Physics-Based Characterization of UXO from Multi-Component TEM data

Newly developed transient electromagnetic (TEM) equipment collects four-dimensional data with three dB/dt components and time variation. Transient data from three receiver loops are recorded at 31 delay times, ranging from 1 to 1900 usec, and transients are recorded 32 times per second, generating high-density data sets. The equipment is flexible, transmitter and receiver loops can be reconfigured to optimize survey results based on expected target characteristics. With such large data sets and variable equipment configuration, modeling plays an important role. Inversion to dipolar models extracts physically meaningful target parameters from high-density TEM data.

Patches of TEM data near target anomalies are parameterized with the widely used anisotropic dipole model. Three orthogonal magnetic dipoles are used to represent target polarizability along three axes. Spatial variation of receiver loop dB/dt across the data patch is projected onto three, generally tilted target axes, compressing data from hundreds of measurement positions and multiple loop orientations into three polarizability values for each transient delay time. Time-dependent target properties are parameterized by fitting a transient-shape model to the polarizability transient for each of the three target axes.

Targets can be characterized by these physics-based model parameters. Transient shape is influenced by target size, shape, conductivity and permeability. Polarizability magnitude is proportional to target volume. Ratios of target-axis polarizability are indicative of target symmetry. We present application of our modeling routines to multi-component data obtained from a demonstration survey at NRL UXO test site located at the Army Research Lab facility at Blossom Point, MD.