. groove.
Figure 1.3-4 shows the 50 per cent 3 mm long O.D. groove signal and the support ring signal. The support ring is located 12 mm (.5 inch) from the 50 per cent groove. Note that it is analyzed correctly as 50 per cent through the wall. Note also that it is analyzed as 4.84 volts, which is very similar to the 4.6 volt analysis in Figure 1.1-2, where it was located far from any support ring.
Figure 1.3-6 shows the 25 per cent 25 mm long O.D. groove signal and the support ring signal. Figure 1.3-7 shows the signal from the support ring superimposed over the 25 per cent O.D. groove. This analyzes incorrectly as 2% for the same reasons that the 25 per cent 3 mm groove was analyzed as NO. However, as the signal is significantly different than the signal for the support plate alone, it should be obvious to the Inspector that there is some form of damage at this support plate.
Figure 1.3-8 shows the 80 per cent pit signal from the big pit standard and the support ring signal in both the differential and absolute impedance plane screens. In the vertical strip chart, you can see both of these signals and the signal from the through wall hole. Note that the vertical amplitude is about the same for all three of these signals. The support plate signal is a larger signal in the impedance plane screens and goes off the screen in the horizontal direction. The total amplitude of the support plate signal is significantly larger than the signal from either of these defects. The support ring is positioned between these two defects.
Figure 1.3-9 shows the Near Field&trade eddy current signal for the support ring centered over the 80 per cent pit. There is very little difference between this signal and the signal from the support plate alone; therefore, it is probable that an Inspector would not notice this defect at the support plate. Having the support ring centered over the 80 per cent pit is the location where it is least likely to be detected. If the support plate is moved so that the pit is close to one edge of the support plate, then it may be detectable.
Section 1.4 Effect of Support Plate on Near FieldTM Eddy Current Testing, Two Frequency with Mix
Section 1.5 Effect of Support Plate on Remote Field Eddy Current Testing, Single Frequency
In a remote field probe, the electromagnetic signal that couples the send coil(s) and receive coil(s) travels for some distance outside the tube and at some distance from the tube. For this reason, the support ring that was described in Section 1.3 is not sufficiently large in the radial direction to properly represent a support plate in a real heat exchanger. In this case, to simulate a support plate, a clearance hole for the tube was drilled through a 150 mm (6 inch) square plate, 6.4 mm (0.25 inch) thick. Four of these were made and stacked together to make a support plate with a thickness of 25 mm (1 inch).
Figure 1.5-1 shows the 25 per cent 3 mm O.D. groove and the support plate signal at separate locations in the tube. The 25 per cent groove is marked with the cursor and analyzed as 25 per cent through the wall. Pay close attention to the shape of the support plate signal in the horizontal strip chart on the left in the strip chart screen. Shortly below where the 25 per cent groove is marked and drawn in red as a defect is the beginning of the support plate signal. Note that the horizontal strip chart moves to the right about one half a division and stays at this location for a period of time. This signal occurs when the leading send coil of this dual send coil probe enters the support plate. The signal then moves rapidly to the right, a total of almost two divisions before moving to the left four divisions. This is when the differential receive coil pair enters, passes through, and exits the support plate. The signal then remains about one half a division to the left of center before returning to center. This is the period of time when the second send coil is on the far side of the support plate or in the support plate. After this send coil exits the support plate, the signal returns to the center point. This lengthened support plate signal is caused when one of the send coils is in or on the opposite side of the support plate from the receive coils, such that its transmitted electromagnetic energy is blocked from the receive coils. If a single transmit coil remote field probe had been used, then when the send coil is on the opposite side from the support plate from the receive coils, the path from the send to the receive coils is blocked, creating a blind spot adjacent to one side of the support plate for a distance of two or three tube diameters, where defects cannot be detected. Because a dual transmit remote field probe performs better, it was chosen for this study. Note also that in this figure, the 25 per cent O.D. groove has a magnitude of 2.8 volts.
In Figure 1.5-2, the support plate is located 12 mm from the 25 per cent O.D. groove. In this case, the analysis of the groove signal shows 14 per cent through the wall from the outside because it has been rotated counterclockwise. Note also that the amplitude of this signal has been reduced to two volts.
