2014 EDDY CURRENT BENCHMARK

Transcription

1 World Federation of N D 2014 EDDY CURRENT BENCHMARK E Centers The World Federation of NDE Centers pleased to announce a new Eddy Current Benchmark Problem for the 2014 Review of Progress in Quantitative NDE Meeting. Experiments were done on plates with a cylindrical coil and a magnetic sensor, providing a direct measurement of the vertical component of the magnetic field.. We invite you to participate in this benchmark and present your results at a special benchmark session during the 2014 Review of Progress in Quantitative NDE meeting being held this year in Boise, Idaho. This study has been conducted by researchers at the Commissariat a l énergie atomique (CEA) in France and the University of Macedonia in Greece. We would like to thank those researchers for their extensive efforts in designing these benchmarks and for obtaining the responses. Continued 1

2 PARTICIPATION IN THE 2014 BENCHMARK SESSION We would like to invite papers that consider the eddy current benchmark outlined here at the next Annual Review of Progress in Quantitative Nondestructive Evaluation (RPQNDE) meeting. This meeting will be held July 20-25, 2014 at the Boise Center in Boise, Idaho. To present a paper at that session, please note that the deadline for submitting an abstract is Monday, April 28, 2014 (mark on your abstract that it is for the benchmark session). Also, please note that the advance registration deadline for the conference is Friday, June 20, For more details of the conference, visit the website at On the following pages is a comprehensive outline of the benchmark problem being considered. You can retrieve this outline and the experimental data files at the Center for NDE, Iowa State University ftp site; ftp://ftp.cnde.iastate.edu For the user name and password for this site please contact Prof. Schmerr. The benchmark data and the complete description of the problems considered are located in the Pub folder on this ftp site in the sub-folder 2014 EC Benchmark. For any questions, please Christophe Reboud at Christophe.REBOUD@cea.fr or Prof. Schmerr at lschmerr@cnde.iastate.edu 2

3 Eddy current benchmark: magnetic field measurements for a coil above plate(s) with cracks Ch. Voulgaraki 1, N. Poulakis 2, R. Miorelli 3, C. Reboud 3 and T. Theodoulidis 1 1 Ch. Voulgaraki and T. Theodoulidis are with University of Western Macedonia, Department of Mechanical Engineering, Bakaloa & Sialvera, Kozani 50100, Greece. xaritini.voulg@gmail.com and theodoul@uowm.gr 2 R. Miorelli and C. Reboud are with CEA LIST, Département Imagerie et Simulation pour le Contrôle, Gif-sur- Yvette 91191, France. roberto.miorelli@cea.fr and christophe.reboud@cea.fr 3 N. Poulakis is with the Technological Educational Institute of Western Macedonia, Department of Electrical Engineering, Koila, 50100, Greece. poulakis@teikoz.gr

4 Introduction This benchmark proposes reference data for eddy current testing configurations corresponding to inspection of flawed planar pieces. The probe used in this benchmark functions in harmonic regime and is made of one emitting cylindrical coil and one receiving magnetic sensor (Hall sensor), which is located at the center of the coil in horizontal directions. This probe setup is interesting since it provides an almost punctual measurement of the vertical component of the magnetic field. Benchmark description A coil and Hall sensor system is inspecting 2 plates, with respective thicknesses of 4 mm and 5 mm, containing artificial cracks. Three cases have been considered: Test Case 1: Plate 4 mm thick with through-the-thickness crack. Test Case 2: Plate 5 mm thick with surface crack. Test Case 3: Plate 5 mm thick with subsurface crack (plate turned upside down). The Hall sensor output is recorded for three frequencies, 1025, 1975 and 3025 Hz, with the probe scan line passing above the crack in two directions, along the crack line (centered above the crack) and across the crack line (passing through the crack center). The coil+hall and plates parameters are given in Tables 1-2, where both coil lift-off and plate conductivities were first measured and then fitted (for greater accuracy) with the method described in [1]. The experimental setup is shown in Figure 1. Figure 1. Experimental setup. [2]

5 The coil is driven by a constant current source and the voltage output of the Hall sensor is recorded by using a lock-in amplifier. Measured signals correspond to the z-component of the magnetic field. Due to the uncertainty of the circuit parameters associated with the coil+hall system at the three different frequencies, a calibration method was used in order to produce experimental results that are directly comparable to theoretical simulation ones. Table 1. Coil and sensor parameters. Inner radius r mm Outer radius r mm Length h 4.0 mm Turns N 407 Lift-off 0.05 mm Free space inductance (experimental) L mh Free space inductance (calculated) L mh Hall sensor lift-off 1.3 mm Table 2. Plates parameters. 4 mm thick 5 mm thick Thickness 4.0 mm 5.0 mm Conductivity 17.5 MS/m 17.0 MS/m Rel. permeability Crack depth 4.0 mm 3.85 mm Crack length mm mm Crack width mm mm Since the analytical model of a coil+hall above an intact plate is well proven and thoroughly validated, it can be used as a reference in order to get a single complex number that translates the voltage measurement to magnetic field value away from the crack (for each measurement configuration). The experimental value for this case is obtained from the first measurement in the coil scan data when the coil is far away from the crack. Hence, numbers given in the following Tables correspond essentially in the combination of the multiplication factor provided by the manufacturer for the conversion between Gauss and Volts, as well as the comparison of the experiment and the theory for the case of an intact plate and the division by the constant current used in the experiments. These numbers (conversion factors) can be used to multiply the experimental complex data to get results that are comparable to theoretical ones. The final values obtained after the multiplication are in mt/a. [3]

