Eddy Current Inspection PDF

Metals Handbook Desk Edition, Second Edition Copyright © 1998 ASM International® J.R. Davis, Editor, p 1275-1281...

Metals Handbook Desk Edition, Second Edition Copyright © 1998 ASM International® J.R. Davis, Editor, p 1275-1281 All rights reserved. DOI: 10.31399/asm.hb.mhde2.a0003234 www.asminternational.org Eddy-Current Inspection / 1275 ~ Magneticfieldstrength(H), Oersted Magneticfieldstrength(H), Oersted 2!t ce -100 -50 0 50 100 -100 -50 0 50 100 Hardness~., ~ 39. H I- 1. H2~rdFnl~s J ~ ===~=~ 10 ~ ~-. 1.0 10 & ,z ':', )) II 8'~ ~ ~H:a2 r ~- -1,0 ~nF~ 010 :~ x 1 -10 (a) A -2.0/ 20 2 0_8 - -20 -8 -4 0 4 8 -8 W Magneticfieldstrength(H), A/m x 103 Magneticfieldstrength(H), A/mx 103 Flux V (a) (b) Fig. 6 Effect of mechanical hardnesson hysteresis loop data. (a) AISI 410 stainlesssteel. (b) SAE 4340 steel (b) mentally indexing the sensor across the raceway. characteristics provide data that yield information Signal ~j Magnetizing fields are applied in the radial and on different material properties. For example, dif- circumferential orientations. Radial field inspec- ferent features of magnetic hysteresis data can tion works best for surface flaws, while circum- be interpreted in terms of heat treatment and mi- ferential field inspection shows greater sensitivity crostructure, plastic deformation, residual stress, i i i l i l i ~ Probe and mechanical hardness. to subsurface flaws. The flux leakage method also is used to inspect An example of the effects of mechanical hard- steel reinforcement in concrete beams. The basic ness on hysteresis data is shown in Fig. 6. These function of the magnetic field disturbance (MFD) data were obtained in the absence of applied ten- inspection equipment is to provide maps of the sile stress. Specimens of different hardness were magnetic field across the bottom and sides of the prepared by tempering at different temperatures. beam. An electromagnet on an inspection cart, The grain size (ASTM No. 7) was the same for (c) which is suspended on tracks below the beam, all four specimens used in the tests. Other data provides a magnetic field that induces magnetiza- show that grain size has little effect on hysteretic Fig. 5 Flux leakage inspection of a bearing race. (a) tion in permeable structures in its vicinity, such behavior for the classes of alloys studied. Magnetization of inner race. (b) Perturbation in The main point illustrated in Fig. 6 is that the the magnetic flux at the surfaceof the inner race. (c) Probe as steel rebars, cables, and stirrups. An array of mechanically harder specimens of the same alloy scanning the surface Hall effect sensors distributed across the bottom are also harder to magnetize; that is, the flux and sides of the beam measures the field pro- density, B, obtained at a large value of H is In addition to systems for inspecting rotation- duced by magnetized structures within the beam. smaller for mechanically harder specimens than ally symmetric cylindrical parts, flux leakage in- If a flaw is present in one of these magnetized for softer specimens. For one alloy, AISI 410 spection is applied to irregular components, such structures, it produces a disturbance of the nor- stainless steel, the hysteresis loop intersects the as gear teeth and artillery projectiles. mal magnetic field pattern associated with the B = 0 axis at larger values of H for the harder The flux leakage method also is used to inspect unflawed beam. specimen than for the softer specimen; that is, the ropes and cables made of strands of ferromag- A flaw, such as a broken wire in a cable or coercive force is greater for the harder material. netic material. One approach is to induce magnet- a fractured rebar, produces a distinctive magnetic However, for SAE 4340 steel, the coercive force ization in the piece by means of an encircling field anomaly that depends on the size of the does not change with hardness. This suggests coil energized by a direct current. With this discontinuity and its distance from the sensor. that, for these two alloys, the saturation flux den- method, the leakage field associated with broken Because the signal shape that results from such sity provides a more reliable measure of hardness strands is measured using a Hall effect probe Or an anomaly is known, flaw detection is enhanced than the coercive force. an auxilliary sensor coil. by searching magnetic field records for specific Certain features of both Barkhausen noise and Another area that uses the flux leakage method signal shapes; that is, those that are characteristic incremental permeability data can be correlated is inspection of rolling-element antifriction bear- of discontinuities in magnetic materials. In the with depth of decarburization. The Barkhausen ings. A schematic illustration of the method as MFD system, this is accomplished by a computer noise method shows a somewhat stronger sensi- applied to an inner bearing race is shown in Fig. program that compares signal shapes with typical tivity to depth, but is useful over a smaller range 5. In this application, the part is magnetized by flaw signal shapes. of depths than the incremental permeability an electromagnet, as indicated in'Fig. 5(a). The Nondestructive Characterization of Materi- method. It can be concluded that both methods race is rotated by a spindle, and the surface is als. Two examples of magnetic methods for mon- are useful. The optimum choice of methods de- scanned using an induction coil sensor. The ac- itoring material properties illustrate the types of pends on the accuracy required and the expected tive portion of the raceway is inspected by incre- tests that can be used. Measurements of magnetic depth of decarburization. Eddy-Current Inspection E D D Y - C U R R E N T INSPECTION is based on nonferromagnetic metals and metal parts. Eddy- netic permeability, and physical dimensions the principles of electromagnetic induction and current inspection is used: (primary factors affecting eddy-current re- is used to identify or differentiate a wide variety sponse) of physical, structural, and metallurgical condi- To measure and identify conditions and prop- To detect seams, laps, cracks, voids, and inclu- tions in electrically conductive ferromagnetic and erties related to electrical conductivity, mag- sions Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/523731/a0003234.pdf by University of British Columbia user  1276 / Nondestructive Testing To sort dissimilar metals and detect differences Tu oLe Inspection c o i i --~..,_. ~----.