Magnetic Field Testing

R.E. Beissner, Southwest Research Institute

Introduction

MAGNETIC FIELD TESTING includes some of the older and more widely used methods for the nondestructive
evaluation of materials. Historically, such methods have been in use for more than 50 years in the examination of
magnetic materials for defects such as cracks, voids, or inclusions of foreign material. More recently, magnetic methods
for assessing other material properties, such as grain size, texture, or hardness, have received increasing attention.
Because of this diversion of applications, it is natural to divide the field of magnetic materials testing into two parts, one
directed toward defect detection and characterization and the other aimed at material properties measurements.

This article is primarily concerned with the first class of applications, namely, the detection, classification, and sizing of
material flaws. However, an attempt has also been made to provide at least an introductory description of materials
characterization principles, along with a few examples of applications. This is supplemented by references to other review
articles.

All magnetic methods of flaw detection rely in some way on the detection and measurement of the magnetic flux leakage
field near the surface of the material, which is caused by the presence of the flaw. For this reason, magnetic testing
techniques are often described as flux leakage field or magnetic perturbation methods. The magnetic particle inspection
method is one such flux leakage method that derives its name from the particular method used to detect the leakage field.
Because the magnetic particle method is described in the article "Magnetic Particle Inspection" in this Volume, the
techniques discussed in this article will be limited to other forms of leakage field measurement.

Although it is conceivable that leakage field fluctuations associated with metallurgical microstructure might be used in
the analysis of material properties, the characterization methods now in use rely on bulk measurements of the hysteretic
properties of material magnetization or of some related phenomenon, such as Barkhausen noise. The principles and
applications of magnetic characterization presented in this article are not intended to be exhaustive, but rather to serve as
illustrations of this type of magnetic testing.

The principles and techniques of leakage field testing and magnetic characterization are described in the two sections that
follow. These sections will discuss concepts and methods that are essential to an understanding of the applications
described in later sections. The examples of applications presented in the third section will provide a brief overview of the
variety of inspection methods that fall under the general heading of magnetic testing.

Magnetic Field Testing

R.E. Beissner, Southwest Research Institute

Principles of Magnetic Leakage Field Testing

Origin of Defect Leakage Fields. The origin of the flaw leakage field is illustrated in Fig. 1. Figure 1(a) shows a
uniformly magnetized rod, which consists of a large number of elementary magnets aligned with the direction of
magnetization. Inside the material, each magnetic pole is exactly compensated by the presence of an adjacent pole of
opposite polarity, and the net result is that interior poles do not contribute to the magnetic field outside the material. At
the surfaces, however, magnetic poles are uncompensated and therefore produce a magnetic field in the region
surrounding the specimen. This is illustrated in Fig. 1(a) by flux lines connecting uncompensated elementary poles.
Fig. 1 Origin of defect leakage fields. (a) Magnetic flux lines of a magnet without a defect. (b) Magnetic flux
lines of a magnet with a surface defect. Source: Ref 1

If a slot is cut in the rod, as illustrated in Fig. 1(b), the poles on the surface of this slot are now also uncompensated and
therefore produce a localized magnetic field near the slot. This additional magnetic field, which is represented by the
extra flux lines in Fig. 1(b), is the leakage field associated with the slot.

Figure 1, although adequate for a qualitative understanding of the origin of leakage fields, does not provide an exact
quantitative description. The difficulty is the assumption that the magnetization remains uniform when the flaw is
introduced. In general, this does not happen, because the presence of the flaw changes the magnetic field in the vicinity of
the flaw, and this in turn leads to a change in magnetization near the flaw. With regard to Fig. 1, this means that the
strengths and orientations of the elementary dipoles (magnets) actually vary from point to point in the vicinity of the flaw,
and this variation also contributes to the flaw leakage field. The end result is that the accurate description of a flaw
leakage field poses a difficult mathematical problem that usually requires a special-purpose computer code for its
solution.

