The objective of this project was to perform numerical analyses and experimental studies to obtain a better understanding about using marine EMI systems to detect and characterize submerged buried munitions.

Technical Approach

The response from a metallic body immersed in a conductive medium is a combination of the eddy current response (ECR) due to currents generated in the target and the galvanic coupling of currents through the body (the current channeling response, CCR). In terrestrial environments only the ECR is important because conduction currents can be ignored in most soil types (conductivity ∼ 0.01 S/m). The measured magnetic fields in the low-frequency EMI regime can be modelled as if the object were in free-space. In contrast, typical seawater environments have a conductivity of approximately 4 - 5 S/m.

For the numerical studies, researchers developed an integral equation for a layered medium that could account for the changing conductivity of the air, water and seabed. The model was extended so researchers could compute both the background and scattered field response from a highly conducting and permeable sphere for dipole and loop transmitters and receivers. They carried out a series of synthetic experiments by considering various factors that might influence EMI signals, including current channeling effects, sea depth, the size of a loop, lateral offset of the receiver, host conductivity, excitation waveform and antenna insulation.

For the marine measurements, researchers built a 2m x 1m x 1m fiberglass frame and encased two receiver cubes in waterproof epoxy. The transmitter loop comprised 12 turns of wire arranged in a 2m by 1m rectangle. A 24V power-supply provided a maximum current of 11.4 amps using a 25 Hz base-frequency with a 50% duty cycle waveform. A series of different measurement campaigns were conducted in sea-water depths of between 2 - 14 m either from a boat or by wading into water adjacent to a sandy beach.


Decay times:  The simulations showed that the CCR from a highly conducting and permeable sphere embedded in the air-sea-sediment is far smaller than the ECR, and decays much faster than the ECR at a rate of t−3. For the interesting time range of 0.1 ms - 25 ms, the CCR contributes little to the target signals and thus can be ignored. At times beyond several hundred microseconds, the ECR response approaches the value for the same object embedded in free space. These numerical observations were confirmed by measurements of an insulated and non-insulated 105 mm projectile at a range of different receiver, transmitter and object offsets.

Depth effect:  When considering a survey close to seafloor, researchers found that the decay rate of the background response is affected by the sea depth, or equivalently by the distance of the sensor from the air-sea interface. The results show that the background response in shallow water decay faster than in a deeper water. In deeper water where the sensor is far away from the air-sea interface, the corresponding background responses asymptotically approaches the response of a half-space. The simulations demonstrated that sea depths do not impact the scattered field response from a buried metallic object. Measurements conducted in water depths of 2 - 14 meters displayed the same general characteristics as the simulations with the response in shallower water (2 m) falling off faster than the measurements at 14 meters. Observed decay rates were between t−5/2 and t−3.

Background Response:  The background responses measured inside and outside of a transmitter loop are very similar and at later times vary little with offset distance from the transmitter center. These aspects of the offset background measurements might be usefully applied to calibrate a system that has an array of sensors.

Geometry:  EMI responses depend upon other geometry survey parameters like the transmitter loop size and stand-off distance. A large transmitter loop can excite a strong scattered response from a buried object but also induces a larger background response. The latter might be removed by subtraction of the fields from widely separated receivers. To test this type of background suppression method researchers considered two possible configurations of receivers: vertical and radial with respect to the transmitters. The experiments showed that the vertical offset configuration can be effective at removing the background signals but only at later times. On the other hand, the results with radially symmetric receivers showed that the differential total responses are identical with the differential scattered responses for all components across all time range. The radial receiving configuration provides an effective way to obtain the scattered response purely from free space models.

Dispersion:  The conductive sea-water medium causes dispersion of electromagnetic fields and can delay and distort the primary and scattered fields. For the primary field the simulations showed significant distortion effects for an ideal rectangular waveform particularly at large offset distances. However, when the exponential rise time of the transmitter loop was considered, the distortion effects became much less significant and can be ignored for the time-ranges and offset distances typically encountered in UXO sensing. These theoretical predictions are supported by flux-gate magnetometer measurements of the primary field distortion.

Pulse Duration:  During the off-time, field dispersion slows as host conductivity increases. Results with a simulated marine TEMTADS system showed that the scattered responses excited from a long pulse of 10 milliseconds are stronger than those from a short pulse of 1 millisecond, in particular for late times. This demonstrates one method to further enhance target signals, which is equally effective and equivalent in the terrestrial realm. Further tests with simulated marine TEMTADS data showed that a terrestrial dipole polarizability model could be used to successfully characterize the scattered data from a buried metallic object in the conducting underwater environment.

Spectral Response:  A concluding set of measurements focused on the spectral content of noise and signals in the underwater environment. Researchers used a 24-bit analog to digital convertor to sample a receiver cube at a rate of 120 kHz. Measurements were conducted on land immediately adjacent to marine measurements collected at a water depth of four meters. The underwater measurements had lower noise across most of the frequency spectrum compared to the on-land measurements. When the transmitter was operating and sufficiently damped by a resistor, the spectrum and off-time behavior on-land and in-water were virtually identical. From these results one can conclude that there are no significant frequency content differences between on-land and in-water.


The results of this limited scope project suggest that underwater detection and characterization of buried metal using EMI based sensor is feasible. Although conductive sea-water introduces complications to the measurement process, there are numerous practical options to mitigate these effects. The conductive sea-water can impact the scattered field from a buried metallic object, but typically only at very early times and at large receiver-to-object offsets. Neither of these conditions are commonly encountered in practice. Thus, apart from the practical considerations of operating underwater, the modelling techniques and methods that have been successfully demonstrated in terrestrial environments can be utilized for marine detection and characterization. The principle remaining challenge is the development and deployment of a practical and effective set of hardware for marine EMI sensing.