New inspection approaches for railway based on Eddy Current More info about this article: http://www.ndt.net/?id=22713 J.L. Lanzagorta 1, J.-M. Decitre 2, F. Nozais 2, I. Aizpurua 1, R. Hidalgo-Gato 1, I. Castro 1, L.Lasa 3 1 IK4-IDEKO, Elgoibar (Gipuzkoa), Spain, jllanzagorta@ideko.es 2 CEA, LIST, F-91191 Gif-sur-Yvette, France. 3 CAF MiiRA, Beasain (Gipuzkoa), Spain. Abstract The inspection stages of railway axles, wheels and wheelsets are critical during the design and implementation of production lines and maintenance workshops. This is because the safety of railway components is a challenge for the railway industry. Taken this into account, Eddy Current technology is standing out as the core of new automatic inspection machines for the automatic surface integrity control of key components. However, the potential use of the Eddy Current inspection in the railway industry is a matter that has not been studied in detail yet. The objective of this paper is to evaluate the Eddy Current technology for the surface inspection of axles. For doing this, a modelling approach by CIVA is presented considering different strategies and new inspection concepts. The simulations have been validated with experimental tests that have demonstrated the advantages of the use of the Eddy Current technology during the surface inspections of the axles. 1. Introduction Assure the safety of components is the most important requirement in the railway industry. This is especially relevant in key components like wheels and axles in which the presence and grow of defects can produce the premature failure of the part. For this reason, non-destructive inspections are mandatory and one of the most important challenges in the railways both in manufacturing and in service. For the surface, the magnetic particle inspection (MT) is the standardized method applied from many years. This process is very well known (1) and relatively easy and practical for low-volume inspections. However, there are important disadvantages associated with MT that has to be considered. The most important problem is the limited probability of detection (POD) of the method, with a complete dependency on the operator for a successful detection of the defects. Additionally, this inspection is very time consuming, not easily automated and produces important impacts from an environmental point of view. In this direction, the railway sector aims to enhance the inspection steps of the process in order to replace old subjective inspection phases by new automatic systems that would yield objective data. Eddy Current method (2) is strongly gaining ground as a reliable, fast and practical method capable of testing complex shapes and high production volumes. In a standard eddy current test, a circular coil carrying current is placed in proximity to the test specimen (which must be electrically conductive). An alternating current in the coil generates a changing magnetic field which interacts with the test specimen and generates Eddy Currents. Variations in the phase and magnitude of these Eddy Currents Creative Commons CC-BY-NC licence https://creativecommons.org/licenses/by-nc/4.0/
can be monitored using a second 'receiver' coil, or by measuring changes to the current flowing in the primary 'excitation' coil. The presence of metallurgical changes like cracks in the material will cause a change in eddy current and a corresponding change in the phase and amplitude of the measured current (3). In the last years, a new technology called Eddy Current Array (ECA) is gaining interest for the detection of surface and sub-surface defects (4,5,6,7). The ECA technology is based on several eddy current coils configuring in a strategic way in the same probe assembly. This technology is used to build flexible sensors (7) with low sensitivity to lift off that allows full inspection of complex geometries (like railway wheels and axles) and it can be readily automated, enabling to have detailed, extensive records of each inspection done. In this work, the application of the ECA technology for axle surface inspection is discussed. An isotropic flexible ECA sensor (8) designed with CIVA is experimentally tested in a real railway axle with machined flaws at different orientations. The results show that the sensor is able to detect the defects independently of their orientations even when the lift-off effect is taken into account. 