Pierre-Yves Joubert

Localization of defects in steam generator tubes using a multi-coil eddy current probe dedicated to high speed inspection

Steam generator (SG) tubing of pressurized water reactor in nuclear plants must be rapidly and accurately checked in order to detect defects in their early stages. In this paper, the authors present a multi-coil eddy current (EC) probe allowing both high speed inspection and circumferential localization of defects in the tube wall. A method of multi-coil EC signal processing, based on a continuous wavelet transform combined with a maximum likelihood diagnosis, is elaborated in order to enhance the detection performances and to provide automatic localization of defects. The inspection of SG tube samples shows good localization performances for defects as small as 10% deep, 15 mm long and 100 μm wide outer diameter notches, of both circumferential and axial orientations.

Wall thickness evaluation of single-crystal hollow blades by eddy current sensor

Advanced high-pressure turbine blades of jet engines are hollow and monocrystalline. The external wall thickness of these blades has to be checked systematically and quickly after manufacturing in order to guarantee the blade strength. Thickness evaluation is made difficult by the presence of internal partitions and by the crystalline anisotropy of the superalloy used in blade manufacturing. In this paper, the authors present the advantages of the eddy current (EC) technique in comparison to other non-destructive evaluation (NDE) techniques for wall thickness evaluation. A dedicated EC sensor was developed and implemented. The thickness evaluation was carried out with a neural network inverse model, and the results show the high accuracy and reliability of the proposed method.

Analytical and Numerical Analyses of a Current Sensor Using Non Linear Effects in a Flexible Magnetic Transducer

A theoretical study and a simulation method are proposed for superparamagnetic current sensors implementing a uniformly wound toroidal core topology. So as to be easy to implement, this sensor topology can be made ∞exible thanks to the use of a core made up of a superparamagnetic powder embedded in a ∞exible plastic matrix. The measurement of DC and AC currents is possible provided that a sinusoidal magnetic fleld excitation is applied to the superparamagnetic transducer. An analytical model is proposed for computing the sensor output signal and we demonstrate that when the detection of the component at the second order harmonic of the excitation frequency is used, the measurement is independent of the conductor position in a given current range. For simulating the dynamic response of the sensor, we propose to combine the analytical model, or a flnite elements model, with a time-discretization method. Furthermore, simulations are carried out considering a ring shaped sensor and the real magnetization characteristics of a superparamagnetic material. Simulations are provided over the (i10kA 10kA) range and for various amplitudes of the excitation signal. The results obtained with the analytical model, which is computationally e-cient, are within 4% to 12.7% from the numerical results.

Study for the non-contact characterization of metallization ageing of power electronic semiconductor devices using the eddy current technique

The ageing of the metallization layers of power semiconductor dies may be the cause of failure of power semiconductor modules. Usual indicators of failure like on-state voltage drops make it difficult to highlight the deterioration of the metallization layer. In this study, we evaluate the relevance of the characterization of power device metallizations by means of the eddy current sensors. Experimental results show the ability to monitor the state and the evolution of the metallization ageing with such a technique.

Pulsed eddy current imaging device for non destructive evaluation applications

In this paper, an original pulsed eddy current (EC) imaging system dedicated to the enhanced non-destructive evaluation (NDE) of large metallic parts is presented. The system aims to combine the advantages of the pulsed eddy current (PEC) sensors with those of EC imaging devices. The system is based on the combination of a PEC inducer, and a magneto-optical (MO) set-up used to measure the magnetic two-dimensional impulse response of the part under evaluation. The obtained data are in the form of a temporal succession of high resolution EC images, enabling both the spatial and temporal EC characterization of the part. The system is implemented here on aluminum assemblies featuring calibrated defects, and enables the rapid characterization of defects according to their position, dimension and depth within the assembly, thank to the obtained spatio-temporal EC data. The presented system opens the way to a new generation of EC NDE imaging devices.

