MFL is an electromagnetic NDT method that magnetizes a ferromagnetic component to saturation and detects flux leakage at metal-loss features using Hall-effect or coil sensors, widely used for tank-floor and pipeline pig inspection.
Definition
Magnetic Flux Leakage (MFL) inspection magnetizes a ferromagnetic component, typically to near magnetic saturation. Where wall thickness is locally reduced — by pitting, corrosion, or grooving — magnetic flux 'leaks' from the surface and is sensed by a Hall sensor or coil array. The leakage amplitude correlates with metal loss.
Technical Context
MFL is the dominant technique for storage-tank floor scanning and in-line pipeline inspection (ILI pigs). It is fast and reliable for general wall loss but less sensitive to circumferentially oriented cracks.
When It Is Used
API 653 tank-floor inspection
Pipeline ILI pigging
Tube and wire rope inspection
Related Standards
API 653, API 1163, ASTM E570.
How it works
A ferromagnetic wall is driven near magnetic saturation. Metal loss reduces the cross-section available to carry that flux, so flux is forced out of the wall and into the surrounding space, where sensors detect it. Signal amplitude relates to the volume of metal missing, not to depth alone.
What it finds
Volumetric metal loss — corrosion, pitting, gouging — quickly and over large areas. Tank floors and pipelines are the classic applications, and in-line inspection tools rely on it.
What it will not find
Tight planar flaws such as cracks aligned with the flux, which remove almost no metal and therefore leak little. Very gradual general thinning can also under-respond because the flux gradient is what generates the signal.
How it is actually done
Wall thickness limits apply, since saturation must be achievable; lift-off, scanning speed and coating thickness all modulate signal amplitude and have to be controlled or compensated. Results are usually confirmed by ultrasonic thickness at the indications found.
Governing codes and standards
API 653 for tank floor examination; API 1163 for in-line inspection system qualification; ASTM E570 for flux leakage examination of ferromagnetic tubing.
Where it goes wrong
Reading amplitude as depth. Signal responds to lost volume, so a broad shallow patch and a narrow deep pit can produce comparable amplitudes — which is exactly why MFL screens and ultrasonics confirms.
Where Magnetic Flux Leakage fits in an inspection programme
A term is only useful when it connects to a decision. Magnetic Flux Leakage appears in written procedures, in technique sheets, and in the records an owner or accreditation body reviews afterwards — which means the way it is defined in your documentation has to match the way it is applied on site. Where the two drift apart, audits find it. Atlantis writes and reviews procedures against the governing codes, trains inspection personnel to apply them, and builds the record-keeping that makes the evidence retrievable years later. Procedure development and code consulting · NDT training and certification · Ask us about your programme.
Related terms
Storage Tank — A storage tank is a low-pressure, large-volume welded container for the storage of liquids — typically atmospheric (API 650), low pressure (API 620), or for water (AWWA D100) — inspected in service per API 653.
API 653 Tank Inspection, Repair, Alteration, and Reconstruction — API 653 is the inspection, repair, alteration, and reconstruction code for above-ground welded steel storage tanks built to API 650, defining periodic external (5 yr), ultrasonic (10 yr), and internal (typically 10 yr) inspection intervals.
Pipeline — A pipeline is a long-distance pipe system for transporting fluids (crude oil, refined products, natural gas, water, chemicals), inspected by in-line inspection (ILI/pigging), direct assessment, and hydrostatic testing under API 1163, ASME B31.4, and B31.8.
Terms like this one appear in three places that matter commercially: the written practice that governs how your personnel are qualified, the procedures and technique sheets that define how an examination is actually performed, and the evidence an auditor or client asks for when they want to know why an inspection was accepted. Getting the terminology right is the easy part; being able to produce the qualification record, the calibration traceability and the procedure revision that applied on the day of the inspection is the part that decides audits.
Magnetic flux leakage detects metal loss by saturating the steel with a magnetic field and sensing the flux that escapes the surface where the section is reduced. Sensors ride over the plate and register the leakage field; the resulting signal indicates that material is missing and roughly how much, but it does not directly measure depth.
That distinction governs how the method is deployed. Leakage amplitude responds to the volume and shape of the loss as well as its depth, so a broad shallow area and a narrow deep pit can produce comparable signals. On tank floors the standard answer is to treat MFL as a screening tool: the scanner covers the floor rapidly and marks every indication, and ultrasonic prove-up then measures remaining thickness at each mark. The floor is inspected in full, but the measurement that feeds the remaining life calculation comes from ultrasonics, not from the leakage signal. In-line pipeline inspection uses the same physics on a much larger scale, with the tool magnetising the pipe wall as it travels. Axial field tools are sensitive to circumferentially oriented and general metal loss, while transverse or circumferential field tools are needed for axially oriented features such as long seam anomalies, which an axial tool can under-call.
Source: API 653 Tank Inspection, Repair, Alteration and Reconstruction for tank floor examination and prove-up; API RP 651 for cathodic protection of aboveground storage tanks; API 1163 In-Line Inspection Systems Qualification; ASNT recommended practice for personnel qualification in magnetic flux leakage.