In Figure 1.5-3, the support plate is located adjacent to the 25 per cent O.D. 3 mm groove. The signal looks like a slightly distorted support plate signal, but the defect detector does trigger on the groove, reporting it as 1.35 volts and 111 degrees. Although the defect depth measurement is indicated as no, it is possible that the Inspector may note that there is some form of damage at this support plate, although he cannot measure it. If the defect detector had not been turned on, or if the eddy current instrument did not have one, then it is possible the Inspector would have missed this defect entirely.
Figure 1.5-4 shows the 50 per cent 3 mm O.D. groove far from a support plate. At 78 degrees and 7.6 volts, it is correctly analyzed as 50 per cent through the wall.
Figure 1.5-5 shows the 50 per cent 3 mm long O.D. groove located 12 mm from the support plate. It is measured at 85 degrees and 40 per cent through the wall. The amplitude has dropped to 4 volts, almost one half of the original. It is incorrectly analyzed at 40 per cent through the wall, but this is still considered as good accuracy.
Figure 1.5-6 is the signal from the 50 per cent 3 mm long O.D. groove adjacent to the support plate. Here it is rotated to 99 degrees and is just over 4.4 volts. Based on the angle, it is measured at 17 per cent through the wall, in error by 33 per cent.
Grooving, which often occurs adjacent to support plates, cannot be accurately measured with the remote field eddy current technique.
Figure 1.5-7 shows the 50 per cent 25 mm (1 inch) long O.D. groove with the support plate located far from the groove. The support plate signal is the green signal (about 2 divisions long) as it has not triggered the defect detector. In the horizontal strip chart, the step before and after the largest portion of the support plate signal which is caused by the support plate blocking the path from the send to the receive coils is readily visible. Note also that similar steps appear before and after the signal for the groove. The step that is seen before and after the 50 per cent groove is caused when the send coils are under this groove. The presence of a large groove like this will interfere with signals from other defects that are within a few tube diameters of the larger defect.
Note that the 50 per cent groove signal is so large that it goes off the screen, even at a screen sensitivity of 2 volts per division.
Figure 1.5-8 is the signal when the support plate is positioned over the 50 per cent wall loss. Despite the presence of the wall loss, the signal has been diminished substantially from the signal in the previous figure that went off the screen. It is true that the defect detector has triggered on this, and it is true that this does not look like a support plate on its own, so the Inspector should realize that something is occurring at the support plate, but may not be able to determine that 50 per cent of the wall has been lost. Damage such as this at support plates is quite common and is caused by tube vibration. This is a very serious defect, but it is difficult to detect with remote field because it is lost under the support plate.
Figure 1.5-9 is the signal from the big pit standard with the support plate located between the through wall hole and the 80 per cent pit. The through wall hole is 12 mm from the support plate. The 80 per cent pit is 40 mm (1.5 inches) from the support plate. By comparison to Figure 1.2-3, the reader should be able to pick out the 20, 40, and 60 per cent pits centered at the center of the screen and reasonably accurately analyzed as 23, 41, and 60 per cent through the wall. The cursor is on the 80 per cent defect, which is measured as 64 per cent through the wall. The through wall hole signal is severely distorted by the presence of the support plate and is that signal that points up and to the left below and to the left of the center of the screen. The defect detector did not detect this defect, even though it has colored a small segment of the defect red. An attentive Operator might see this and realize that something has occurred to the tube, but it would be difficult to determine what sort of damage it is. Note that under the word Mixer, only four of the five defects in this tube are listed.
Figure 1.5-10 is the signal from the big pit standard with the support plate located over the 80 per cent pit. This positions the support plate 60 mm (2.4 inches) from the 60 per cent pit and also 60 mm from the through wall hole. In the list of defects, the hole is reported as 94 per cent through the wall and the 60 per cent pit (highlighted by the cursor) is reported as 52 per cent. The 80 per cent signal is not detected, as it is located under the support plate. Any of these defects from the big pit calibration tube including the hole would not be detected if they are within a few tube diameters of the support plate.
Section 1.6 Effect of Support Plate on Remote Field Eddy Current Testing, Two Frequency with Mix