I recently received an e-mail from the NDT Department at a Chemical Plant. They were concerned about the accuracy of eddy current testing of heat exchanger tubes because they had recently used a Sub-Contractor to provide inspection services for several of their heat exchangers. The Contractor had reported on large numbers of internal pits ranging in depth from 40 per cent through the wall all the way to 80 per cent through the wall. The Chemical Company removed a few tubes with the worst indications and discovered that they were not pits at all. After consulting with the Sub-Contractor, it was concluded that the Sub-Contractor had mistakenly identified magnetic deposits in the tube as being I.D. pits. The person who sent me the e-mail was actually suggesting that if eddy current testing is this easily fooled, then eddy current testing of heat exchanger tubes was useless and I should close down Eddy Current Technology Incorporated.
I responded that a competent eddy current Inspector would know how to distinguish between deep internal pits and magnetic inclusions by using a multi-frequency test; in fact, one of the reasons that I started Eddy Current Technology Incorporated was provide multi-frequency eddy current instruments to industry that needed more and better information about the condition of their heat exchanger tubes.
To show how easily this confusion can be made, Figure 1 shows the signal from the through wall hole in an ASME Calibration Tube and Figure 2 shows the signal of a magnetic inclusion in the same tube (or is it the other way around?). It can be seen that in Figure 1, the through wall hole is correctly analyzed as 100 per cent through the wall (6 th line of text in the menu), and in Figure 2, the signal from the magnetic inclusion where there is little or no wall loss has been analyzed as 98 per cent through the wall.
An Inspector may choose not to investigate any further and just report this as a 98 per cent I.D. pit, recommending that the tube be removed from service, either by plugging or by replacement.
The way to distinguish between a serious internal pit and a mundane magnetic inclusion in the tube is with a multi-frequency test.
First, I would like to review the signals that you get from a multifrequency test of an ASME calibration tube. In Figure 3, the eddy current instrument is set somewhat differently than I have used in previous articles. As before, Channel 1 is in the upper left hand corner and is 2SDF (two times the Skin Depth Frequency). This is considered the primary inspection frequency, and it was used in Figures 1 and 2 above. Channel 2 is set to 1SDF and is in the upper right hand corner of the screen. Channel 3 is in the lower left hand signal. This is an absolute channel, and it is not important for this discussion. Channel 4 is set to one half SDF, and is in the lower right hand corner of the screen.
In past articles,the output from the Mixer Channel was normally in this location. In the screens for Channels 1, 2, and 3, you can see the defect signals from the 100, 80, 60, 40, and 20 per cent O.D. defects in the ASME Calibration Tube. In the highest frequency channel where the skin effect is the greatest, there is a large phase angle spread between the through wall defect (approximately 45 degrees clockwise from horizontal), and the shallowest defect (because the skin effect is more predominant at this higher frequency). At the lowest frequency in the lower right hand corner of the screen, the phase angle spread between the through wall hole and the 20 per cent defect is much less.
Figure 4 shows the signal from the through wall hole in all four channels.
Figure 5 shows the signal from the magnetic inclusion in all four channels. In Channel 1, which is the primary inspection frequency, the magnetic inclusion is analyzed as 98 per cent through the wall from the inside. The real clue that this is a magnetic inclusion and not a through wall defect is in Channel 4, the low frequency or confirmation channel, where it can be seen that the magnetic inclusion signal has been rotated to look like a shallow O.D. defect in that channel.
In Figures 4 and 5, the eddy current instrument has been set to give the per cent wall loss of the defect signal for all available channels. This information appears just below the word Channel 1A in the menu, where indications are given for Channels 1 through 4 (C1 through C4) and Mixers 1 and 2 ( M1 and M2). Note that the analysis for Channels 1, 2, 4 and M1 are 100, 99, 96 per cent outside and 98 per cent inside. The maximum error is 4 per cent. With such good agreement between the analysis in the various channels, the Inspector can be confident that this is truly a defect. By comparison, in Figure 5 the analysis in these Channels varies from 98 per cent inside for Channel 1 to 13 per cent outside for Channel 4, the confirmation channel. Since the confirmation channel is so far from agreeing with the primary channel, it is concluded that this is not a defect and in fact, this is the signature from a magnetic inclusion.
Real tube wall loss, such as an I.D. pit, will analyze with relatively similar per cent wall loss indications in all frequency channels, whereas signals from another source, such as internal magnetic inclusions, will analyze at significantly different per cent wall loss values in the various frequency channels. An eddy current Inspector must analyze all eddy current signals at all available frequencies in order to confirm whether or not the signal originates from a wall loss defect.
It is useful to have on the screen the analysis of all frequency channels so that this determination can easily be made.