Quantitative NMR Interpretation

Abstract

Abstract Advancements in nuclear magnetic resonance (NMR) logging have provided important information on formation productivity, pore size and fluid type. This information has been most easily used as a stand-alone NMR interpretation or as overlays of NMR data on other log data. Attempts at using NMR data quantitatively with traditional porosity measurements have been more difficult. The difficulty originates from the need to combine measurements that are insensitive to pore structure and pore fluid with NMR measurements that are sensitive to those formation properties. For example, even in a perfectly clean freshwater sand, an NMR tool may produce a porosity that is different from density and neutron porosity. The differences have been attributed to either the lack of an early-time signal from small pores associated with silt and clay size grains or incomplete fluid polarization. The lack of an early-time signal has been addressed with the latest generation tools; however, incomplete polarization still needs to be addressed. Polarization would not be a problem if NMR data were acquired with a sufficient wait time to guarantee complete polarization. The penalty for such a guarantee is long acquisition times and increased costs. In addition, accurate knowledge of the formation's T1 is needed to determine the correct wait time. Such knowledge is often not available. Current NMR tools include an approximate polarization correction based on the measured T2 and an estimate of T1/T2 ratio. A more rigorous method is proposed that replaces the welisite correction with one dependent on the actual volumes the tool is measuring. To do this, the data are reprocessed without a polarization correction. Calculations are made of NMR tool-response parameters for different pore sizes and pore fluids. These calculations are a function of the hydrogen index and T1 of the individual formation components; inputs to the parameter calculations are pressure, temperature, oil viscosity, water salinity and acquisition wait time. The outputs are the tool's response to fine grain, coarse grain and vuggy porosities; gas; oil; and oil-base mud filtrate. The resultant parameters, NMR data and other log data are entered into a simultaneous solver for the final result. The advantage of this approach is that it allows the polarization correction to be adjusted according to the individual volumes present. it also allows incomplete polarized NMR data acquired at increased logging speeds to be appropriately corrected. Examples using this approach on formations with light hydrocarbon zones and vugs in water-base and oil-base muds are examined. The results for all cases are self-consistent including resistivity measurements, which are also sensitive to pore size and pore fluids, confirming the viability of the technique. P. 227