Adapting the type-curve approach for estimating reservoir rock permeability

This study introduces a type-curve technique for estimating rock permeabili ty from capillary pressure measurements. The Brooks and Corey model in conj unction with the Wyllie and Gardner model were the basis for the developmen t of this type-curve approach. The Wyllie and Gardner permeability model wa s used as the starting point for obtaining an analytical expression for per meability as a function of displacement pressure (P-d), pore geometrical fa ctor (lambda), interfacial tension (gamma), fraction of pore volume availab le for flow (phi*), and porous medium configuration parameters (beta and n) . The last parameters (beta and n) are assumed to be functions of porosity and irreducible water saturations, respectively. Parameters lambda and P-d are determined by matching the actual rock capillary pressure versus wettin g-phase saturation (P-c-S-w) profile with that obtained from the Brooks and Corey model. Capillary pressure data for 62 rock samples representing a variety of sands tone and carbonate facies were used to test this approach. Permeability (k) of these samples ranged from less than 7 md to over 2500 md, and porosity (phi) ranged from 12 to 32%. In general, laboratory-measured permeability v alues matched reasonably well estimated permeability values using the propo sed type-curve approach. An average absolute relative error of 7% was obtai ned with a standard deviation of 6%. For carbonate rocks characterized by a bimodal pore-size distribution, the capillary pressure data match in some cases two distinct type-curves rather than one. The use of an average pore geometrical factor for these cases is recommended, since it gives an accura te estimate of permeability. The use of this type-curve approach is restric ted to air-brine drainage capillary data. This approach improves considerab ly the permeability evaluation from capillary pressure measurements, since it is based on the whole capillary pressure profile data rather than a sing le point and since it does not require an advance knowledge of fudge factor s related to rock pore geometry. The approach is direct and is not plagued by any trial-and-error procedures.