2. Experimental Experimental scans have been performed in the axle pattern (material E41T) showed in Figure 1 a). The axle has been inspected with an Isotropic ECA sensor (8) connected to the ECTANE 2 test instrument. The ECA scanning parameters are defined in Table 1. Three different zones have selected in the axle for the inspections, see Figure 1 b). The zone A corresponds to the flat region of the axle, the Zone B to the fillet and the zone C to the transition. In each zone, several groups of flaws have been mechanically produced with different sizes and orientations as described also in Table 1. Each group of cracks have different orientations with respect to the principal axis of the axle. The zone A has been inspected semi-automatically with a mechanical arm moving the sensor in the direction of the principal axis of the axle. The zones B and C has been inspected manually in the radial direction. The lift-off effect has also been studied adding tapes up to a total thickness of 0.22 mm. For designing the pattern of the sensor, the last version of the software CIVA 2017 has been used for the simulations. A B C (a) (b) Figure 1. Axle pattern with machined cracks inspected during the experimental tests (a) and the experimental set-up (b). 2
Table 1. Experimental set-up Scanning Nominal lift-off (mm) Probe drive Frequency (khz) LO (V peak) 0.06 mm 1000 1.5 V Flaws: Parallelepiped notch Zone Length (mm) l Width (mm) w Depth (mm) d Orientation with respect to the principal axis of the axle 2 A 0 10 45 70 90 0.3 0.25 B 1 0 45 70 90 C 0 10 45 70 90 3. Results and discussion The proposed isotropic pattern, optimized by CEA, is composed with 3 coaxial coils, see Figure 2. A transmitting coil is etched on a first layer of the kapton film and two receiving coils are etched on the other layer of the film. These two receiving coils are connected in differential mode. Figure 2. Isotropic pattern using in the CIVA simulations. Following a principle of optimization described in (8), it is possible to adjust different internal and external diameters of the three coils in order to obtain good behaviour towards the lift-off. Furthermore, it is interesting to obtain a very small signal in absence of defect, as a classical differential sensor. Figure 3 gives a cartography obtained by simulation with the isotropic optimized pattern for a 2mm long defect. The external diameter of emitting coil is 2.1mm. At right, it is the vertical signal on transverse defect for nominal lift-off (e = 50μm, black line) and for higher lift-off (e = 100μm, red line). Note that both curves are almost superposed due to the weak influence of the lift-off on EC signals. 3
Figure 3. Simulation results with isotropic pattern for 2mm long defects. The results of the inspections at different zones are showed in Figure 4. The experimental data are obtained with a flexible EC array with 48 identical isotropic patterns. The pitch between two consecutive elements in the C-Scan is equal to 700µm. This pitch is sufficient for detection of all notches, but it would be possible to reduce the pitch in order to obtain a better estimation of the amplitude and shape on each defect. For the A zone, the C-Scan shows that the sensor is able to detect 2 and 1 mm defects independently of the orientation of the crack see Figure 4 a) and b). The effect of lift off does not affect the detectability of the defects as can be seen on Figure 4 c). The same results are obtained in the fillets (zone B) and in the transition region (Zone C) for flaws with 1 mm in length. The amplitude of the signal and the signal noise ratio (SNR) are also represented in Figure 4. The amplitude is not perfectly identical on all notches, notably on the longitudinal notches, because they are not necessarily positioned just above an element. Zone A, l=2 mm, Lift off = 0.06 mm 0,5V 1 0,7V 1,2V 1,1V 1,1V 20dB Zone A, l=1 mm, Lift off = 0.06 mm (a) 27d B Zone A,l=1 mm, Lift off = 0.22 mm 0,9V 25dB 1 0,64V 1,2V (b) 1,26V 1,33V Zone B, l=1 mm, Lift off = 0.06 mm 0,12V 7dB 1 0,16V 10dB 0,27 V 14dB 0,3V 15dB 0,3V 15dB (c) Zone C, l=1 mm, Lift off = 0.06 mm 1,2V 1,27V 1,1V 1,29V 0,79V 24dB 0,74V (d) (e) Figure 4. C-Scans obtained in the different inspections. 0,48V 20dB 1 0,69V 0,85V 25dB It will be possible to use a smaller pitch to obtain more homogeneous amplitude. As some coils are connected in serial, it is possible to obtain signal due to a coupling 4
