· Table of Contents
· Methods & Instrumentation
A New Eddy Current Surface Probe without Lift-off NoiseHiroshi HOSHIKAWA and Kiyoshi KOYAMA
Izumicho Narashino Chiba 275-8575, Japan.
The authors propose a new surface eddy current probe that generates no lift-off noise in principle. The probe does not detect the eddy current induced by the exciting coil when the test material has no flaws and detects the eddy current generated only by flaws in the test material. The flaw signal phase changes according to flaw depth and changes little by the length and direction of the flaws. The authors believe that the new probe makes eddy current testing more quantitative in flaw depth evaluation than the conventional pancake coil probes.
Conventional eddy current probes detect the change of the eddy current induced by the exiting coil or the change of the test coil impedance. Since the eddy current induced by the probes varies by the lift-off of the probe from the test material, the lift-off noise is unavoidable for those conventional probes. Since the large lift-off noise changes the signal phase much from the probe, the signal phase can hardly be used to evaluate flaws. As a result, the conventional eddy current surface probes usually use only the amplitude of flaw signals to evaluate flaws. However, the signal amplitude changes not only by the depth of flaws but also by the length and width. Consequently, eddy current testing by the conventional surface probes has hardly been considered as a quantitative method for evaluating the depth of surface flaws. Although some probes with little lift-off noise have been developed in recent years [1-2], eddy current testing requires more probes with little lift-off noise to be developed for more reliable nondestructive testing.
The authors have noticed that lift-off noise in eddy current testing is avoidable if the probe picks up only the eddy current induced by flaws but not directly by the exiting coil. Thus the authors have devised a new probe that is composed of a circular exciting coil and a rectangular tangential detecting coil arranged at the center of the exciting coil. The exciting coil induces axi-symmetrical eddy current in the test material and the detecting coil generates a signal only when some eddy current flows along the detecting coil. The probe generates no signal as long as the induced eddy current is kept axi-symmetrical. Thus the probe is lift-off noise free.
The finite element analysis predictions and experimental results have indicated that the phase of the figure eight like flaw signal patterns obtained by the new probe is almost proportional to the flaw depth in spite of the obvious changes of the signal amplitude by the flaw length. Since the probe actually generates only very little lift-off noise, the phase of the flaw signal can be utilized to evaluate the depth of flaws. The authors believe that the new probe can make the eddy current testing quantitative in evaluating the depth of flaws.
Conventional pancake coil probes detect flaws in the test material by picking up the variation of the eddy current by flaws. Many works on coil impedance analysis have been conducted and have contributed a great deal to the development of eddy current testing [3-4]. Since the variation of the probe lift-off from the test material changes the eddy current induced in the test material, the conventional pancake coil probes necessarily suffer from large lift-off noise. The lift-off noise is unavoidable so long as the probes pick up the variation of the eddy current induced by the test coil or the test coil impedance.
Since the large lift-off noise causes the phase of the probe signal to change a great deal, the signal phase can hardly have been utilized to evaluate flaws in the eddy current testing by the conventional probes, except for the tube inspection by inner bobbin probes where the lift-off noise is mostly minimal. The signal phase information in the eddy current testing by the surface probes is usually eliminated in the signal processing to suppress the lift-off noise. Thus only the flaw signal amplitude has usually been used to evaluate flaws by surface probes. However, since the signal amplitude changes not only by the depth of flaws but also by the length and width, eddy current testing has been considered as a non-quantitative method in evaluating depth of flaws.
The authors have conceived the following two notions in order to design a new probe that eliminates lift-off noise.
Based on the above notions, the authors have devised a new eddy current surface probe that is composed of a circular pancake exciting coil and a tangential detecting coil as shown in FIGURE 1. The probe was named as probe because its two coils look like the Greek alphabet when seen from the top. The exciting coil induces axi-symmetric eddy current in the test material with no eddy current circulating across the exciting coil circle when there is no flaw in the test material. The detecting coil picks up only the eddy current flowing in the test material under across the exciting coil circle. The new probe is lift-off noise free because the lift-off of the probe from the material does not cause any eddy current to circulate across the exciting coil circle. Thus lift-off noise is avoidable by detecting only the eddy current generated by flaws but not detecting the eddy current induced directly by the exciting coil. The probe is self-nulling because the detecting coil generates no signal as long as the eddy current flows axi-symmetrically.
|Fig 1: Structure of probe.|
|Fig 2: Eddy current and flaw signal generated by probe.|
When there is a flaw in the test material under across the exciting coil circle, some of the eddy current circulates along the flaw across the circle although most of the eddy current circulates beneath the surface flaw. Now let us consider the probe scanning over perpendicularly to a flaw as shown in FIGURE 2. The tangential detecting coil picks up the parallel eddy current component to itself. When a flaw is above the detecting coil in the figure, the detecting coil generates a plus signal. The detecting coil generates no signal when a flaw is at the center of the probe. When a flaw proceeds below the detecting coil in the figure, the detecting coil generates a minus signal. Thus the probe generates a figure eight like signal pattern as the probe scans over a flaw. Since the probe generates only minimal lift-off noise, the authors have expected that the probe lift-off does not influence much on the phase of the flaw signal and that the signal phase can be used for evaluating the depth of surface flaws.