What magnetic flux leakage tells you, and what it does not
Question
MFL answer
What is actually needed
Is there metal loss here?
Yes, reliably and quickly over large areas
MFL alone is sufficient
How deep is it?
An estimate influenced by volume and shape
Ultrasonic prove-up at the indication
Is the loss top side or soil side?
Distinguishable with a top-side discrimination sensor
Confirmed at prove-up
Is it a crack?
Poorly - MFL is a metal loss method
Ultrasonic, ACFM or a dedicated crack tool
Axially oriented seam anomaly?
Under-called by an axial field tool
Transverse or circumferential field tool
Remaining thickness for the calculation
Not a direct measurement
Ultrasonic thickness at the governing location
MFL finds and locates; ultrasonics measures. Treating a leakage amplitude as a thickness is the central misuse of the method.
Why saturation level decides whether the survey means anything
The physics depends on driving the steel close to magnetic saturation. Below saturation the relationship between section loss and leakage field is weak and inconsistent, so the same defect produces different signals depending on how well the magnet is coupling.
Coupling is where field conditions intrude. Plate thickness beyond the tool's design range, heavy scale or debris holding the magnet off the surface, and coating thickness all reduce the field in the steel. A floor scanner run on plate thicker than it was designed to saturate will produce a clean-looking survey that is simply insensitive.
This is why the scanner's thickness range and the surface preparation standard belong in the procedure and in the report. A survey without them recorded cannot be assessed afterwards for whether it was capable of finding what it was looking for.
Tank floors: what a compliant MFL campaign actually delivers
Under API 653 the floor examination has to establish the minimum remaining thickness so that the floor's remaining life and any repair requirement can be determined. MFL contributes coverage; ultrasonics contributes the number.
A defensible campaign therefore records the scan coverage achieved including the areas the scanner could not reach — the critical zone at the shell-to-floor junction, around sumps and columns, and under any obstruction — because those unscanned areas need their own examination rather than an assumption. Coverage claimed rather than recorded is a recurring audit finding.
Top side versus soil side discrimination matters commercially as much as technically. Soil side loss points toward cathodic protection and under-tank environment, top side loss toward product and water bottoms, and the two lead to entirely different remediation. A survey that reports loss without discriminating leaves that decision unsupported.
In-line inspection: orientation is the thing to get right
An MFL in-line tool magnetises the pipe wall as it travels and reads leakage through sensor arrays around the circumference. The field direction determines what it sees well. An axial field responds strongly to features that interrupt it — general corrosion and circumferentially oriented loss — and weakly to a narrow axial feature aligned with the field.
That is why long seam anomalies and axially oriented crack-like features need a transverse or circumferential field tool, or a different technology altogether. Running an axial tool and concluding the seam is sound is a known false-negative path.
Tool performance specifications are qualified under API 1163, which is the reference for what a stated detection threshold and sizing accuracy actually mean. Reading the vendor specification without reference to the qualification basis behind it is how unrealistic expectations get written into integrity plans.
How does magnetic flux leakage work?
The steel is magnetised close to saturation. Where the section is reduced by metal loss, the flux cannot all remain in the steel and some leaks out through the surface, where sensors riding over the plate detect it. The signal indicates that material is missing and approximately how much.
Does MFL measure the depth of corrosion?
Not directly. Leakage amplitude responds to the volume and shape of the loss as well as its depth, so a broad shallow area and a narrow deep pit can give comparable signals. Depth comes from ultrasonic prove-up at the indication, which is what feeds the remaining life calculation.
Why is MFL used on tank floors instead of ultrasonics alone?
Coverage. A floor is a large area and ultrasonic thickness measurement point by point would be impractical across all of it. MFL scans the floor rapidly and marks every indication; ultrasonics then measures remaining thickness only where it matters, giving full coverage and defensible numbers together.
Can MFL detect cracks?
Poorly. It is a metal loss method, and a tight planar crack displaces little material and may not disturb the flux enough to produce a reliable signal. Crack detection needs ultrasonics, alternating current field measurement, magnetic particle, or a purpose-built in-line crack detection tool.
What is the difference between axial and transverse field MFL tools?
Field orientation determines sensitivity. An axial field tool responds well to general and circumferentially oriented metal loss but under-calls narrow axially oriented features aligned with the field. Long seam anomalies therefore require a transverse or circumferential field tool rather than an axial run.
What limits MFL sensitivity in the field?
Anything that stops the steel reaching saturation — plate thicker than the tool's design range, heavy scale or debris holding the magnet off the surface, and excessive coating thickness. A survey run outside the tool's capability looks clean because it is insensitive, which is why the thickness range and surface standard belong in the report.
Frequently asked
Does MFL distinguish soil-side from top-side corrosion on a tank floor?
With a top-side discrimination sensor, yes, and the distinction matters because soil-side loss points toward cathodic protection and the under-tank environment while top-side loss points toward product and water bottoms. The two lead to different remediation, so a survey that does not discriminate leaves that decision unsupported.
How is in-line MFL tool performance qualified?
Under API 1163, which sets out how detection thresholds, sizing accuracy and confidence levels for in-line inspection systems are established and validated. A stated specification is only meaningful in relation to that qualification basis.