-..~ Each and all of these changes can have an in their composition, microstructure, and other effect on the exciting coil and other coil or coils properties (such as grain size, heat treatment, used to sense the electromagnetic field adjacent and hardness) to a part. The effects most often used to monitor To measure the thickness of a nonconductive the condition of the part being inspected are the coating on a conductive metal, or the thickness electrical impedance of the coil and the induced of a nonmagnetic metal coating on a magnetic (a) voltage of either the exciting coil or other adja- metal cent coil or coils. Eddy-current systems vary in complexity de- Because eddy-current inspection is an electro- pending on individual inspection requirements. magnetic-induction technique, it does not require However, most systems must provide for the fol- direct electrical contact with the part being in- lowing functions: spected. The eddy-current method is adaptable to high-speed inspection, and because it is nondes- Excitation of the inspection coil with one or tructive, it can be used to inspect an entire pro- more frequencies duction output if desired. The method is based Modulation of the inspection-coil output signal on indirect measurement, and the correlation be- by the part being inspected tween instrument readings and the structural Processing of the inspection-coil signal prior characteristics and serviceability of parts being to amplification inspected must be carefully and repeatedly estab- Amplification of the inspection-coil signals lished. (b) Detection or demodulation of the inspection- Eddy-current inspection is extremely versatile, coil signal, usually accompanied by some anal- which is both an advantage and a disadvantage. Fig. 1 Two common types of inspection coils and the ysis or discrimination of signals, which can be The advantage is that the method can be applied patterns of eddy-current flow.generated by the performed by a computer to many inspection problems provided that the exciting current in the coils. Solenoid-type coil is applied Display of signals on an instrument such as a physical requirements of the material are compat- to cylindrical or tubular parts; pancake-type coil, to a flat meter, an oscilloscope, an oscillograph, and a ible with the inspection method. However, in surface. (a) Solenoid-type coil. (b) Pancake-type coil strip-chart recorder; or recording of signals on many applications, the sensitivity of the method paper punch tape and magnetic tape to m a n y inherent material properties and charac- Handling of the part being inspected and sup- teristics can be a disadvantage. Some variables currents in the part depends on the electrical port of inspection-coil assembly in a material that are not important in terms of characteristics of the part, the presence or ab- material or part serviceability can cause instru- sence of flaws and other discontinuities in the Elements of a typical inspection system are ment signals that mask critical variables or are part, and the total electromagnetic field within shown schematically in Fig. 3. The particular ele- mistakenly interpreted to be caused by critical the part. ments in Fig. 3 are for a system developed to variables. The change in flow of eddy currents caused inspect bar or tubing. The generator supplies ex- Eddy-Current vs. Magnetic Inspection Meth- by the presence of a crack in a pipe is shown citation current to the inspection coil and a syn- ods. In eddy-current inspection, eddy currents in Fig. 2. The pipe travels along the length of chronizing signal to the phase shifter, which pro- create their own electromagnetic field, which is the inspection coil, as shown. In section A - A in vides switching signals for the detector. The sensed either through the effects of the field on Fig. 2, no crack is present and the eddy-current loading of the inspection coil by the part being the primary exciting coil or by means of an inde- flow is symmetrical. In section B-B, where a inspected modulates the electromagnetic field of pendent sensor. In nonferromagnetic materials, crack is present, the eddy-current flow is the coil. This causes changes in the amplitude the secondary electromagnetic field is derived ex- impeded and changed in direction, causing signif- and phase of the inspection-coil voltage out- clusively from eddy currents. However, with fer- icant changes in the associated electromagnetic put. romagnetic materials, additional magnetic effects field. The condition of the part can be monitored The output of the inspection coil is fed to the occur that usually are of sufficient magnitude to by observing the effect of the resulting field on amplifier and detected or demodulated by the de- overshadow the basic eddy-current effects from the electrical characteristics of the exciting coil, tector. The demodulated output signal, after some electrical conductivity only. These magnetic ef- such as its electrical impedance, induced voltage, further filtering and analyzing, is then displayed fects result from the magnetic permeability of and induced currents. Alternatively, the effect of on an oscilloscope or a chart recorder. The dis- the material being inspected, and can be virtually the electromagnetic field can be monitored by played signals, having been detected or demodu- eliminated by magnetizing the material to satura- observing the induced voltage in one or more lated, vary at a much slower rate, depending on tion in a static (direct-current) magnetic field. other coils placed within the field near the part (a) the rate of changing the inspection probe from W h e n the permeability effect is not eliminated, being monitored. one part being inspected to another, (b) the speed the inspection method is more correctly catego- at which the part is fed through an inspection rized as electromagnetic or magnetoinductive in- coil, or (c) the speed with which the inspection spection. Inspection coil--~ B ~,,z~;>,~ ~ coil is caused to scan past the part being in- spected. Principles of Operation Functions of a Basic System. The part to be inspected is placed within or adjacent to an elec- or to ,no- II1 r v V trical coil in which an alternating current is flow- ing. As shown in Fig. 1, the alternating current, Oscilloscope..