between elements. This could be reduced using a higher number of time-slots or channels. The SNR has been represented in function of the orientation of the notch in Figure 5. When the nominal lift off (0.06 mm) is used, the SNR is higher than 20dB in all of the studied cases. The increase of the lift off to 0.22 mm produces a drop of the SNR to values closed to 10-15 db. However, the SNR is always higher than 3 db independently of the lift off, inspection zone, length and orientation of the notch. Therefore, the experimental results validate the predictions of the CIVA model. The signal resulting from the isotropic sensor is independent of the orientation of the crack and it has weak influence to the lift-off on the EC signals. Figure 5. SNR at all of the studied configurations. Automatic systems for the inspection of railway axles could be produced using the type of sensor prototype of the present study. One of the most important features of this sensor is the excellent lift behaviour that it provides. In fact, this weak dependency on lift off has been the result of the optimization method based on simulation tools integrated in the platform CIVA (9,10). Good lift-off control is crucial in order to automatize the inspection technique, because it increases the detectability, greatly reducing the occurrence of false positives and increasing the self-life of the sensor. These could be the base for designing a new generation of machines for railway axle inspection. The flexibility and adaptability of the sensor is another important characteristic in order to guarantee the correct inspection of the fillets in which the POD has to be maximized. The isotropic behaviour of the sensor is also a key aspect in order to detect all possible flaws, regardless of their relative orientation on the axle. Eddy Currents Array technology are now considered in a new standard published in 2017 (11), but the current normative used in the surface integrity inspection of axles during production only refers to MT (12,13). However, taking into account the requirements and acceptance criteria of the actual standards with respect to surface integrity inspection, the isotropic sensor used in this work fulfils these requirements and could be the base for a new generation of machines for railway axle s inspection. 5
4. Conclusions CIVA simulation models have been developed to design Eddy Current probes. The isotropic flexible Eddy Current pattern has been optimized to obtain a weak sensitivity to lift-off and good SNR. A flexible array probe based on such an elementary pattern has been designed and manufactured. With this isotropic sensor, different inspections have been performed in a real railway axle section, with cracks machined at different orientations. The sensor is able to detect the target cracks independently on the size, orientation and axle s zone. This clearly demonstrates that this sensor is appropriate for detecting omnidirectional cracks with a minimum size of 1 mm in railway axles even in the fillets. Further research work is necessary to demonstrate that the inspection could be done automatically in order to suit the specific requirements of railway axles. Acknowledgements This work has been partially done in the framework of the Proyecto AxIS - Tecnologías avanzadas de Alta fiabilidad para maximizar la vida útil, seguridad y disponibilidad de los vehículos ferroviarios para su servicio. RETOS- COLABORACIÓN 2016 (RTC-2016-4813-4). References 1. D. Lovejoy, Magnetic Particle Inspection: A Practical Guide. David Lovejoy: Springer, 1993. 2. Janousek L, Capova K, Yusa N, Miya K. Multiprobe inspection for enhancing sizing ability in eddy current nondestructive testing. IEEE Trans. Magn. 2008;44:1618 1621. 3. Hashizume H, Yamada Y, Miya K, Toda S, Morimoto K, Araki Y, Satake K, Shimizu N. Numerical and experimental analysis of eddy current testing for a tube with cracks. IEEE Trans. Magn. 1992;28:1469 1472. 4. M. Wright, Eddy Current Array Technology, 1st Edition 2014, Eclipse Scientific Products Inc. 5. Gilles-Pascaud C., Vacher F., Decitre JM., Cattiaux G., EC Array Probe Development For Complex Geometries, 5th International Conference on Non Destructive Evaluation, San Diego, July 2006. 6. Marchand B, Decitre J-M, Casula O. Innovative flexible eddy current probes for the inspection of complex parts, World Conf. on NDT, 2012. 7. J.L.Lanzagorta1, R. Leclerc, E. Grondin, L.Lasa, A. Landaberea, X.Alzaga, Eddy Current Array technology for the surface inspection of railway axles 18th International Wheelset Congress. November 7-11,2016 Chengdu, China. 8. J-M. Decitre, Optimization of Flexible Eddy Current Patterns with Low Sensitivity to Lift-Off, Studies in Applied Electromagnetics and Mechanics, Volume 42, 2017, 275-281 9. Extende, CIVA 11 - The simulation platform for NDE, NDT.net, 2013. 10. Reboud C. and Theodoulidis T., Field computations of inductive sensor with various shapes for semianalytical ECT simulation, Studies in Applied 6
Electromagnetics and Mechanics, Electromagnetic Non-Destructive Evaluation (XV), 2012. 11. ISO 20339:2017: European Standard : Non-destructive testing -- Equipment for eddy current examination -- Array probe characteristics and verification 12. IS06933-1986(E): International Standard: Railway rolling stock material Magnetic particle acceptance testing. 13. EN 13261:2009: European Standard: Railway applications- Wheelsets and bogies- Axles-Product requirements. 7