Neel Effect Toroidal Current Sensor

In this paper, a configuration for a Neel effect ac-dc current sensor designed with second harmonic sensing in view is proposed. This configuration relies on the use of two toroidal superparamagnetic cores wound in opposite direction with respect to each other and connected in series. An analytical model is developed, which shows that the component of the output signal at the excitation frequency, which is useless, should be removed provided that the windings are symmetrical. The latter principle is validated experimentally and a prototype device is characterized for different excitation conditions. Open loop dc and ac current measurements are carried out over the [-110 A, 110 A] range.

Regularized inversion of a distributed point source model for the reconstruction of defects in eddy current imaging

Purpose – The inverse problem in the eddy current (EC) imaging of metallic parts is an ill‐posed problem. The purpose of the paper is to compare the performances of regularized algorithms to estimate the 3D geometry of a surface breaking defect.Design/methodology/approach – The forward problem is solved using a mesh‐free semi‐analytical model, the distributed point source method, which allows EC data to be simulated according to the shape of the considered defect. The inverse problem is solved using two regularization methods, namely the Tikhonov (l2) and the 3D total variation (tv) methods, implemented with first‐ and second‐order algorithms. The inversion performances were evaluated in terms of both mean square error (MSE) and computation time, while considering additive white and colored noise, respectively, standing for acquisition errors and model errors.Findings – In presence of colored noise, the authors found out that first‐ and second‐order methods provide approximately the same result according ...

A novel modeling of surface breaking defects for eddy current quantitative imaging

A simplified electromagnetic modeling is proposed for the eddy current imaging of surface breaking defects. The model assumes that the interactions between the eddy currents and a defect are equivalent to current sources placed inside the defect. The use of the distributed point source method is thus suitable for implementing the model. An example is provided considering the case of aeronautical fastener holes. The results given by the proposed model are confronted with experimental data obtained with an eddy currents imaging probe as well as with finite elements modeling simulations and good agreement is shown.

Enhancement of open-crack detection in flying-spot photothermal non-destructive testing using physical effect identification

This paper considers the images provided by a photothermal camera (flying-spot camera) dedicated to open-crack detection. In this type of active thermography, both thermal and optical effects contribute to the elaboration of photothermal images. Here the thermal effect is relative to the presence of open-cracks and the optical effects are due to surface conditions. In the case of open-cracks detection, the optical effects induce high magnitude perturbation signals, possibly masking the presence of open-cracks. In this contribution a signal processing method is proposed in order to identify both thermal and optical effects separately. The method lies uses multiple principal component analysis combined with a continuous wavelet transform. It is used to enhance the open-crack detection for the inspection of an industrial mock-up showing open-cracks and various surface conditions. The enhancement of the detection performance is characterized thanks to Receiver Operating Characteristic curves. The proposed method shows high detection performances and could be extended to a classification scheme.

A Differential DPSM Based Modeling Applied to Eddy Current Imaging Problems

This paper deals with an innovative implementation of a semi-analytical modeling method, called the Distributed Points Source Method (DPSM), in the case of an eddy current problem. The DPSM has already shown great potentialities for the versatile and computationally efficient modeling of complex electrostatic, electromagnetic or ultrasonic problems. In this paper, we report a new implementation of the DPSM, called differential DPSM, which shows interesting prospects for the modeling of complex eddy current problems such as met in the non-destructive imaging of metallic parts. In this paper, the used eddy current imaging device is firstly presented. It is composed of an eddy current (EC) inducer and a magneto optical set-up used to translate the magnetic field distribution appearing at the surface of the imaged part, into a recordable optical image. In this study, the device is implemented for the time-harmonics (900 Hz) imaging of a two-layer aluminum based assembly, featuring surface-breaking and buried defects. Then, the basics of the time-harmonics DPSM modeling are recalled, and the differential approach is presented. It is implemented for the modeling of the interactions of the eddy current imaging device with the considered flawed assembly in the same operating conditions as the experimental implementation. The comparison between experimental and computed data obtained for millimetric surface and buried defects is presented in the form of complex magnetic cartographies and Lissajous plots. The obtained results show good agreement and open the way to the modeling of complex EC problems. Furthermore, the low computational complexity of the differential DPSM modelings makes it promising to consider for the solving of EC inverse problems.