~ _ called the exciting current, causes eddy currents ~-Pipe to flow in the part as a result of electromagnetic Inspection c o ~ Inspection co,l.~ induction. These currents flow within closed loops in the part, and their magnitude and timing (or phase) depend on (a) the original or primary field established by the exciting currents, (b) the electrical properties of the part, and (c) the elec- tromagnetic fields established by currents flowing within the part. The electromagnetic field in the region in the part and surrounding the part depends on both Section A-A Section B-B Fig. 3 Principal elements of a typical system for eddy the exciting current from the coil and the eddy Fig. 2 Effect of a crack on the pattern of eddy-current current inspection of bar or tubing. See descrip- currents flowing in the part. The flow of eddy flow in a pipe tion in text. Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/523731/a0003234.pdf by University of British Columbia user  Eddy-Current Inspection / 1277 20 , I I # C/nciuctivity _\ ~ Conductivity saurQon "~'~- \ iO0*/,, I0 I / ~ Soturation 0% ~ lair) %' ~a'~ 2 %-.~_\ IV'",, 0 20 40 60 1 80 Moqnetizinqforce (H), oersteds tO0 t20 Lift-off J Fig. 5 Magnetization curves for annealed commercially Coil resistance pure iron and nickel Fig. 6 Impedance-plane diagram showing curves for electrical conductivity and lift off. Inspection fre- quency is 100 kHz. 2, Electrical Conductivity All materials have a characteristic resistance to ceptibilities, and have very high and variable per- the flow of electricity. Those with the highest 100%, IACS-~ resistivity are classified as insulators; those hav- meabilities. Magnetic permeability is not a constant for a ing intermediate resistivity are classified as semi- given material, but depends on the strength of conductors; and those having low resistivity are the magnetic field acting on it. For example, con- Coil resistance ,, classified as conductors. Conductors, which in- sider a sample of steel that has been completely clude most metals, are of greatest interest in Fig. 4 Typical impedance-plane diagram derived by eddy-current inspection. The relative conductivi- demagnetized and then placed in a solenoid coil. placing an inspection coil sequentially on a se- As current in the coil is increased, the magnetic ties of common metals and alloys vary over a ries of thick pieces of metal, each with a different Interna- field associated with the current increases How- tional Annealed Copper Standard (IACS) electrical resis- wide range. ever, the magnetic flux within the steel increases tance or conductivity rating. The inspection frequency is Capacity to conduct current is measured in rapidly at first and then levels off so that an 100 kHz. terms of either conductivity or resistivity. In additionally large increase in the strength of the eddy-current inspection, measurement often is magnetic field results in only a small increase in based on IACS. In this system, the conductivity flux within the steel. The steel sample achieves of annealed, unalloyed copper is arbitrarily rated a condition known as magnetic saturation. Operating Variables at 100%, and the conductivities of other metals The curve showing the relation between mag- The principal operating variables encountered and alloys are expressed as percentages of this netic-field intensity and the magnetic flux within in eddy-current inspection include coil imped- standard. Thus, the conductivity of unalloyed alu- the steel is known as a magnetization curve ance, electrical conductivity, magnetic permeabil- minum is rated 61% IACS, or 61% that of unal- Magnetization curves for annealed commercially ity, lift-off and fill factors, edge effect, and skin loyed Copper. Table 1 gives the resistivities and pure iron and nickel are shown in Fig. 5. The effect. IACS conductivity ratings of several common magnetic permeability of a material is the ratio metals and alloys. between the strength of the magnetic field and the amount of magnetic flux within the material. Coil Impedance Magnetic Permeability As shown in Fig. 5, at saturation (where there is no appreciable change in induced flux in the When direct current flows in a coil, the mag- Ferromagnetic metals and alloys, including material for a change in field strength) the per- netic field reaches a constant level and the elec- iron, nickel, cobalt, and some of their alloys, meability is nearly constant for small changes in trical resistance of the wire is the only limitation concentrate the flux of a magnetic field They field strength. to the flow of current. However, when alternating are strongly attracted to a magnet and an electro- Magnetic permeability of the material being in- current flows in a coil, two limitations are im- magnet, have exceedingly high and variable sus- spected strongly influences the eddy-current re- posed: the alternating-current resistance of the sponse. Consequently, the techniques and condi- wire and a quantity known as inductive reactance tions used for inspecting magnetic materials (XL). Table 1 Electrical resistivity and differ from those used to inspect nonmagnetic Impedance usually is plotted on an impedance- conductivity of several common metals materials. plane diagram. In such a diagram, resistance is and alloys plotted along one axis and inductive reactance Resistivity, Conductivity, (or inductance) along the other axis. Because Metal or alloy I t f l , mm %IACS "Lift-Off" Factor each specific condition in the material being in- When a probe inspection coil, attached to a spected can result in a specific coil