Magnetic flux leakage examination (MFL) tank bottom scanning is the accepted nondestructive test for under-side corrosion detection. There are two approaches to using MFL. One method is thresh-holding that alarms, stops, maps or is a combination of the three. Another method is to have an MFL scanner provide a real-time display of all signals regardless of size or strength of the signal received.
Signals are affected by the following main variables:
1. The amplitude is relative to the remaining wall thickness.
2. The consistencies of the magnetic properties of the material being tested.
3. The depth of the corrosion pit on the underside of the material.
4. The diameter of the corrosion pit on the underside of the material.
5. The orientation of the corrosion pit on the underside of the material where the length and width of the pit are not the same.
6. The distance of the sensor bar and/or magnets are from the surface of the material, which in a real-world tank bottom varies due to contour of the bottom plates.
7. The interaction of corrosion pits close to each other.
8. The interaction of the above variables with each other.
The signal of MFL units is volumetric. The MFL signal is the combination of all variables. One MFL manufacturer has a functioning plate that is used to confirm the unit is operating correctly. A one-quarter-inch plate has two simulated defects — a one-half-inch diameter simulated pit with a 40-percent wall loss and a three-sixteenth-inch hole. Both defects produce the same level of signal. Therefore, if a thresh-hold is set for a one-half-inch diameter pit greater than 40-percent loss, a three-sixteenth-inch hole would not be found.
Thresh-holding is only for one variable; however, it is obvious there are many variables creating the MFL signal. It is not reliable to set a signal height (thresh-hold) for determining the depth of an underside corrosion pit. If an MFL scanner has stop on defect or mapping tool and a real-time display, it is my opinion thresh-holding can be a viable option only to aid the operator and not to erroneously determine pit depth.
Real-time display MFL rely on the operator to determine what is recordable and to test by ultrasound for the remaining wall thickness. With experience, a qualified MFL real-time display unit operator can reliably scan a tank bottom. Any signal above what is called the noise level of the scanning operation is subject to further testing.
Corrosion pitting is one of three types: 1. lake type, 2. localized pitting or 3. bacteriological. Lake type pits give a wide signal and, typically due to the width, create a high-amplitude signal. Localized pits typically give a narrow signal with varying amplitude of signal. Bacteriological pitting tends to be very small in diameter and very deep. The signal from bacteriological pitting is typically very narrow and very low amplitude. Understanding the creation of the amplitude of a MFL signal concludes a thresh-hold for amplitude will not work. Since the variables have an effect on other variables, I do not believe there is any instrument that is capable of accurately accounting for the variables that gives the operator only the depth of the underside corrosion pit.
There are many in the storage tank industry who desire to have an instrument that will only determine the depth of underside pits regardless of the size, shape or lift-off of the scanner. I, too, would like to have such a unit; however, there is currently no unit that can account for all the variables in MFL signals. It would be desirable to find only the depth of an underside corrosion pit that needs to be repaired; however, the unit that can do this is not possible given today’s technology. Therefore, real-time display units are the most reliable means of performing MFL tank bottom scans.
For more information, contact Dan Boley at danb@djainspection.com.
