Noise analysis and cancellation for the underground application of magnetic resonance using a multi-component reference antenna – Case study from the rock laboratory of Mont Terri, Switzerland

Abstract

Because of their principle capability to non-invasively detect dangerous water bodies and water-filled decompaction zones, underground magnetic resonance sounding (MRS) measurement are expected to have a great potential for the safety of underground facilities. However, a general applicability of underground MRS is questionable. The small antenna sizes reduce the sensitive volume and thus the signal-to-noise ratio. Additionally, the receivers must be installed in the nearest vicinity (i.e. reactive near field) of electromagnetic (EM) noise-emitting installations necessary for the logistics and safety inside the tunnels. This study focusses on the characterisation of underground EM noise characteristics including their spatial and temporal variability as well as proper strategies for noise cancellation. Passive EM measurements using conventional MRS equipment were conducted in the rock laboratory of Mont Terri, Switzerland. The reference-based noise cancellation (RNC) method yielded the best performance, given that all orthogonal spatial components of the noise field were considered and combined on the remote multi-component reference antenna. The multiple coherence between the primary and the reference channels is highest and stable against temporal and spatial noise variations, if the reference antenna combines square and figure-of-eight coils. Using a primary figure-of-eight coil, final noise levels of <0.1 nV/m2 were achieved after a common averaging time. An MRS noise level at this magnitude allows the detection of non-clay-bound and therefore dangerous water deposits inside the shale formation. This is demonstrated with data from a surface MRS experiment in an urban test area, which exhibits similar noise characteristics as the Mont-Terri rock laboratory. Under such conditions, the underground MRS method would be able to provide at least a yes/no criterion for further activities to maintain on-site safety. Future research should focus on strategies for a data acquisition that allow at least a rough 3-D mapping of potential water reservoirs.