|NDT.net June 2005 Vol. 10 No.6|
Time domain lift-off compensation method for eddy current testingCatalin Mandache*, Mike Brothers
National Research Council of Canada, Institute for Aerospace Research, Ottawa, ON, Canada
Defence R & D Canada, Air Vehicle Research Section, Ottawa, ON, Canada
*Corresponding Author Contact:
Email: firstname.lastname@example.org, Internet: iar-ira.nrc-cnrc.gc.ca/main_e.html
1200 Montreal Rd., Bldg. M-14, Ottawa, ON, Canada, K1A 0R6; phone: 613-991-2132
AbstractThis paper discusses a new lift-off compensation method for sinusoidal, time domain represented eddy current inspection. Time domain representations of both pulsed and sinusoidally excited eddy current responses display stationary points when the probe-specimen coupling is varied. Time gating on these points is used for the evaluation of material discontinuities, based on the signal amplitude within the gate. As an example, a simulated second layer crack in a two-layer aluminum lap joint is investigated. The results obtained with established techniques for eddy current lift-off compensation are compared with the lift-off point of intersection method for time domain sinusoidal eddy current. The proposed technique simplifies the interpretation of eddy current signals and the resulting C-scan image proves superior in quality to those provided by conventional and pulsed eddy current techniques.
1. Introduction and BackgroundDespite eddy currents’ popularity as a non-destructive technique for inspection of electrically conductive materials, the method suffers because of data interpretation difficulties. This challenge is mainly due to the non-linear spatial and temporal distributions of the electromagnetic fields in the conductive material, making it difficult to discriminate between possible specimen variables that could simultaneously affect the response signal. The test piece characteristics (electrical conductivity, magnetic permeability, thickness, etc.) and the investigation probe properties (type of coils arrangement and geometry) further complicate the discontinuity signature in an eddy current testing signal response.
One of the most important factors affecting the eddy current signals is the coupling between the specimen and the investigation probe, which translates as the probe-to-specimen separation or lift-off. In real testing situations, applied non-conductive coatings, painting and geometry variations of the specimen result in a locally varying lift-off. Lift-off elimination methods help in the interpretation of eddy current signals as a function of material discontinuities.
The ingenious representation of eddy current testing on the impedance plane diagram is able to separate between various material discontinuities. This method uses sinusoidally excited eddy currents and plots the normalized inductive reactance of the probe as a function of its resistance. In this plane, the locus variation with the electrical conductivity of the material follows the curve indicated in figure 1.a . The direction of movement of the signal from the conductivity curve reveals the possible type of discontinuity in the material under investigation, as shown in figure 1.a. However, this representation is not easy to interpret and relies heavily on the experience and intuition of the inspector.
Constraining the signal variation due to the lift-off change to the negative horizontal axis minimizes the varying probe-specimen coupling effect; in this way, other specimen.discontinuities may be revealed, as schematically shown in figure 1.b. The extent and depth of a crack are estimated based on the amplitude and phase, respectively, of the eddy current signals.
Pulsed eddy current inspection techniques represent a spin-off from the conventional technique explained above. In this case, the magnetic field is produced by a square wave rather than a sinusoidal excitation. The biggest advantage of this type of testing is the broadband frequency excitation that permits the simultaneous evaluation of the material at different depths [2-4]. Time domain representation of the transient signals has shown that there is a crossing point whose time and voltage coordinates are independent of the lift-off variation , as depicted in figure 2. This crossing point is called lift-off point of intersection, or, on short, LOI point [6,7]. Eddy current specimen mapping (or C-scan imaging) of the voltage amplitude while gating at the LOI time showed very promising results for materials characterization  and for finding specimen discontinuities [6,7,9].
Based on the successful use of this time domain representation of pulsed eddy currents and taking into account that any pulse can be represented as a summation of sinusoidal.waveforms, it has been previously shown that the lift-off point of intersection also exists for harmonically excited eddy currents , as depicted in figure 3. For distinction from the LOI feature of the pulsed eddy currents, the lift-off independent crossing point in sinusoidally represented eddy currents is called the harmonic LOI. The amplitude and the phase of the response signal (measured here with respect to the excitation signal) change in such a way that the sinusoidal waveforms always cross at the same point, regardless the degree of probe-specimen coupling.
2. Experimental detailsThe three methods of lift-off compensation described in the previous section (impedance plane diagram, transient LOI, and harmonic LOI) are used to examine the same specimen, with the same probe, and at the same testing frequency.
The test piece inspected in this study consisted of a simulated crack in the second layer of a two-layer aluminum structure. Each aluminum layer was 1mm thick. The crack length was equal to 6mm and through the wall of the second layer. The area investigated was equal to 25.4mm x 25.4mm.