Experimental Techniques. One of the first considerations in the experimental application of magnetic leakage field
methods is the generation of a suitable magnetic field within the material. In some ferromagnetic materials, the residual
field (the field that remains after removal of an external magnetizing field) is often adequate for surface flaw detection. In
practice, however, residual magnetization is rarely used because use of an applied magnetizing field ensures that the
material is in a desired magnetic state (which should be known and well characterized) and because applied fields provide
more flexibility (that is, one can produce a high or low flux density in the specimen as desired.

Experience has shown that control of the strength and direction of the magnetization can be useful in improving flaw
detectability and in discriminating among different types of flaws (Ref 1, 2, 3, 4, 5, 6, 7, 8, and 9). In general, the
magnitude of the magnetization should be chosen to maximize the flaw leakage field with respect to other field sources
that might interfere with flaw detection; the optimum magnetization is usually difficult to determine in advance of a test
and is often approached by trial-and-error experimentation. The direction of the field should be perpendicular to the
largest flaw dimension to maximize the effect of the flaw on the leakage field.

It is possible to generate a magnetic field in a specimen either directly or indirectly (Ref 10, 11, 12). In direct
magnetization, current is passed directly through the part. With the indirect approach, magnetization is induced by placing
the part in a magnetic field that is generated by an adjacent current conductor or permanent magnet. This can be done, for
example, by threading a conductor through a hollow part such as a tube or by passing an electric current through a cable
wound around the part. Methods of magnetizing a part both directly and indirectly are illustrated schematically in Fig. 2.
Fig. 2 Methods of magnetization. (1) Head-shot method. (b) Magnetization with prods. (c) Magnetization with a
central conductor. (d) Longitudinal magnetization. (e) Yoke magnetization

The flaw leakage field can be detected with one of several types of magnetic field sensors. Aside from the use of
magnetic particles, the sensors most often used are the inductive coil and the Hall effect device.

The inductive coil sensor is based on Faraday's law of induction, which states that the voltage induced in the coil is
proportional to the number of turns in the coil multiplied by the time rate of change of the flux threading the coil (Ref 13).
It follows that detection of a magnetostatic field requires that the coil be in motion so that the flux through the coil
changes with time.

The principle is illustrated in Fig. 3, in which the coil is oriented so as to sense the change in flux parallel to the surface of
the specimen. If the direction of coil motion is taken as x, then the induced electromotive force, E, in volts is given by:
where N is the number of turns in the coil, A is its cross-sectional area, and B is the flux density, in Gauss, parallel to the
surface of the part. Thus, the voltage induced in the coil is proportional to the gradient of the flux density along the
direction of coil motion multiplied by the coil velocity. Figure 4 shows the flux density typical of the leakage field from a
slot, along with the corresponding signal from a search coil oriented as in Fig. 3.

Fig. 3 Flux leakage measurement using a search coil. Source: Ref 13

 Fig. 4 Leakage flux and search coil signal as a function of position. Source: Ref 13

Unlike the inductive coil, which provides a measure of the flux gradient, a Hall effect sensor directly measures the
component of the flux itself in the direction perpendicular to the sensitive area of the device (Ref 1). Because the response
of a Hall effect sensor does not depend on the motion of the probe, it can be scanned over the surface to be inspected at
any rate that is mechanically convenient. In this respect, the Hall device has an advantage over the coil sensor because
there is no need to maintain a constant scanning speed during the inspection. On the other hand, Hall effect sensors are
more difficult to fabricate, are somewhat delicate compared to inductive coil sensors, and require more complex
electronics.

Other magnetic field sensors that are used less often in leakage field applications include the flux gate magnetometer (Ref
14), magnetoresistive sensors (Ref 15), magnetic resonance sensors (Ref 16), and magnetographic sensors (Ref 17), in
which the magnetic field at the surface of a part is registered on a magnetic tape pressed onto the surface.

Analysis of Leakage Field Data. In most applications of the leakage field method, there is a need not only to detect
the presence of a flaw but also to estimate its severity. This leads to the problem of flaw characterization, that is, the
determination of flaw dimensions from an analysis of leakage field data.

