This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 201358, “Quality Control of NMR Measurements in Unconventional Rocks Using Hilbert Transforms and Semilogarithmic Derivatives,” by Pierre Aérens, SPE, David Medellin, and Carlos Torres-Verdin, SPE, The University of Texas at Austin, prepared for the 2020 SPE Annual Technical Conference and Exhibition, originally scheduled to be held in Denver, 5-7 October. The paper has not been peer reviewed.
Nuclear magnetic resonance (NMR) measurements are extremely valuable in the assessment of fluid-flow properties of rocks. However, inverted transverse-relaxation time (T2) distributions are often biased. The authors of the complete paper introduce and compare two quality-control approaches based on two different signal-processing practices: the semilogarithmic derivative and the Hilbert transform. This work provides the basis of effective quality-control methods for NMR measurements for the petrophysical interpretation of rocks with complex pore-size distributions. The fast and reliable quality control of estimated T2 distributions is not affected by inversion artifacts, relying only on unfiltered, raw data.
Introduction
Borehole NMR measurements are commonly riddled with excessive noise that precludes high-resolution assessments of rock and fluid properties, especially in unconventional rock. Inverted T2 distributions can be biased by positivity constraints and the regularization method used in the inversion, making it difficult to determine whether the estimated pore-size distributions are reliable or byproducts of the inversion or regularization. Quality control of inversion results is essential for the robust quantitative interpretation of proton magnetization decays.
The complete paper presents a mathematical analysis for the two data-quality assessment methods and the work flow to estimate pseudo-T2 distributions on the basis of time-decay proton magnetizations. The estimation and assessment methods are then applied to both synthetic and experimental data sets to verify their validity and reliability. The laboratory examples are detailed in the complete paper.
Validation With Synthetic Cases
The authors apply the derivative and Hilbert methods to two synthetic examples for a total of three cases. Both synthetic examples comprise three exponential magnetization decays, with amplitudes equal to A1=0.5, A2 =0.2, and A3=0.3. The first example is referred to as ideal because it is not contaminated with noise and because the authors chose the T2 peak times separated by three orders of magnitude at 1 µs, 1 ms, and 1000 ms, respectively. The second example showcases a more-realistic choice of T2 peak values, namely 0.1, 10, and 200 ms, respectively.
