The overall objective of this project was to design, build, and demonstrate an underwater advanced time-domain electromagnetic (TEM) system for cued classification of munitions in the underwater environment. The phased approach consisted of initial design and modeling (Phase 1), engineering design and construction (Phase 2), underwater evaluation of the system in a freshwater pond (Phase 3), and demonstration of the system at a saltwater site (Phase 4).
The technology designed and demonstrated was an underwater version of the Geometrics MetalMapper system, which has been well described, validated, and demonstrated through several SERDP and ESTCP projects. The project’s design consisted of:
A diagram of the system, as tested in the freshwater pond, is provided as Figure 1.
As in the 2016 Freshwater Demonstration at Panama City, background-subtracted target data were inverted to estimate target polarizabilities using the UX-Analyze dipole fit algorithm. Malfunctioning data channels were not included, and the first 18 timegates (t < 0.132 ms) were not used. The calculated polarizabilities were compared with free-air polarizabilities using the UX-Analyze classification algorithm.
Table 2 presents all Saltwater Demonstration Test Area results for the eleven test items, including:
The effects of test item size and test item location may be seen in the results. Larger test items and test items closer to the center of the EM system array generally had higher fit coherence and library match metrics. Test item location #4 (at the outer edge corner of the EM system array) had the lowest average fit coherence and library match metrics for all items (0.863 and 0.581, respectively). Positioning uncertainty of EM system to the emplaced test item applies to all test item locations.
To maintain some consistency with performance of the standard TEMTADS 2x2 and Metal Mapper (which have 48 data channels), the tabulated SNR is the average over the strongest 48 channels in the EM system. Fifty-eight of the 66 target measurements (88%) have fit coherence >0.8. This is suspected to be a signal to noise issue. Fifty-three of 53 (100%) measurements with SNR > 30 have fit coherence >0.8. Only 40 of the targets (61%) have a library match metric >0.9. Restricted to targets with SNR >30, the percentage of targets with a library match metric increases to 74% (39 of 53). There is no indication that having the array in seawater affects its classification performance. By way of example, the figure below shows 60mm polarizabilities measured in seawater (symbols) and the corresponding free-air library polarizabilities. For this target the SNR is 162, the fit coherence is 0.9989, and the library match metric is 0.95.
The underwater EM system demonstrated is custom-built by Geometrics. The hardware and software are based on the commercially-available Metal Mapper. Though the demonstration was successful, no subsequent demonstrations are planned to progress the EM system from a prototype to commercially available equipment.
Data collection times and transit times between cued data collection locations are minimal compared to the time required to launch and recover the EM system. Improvements to deployment methods will improve overall efficiency, production rates, and cost effectiveness.
Vessel navigation and dive operations for underwater cued Advanced Geophysical Classification (AGC) data collection are significantly different from traditional underwater geophysical operations (e.g., Digital Geophysical Mapping), and require specialized skills and equipment. For example, real time data review is best implemented with continuous communication between a topside Data Team and the Dive Team.
Divers must be able to safely transport the EM system between cued data collection locations, typically in low visibility settings. The size and weight of the EM system in its current configuration requires two divers to transport with both hands. Reducing the number of transmitters and receivers in the EM system array is possible and would reduce its dimensions and drag, but would also reduce the effective footprint of the cued data collection and the ‘positioning error budget’ for effectively illuminating the subsurface item. This may be a permissible reduction in EM system performance if positional data are incorporated into cued AGC data collection.
The Lighter, Amphibious Resupply, Cargo (LARCs) used in the Saltwater Demonstration are larger than necessary and are the property of U.S. Army Corps of Engineers Field Research Facility (FRF). Smaller vessels can deploy the current EM system and vessel requirements must be considered for future underwater EM systems. Commercially-available vessels (e.g., pontoon boat) could be used or modified to launch and recover the EM system.
The finite lengths of the divers’ surface supplied air umbilical cords and the EM system’s data cables limit the workable configurations between the two LARCs, the divers, and the topside data acquisition computer. The distance between the LARCs must be far enough to prevent collision, but less than the length of the umbilical cord so that the divers can assist with launch and recovery. The data acquisition computer, whether on a vessel or fixed (e.g., on the FRF pier during this demonstration), must be close enough to the divers and EM system to not stretch or break the data cable (currently 70m). This may be improved by placing the data acquisition computer on a vessel.