Applying Teflon tape of different thickness and covering equal areas on the top of the specimen simulated four different levels of probe-specimen coupling, corresponding to applied lift-off values of 0.45mm, 0.30mm, 0.15mm, and 0.00mm (i.e. no applied lift-off). This situation is represented in figure 4.a. The scanning was performed with a resolution of 0.25mm on both axes and the scanning direction was perpendicular to the crack, as shown in figure 4.b.
The same eddy current probe (T/R P11706, 1-15kHz, manufactured by R/D Tech) was used in all inspection situations. This probe consisted in two concentric coils, the outer one used for signal excitation, while the shielded, inner coil was used for signal pick-up. The testing frequency of the signal was set to 7.5kHz, a value that corresponds to a standard depth of.penetration of 1.38mm. In the case of pulsed eddy current, the pulse width was set equal to half of the period corresponding to the sinusoidally excited eddy current, i.e. 66.7 µs.
For harmonic eddy current testing, two signal visualization procedures were used: one on the conventional impedance plane diagram (as seen in figure 1), and the other by time domain representation of the sinusoidal waveform (as seen in figure 3). For the pulsed eddy current inspection, only the time domain analysis was performed (figure 2).
The lift-off elimination method for conventional eddy current was discussed earlier (figure 1), when impedance plane diagram is used to represent the signal response. For time domain harmonic and pulsed eddy current, the lift-off compensation procedure is described as follows. The time coordinate of the crossing point is determined visually by examining the overlapped signal responses obtained for a plurality of lift-off values. A time gate is set on this lift-off independent point (see figures 2 and 3), and the amplitude of the signal at this time gate is used for the C-scan mapping of the specimen. A typical time gate has a length of about 0.5% of the period of the waveform.
3. ResultsAll three eddy current methods described earlier were applied for the inspection of the two-layer sample containing a second layer crack shown in figure 4. Although the same probe and sample had been used for every investigation method, the auxiliary instrumentation was partially different in each case. For this reason, in order to facilitate comparison, the C-scan images were normalized on a scale from 0 to 1.
3.1. Conventional eddy current testingBy the procedure explained in the ‘Introduction and Background’ section, on the impedance plane diagram the lift-off variation is constrained on the horizontal axis, while the vertical variation of the locus is used for defect detection (see figure 1.b). For determination of crack depth, the phase of this signal would have been investigated. The C-scan image of the specimen is shown in figure 5. A small background gradient is still observed in figure 5, meaning that the.lift-off was not completely eliminated by this method. However, the defect is evident due to the concentration of the eddy current flow at the extremities of the crack.
3.2. Pulsed eddy current testingThe LOI time is determined before proceeding with the surface scanning by simultaneously recording and displaying a plurality of transient responses corresponding to various probe-specimen coupling (as schematically shown in figure 2). This is done by manually moving the probe away from the sample surface while the signal is continuously recorded . A time gate is placed at this time coordinate and the C-scan image of the specimen is obtained by plotting the maximum signal amplitude within the time gate placed at the LOI point. The result obtained in this case is shown in figure 6; no background gradient is observed. This procedure is known as a very efficient lift-off compensation method [6,7].
3.3. Time domain sinusoidal eddy current testingSampling of the signal amplitude at the experimentally pre-determined harmonic LOI time (figure 3), similar to the procedure described above for the pulsed eddy current testing, yielded the C-scan image shown in figure 7. This result is at least comparable in quality to the one obtained by the pulsed eddy current technique (figure 6).
The outcome of the time domain representation of sinusoidal eddy current investigation of the second layer crack in a two-layer aluminum structure validates the existence of a lift-off independent crossing point and proves its effectiveness as a lift-off compensation method.
4. Summary and ConclusionThis study was motivated by the long-known lift-off independence of the crossing points in pulsed eddy current testing [5-9]. It was shown that the lift-off point of intersection is also a feature of the conventional (sinusoidal) time domain represented eddy current. The method was evaluated by inspecting a simulated two-layer lap joint containing a second layer crack by conventional/impedance, pulsed, and conventional/time domain eddy current methods.
Comparison of the resulting C-scan images obtained by these three methods shows that the harmonic LOI feature could be used for lift-off compensation.
Moreover, since any sinusoidal waveform may be easily reproduced based on only two of its parameters, amplitude and phase, the method proposed in this paper presents the advantage of small computer storage space. When large surfaces are inspected, instead of saving the entire response pattern, which usually consists in thousands of time steps for each spatial coordinate of the test, the inspector might choose to save only the amplitude and phase of the signal. The data may be reconstructed at a later stage, when different inspection characteristics may be determined, such as signal amplitude, relative phase, LOI time, LOI voltage, etc.
Although the quantification of the crack size and depth is beyond the purpose of this paper, it is apparent from figures 5-7, that the crack length (6mm) might be determined as the horizontal axis separation between the peak values on the C-scans.
The potential shown here by the ‘harmonic lift-off point of intersection’ in sinusoidally, time domain represented eddy currents will be evaluated for other types of material discontinuities.