6 Test case 1: plate 4 mm thick with through-the-thickness crack. f [Hz] Along (crack center at x = 30 mm) Across (crack center at x = 20 mm) i i i i i i Test case 2: plate 5 mm thick with surface crack. f [Hz] Along (crack center at x = 35 mm) Across (crack center at x = 20 mm) i i i i i i Test case 3: plate 5 mm thick with subsurface crack. f [Hz] Along (crack center at x=35 mm) Across (crack center at x = 20 mm) i i i i i i Each txt file has a self-explaining name and consists of three columns, the first one corresponds to the distance covered by the coil+hall system, the second is the real part of the magnetic field and the third one is the imaginary part of the magnetic field. The field values are provided in Gauss and also correspond to the coil current amplitude provided in the first line of the txt file. This value is not required when the conversion factor of the Tables is used (and the values of the magnetic field are provided in mv/a). The magnetic field change that corresponds to the crack signal can be deduced by subtracting the data from the first measurement point in the file. In the next sections, experimental data obtained in each case after these two operations (calibration and balancing) are represented, as well as illustrations of the three configurations. References [1] Harrison D.J., Jones L.D., Burke S.K., "Benchmark problems for defect size and shape determination in eddy-current nondestructive evaluation", Journal of Nondestructive Evaluation, 15(1), pp.21-34, [2] Li Y., Theodoulidis T., Tian G.Y., "Magnetic field-based eddy-current modeling for multilayered specimens ", IEEE Transactions on Magnetics, 43(11), pp , [4]

7 Test Case 1: Plate 4 mm thick with through-wall crack (a) (b) Figure 2. ECT inspection problem: (a) 3D-view, (b) Top and side views. Figure 3. Experimental data obtained at the frequency of 1025 Hz and corresponding to files Case_1_Across_1025Hz_4mm.txt and Case_1_Along_1025Hz_4mm.txt. Signals are plotted versus the coil movement on the left and in the impedance plane on the right. [5]

8 Figure 4. Experimental data obtained at the frequency of 1975 Hz and corresponding to files Case_1_Across_1975Hz_4mm.txt and Case_1_Along_1975Hz_4mm.txt. Signals are plotted versus the coil movement on the left and in the impedance plane on the right. Figure 5. Experimental data obtained at the frequency of 3025 Hz and corresponding to files Case_1_Across_3025Hz_4mm.txt and Case_1_Along_3025Hz_4mm.txt. Signals are plotted versus the coil movement on the left and in the impedance plane on the right. [6]

9 Test Case 2: Plate 5 mm thick with surface crack a) (b) Figure 6. ECT inspection problem: (a) 3D-view, (b) Top and side views. Figure 7. Experimental data obtained at the frequency of 1025 Hz and corresponding to files Case_2_Across_1025Hz_5mm_surface.txt and Case_2_Along_1025Hz_5mm_surface.txt. Signals are plotted versus the coil movement on the left and in the impedance plane on the right. [7]

10 Figure 8. Experimental data obtained at the frequency of 1975 Hz and corresponding to files Case_2_Across_1975Hz_5mm_surface.txt and Case_2_Along_1975Hz_5mm_surface.txt. Signals are plotted versus the coil movement on the left and in the impedance plane on the right. Figure 9. Experimental data obtained at the frequency of 3025 Hz and corresponding to files Case_2_Across_3025Hz_5mm_surface.txt and Case_2_Along_3025Hz_5mm_surface.txt. Signals are plotted versus the coil movement on the left and in the impedance plane on the right. [8]

11 Test Case 3: Plate 5 mm thick with subsurface crack (a) (b) Figure 10. ECT inspection problem: (a) 3D-view, (b) Top and side views. Figure 11. Experimental data obtained at the frequency of 1025 Hz and corresponding to files Case_3_Across_1025Hz_5mm_subsurface.txt and Case_3_Along_1025Hz_5mm_subsurface.txt. Signals are plotted versus the coil movement on the left and in the impedance plane on the right. [9]

12 Figure 12. Experimental data obtained at the frequency of 1975 Hz and corresponding to files Case_3_Across_1975Hz_5mm_subsurface.txt and Case_3_Along_1975Hz_5mm_subsurface.txt. Signals are plotted versus the coil movement on the left and in the impedance plane on the right. Figure 13. Experimental data obtained at the frequency of 3025 Hz and corresponding to files Case_3_Across_3025Hz_5mm_subsurface.txt and Case_3_Along_3025Hz_5mm_subsurface.txt. Signals are plotted versus the coil movement on the left and in the impedance plane on the right. [10]