METHOD OF DETERMINING FORMATION GRAIN SIZEDISTRIBUTION FROM ACOUSTIC VELOCITIES AND NMRRELAXATION TIME SPECTRUM

摘要

Formation grain size distribution is important inreservoir sand management, perforation, optimizingfracturing strategy, and predicting depositional anddiagenetic facies and hydraulic units. This paperdescribes a new method of using acoustic velocities andNMR relaxation time spectrum to determine parametersof formation grain size distribution.Grain size distribution is often approximated by normal,log-normal, binary, or more general types ofdistribution, e.g., Weibull distribution. In most cases,however, it can be characterized by two numbers: meangrain size and sorting, which describe, respectively, themeasures of the center and the spread of a grain sizepopulation. The influence of sorting on acousticproperties of porous media is well known; however,there is no direct relationship between acousticproperties and mean grain size. NMR T_1 and T_2relaxation time spectra of wetting phase are directlyrelated to the distribution of pore volume-to-surfaceratio, which, in turn, depends on the mean grain size.On the other hand, there is no straightforwardcorrelation between NMR relaxation time and sorting.Moreover, neither NMR nor acoustic data alone canresolve effects due to heterogeneity in mineralogy andheterogeneity in pore geometry (sorting, compaction,cementation, etc).The approach presented in this paper is based on thepredictions of NMR T_2 spectrum and acoustic velocitiesin numerical model rocks having different grain sizedistribution. Pore geometry of these model rocks isdefined by geologic and mineralogical informationabout the formation. This information is supplied by amineral reading from a logging tool, from drillingcuttings, or from prior knowledge based on core orlogging data of a similar field. Based on the results ofthese predictions, we show that NMR relaxation timespectrum allows estimating mean grain size, and thatcompressional and shear acoustic velocities providemeans to compute sorting parameter. The results arecompared with the sandstone core measurements dataand are found to be consistent with them.