impedance, Silver 16.3 105 Copper, annealed 17.2 100 suitable inspection instrument, is energized in air, each condition corresponds to a particular point Gold 24.4 , 70 it produces an indication even if there is no con- on the impedance-plane diagram. For example, if Aluminum 28.2 61 ductive material in the vicinity of the coil. The a coil is placed sequentially on a series of thick Aluminum alloys initial indication starts to change as the coil is pieces of metal, each having a different resistiv- 606t-T6 41 42 moved closer to a conductor. Because the field ity, each piece causes a different coil impedance 7075-T6 53 32 of the coil is strongest close to the coil, the indi 5 and corresponds to a different point on a locus in 2024-T4 52 30 cated change on the instrument continues to in- the impedance plane. The curve generated might Magnesium 46 37 crease until the coil is directly on the conductor. resemble that shown in Fig. 4, which is based on 70-30 brass 62 28 Phosphor bronzes 160, l1 These changes in indication with changes in International Annealed Copper Standard (IACS) spacing between the coil and the conductor, or Monel 482 3.6 conductivity ratings. Other curves are generated Zirconium 500 3.4 part being inspected, are called "'lift off." The for other material variables, such as section Zircaloy-2 720 2.4 lift-off effect is so pronounced that small varia- thickness and types of surface flaws. Titanium 548 3.1 tions in spacing can mask many indications re- By use of more than one test frequency, the Ti-6AI-4V alloy 1720 1.0 sulting from the condition or conditions of pri- impedance planes can be manipulated to accept Type 304 stainless steel 700 2.5 mary interest. Consequently, it usually is a desirable variable (in flaws) and reduce the Inconel 600 980 1.7 necessary to maintain a constant relationship be- effects of undesirable variables--that is, lift-off Hastelloy X 1150 1.5 tween the size and shape of the coil and the size Waspaloy 1230 1.4 and/or dimensional effects (see Fig. 3). and shape of the part being inspected. Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/523731/a0003234.pdf by University of British Columbia user  1278 / Nondestructive Testing The change of coil impedance with lift-off can of coils for given applications. In general, it is Inspection Frequencies be derived from the impedance-plane diagram not advisable to inspect any closer than 3.2 mm shown in Fig. 6. When the coil is suspended in The inspection frequencies used in eddy-cur- (~/8 in.) from the edge of a part. air away from the conductor, impedance is at a rent inspection range from about 60 Hz to 6 One alternative for inspection near an edge point at the upper end of the curve at far left in MHz. Most inspection of nonmagnetic materials with minimal edge effect is to scan in a line Fig. 6. As the coil approaches the conductor, the is performed at a few kilohertz. In general, lower parallel to the edge. Inspection can be carried impedance moves in the direction indicated by frequencies are used to inspect magnetic materi- out by maintaining a constant probe-to-edge rela- the dashed lines until the coil is in contact with als. However, the actual frequency used in any tionship, but each new scan-line position requires the conductor. When contact occurs, the imped- specific eddy-current inspection depends on the adjustment of the instrument. Fixturing of the ance is at a point corresponding to the impedance thickness of the material being inspected, the re- probe is recommended. of the part being inspected, which in this in- quired depth of penetration, the degree of sensi- stance, represents its conductivity. The fact that tivity or resolution required, and the purpose of Skin Effect the inspection. the lift-off curves approach the conductivity curve at an angle can be used in some instru- Eddy currents are not uniformly distributed Selection of inspection frequency normally is ments to separate lift-off signals from those re- throughout a part being inspected; rather, they a compromise. For example, penetration should are densest at the surface immediately beneath be sufficient to reach subsurface flaws that must sulting from variations in conductivity or some other parameter of interest. the coil and become progressively less dense with be detected, and to determine material condition Although lift off can be troublesome in many increasing distance below the surface. The con- (such as case hardness). Although penetration is centration of eddy currents at the surface of a greater at lower frequencies, it does not follow applications, it can be also be useful. For exam- ple, using the lift-off effect, eddy current instru- part is known as "skin effect." At some distance that the lowest possible frequency should be below the surface of a thick part, there essen- used. Unfortunately, as the frequency is lowered, ments are excellent for measuring the thickness of nonconductive coatings, such as paint and an- tially are no currents flowing. The depth of eddy- the sensitivity to flaws decreases somewhat and odized coatings, on metals. current penetration should be considered for the speed of inspection could be reduced. thickness measurements and for detection of sub- Typically, the highest possible inspection fre- surface flaws. quency that still is compatible with the penetra- Fill Factor Figure 7 shows how the eddy current varies as tion depth required is selected. The choice is rel- In an encircling coil, a condition comparable a function of depth below the surface. The depth atively simple when only surface flaws must be to lift-off is known as "fill factor." It is a mea- at which the density of the eddy current is re- detected, in which case frequencies up to several sure of how well the part being inspected fills duced to about 37% of the density at the surface megahertz can be used. However, when flaws at the coil. As with lift off, changes in fill factor is defined as the standard depth of penetration. some considerable depth below the surface must resulting from factors such as variations in out- This depth depends on the electrical conductivity be detected, or