Wuliang Yin

A planar EMT system for the detection of faults on thin metallic plates

This paper describes a novel planar electromagnetic tomography system for the detection of conductivity inhomogeneity on a metallic plate. The proposed system differs from traditional electromagnetic inductance tomography (EMT) systems in its spatial arrangements of coils. Sensor coils are distributed to form a circular array with their axes not parallel but perpendicular to the plate under inspection. The forward solution for the sensor array next to a homogeneous conductive plate is based on the analytical solution provided by Cheng. The sensitivity matrix for a prototype sensor was computed by numerical evaluation of the analytical solution. For the inverse solution, a modified Newton?Raphson method was used to adjust the conductivity distribution to fit a set of inductances measured from the sensor array in a least-squared sense. Frequency- dependent sensitivity analysis was performed to find an optimum testing frequency. The far-field and near-field effects in electrical tomography are discussed. Good estimates for the conductivity distribution were obtained at the optimum frequency. Experimental tests were performed by taking the difference in mutual inductance of the coil pairs when placed next to a homogeneous reference conductor and next to a conductor with faults. Inverse results based on experimental data verified this method.

Simultaneous measurement of distance and thickness of a thin metal plate with an electromagnetic sensor using a simplified model

This paper presents a simplified model which can describe the inductance change when an air-core coil is placed next to a thin nonmagnetic metallic plate. The model has two independent parameters and is valid for a range of thickness, conductivity, and lift-offs. Use of this new relationship provides a fast and accurate method to measure the distance and thickness simultaneously. Measurements made for a sample coil next to thin copper and aluminum plates of various thicknesses verified the theory and the proposed method.

Noncontact Characterization of Carbon-Fiber-Reinforced Plastics Using Multifrequency Eddy Current Sensors

The characterization of carbon-fiber-reinforced plastics (CFRPs) using multifrequency eddy current sensors is presented in this paper. Three sensors are designed for bulk conductivity measurements, directionality characterization, and fault detection and imaging of unidirection, cross-ply, and impact-damaged CFRP samples. Analytical and finite-element (FE) models describing the interaction of the sensors with the CFRP plate samples are developed to provide an explanation of, and physical insights into, the measured results and observed phenomena. A signal processing method is developed to compensate for the variation in lift-off during the measurements.

Simultaneous Non-contact Measurement of Water Level and Conductivity

Measurement of water level and conductivity is important in many applications, including geophysics, ocean engineering, and the oil industry, with one example being gas/oil/water separation. Noncontact measurement is ideal for those applications. Capacitance, radar, ultrasonic, and laser techniques have been used in measuring the level, but simultaneous measurements of level and conductivity have not been reported. In this paper, we developed a novel inductive sensor, which uses two coils of different sizes to simultaneously infer conductivity and level based on a simplified model deduced from an analytical solution. The simplified model is verified with both simulation and experimental data. The experimental results have shown that the measurement accuracy for both level and conductivity is within 3%.

Analysis of the Lift-off Effect of Phase Spectra for Eddy Current Sensors

This paper presents an analytical model that describes the inductance change when a double air-cored coil sensor is placed next to a conducting plate. Analysis of the analytical model reveals that the phase signature of such a sensor is virtually liftoff independent. This finding is verified by numerical evaluations. This paper also finds that the phase signature of a ferrite U-cored sensor can be approximated by that of a double air-cored sensor of similar size and, therefore, possesses a similar liftoff-independent property. Measurements made with a sample U-cored sensor next to plates of nonmagnetic and magnetic materials verified the theoretical results.

A pre-iteration method for the inverse problem in electrical impedance tomography

In this paper, the relationship between Landweber iteration and the generalized inverse has been analyzed, and it is shown that Landweber iteration is, in fact, a modification of the generalized inverse that is implemented by using the iteration solution. At the same time, a new reconstruction algorithm, termed the pre-iteration method, is proposed, which splits the image reconstruction process into two steps that are the off-line pre-iteration and the on-line one-step reconstruction. The speed of image reconstruction is remarkably increased.

Imaging the continuous conductivity profile within layered metal structures using inductance spectroscopy

This paper presents an inverse method for determining the conductivity distribution of a flat, layered conductor using a multifrequency electromagnetic sensor. Eddy-current sensors are used in a wide range of nondestructive testing applications. Single-frequency sensors are very common; however, the potential of an eddy-current sensor with spectroscopic techniques offers the ability to extract depth profiles and examine more fully the internal structure of the test piece. In this paper, the forward solution for a small right-cylindrical air-cored coil placed next to a layered conductor is based on the analytic solution provided by the transfer matrix approach. For an inverse solution, a modified Newton-Raphson method was used to adjust the conductivity profile to fit a set of multifrequency inductances in a least-squared sense. The approximate Jacobian matrix (sensitivity matrix) was obtained by the perturbation method. Numerical results of the forward solution are provided for cases of step, continuous conductivity profiles. Good estimates for the conductivity profile were obtained. Experimental eddy-current tests are performed by taking the difference in inductance of the coil when placed next to a reference conductor and next to a layered conductor over the range 100 kHz - 1 MHz. Inverse results based on experimental and simulated data verified this method.

Thickness Measurement of Metallic Plates With an Electromagnetic Sensor Using Phase Signature Analysis

This paper presents a simple model that can predict the phase signature of eddy current sensors over nonmagnetic metallic plates. In contrast to the magnitude of the eddy current signal, which generally significantly decreases with an increase in liftoff, the phase signature appears to be virtually immune to liftoff variations and, therefore, can be more reliably used in field eddy current testing. A parameter that encapsulates the coil geometry factors can be obtained through calibration with samples of known properties. This geometry parameter can then be used in the simple model to predict the phase signature of this sensor over any plate with any thickness. The use of this new model provides a fast and accurate method to measure plate thickness. Measurements and finite-element (FE) simulations made for two sample coils next to copper and aluminum plates of various thicknesses verified the theory and the proposed method.

The Design of a Digital Magnetic Induction Tomography (MIT) System for Metallic Object Imaging Based on Half Cycle Demodulation

This paper presents several important aspects of a highly integrated, Field Programmable Gate Array (FPGA)-based digital Magnetic Induction Tomography (MIT) system for metallic object imaging applications. Excitation signal generation, receiving signal demodulation, and channel multiplexing control are all implemented inside a Xilinx FPGA (Spartan III). A novel digital demodulation method using only half of the signal cycle is proposed and implemented, which improves the operation speed of the system by two folds. A geometrical scaling relationship has been discovered for MIT sensors, and a particular design has been implemented. Two imaging reconstruction algorithms (Tikhonov Regularization and Total Variation method) are applied to experimental data. The performance of the system has been verified.

The Development of a Multi-frequency Electromagnetic Instrument for Monitoring the Phase Transformation of Hot Strip Steel

This paper describes the development of an instrument to analyze the phase transformations of hot strip steel using an electromagnetic sensor. The sensor exploits variations in the electrical conductivity and magnetic permeability of the steel to monitor microstructure evolution during processing. The sensor is an inductive device based on an H-shaped ferrite core, which is interrogated with a multifrequency impedance analyzer containing a digital signal processor. Online fast Fourier transform was performed to abstract the multifrequency inductance changes due to the microstructural evolution of the sample. An overview of the instrument and measurements from a range of carbon steel samples are presented. The results verify the ability of the instrument both to monitor the microstructural changes and to reject variations in liftoff distance between the sensor and the hot strip. It is believed that this is the first time this result has been reported with tests on hot steel samples