Imaging for 3D Eddy Current Nondestructive Evaluation Pasquale Buonadonna Sponsored by: INFM

Transcription

1 59 Imaging for 3D Eddy Current Nondestructive Evaluation Pasquale Buonadonna Sponsored by: INFM Introduction Eddy current (EC) inspection is based on the principles of electromagnetic induction and is used to identify or differentiate among a wide variety of physical, structural, and metallurgical conditions in electrically conductive ferromagnetic and non ferromagnetic metals and metal parts [1, 2]. EC inspection can be used to detect seams, laps, cracks, voids, and inclusions. Because EC are created using an electromagnetic induction technique, the inspection method does not require direct electrical contact with the part being inspected. The EC method is adaptable to high-speed inspection and, because it is nondestructive, can be used to inspect an entire production output if desired. The method is based on indirect measurement, and the correlation between the instrument readings and the structural characteristics of the parts being inspected must be carefully and repeatedly established. Eddy current testing methods Advantages and limitations EC inspection is extremely versatile, which is both an advantage and disadvantage. The advantage is that the method can be applied to many inspection problems provided the physical requirements of the material are compatible with the inspection method. In many applications, however, the sensitivity of the method to the many properties and characteristics inherent within a material can be a disadvantage; some variables in a material that are not important in terms of material or part serviceability may cause instrument signals that mask critical variables or are mistakenly interpreted to be caused by critical variables. Principles of operation The EC method of inspection is dependent on the principles of electromagnetic induction for inducing EC within a part placed within or adjacent to one or more induction coils. Changes in coupling between the induction coils and the part being inspected and changes in the electrical characteristics of the part cause variations in the loading and tuning of the generator. The system used in EC inspection is usually operated at very low power levels to minimize the heating losses and temperature changes. Even, in the EC system, electrical loading changes caused by variations in the part being inspected, such as those caused by the presence of flaws or dimensional changes, are monitored by electronic circuits. In EC inspection the selection of operating frequency is largely governed by the skin effect. This effect causes the EC to be concentrated toward the surfaces adjacent to the coils carrying currents that induce them. Skin effect becomes more pronounced with increase in frequency. The coils used in EC inspection necessitate special inspection coil arrangements to facilitate the monitoring of the electromagnetic field in the vicinity of the part being inspected. Functions of a basic system The part to be inspected is placed within or adjacent to an electric coil in which an alternating current is flowing. As shown in Fig. 1a, this alternating current, called the exciting current, causes EC to flow in the part as a result of electromagnetic induction. These currents flow within closed loops in the part, and their magnitude and timing (or phase) depend on: - The original or primary field established by the exciting currents - The electrical properties of the part - The electromagnetic fields established by currents flowing within the part The electromagnetic field in the material region and its surroundings depends on both the exciting current from the coil and EC flowing in the part. The flow of EC in the part depends on: - The electrical characteristics of the part - The presence or absence of flaws or other discontinuities in the part - The total electromagnetic field within the part The change in flow of EC caused by the presence of a crack in a sample is shown in Fig. 2. From Fig. 1b it is seen that the electromagnetic field surrounding a part depends partly on the properties and characteristics of the part. Finally, the condition of the part can be monitored by observing the effect of the resulting field on the electrical characteristics of the exciting coil, such as its electrical impedance, induced voltage, or induced currents. Alternatively, the effect of the electromagnetic field can be monitored by observing the

