Mining Publication: Adaptive-Noise-Cancellation Techniques for Through-the-Earth Electromagnetics: Volume III

An electromagnetic (EM) system for the location of trapped mine workers at coal-mine (300 m) depths has been developed. The use of such a system in deep mines (1-km depth) requires transmission at extremely low frequencies (ELF) and a reduction in the effective noise level beyond that which can be obtained by acceptable integration times. Automatic-noise-cancellation (ANC) techniques reduce the effective noise level by combining several received signals that contain correlated noise. Such signals can be obtained either from a remote magnetic antenna or a local electric antenna. During Phase III of this program, real ELF noise data were acquired and used to test the ANC algorithms developed during Phase II. The ELF noise data were obtained from analog recordings made by Develco under another Bureau of Mines contract. Digitization was accomplished by an A/D converter that was added to a GMRR computer. Phase III tasks included preparation of software for digitization, computer-computer communication, data processing, and statistical analysis. Useable data were obtained from eleven half-hour tapes containing noise from one axis of the primary antenna and three axes of the remote antenna. The overall average improvement in SNR was 4.1 dB. However, the average improvements for the individual X, Y, and Z axes were 4.2, 11.9, and -1.9 dB, which roughly follow the observed noise power. Nonlinear processing was not effective with this noise.