| Popis: |
Abstract A permeability estimator is constructed from the lambda parameter, A, which is the size of dynamically connected pores. For simple pore geometries at high permeabilities, A should theoretically be proportional to the pore volume-to-surface area ratio. Further, at high permeabilities, the permeability is proportional to A2/F, where F is the formation factor. Core data are used to establish the proportionality constant. For permeabilities less than about 100 mD, this formula overestimates the actual permeability by a factor related to the original estimate. Thus, a corrected second estimate can be readily made, and the combined estimate is shown to match measured permeabilities on a diverse set of sands and shaly sands with a correlation coefficient of 0.89 for log k. The estimate can also be made using NMR relaxation time data, instead of the surface-to-pore volume ratio, with a correlation coefficient of 0.92 for the same data set. The estimate can also be made from mineralogy or lithology data, in which each mineral or lithologic fraction is assumed to have a fairly constant specific surface area. Examination of the data suggests that only about 10% of the total clay mineral surface area normally influences fluid flow. For Fontainebleau quartz arenites, the resulting estimate has a correlation coefficient of 0.99 with the measured permeabilities. For Utah sands and shaly sands, the mineral-based estimate has a correlation coefficient of 0.84 with the measured permeabilities; the lithology-based estimate had the same correlation coefficient. Using the lithology-based form of the equation, accurate permeability logs are derived from geochemical logs in Utah and from conventional logs in Venezuela.. P. 777 |