2 60 induced voltage in one or more other coils placed within the field near the part being monitored. Each and all of these changes can have an effect on the exciting coil or other coil or coils used for sensing the electromagnetic field adjacent to a part. The effects most often used to monitor the condition of the part being inspected are the electrical impedance of the coil or the induced voltage of either the exciting coil or other adjacent coil or coils. EC systems vary in complexity depending on individual inspection requirements. However, most systems provide for the following functions: - Excitation of the inspection coil - Modulation of the inspection coil output signal by the part being inspected - Processing of the inspection coil signal prior to amplification - Amplification of the inspection coil signals - Detection or demodulation of the inspection coil signal, usually accompanied by some analysis or discrimination of signals - Display of signals on a meter, an oscilloscope, an oscillograph, or recording of signal data on magnetic tape or other recording media - Handling of the part being inspected and support of the inspection coil assembly or the manipulation of the coil adjacent to the part being inspected. Operating variables The principal operating variables encountered in EC inspection include coil impedance, electrical conductivity, magnetic permeability, lift-off and fill factors, edge effect, and skin effect. Some of these variables, particularly important for the applications of this work, will be discussed in the next section. Coil impedance When direct current is flowing in a coil, the magnetic field reaches a constant level, and the electrical resistance of the wire is the only limitation to current flow. However, when alternating current is flowing in a coil, two limitations are imposed: - The ac resistance of the wire, - A quantity known as inductive reactance, X L The ac resistance of an isolated or empty coil operating at low frequencies or having a small wire diameter is nearly the same as the dc resistance of the wire of the coil. The ratio of ac resistance to dc resistance increases as either the frequency or the wire diameter increases. In the discussion of EC principles, the resistance of the coil wire is often ignored, because it is nearly constant. It varies mainly with wire temperature and the frequency and spatial distribution of the magnetic field threading the coil. Inductive reactance, X L, is the combined effect of coil inductance and test frequency and is expressed in ohms. Total resistance to the flow of alternating current in a coil is called impedance,, and comprises both ac resistance,, and inductive reactance, X L. The 2 2 impedance can be expressed as = + X L, where X L = 2π fl0, f is the test frequency (in Hertz), and L 0 is the coil inductance (in Henrys). When a metal part is placed adjacent to or within a test coil, the electromagnetic field threading the coil is changed as result of EC flow in the test object. In general, both the ac resistance and inductive reactance of the coil are affected. The resistance of the loaded coil consists of two components, namely, the ac resistance of the coil wire and the apparent, or coupled, resistance caused by the presence of the test object. Changes in these components reflect conditions within the test object. Impedance is usually plotted on an impedance-plane diagram. In the diagram, resistance is plotted along one axis and inductive reactance along the other axis. Because each specific condition in the material being inspected may result in a specific coil impedance, each condition may correspond to a particular point on the impedance-plane diagram. For example, if a coil were placed sequentially on a series of thick pieces of metal, each with a different resistivity, each piece would cause a different coil impedance and would correspond to a different point on a locus in the impedance plane. Other curves would be generated for other material variables, such as section thickness and type surface flaws. Fig. 2a shows a simple equivalent circuit of an inspection coil and the part being inspected. The coil is assumed to have inductance, L 0, and negligible resistance. The part being inspected consists of a very thin tube having shunt conductance, G, closely coupled to the coil. When an alternating current is caused to flow into the system under steady-state conditions, some energy is stored in the system and returned to the generator at each cycle and some energy is dissipated or lost as heat each cycle. The inductive-reactance component, X L, of the impedance,, of the circuit is proportional to the energy stored per cycle, and the resistance component,, of the impedance is proportional to the energy dissipated per cycle. The impedance,, is equal to the complex ratio of the applied voltage, E, to the current, I, in