when flaw depth and size must side diameter must be controlled because small and magnetic permeability of the material and on be determined, low frequencies must be used at changes can produce large indications. The lift- the frequency of the magnetizing current. Depth the expense of sensitivity. off curves shown in Fig. 6 are very similar to of penetration decreases with increases in con- In inspection of ferromagnetic materials, rela- those for changes in fill factor. For a given lift- ductivity, permeability, and inspection frequency. tively low frequencies typically are used because off or fill factor, the conductivity curve shifts to The standard depth of penetration can be calcu- of the low penetration in these materials. Higher a new position, as indicated in Fig. 6. Fill factor lated from the equation: frequencies can be used when it is necessary to can sometimes be used as a rapid method to inspect for surface conditions only. However, check variations in outside-diameter measure- S = 1980 ~/~ even the higher frequencies used in these applica- ments in rods and bars. tions still are considerably lower than those used where S is standard depth of penetration, in to inspect nonmagnetic materials for similar con- Edge Effect inches; p is resistivity, in ohm-centimeters; ix is ditions. magnetic permeability (1 for nonmagnetic materi- When an inspection coil approaches the end or als); and f is inspection frequency, in hertz. Fig- edge of a part being inspected, eddy currents are ure 8 shows the standard depth of penetration, Inspection Coils distorted because they are unable to flow beyond as a function of inspection frequency, for several The inspection coil is an essential part of every the edge of a part. The distortion of eddy currents metals of various electrical conductivities. eddy-current inspection system. The shape of the results in an indication known as "edge effect." Because the magnitude of the effect is very large, it limits inspection near edges. Unlike lift-off, '° F-..l > '>,L'~ little can be done to eliminate edge effect. A reduction in coil size lowers the effect somewhat, ~["~ ; I FGra~ hire [ but there are practical limits that dictate the sizes I,i- [rT,t0ni0m r-,-zJ 1.20. I lil Stoin,essstee, "'~1 li I 'b-L~Z~I r'~l /I I i li J [ 1,00, 0.80 /-Stondard depth of penetro!ion \/ /where density of eddy current = "6 0.60 3 7 % of density at surface 0.01 _ _ Ingot iron - / ' I -'-.t-'~l l I l'~-~l-'~'k] !I ~.~k.. I 7"" \ £3 I[[] I I"~"~'~.1 r ill h-z-.J ~ J,/I [l"N~"ff,~i ill 0.40 o.oo I l l[tl til'l 0.20 I I II/ ~'-~-~L.Jl ~ I' '~'~ I Jill '~' ~ I IfI Copper-/ 1" ' ~ E ~ I Ill ~ r"N o ~"'~ ~---- 0 I 2 5 Deplhbelowsurfoee 4 5 6 o.ooo I liJ I Stondorddepthof ~enetration I0 I0 z I0 3 10 4 10 5 i0 6 IO T I0 e Frequency, Hz Fig. 7 Variation in density of eddy current as a function of depth below the surface of a conductor, Fig. 8 Standard depths of penetration as a function of frequencies used in eddy-current inspection for several metals known as skin effect of various electrical conductivities Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/523731/a0003234.pdf by University of British Columbia user  Eddy-Current Inspection / 1279 iruOCk\ circling coil is used to inspect tubing and bar for short discontinuities, best resolution is obtained with a short coil. On the other hand, a short coil has the disadvantage of being sensitive to the position of the part in the coil. Longer coils are not as sensitive to position of the part, but are not as effective in detecting very small disconti- nuities. Small-diameter probe coils have greater resolution than larger ones but are more difficult to manipulate and are more sensitive to lift-off variations..... ZE d d y - c u r r e n t flow (a) (b) (c) (d) Eddy-Current Instruments Fig. 9 Types and applications of coils used in eddy-current inspection. (a) Probe-type coil applied to a flat plate for A simple eddy-current instrument, in which the crack detection. (b) Horseshoe-shape, or U-shape, coil applied to a flat plate for laminar-flaw detection. /c) voltage across an inspection coil is monitored, is Encircling coil applied to a tube. (d) Internal, or bobbin-type, coil applied to a tube shown in Fig. ll(a). This circuit is adequate to measure large lift-off variations, if accuracy is not of great importance. A circuit designed for inspection coil depends to a considerable extent type commonly are used in sort applications. Fix- greater accuracy is shown in Fig. ll(b). This in- on the purpose of the inspection and on the shape tures are used to maintain a constant geometrical strument consists of a signal source, an imped- of the part being inspected. In inspection for relationship between coil and part. ance bridge with dropping resistors, an inspection flaws, such as cracks and seams, it is essential An absolute coil arrangement is not a good coil in one leg, and a balancing impedance in that the flow of the eddy currents be as nearly method in many applications. For example to in- the other leg. The differences in voltage between perpendicular to the flaws as possible to obtain spect tubing, an absolute arrangement indicates the two legs of the bridge are measured by. an a maximum response from the flaws. If the eddy- dimensional variations in both outside diameter alternating-current voltmeter. Alternatively, the current flow is parallel to flaws, there is little or and wall thickness even though such variations balancing impedance in the leg opposite the in- no distortion of the currents, and, therefore, very can be well within allowable limits. To avoid spection coil can be a coil identical to the inspec- little reaction on the inspection coil. this problem, a differential coil arrangement such tion coil, as shown in Fig. 11(c), or it can have Probe and Encircling Coils. Of the almost in- as that shown in Fig. 10(b) can be used. Here, a reference sample in the coil, as shown in Fig. finite variety of coils used in eddy-current in- the two coils compare one section of the tube 11(d). In the latter, if all the other components spection, probe coils and encircling coils are the with an