The most widely used method of flaw characterization is based on the assumptions that the leakage field signal amplitude
is proportional to the size of the flaw (which usually means its depth into the material) and that the signal amplitude can
therefore be taken as a direct measure of flaw severity. In situations where all flaws have approximately the same shape
and where calibration experiments show that the signal amplitude is indeed proportional to the size parameter of concern,
this empirical method of sizing works quite well (Ref 18).

There are, however, many situations of interest where flaw shapes vary considerably and where signal amplitude is not
uniquely related to flaw depth, as is the case for corrosion pits in steel tubing (Ref 19). In addition, different types of
flaws, such as cracks and pits, can occur in the same part, in which case it becomes necessary to determine the flaw types
present as well as their severity. In such cases, a more careful analysis of the relationship between signal and flaw
characteristics is required if serious errors in flaw characterization are to be avoided.
One of the earliest attempts to use a theoretical model in the analysis of leakage field data was based on the analytic
solution for the field perturbed by a spherical inclusion (Ref 20, 21). Two conclusions were drawn from this analysis.
First, when one measures the leakage flux component normal to the surface of the part, the center of the flaw is located
below the scan plane at a distance equal to the peak-to-peak separation distance in the flaw signal (Fig. 5), and second, the
peak-to-peak signal amplitude is proportional to the flaw volume. A number of experimental tests of these sizing rules
have confirmed the predicted relationships for nonmagnetic inclusions in steel parts (Ref 21).

Fig. 5 Dependence of magnetic signal peak separation (a) on the depth of a spherical inclusion (b)

Further theoretical and experimental data for spheroidal inclusions and surface pits have shown, however, that the simple
characterization rules for spherical inclusions do not apply when the flaw shape differs significantly from the ideal sphere.
In such cases, the signal amplitude depends on the lateral extent of the flaw and on its volume, and characterization on the
basis of leakage field analysis becomes much more complicated (Ref 19, 22).

Finally, there has been at least one attempt to apply finite-element calculations of flaw leakage fields to the development
of characterization rules for a more general class of flaws. Hwang and Lord (Ref 23) performed most of their
computations for simple flaw shapes, such as rectangular and triangular slots and inclusions, and from the results devised
a set of rules for estimating the depth, width, and angle of inclination of a flaw with respect to the surface of the part. One
of their applications to a flaw of complex shape is shown in Fig. 6.
Fig. 6 Characterization of a ferrite-tail type of defect. The dashed line shows the flaw configuration estimated
from the leakage field data.

The promising results obtained from the finite-element work of Hwang and Lord, as well as the analytically based work
on spheroidal flaws, suggest that the estimation of flaw size and shape from leakage field data is feasible. Another
numerical method potentially applicable to flux leakage problems is the boundary integral method, which may prove
useful in flaw characterization. Unfortunately, much more work must be done on both the theoretical basis and on
experimental testing before it will be possible to analyze experimental leakage field data with confidence in terms of flaw
characteristics.