3 61 accordance with Ohm's law. The term complex is used to indicate that, in general, the alternating current and voltage do not have the same phase angle. Figs. 2b to (d) show three impedance diagrams for three conditions of the equivalent circuit in Fig. 2a. When only the coil is present the circuit impedance is purely reactive; that is, =X L =ωl=2πfl, as shown in Fig. 2b. When only the conductance of this equivalent circuit is present (a hypothetical condition for an actual combination of inspection coil and part being inspected), the impedance is purely resistive; that is, = 1 G =, as shown in Fig. 2c. When both coil and conductance are connected, the impedance has both reactive and resistive components in the 2 2 general instance, and the impedance = + X L, as shown in Fig. 2d. Here, is the series resistance and X L is the series reactance. An angle, θ, is associated with the impedance,. This angle is a function of the ratio of two components of the impedance, and XL. In Fig. 2d, this angle, θ, is about 45. Lift-off factor When a probe inspection coil, attached to a suitable inspection instrument, is energized in air, it will give some indication even if there is no conductive material in the vicinity of the coil. The initial indication will begin to change as the coil is moved closer to a conductor. Because the field of the coil is strongest close to the coil, the indicated change on the instrument will continue to increase at a more rapid rate until the coil is directly on the conductor. These changes in indication with changes in spacing between the coil and the conductor, or part being inspected, are called lift-off. The lift-off effect is so pronounced that small variations in spacing can mask many indications resulting from the condition or conditions of primary interest. Consequently, it is usually necessary to maintain a constant relationship between the size and shape of the coil and the size and shape of the part being inspected. The lift-off effect also accounts for the extreme difficulty of performing an inspection that requires scanning a part having a complex shape. Although troublesome in many applications, lift-off can also be useful. For example, with the lift-off effect, EC instruments are excellent for measuring the thickness of nonconductive coatings, such as paint and anodized coatings on metals. Edge effect When an inspection coil approaches the end or edge of a part being inspected, the EC are distorted because they are unable to flow beyond the edge of a part. The distortion of EC results in an indication known as edge effect. Because the magnitude of the effect is very large, it limits inspection near the edges. Unlike lift-off, little can be done to eliminate edge effect. A reduction in coil size will somewhat lessen the effect, but there are practical limits that dictate the sizes of coils for given applications. Skin effect In addition to the geometric relationship that exists between the inspection coil and the part being inspected, the thickness and shape of the part itself will affect EC response. EC are not uniformly distributed throughout a part being inspected; rather, they are densest at the surface immediately beneath the coil and become progressively less dense with increasing distance below the surface a phenomenon known as the skin effect. At some distance below the surface of a thick part there will be essentially no currents flowing. The depth at which the density of the EC is reduced to a level about 37% of the density at the surface is defined as the standard depth of penetration. This depth depends on the electrical conductivity and magnetic permeability of the material and on the frequency of the magnetizing current. Depth of penetration decreases with increasing conductivity, permeability, or inspection frequency. The standard depth of penetration can be calculated by: S =1980 ρ / µf where S is the standard depth of penetration (in inches), ρ is the resistivity (in ohm-centimeters), µ is the magnetic permeability (1 for nonmagnetic materials), and f is the inspection frequency (in hertz). esistivity, it should be noted, is the reciprocal of conductivity. The EC response obtained will reflect the work piece material thickness. It is necessary, therefore, to be sure that either the material has a constant thickness or is sufficiently thick so that the EC do not penetrate completely through it. It should be remembered that EC do not cease at the standard depth of penetration but continue for some distance beyond it. Normally, a part being inspected must have a thickness of at least two or three standard depths before thickness ceases to have a significant effect on EC response. By properly calibrating an EC instrument, it is possible to measure material thickness because of the varying response in thickness.