adjacent section. When the two sections in the bridge are identical, a signal occurs only most common. A probe-type coil typically is used are the same, there is no output from the pair when the inspection-coil impedance deviates to inspect a flat surface for cracks at an angle of coils and no indication on the eddy-current from that of the reference sample. to the surface because this type of coil induces instrument. Gradual dimensional variations There are other methods to achieve bridge bal- currents that flow parallel to the surface, and within the tube or gross variations between indi- ance, such as varying the values of resistance of therefore across a crack as shown in Fig. 9(a). vidual tubes are not indicated, whereas disconti- the resistor in the upper leg of the bridge and Conversely, a probe-type coil is not suitable to n u i t i e s - w h i c h normally occur abruptly--are one in series with the balancing impedance. The detect a laminar type of flaw. For such a discon- very apparent. In this way, it is possible to have most accurate bridges can measure absolute tinuity, a U-shape, or horseshoe-shaped coil such an inspection system that is sensitive to flaws impedance to within 0.01%. However, in eddy- as the coil shown in Fig. 9(b) is satisfactory. and relatively insensitive to changes that nor- current inspection, it is not how an impedance To inspect tubing and bar, an encircling coil mally are not of interest. bridge is balanced that is important, but rather (Fig. 9c) generally is used because of comple- Sizes and Shapes. Inspection coils are made how it is unbalanced by the effects of a flaw. mentary configuration and because of the testing in a variety of sizes and shapes. Selection of a Another type of bridge system is an induction speeds that can be achieved. However, an encir- coil for a particular application depends on the bridge, in which the power signal is transformer cling coil is sensitive only to discontinuities that type of discontinuity. For example, when an en- coupled into an inspection coil and a reference are parallel to the axis of the tube and bar. The coil. In addition, the entire inductance-balance coil is satisfactory for this particular application system is placed in the probe, as shown in Fig. because most discontinuities in tubing and bar /-Reference I J 12. The probe consists of a large transmitter (or are parallel to the major axis as a result of the driver) coil and two small detector (or pickup) manufacturing process. If it is necessary to locate coils wound in opposite directions as mirror im- discontinuities that are not parallel to the axis, a ages of each other. An alternating current is sup- probe coil must be used, and either the coil or the plied to the large transmitter coil to generate a part must be rotated during scanning. To detect magnetic field. If the transmitter coil is not in discontinuities on the inside surface of a tube, the vicinity of a conductor, the two detector coils an internal, or bobbin-type, coil (Fig. 9d) can be detect the same field, and the net signal is zero used. An alternative is to use an encircling coil because they are wound in opposition to each with a depth of penetration sufficient to detect other. However, if one end of the probe is placed flaws on the inside surface. The bobbin-type coil, near a metal surface, the field is different at the similar to the encircling coil, is sensitive to dis- (a) two ends of the probe, and a net voltage appears continuities that are parallel to the axis of the across the two coils. The resultant field is the tube or bar. Reference sum of a transmitted signal, which is present all Multiple Coils. In many eddy-current inspec- coil the time, and a reflected signal due to the pres- tion setups, two coils are used. The two coils ence of a conductor (the metal surface). This coil typically are connected in a series-opposing ar- arrangement can be used both as a probe and as rangement so there is no output from the pair an encircling coil (see Fig. 12). when their impedances are the same. Pairs of Readout Instrumentation. An important part coils can be used in either an absolute or a differ- of an eddy-current inspection system is the in- ential arrangement (see Fig. 10). In the absolute coil strument used for a readout. The readout device arrangement (Fig. 10a), a sample of acceptable can be an integral part of the system, an inter- (b) changeable plug-in module, and a solitary unit material is placed in one coil, and the other coil is used for inspection. In this manner, the coils Fig. 10 Absolute and differentia[ arrangements of mul- connected by cable. The readout instrument compare an unknown against a standard, the dif-. tiple coils used in eddy-current inspection. (a) should be of adequate speed, accuracy, and range ferences between the two (if any) are indicated Absolute coil arrangement. (b) Differential coil arrange- to meet the inspection requirements of the sys- by a suitable instrument. Arrangements of this ment tem. Frequently, several readout devices are used Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/523731/a0003234.pdf by University of British Columbia user  1280 / Nondestructive Testing Resisfor ! R R R R R R R lion R R R Inspectiotl Test Test coil Test sTmeptle ~ ,mB°p~a~" c'~e sample~........ Icing sampl sample _ \./"I" =='::' _ ~ Reference Ground Ground Ground Ground Ground Ground Ground Ground Grounq sample (a) (b) (c) (d) Fig. 11 Four types of eddy-current instruments. (a) A simple arrangement, in which voltage across the coil is monitored. (b) Typical impedance bridge. (c) Impedance bridge with dual coils. (d) Impedance bridge with dual coils and a reference sample in the second coil ~] Detector Analog meters give a continuous reading over catalog the data, print summaries of the result, (pick-up) an extended range. They are fairly rapid (with and store all data on tape for reference in fu- ~ coils a frequency of about 1 Hz), and the scales can ture scans. II I be calibrated to read parameters directly. The It I accuracy of these devices is limited to about 1% of full scale. They can be used to set the limits on alarm lights, sound alarms, and kick- Discontinuities Detectable out relays. by Eddy-Current Inspection Digital meters are easier to read and can have B a s i c a l l y , any discontinuity that appreciably greater ranges than analog meters. Numerical alters the normal flow of eddy currents can be Transmitter values are easily read without extrapolation, detected by eddy-current inspection. With encir- (driver) coil but fast trends of changing readings are more cling-coil inspection of either solid cylinders or difficult to interpret. Although many digital tubes, surface discontinuities having a combina- Arrangement of probe and workpiece meters have binary coded decimal (bcd) out- tion of predominantly longitudinal and radial di- put, they are relatively slow. mensiorial components are readily detected. X-Y plotters can be used to display imped- When discontinuities of the same size are located ance-plane plots of the eddy-current response. beneath the surface of the part being inspected Transmitter Detector(plck They are very helpful in the design and set up at progressively greater depths, they become in- (driver) up)coil(1 of 2) coil of eddy-current, bridge-unbalance inspections creasingly difficult to detect, and can be detected and in discriminating against undesirable vari- at depths greater than 13 mm (1/2 in.) only with ables. They also are useful to sort out inspec- special equipment designed for this purpose. _x_ ~ tion results. They are fairly accurate and pro- Conversely, laminar discontinuities such as Ground Workpiece (conductor) Ground vide a permanent copy. those in welded tubes might not alter the flow X-Y storage oscilloscopes are very similar to of the eddy currents enough to be detected unless Wiring schematic for probeand workpiece X-Y plotters but can acquire signals at high the discontinuity breaks either the outside or in- speed. However, the signals have to be pro- side surfaces, or unless it produces a discontinu- Fig. 12 Induction-bridge probe in place at the surface cessed manually, and the screen can quickly ity in the weld from upturned fibers caused by of a workpiece. Schematic shows how power become cluttered with signals. In some instru- extrusion during welding. A similar difficulty signal is transformer coupled from a transmitter coil into two detector coils--an inspection coil (at bottom) and a ments, high-speed X-Y gates can be displayed could arise in trying to detect a thin planar dis- reference coil (at top). and set on the screen. continuity that is oriented substantially perpen- Strip-chart recorders furnish a fairly accurate dicular to the axis of the cylinder. (about 1% of full scale) recording at reason- Regardless of the limitations, a majority of ob- in a single inspection system. A list of more ably high speed (about 200 Hz). However, jectionable discontinuities can be detected by common types of readout, in order of increasing once on the chart, the data must be read by eddy-current inspection at high speed and at low cost and complexity, follows: an operator. Several channels can be recorded cost. Some of the discontinuities that are readily simultaneously, and the record is permanent. detected are seams, laps, cracks, slivers, scabs, Alarm lights alert the operator that a test-pa- Magnetic-tape recorders are fairly accurate and pits, slugs, open welds, miswelds, misaligned rameter limit has been exceeded. capable of recording at very high speed (10 welds, black and gray oxide weld penetrators, Sound alarms serve the same purpose as alarm MHz). Moreover, the data can be processed by pinholes, hook cracks, and surface cracks. lights but free the attention of the operator to automated techniques. R e f e r e n c e Samples. A basic requirement for allow manipulating the probe in manual scan- Computers. The data from several channels can eddy-current inspection is a reliable, consistent ning. be fed directly to a high-speed computer, means to set tester sensitivity to the proper level Kick-out relays activate a mechanism that au- either analog or digital, for on-line processing. each time it is used. A standard reference sample tomatically rejects and marks a part when a The computer can separate parameters and cal- must be provided for this purpose. Without this test parameter is exceeded. culate the variable of interest and significance, capability, eddy-current inspection is of little @ @ illed or @ illed or @.=d :al discharge al discharge ~e notch )ngitudinol notch transverse notch I Fig. 13 Several fabricated discontinuities used as reference standards in eddy-current inspection. ASTM standards for eddy-current testing include E 215 (aluminum alloy tube), E 376 (measurement of coating thickness), E 243 (copper and copper alloy tube), E 566 (ferrous metal sorting), E 571 (nickel and nickel alloy tube), E 690 (nonmagnetic heat-exchanger tubes), E 426 (stainless steel tube), and E 309 (steel tube). Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/523731/a0003234.pdf by University of British Columbia user  Microwave Inspection / 1281 value. In selecting a standard reference sample, The type of reference discontinuities that must and (e) they should produce an indication on the the usual procedure is to select a sample of prod- be used for a particular application are specified eddy-current tester that closely resembles those uct that can be run through the inspection system (for example, by ASTM and API). In selecting produced by the natural discontinuities. without producing appreciable indications from reference discontinuities, some of the major con- Figure 13 shows several discontinuities that the tester. Several samples might have to be run siderations are: (a) they must meet the required have been used for reference standards, these in- before a suitable one is found; the suitable one specification, (b) they should be easy to fabri- clude a filed transverse notch, milled or electrical then has reference discontinuities fabricated into cate, (c) they should be reproducible, (d) they discharge machined longitudinal and transverse it. should be producible in precisely graduated sizes, notches, and