Magnetic flux leakage NDE
In the MFL-NDE, magnetic flux leaks over from the surface
defects in the presence of a strong static magnetic field. The leakage
can be sensed by hall probes and the defect size and
shape can be quantitatively identified. The screenshot shows the
finite element simulation of this phenomena, which can estimate the
leakage flux given the geometry of the defect (the forward model). The
application allows one to study the effect of variations in material
properties, input variables like current, the defect geometry etc. in
the process. The flux leakage is shown as a intensity chart. Various
defect shapes can be simulated by interactively selecting elements from
the finite element grid.
- Overview of computational electromagnetic NDE forward models
(research at Center of Non destructive evaluation, IIT Madras [PPT])
Eddy current
NDE/ signal inversion
<>
This application is used as a reference defect signal generator model for eddy current NDE methods. The analysis involves a non-linear axisymmetric electromagnetic finite element formulation. The package includes a mesh generator and graphical interface for creating customized defects on the geometry for generating a variety of defect signals. We also have capabilities for simulating remote-field eddy current and pulsed eddy current NDE defect signals. We have employed this simulator for optimizing coil dimensions for flaw sizing in pipes and for addressing the WFNDE benchmark problems.
A study of literature pertaining to the use of eddy
current sensor for study of peened samples reveals the extensive use of
apparent conductivity versus the frequency curve. Simulations show that
apparent conductivities are not indicative of actual conductivity
gradients because of inherent constant conductivity approximation at
every frequency. This work focuses on facilitating the conversion of
the multi-frequency apparent conductivity data to the conductivity
gradients through a novel inversion scheme. The method does not depend
on the sensor coil parameters and is robust enough to accommodate for
the measurement uncertainties. My inversion software (shown above) uses
a multi-layer axisymmetric finite element model as the forward model
and uses an optimal skin depth approximation for isolating the integral
effects of the conductivity gradients on the multi-frequency apparent
conductivity measurements.
- Veeraraghavan S., Balasubramaniam, K., “A multi frequency eddy current inversion method for water peened specimens”, Review of Progress in Quantitative NDE-2003, Green Bay, Wisconsin, U.S.A. AIP Conf. Proc. 700, 651 (2004)[PDF]
- Sundararaghavan, K. Balasubramaniam," A multi-frequency eddy current inversion method for characterizing conductivity gradients on water jet peened components. NDT&E International Journal. Vol. 38(7), 541-547, 2005. (Figured in TOP25 articles in ScienceDirect) [LINK]
- Overview of computational electromagnetic NDE forward models (research at Center of Non destructive evaluation, IIT Madras [PPT])
Ultrasound imaging
A "Flock of Birds" position sensor was integrated with an ultrasound transducer to create a portable ultrasound scanner. Some of software component screenshots of my program are shown above. The scan data is recreated as a tomographic image automatically by updating the pixel data at the position of the scanner. This is done online and the warping of the data over the components tested can be done offline.
- Overview of this work at Michigan state university [PPT])