FIGURE 3 shows the sizes of the probe and the brass plates with an electric discharge machined slit flaw of different depths, lengths, and widths used for the experiments. The test frequency of 20 kHz was chosen to make the ratio of the skin depth of the eddy current induced in the material to the plate thickness equal to 1.5. The exciting coil alone was also used to conduct experiments of the conventional pancake coil probe.
Fig 3: Experimental setup and test conditions.
Exciting coil: outer diameter 7mm , cross section of windings 1x1mm2.
Detecting coil: length 5mm , height 3mm , cross section of windings 1x1mm2. Brass plate: 160x160x1.5mm3. Test frequency: 20kHz.
Experiments were conducted to generate lift-off noise by inserting thin papers with different thicknesses between the probe and the test material. FIGURE 4(a) indicates that the conventional pancake coil probe generates far larger lift-off noises than the flaw signals. On the other hand, FIGURE 4(b) indicates that the probe generates far larger flaw signals than lift-off noises. Thus it is obvious that the probe generates flaw signals of much higher signal to noise ratio than conventional probes. FIGURE 4 (b) also indicates that the flaw signal phase does not fluctuate much by lift-off.
|Fig 4: Flaw signals and lift-off noises.|
FIGURE 5 shows flaw signal patterns for front surface flaws and back surface flaws with different depths obtained by the probe. The figure indicates that the amplitude and the phase of the flaw signals change according to the depths of front surface flaws and back surface flaws. Thus the probe has the possibility of evaluating the depth of flaws based on the flaw signal phase.
|Fig 5: Flaw signal patterns for different flaw depths by probe|
FIGURE 6 shows flaw signal patterns obtained by the probe for flaws of the same depth and different lengths. The figure indicates that the flaw length changes a lot the amplitude of the flaw signal. As a result, it is obvious that the flaw signal amplitude is not effective to estimate flaw depth. On the other hand, FIGURE 6 indicates that the signal phase does not change by flaw length. Thus the probe provides the phase information that indicates flaw depth without much influence from flaw length.
|Fig 6: Flaw signal patterns for different flaw lengths by probe.||Fig 7: Flaw signal patterns for different flaw widths by probe|
FIGURE 7 shows flaw signal patterns obtained by the probe for flaws of the same depth and different widths. The figure indicates that the flaw width changes the amplitude of signals a lot but keeps the phase almost constant.
FIGURE 8 shows the flaw signal patterns obtained by the probe for different angles of a flaw with respect to the pick-up coil of the probe. The figure indicates that the angle changes the amplitude of the signal but not the phase.
The experimental results from FIGURE 5 to FIGURE 7 indicate that signal phase by the probe is a reliable parameter to estimate the depth of flaws in eddy current testing and the signal amplitude is not a proper parameter because it is influenced largely by the length, width, and angle of a flaw.
|Fig 8: Flaw signal patterns for different angles of a flaw with respect to the pick-up coil of probe.|
FIGURE 9 shows the signal patterns of different flaw depths with the normalized amplitude. The figure indicates that the depth of front surface flaw lags the flaw signal phase and the depth of back surface flaw leads the phase.
From the signal patterns shown in FIGURE 9, the authors have derived the flaw depth evaluation curve based on the signal phase as shown in FIGURE 10. Thus flaw depth can be evaluated by applying flaw signal phase to the curve in FIGURE 10 without much influence from the variations of flaw length and width and also from the angle of the flaw with respect to the pick-up coil. The relation between signal phase and flaw depth is just the same as the one known in the eddy current tubing inspection by the inner bobbin coil probe. The authors believe that the flaw depth evaluation method based on the signal phase improves the accuracy of flaw depth evaluation in eddy current testing.
|Fig 9: Flaw signal patterns with normalized amplitude by different flaw depths.||Fig 10: Flaw depth evaluation curve based on the flaw signal phase.|
The authors have devised a new eddy current probe that has little lift-off noise and provides phase information on surface flaw depth. The signal phase is kept almost constant by the variations of the length, width, and angle of a flaw. The authors believe that the probe will make eddy current testing more quantitative than the conventional probes. The impedance of the exciting coil can also be used to monitor the probe lift-off in order to avoid the probe not detecting flaws in the material.
|© AINDT , created by NDT.net|||Home| |Top||