drilled holes. Microwave Inspection MICROWAVES (or radar waves) are a form The complete microwave system can be made Fixed-frequency standing waves of electromagnetic radiation with wavelengths from solid-state components so that it will be Fixed-frequency reflection scattering between 1000 cm and 1 mm in free space. Be- small, rugged, and reliable. Microwave holography cause microwaves have wavelengths that are 104 Microwaves can be used for locating and siz- Microwave surface impedance to 105 times longer than those of light waves, ing cracks in materials if the following consid- Microwave detection of stress corrosion microwaves penetrate deeply into materials, with erations are followed. First, the skin depth at the depth of penetration dependent on the con- microwave frequencies is very small (a few Each of these techniques uses one or more of ductivity, permittivity, and permeability of the micrometers), and the crack is detected most the several processes by which materials can in- materials. Microwaves are also reflected from sensitively when the crack breaks through the teract with microwaves, namely, reflection, re- any internal boundaries and interact with the surface. Second, when the crack is not through fraction, scattering, absorption, and dispersion. the surface, the position of the crack is indi- The basic components of the transmission tech- molecules that constitute the material. For exam- ple, it was found that the best source for the cated by a detection of the high stresses in the nique are shown schematically in Fig. 1. thickness and voids in radomes was the micro- surface right about the subsurface crack. Fi- Thickness measurements can be made with waves generated within the radomes. Both con- nally, microwave crack detection is very sensi- microwave techniques on both metallic and non- tinuous and pulsed incident waves were used in tive to crack opening and to the frequency metallic materials. For metals, two reflected used. Higher frequencies are needed for the waves are used from two waveguide arms that these tests, and either reflected or transmitted smaller cracks. If the frequency is increased differ in length b y an integral number of half- waves were measured. sufficiently, the incident wave can propagate wavelengths for detector output null. This mea- One of the first important uses of microwaves in nondestructive evaluation (NDE) was for com- into the crack, and the response is then sensi- surement is made using the standing wave tech- tive to crack depth. nique. When the wave is incident on a metal ponents such as waveguides, attenuators, cavitie s, antennas, and antenna covers (radomes). Subse- (electrically conductive), most of the wave is re- quently, microwave inspection methods were de- flected; only a small amount is transmitted (re- Limitations. The use of microwaves is in some veloped for: fracted). The transmitted wave is highly attenu- cases limited by their inability to penetrate ated in the metal within the first skin depth. For deeply into conductors or metals. This means that nonmetallic materials (electrically nonconduc- Evaluation of moisture content in dielectric nonmetallic materials inside a metallic container tive), the reflected wave is much smaller than materials cannot be easily inspected through the container. the incident wave, so that any standing wave that Thickness measurements of thin metallic coat- Another limitation of the lower-frequency micro- does develop does not have a large amplitude. ings on dielectric substrates waves is their comparatively low power for re- Detection of Discontinuities. Discontinuities Detection of voids, delaminatrons, macroporos- solving localized flaws. If a receiving antenna of such as cracks, voids, delaminations, separations, ity, inclusions, and other flaws in plastic or practical size is used, a flaw with effective di- and inclusions predominantly reflect or scatter ceramic materials. mension that is significantly smaller than the electromagnetic waves. Wherever these types of wavelength of the microwaves used cannot be flaws occur, there is a more or less sharp bound- Advantages. In comparison with ultrasonic in- completely resolved (that is, distinguished as a ary between two materials having markedly dif- spection and x-ray radiographic inspection, the separate, distinct flaw). The shortest wavelengths ferent velocities for electromagnetic waves. At advantages of inspection with microwaves are as for which practical present-day microwave appa- these boundaries, which are usually thin com- follows: ratus exists are of the order of 1 mm (0.04 in.). pared to the wavelength of electromagnetic radia- However, the development of microwave sources Broadband frequency response of the coupling with wavelengths of 0.1 mm (0.004 in.) are pro- antennas ceeding rapidly. Consequently, microwave in- Efficient coupling through air from the anten- nas to the material spection for the detection of very small flaws is not suited for applications in which flaws are wove-\ ~ Test piece equal to or smaller than 0.1 mm (0.004 in,). Sub- ,..... ,,,,Zo',o Recte~~ino9 No material contamination problem caused by the coupling surface cracks can be detected by measuring the Microwaves readily propagate through air, so surface stress, which should be much higher in successive reflections are not obscured by the the surface above the subsurface crack. hg;CnreOrW ~vre first one. Techniques of Microwave Inspection. The Information concerning the amplitude and following general approaches have been used in the development of microwave nondestructive in- Reference '-- \ phase of propagating microwaves is readily ob- tainable. spection: No physical contact is required between the transmitted wove measuring device and the material being mea- Fixed-frequency, continuous-wave transmis- Phase ~-Incident wave sured; therefore, the surface can be surveyed sion detector rapidly without contact. Swept-frequency, continuous-wave transmis- The surface can be scanned in strips merely sion , In-phase t Quadrature by moving the surface or by scanning the sur- Pulse-modulated transmission output