4 62 Eddy current inspection coil, instrumentation and readout instrumentation. Appropriate coil selection is the most important part of solving an eddy current application; no instrument can achieve much if it doesn t get the right signal from the probe. Coil designs can be split into three main groups: surface probes used mostly with probe axis normal to the surface. In addition to the basic pancake coil, this includes pencil probes and special-purpose surface probes such as those used inside a fastener hole. Encircling coils are normally used for in-line inspection of round products (the product to be tested is inserted through a circular coil) and ID probes are normally used for in-service inspection of heat exchangers (the probe is inserted into the tube); normally ID probes are wound with the coil axis along the centre of the tube. To this point we have only discussed eddy current probes consisting of a single coil. These are commonly used in many applications and are commonly known as absolute probes because they give an absolute value of the condition at the test point. Absolute probes are very good for metal sorting and detection of cracks in many situations, however they are sensitive also to material variations, temperature changes etc. Another commonly used probe type is the differential probe which has two sensing elements looking at different areas of the material being tested. The instrument responds to the difference between the eddy current conditions at the two points. Differential probes are particularly good for detection of small defects and are relatively unaffected by lift-off (although the sensitivity is reduced in just the same way), temperature changes and (assuming the instrument circuitry operates in a "balanced" configuration) external interference. Fig. 3 shows a typical response from a differential probe. Note the characteristic "figure eight" response as the first probe element, then the other one, move over the defect. In general the closer the element spacing the wider the "loop" in the signal. Lift-off should be cancelled out assuming that the probe is perfectly balanced, but there will still be a "wobble" response as the probe is moved and tilted slightly. eflection or driver pick-up probes have a primary winding driven from the oscillator and one or more sensor windings connected to the measurement circuit. Depending on the configuration of the sensor windings reflection probes may give a response equivalent to either an absolute or differential probe (Fig. 4). The EC instrument type used here is based on the bridge unbalance system. It consists of a signal source, an impedance bridge with dropping resistors, an inspection coil in one leg, and a balancing impedance in the other leg (Fig. 5a). The differences in voltage between the two legs of the bridge are measured by an ac voltmeter. Alternatively, the balancing impedance in the leg opposite the inspection coil may be a coil identical to the inspection coil, as shown in Fig. 5b, or it may have a reference sample in the coil, as shown in Fig. 5c. In the latter, if all the other components in the bridge were identical, a signal would occur only when the inspection coil impedance deviated from that of the reference sample. Another type of bridge system is the induction bridge or driver pick-up configuration, in which the power signal is transformer-coupled from a transmitter coil into two detector coils: an inspection coil and a reference coil. In addition, the entire inductance-balance system is placed in the probe, as shown in Fig. 5c. The probe consists of a large transmitter, or driver, coil, and two small detectors, or pickup, coils wound in opposite directions as mirror images to each other. An alternating current is supplied to the large transmitter coil to generate a magnetic field. If the transmitter coil is not in the vicinity of a conductor, the two detector coils detect the same field, and because they are wound in opposite directions to each other, the net signal is zero. However, if one end of the probe is placed near a metal surface, the field is different at the two ends of the probe, and a net voltage appears across the two coils. The resultant field is the sum of a transmitted signal, which is present all the time, and a reflected signal due to the presence of a conductor (the metal surface). The intensity of the transmitted signal decreases rapidly as the distance between the coil and conductor is increased, and the intensity of the reflected wave does the same. The detector coil nearer the conductor detects this reflected wave, but the other detector coil (the reference coil) does not, because the amplitude of the wave has greatly decreased in the distance from the reflecting metal surface to the rear detector coil. The magnitude and phase obtained for a system such as this are similar to those in a bridge unbalance system with the reference coil in air. An important part of an EC inspection system is the instrument used for a readout. The readout device can be an integral part of the system, an interchangeable plug-in module, or an external computer. Eddy current responses of a single coil may be conveniently described by reference to the "impedance plane". This is a graphical representation of the complex probe impedance where the abscissa (x value) represents the resistance and the ordinate (y value) represent the inductive reactance. It is worth noting that, while the general form of the impedance plane remains the same, the details are unique for a particular probe and frequency. The display of a typical CT eddy current instrument represents a window into the impedance plane, which can be rotated and zoomed to suit the needs of the application.

5 63 For example in the impedance plane diagram of Fig. 6a, a rotated detail of the probe on aluminum area would appear as in Fig. 6b. This shows the display when moving over a series of simulated cracks of varying depths. In the example shown both the amplitude and the phase of the response from the different sized cracks varies. Eddy Current Imaging EC scanning was carried out on an Al alloy plate of size 127x25x6 mm with 4 notches made by EDM with width mm and different depth (Fig. 7). The scan area comprised notch B (0.5 mm depth) and notch C (0.25 mm depth); testing parameters are in Tab. 1. Sample Probe Step Area Scan from the side of the defects: KA 2-1 (800 khz) 0.4 mm x = 8 mm notches B and C. y = 44 mm Tab. 1 Experimental conditions for the EC scanning. Prior to scanning, the EC instrumentation was appropriately set up. The impedance plane diagram was rotated on the instrument display so that the probe response, as it moved over a defect, is given only by an amplitude variation along the Y axis whereas no appreciable phase variation is observed. During scanning, the instrument analog output proportional to the signal amplitude (Y component) was fed to a computer, digitized and stored as a 2D numerical array. Each row in the array contains the numerical values of the detected signal for each material interrogation point with constant step during one scanning line. The set of scanning lines makes up the entire X-Y scan. The 2D numerical array is normalized and images with 128 gray tones or pseudo colors are created using the Image Processing Toolbox of Matlab [3] (Fig. 8). In Fig. 8, the location of the notches is identified by a vertical stripe in lighter gray tones (Fig. 8a) or different pseudo colors (Fig. 8b) in comparison with the base material. EC NDE allows to obtain 3D information on the material structure under examination. The sensor signal, proportional to the magnetic field variation due to the presence of a defect (notch), is influenced by the notch extension in the thickness direction (notch depth). By examining the images in Fig. 8, it is possible to discriminate the notch depth on a qualitative basis: a higher depth corresponds to a wider vertical stripe with higher values of gray tone intensity or pseudo color values. In order to obtain quantitative information on notch depth, 3D EC images can be created by associating the values of gray tone (or pseudo color) intensity in the 2D matrix to values on the -axis of a 3D diagram (Fig. 9). The 3D wire net (Fig. 9a) and color graded surface (Fig. 9b) representations indicate the presence of notches through peaks of different height on a horizontal surface which represents the base material. Because of the sensor signal detected during the scan, ie impedance, the peak height is proportional to the notch depth: as a matter of fact, the higher peak, corresponding to notch B with depth 0.5 mm, is twice as high as the lower peak which corresponds to notch C with depth 0.25 mm. This capability of EC NDE of providing quantitative information of defect extension in the thickness direction is particularly interesting because it is not as effective in other NDE approaches. Fig. 10 reports the 3D ultrasonic image of a C scan with time gate on the back echo carried out on the same sample [4]. As can be seen, the information on notch depth, though present, is not as quantitatively dependable as the information from EC scanning. Future Work Imaging methods for EC NDE will be utilized to comparatively assess the capabilities of conventional and innovative EC sensors in the NDE of metal alloys and composite materials. eferences [1] Cecco, V. S., Feanklin, E. M., Houserman, H. E., Kincaid, T. G.,Pellicer, J., Hagemaier,D., 1989, Eddy Current Inspection, in Nondestructive Evaluation and Quality Control, Metals handbook, Vol. 17, ASM International, USA [2] McMaster,. C., McIntre, P., Mester, M. L., 1986, Electromagnetic Testing, Nondestructive Testing Handbook, Vol. 4, Published by the American Society for Nondestructive Testing, USA [3] Thompson, C.M., Shure, L., 1993, Image Processing TOOLBOX, The Math Works Inc., Natick, MA, USA [4] Duraccio,., 1997, NDE Techniques through Ultrasonic and Eddy Currents, Graduation Thesis, Dept. of Materials and Production Engineering, University of Naples Federico II

