Fundamentals of Borehole NMR Data Processing and Inversion

Abstract

Borehole NMR tools usually acquire basic Carr–Purcell–Meiboom–Gill (CPMG) echo trains consisting of several hundred to a few thousand echoes. Echo train data should be converted to a 1D distribution (T1 and T2), 2D distribution (D–T2 and T1–T2) or 3D distribution (T1–D–T2) through Inverse Laplace Transformation (ILT) to further obtain petrophysical parameters and identify the fluid properties of the reservoir. Meanwhile, data compression is normally employed in multidimensional (2D and 3D) NMR inversion procedures to avoid huge memory usage and slow computational speeds. Furthermore, highly resolved solutions of inverted echo data related to signal-to-noise ratios can be significantly improved using proper inversion methods and denoising techniques. A better understanding of the data processing and corresponding inversion methods will be very helpful for the data interpretation and application.