REMOVAL OF HARMONIC NOISE FROM SURFACE NUCLEAR MAGNETIC RESONANCE MEASUREMENTS IN THE FREQUENCY DOMAIN

摘要

Electromagnetic noise, particularly constant frequency sources such as powerlines, is widely recognized as an obstacle to making surface nuclear magnetic resonance (NMR) measurements. This study introduces a new method of removing harmonic noise by exploiting the frequency domain symmetry of surface NMR signals. Information on one side of the spectrum may be used to reconstruct other portions of the spectrum corrupted by frequency-domain noise peaks. This method supplements the existing NMR processing workflow and is applicable after despiking, coherent noise cancellation, and stacking. Synthetic and field data examples show that symmetry based noise removal reduces the effects of harmonics. In one modeling example, symmetry based noise removal improved signal to noise ratio in the data by 10%. This improvement had noticeable effects on inverted parameters including water content and the decay constant T2* . After harmonics are removed, aquifer boundaries and water content become more accurate and fewer spurious water content spikes appear within aquifers. This is especially useful for resolving multi-layered structures. Within T2* profiles, estimates are more accurate after harmonics are removed, especially in the lower half of profiles. This symmetry-based correction has several strengths including simplicity, a strong grounding in the mathematical theory describing NMR signals, no introduction of errors into the dataset, and no requirement of prior knowledge about the harmonics.