6 64 EC Coil EC Coil Crack Fig. 1 EC flowing in the part as a result of electromagnetic induction, change in the flow of EC caused by a crack. Ohm's law I = E, E = I, = E I Inductive reactance (XL Alternating current (I) Voltage (E) G = 11 1 = G esistance () (c) G L0 Inductive reactance (XL Inductive reactance (XL esistance () L =XL = ωl = 2πfL 45.0 θ G esistance () L = XL1 2 Fig. 2 Simplified equivalent circuit of an EC coil and the part being inspected;, (c) and (d) three impedance diagrams for 3 conditions of the equivalent circuit. = series resistance; f = frequency; L 0 = coil inductance; ω = 2πf; G = shunt conductance; = impedance. (d) Fig. 3 Typical response from a differential probe.

7 65 Fig. 4 Different configurations of the sensor windings in a reflection probe: differential reflection probe and absolute reflection probe. Test sample Voltmeter V Inspection coil Balancing impedanc Test sample Voltmeter V Inspection coil Coil (balancing impedance) Test sample Coil (balancing impedance) V Voltmeter Inspection coil eference sample Trasmitter (drive) coil Workpiece (conductor) (c) (d) Detector (pickup) coil (1 of 2) Fig. 5 Three types of EC instruments: typical impedance bridge; impedance bridge with dual coils; (c) impedance bridge with dual coils and a reference sample in the second coil; (d) induction bridge or driver pick-up configuration. X L Ferrite Crack in Steel Steel Probe in Air MAGNETIC Titanium NON-MAGNETIC Crack in Aluminium Lift-Off Aluminium Increasing conductivity of Test Sample Copper Probe Coil Impedance - Fig. 6 Impedance plane diagram; rotated detail of the probe on aluminum area when moving over a series of cracks with different depths. D C B A Fig. 7 Al alloy plate (127 x 25 x 6 mm) with four EDM notches with some width mm and different depths : A = 1 mm; B = 0.5 mm; C = 0.25 mm; D = mm.

8 66 notch C notch B notch C notch B Fig. 8 2D images from EC scanning of the AL alloy plate: gray tones; pseudo-colors. Fig. 9 3D representations from the EC scanning of the Al alloy plate: wirenet; color graded surface. Fig. 10 3D ultrasonic image from a C-scan carried out on the Al alloy from the side opposite to the defects.