References cited in this section

1. R.E. Beissner, G.A. Matzkanin, and C.M. Teller, "NDE Applications of Magnetic Leakage Field Methods,"
Report NTIAC-80-1, Southwest Research Institute, 1980
2. G. Dobmann, Magnetic Leakage Flux Techniques in NDT: A State of the Art Survey of the Capabilities for
Defect Detection and Sizing, in Electromagnetic Methods of NDT, W. Lord, Ed., Gordon and Breach, 1985
3. P. Höller and G. Dobmann, Physical Analysis Methods of Magnetic Flux Leakage, in Research Techniques
in NDT, Vol IV, R.S. Sharpe, Ed., Academic Press, 1980
4. F. Förster, Magnetic Findings in the Fields of Nondestructive Magnetic Leakage Field Inspection, NDT Int.,
Vol 19, 1986, p 3
5. F. Förster, Magnetic Leakage Field Method of Nondestructive Testing, Mater. Eval., Vol 43, 1985, p 1154
6. F. Förster, Magnetic Leakage Field Method of Nondestructive Testing (Part 2), Mater. Eval., Vol 43, 1985,
p 1398
7. F. Förster, Nondestructive Inspection by the Method of Magnetic Leakage Fields. Theoretical and
Experimental Foundations of the Detection of Surface Cracks of Finite and Infinite Depth, Sov. J. NDT, Vol
 18, 1982, p 841
8. F. Förster, Theoretical and Experimental Developments in Magnetic Stray Flux Techniques for Defect
Detection, Br. J. NDT, Nov 1975, p 168-171
9. J.R. Barton and F.N. Kusenberger, "Magnetic Perturbation Inspection to Improve Reliability of High
Strength Steel Components," Paper 69-DE-58, presented at a conference of the Design Engineering
Division of ASME, New York, American Society of Mechanical Engineers, May 1969
10. R.C. McMaster, Nondestructive Testing Handbook, Vol II, Section 30, The Ronald Press Company, 1959
11. H.J. Bezer, Magnetic Methods of Nondestructive Testing, Part 1, Br. J. NDT, Sept 1964, p 85-93; Part 2,
Dec 1964, p 109-122
12. F.W. Dunn, Magnetic Particle Inspection Fundamentals, Mater. Eval., Dec 1977, p 42-47
13. C.N. Owston, The Magnetic Leakage Field Technique of NDT, Br. J. NDT, Vol 16, 1974, p 162
14. F. Förster, Non-Destructive Inspection of Tubing and Round Billets by Means of Leakage Flux Probes, Br.
J. NDT, Jan 1977, p 26-32
15. A. Michio and T. Yamada, Silicon Magnetodiode, in Proceedings of the Second Conference on Solid State
Devices (Tokyo), 1970; Supplement to J. Jpn. Soc. Appl. Phys., Vol 40, 1971, p 93-98
16. B. Auld and C.M. Fortunko, "Flaw Detection With Ferromagnetic Resonance Probes," Paper presented at
the ARPA/AFML Review of Progress in Quantitative NDE, Scripps Institution of Oceanography, July 1978
17. F. Förster, Development in the Magnetography of Tubes and Tube Welds, Non-Destr. Test., Dec 1975, p
304-308
18. W. Stumm, Tube Testing by Electromagnetic NDE Methods-1, Non-Destr. Test., Oct 1974, p 251-256
19. R.E. Beissner, G.L. Burkhardt, M.D. Kilman, and R.K. Swanson, Magnetic Leakage Field Calculations for
Spheroidal Inclusions, in Proceedings of the Second National Seminar on Nondestructive Evaluation of
Ferromagnetic Materials, Dresser Atlas, 1986
20. G.P. Harnwell, Principles of Electricity and Magnetism, 2nd ed., McGraw-Hill, 1949
21. C.G. Gardner and F.N. Kusenberger, Quantitative Nondestructive Evaluation by the Magnetic Field
Perturbation Method, in Prevention of Structural Failure: The Role of Quantitative Nondestructive
Evaluation, T.D. Cooper, P.F. Packman, and B.G.W. Yee, Ed., No. 5 in the Materials/Metalworking
Technology Series, American Society for Metals, 1975
22. M.J. Sablik and R.E. Beissner, Theory of Magnetic Leakage Fields From Prolate and Oblate Spheroidal
Inclusions, J. Appl. Phys., Vol 53, 1982, p 8437
23. J.H. Hwang and W. Lord, Magnetic Leakage Field Signatures of Material Discontinuities, in Proceedings of
the Tenth Symposium on NDE (San Antonio, TX), Southwest Research Institute, 1975

Magnetic Field Testing

R.E. Beissner, Southwest Research Institute

Principles of Magnetic Characterization of Materials

Metallurgical and Magnetic Properties. The use of magnetic measurements to monitor the metallurgical properties
of ferromagnetic materials is based on the fact that variables such as crystallographic phase, chemical composition, and
microstructure, which determine the physical properties of materials, also affect their magnetic characteristics (Ref 24, 25,
26). Some parameters, such as grain size and orientation, dislocation density, and the existence of precipitates, are closely
related to measurable characteristics of magnetic hysteresis, that is, to the behavior of the flux density, B, induced in a
material as a function of the magnetic field strength, H.

This relationship can be understood in principle from the physical theory of magnetic domains (Ref 27). Magnetization in
a particular direction increases as the domains aligned in that direction grow at the expense of domains aligned in other