MFL inspection systems detect possible defects by measuring and analyzing changes in flux leakage signals from a base level. The extra separation caused by wear plates, internal, deposits and liners affects both signal and base levels, thus affecting detection and sizing capability.

The presence of a liftoff has two effects. First, the amount of the air-coupled field increases. The sensor measures a higher base signal level for a sensor with a large lift-off than for a sensor with a small liftoff. Second, there is a change in signal level at defects. The additional separation decreases the amplitude and increases the length or duration of the signal.

The net result of these two effects is less sensitivity to smaller defects; the defect signal may be less than the increase in the air coupled field. In the above figure, the effects are shown conceptually for a pipe with and without a liner.

The flux leakage fields from many defects and degrees of liftoff were examined for the Gas Research Institute. The purpose of the analysis was to evaluate the effect of wear plates that protect the sensor on long distance inspection runs. The parameters that were evaluated were:

External spherical pit


- 0.5 inch wall-thickness
- 33% deep
- 0.5 inch diameter

Six liftoffs (0.08, 0.16, 0.24, 0.32, 0.40, 0.48)
Field levels forced to 100 gauss, independent of liftoff.

One important variable, magnetizer liftoff, was held constant in this analysis. A magnetizer designed for a specific liftoff distance could minimize the adverse effects of the air-coupled field.

For the case shown above, the minimal liftoff signal (0.08 inches) is approximately 57 gauss above the base magnetization level, while the maximum lift off signal (0.48 inches) is only 7 gauss above the base magnetization level. This represents a factor of 8 reduction in signal strength. The typical noise level due to magnetic property and sensor liftoff variation is nominally between one to two gauss. So, the signal-to-noise ratio varied from dropped 30 for small liftoff to about 4 for large liftoff. A small signal-to-noise ratio usually indicates a signal that is difficult to analyze.

The above plot indicates that an MFL system designed to inspect with a large liftoff may be able detect relatively small defects, but it would have difficulty sizing them. The noise levels and signal quality would be depend on liftoff uniformity and the sensor type. The ability to estimate the depth of defect decreases with increasing liftoff. Depending on defect geometry, a small change in liftoff may cause a large change signal amplitude.

In addition, magnetization systems not designed to inspect at large liftoff may introduce other sources of error, as discussed next.

Experimental data using magnetizers and sensors that represent commercial inspection systems were acquired at the Pipeline Simulation Facility using the linear test rig. A reference metal-loss defect and four defects with a variation of one dimensional parameter were machined in the pipe with the dimensions shown below. These defects were on the external surface of the pipe.

Description	Depth (percent)	Length (inches)	Width (inches)
Reference	50	2.0	3.0
Deep	80	2.0	3.0
Shallow	20	2.0	3.0
Wide	50	2.0	6.0
Long	50	4.0	3.0

The signals from the five external defects acquired at three liftoff separations:

• contact,
• 0.40 inches (10mm)
• 0.80 inches (20mm)

The experimental results depend on magnetizer variables. For these experiments, the separation between the north and south poles (the pole spacing) was 14 inches, which is considered wide for most pigging applications. This magnetizer would be restricted to inspection of pipelines with bend radius of greater than 5 pipe diameters (5D bend) in a 24-inch pipeline.

The results are shown above. The data at the 0.40 (10mm) liftoff show:

• A slight increase in the air coupled field
• A decrease is signal amplitude that is dependent on defect geometry. For these defects, the absolute peak amplitude is conserved for the smallest defect while largest drop occurred for defects with the largest amplitude.
• The signal shape was only slightly degraded, which would lead to a only a small error in the estimation of defect length

The data at the 0.80 (20mm) liftoff show:

• A significant increase in the air coupled field
• The near loss of signal from a 20 percent deep, 2 inch long, 3 inch wide machined area at the highest liftoff.
• A general degradation of the signal shape, which would lead to errors in the estimation of defect length

The effects of pole spacing on these results are difficult and confounding. A wide pole spacing decreases the air-coupled field and lowers the applied magnetization level in the pipe. Reducing the pole spacing enables the passage of tight bends and increases the applied magnetization level, which would increase the signal levels and aid in detection. It also increases the air-coupled field, though, which could mask defect signals. The net result depends on the specific tool design.

The effect of pole spacing is also influenced by inspection velocity. Usually, a wide pole spacing is needed necessary for high inspection speeds to reduce the effects of velocity. For inspections where tool velocity is not expected to he high (faster than 4 mph), lift off can be increased and pole reduced.