The application of Gaussian distribution deconvolution method to separate the overlapping signals in the 2D NMR map

Abstract

The 2D NMR (T1-T2) mapping technique, which can be used to separate different proton populations from various sources (hydroxyls, solid organic matter, free water, and free HC) has gained attention in petroleum industry. To separate proton contributions, a fixed straight line is commonly employed to separate different regions representing proton sources on the map. However, some of these regions (Region 1 and 2) might overlap which makes extracting the NMR signal amplitude from these regions inaccurate. In order to solve this issue, in this study, we applied the Gaussian distribution deconvolution method to separate the T1 and T2 relaxation distributions and then derived the signal amplitude of each region instead of following the common fixed line approach. Next, we employed this method to analyze several shale samples from the literature and compared the results following both methods to verify our methodology. Finally, samples from the Bakken Shale were studied to separate signals from Region 1 and Region 2 and correlated the results with geochemical properties that were obtained from programmed (Rock Eval) pyrolysis. Results demonstrated an improvement in their relation when our approach is employed compared to the fixed line technique to differentiate signal from overlapping regions. This means the Gaussian distribution deconvolution method can be used with confidence to provide us with more accurate petrophysical and geochemical understanding of complex formations.