Accordingly with Welge formulation, to obtain relativernpermeability curves at laboratory, all calculations are made onrna selected point of the sample. Usually this point is located atrnthe outlet face of the sample, where production rates arerndirectly measured. As a result, relative permeability curves arernreported as function of point saturations and not as function ofrnaverage saturations. Laboratory curves are then adapted tornreproduce reservoir behavior, usually through the derivation ofrnpseudo functions. With usual methodologies, this pseudorncurves are also expressed as function of point saturations andrnintroduced in numerical simulators. In spite of this procedure,rnnumerical simulators perform their calculations using thernaverage water saturation at every grid block. Although pointrnand average saturations are expected to be the same atrninfinitely small grid size, this is not the case with coarse arealrnsimulation grids.rnIn this paper, water-cut as a function of produced oil isrnanalyzed for a linear case. Several cases are developed usingrnrelative permeability curves defined as function of pointrnsaturations for different grid sizes, and as a function ofrnaverage water saturation. It is shown that only curves withrnaverage values give reliable data.rnAs a result of this work, an advantageous methodology torntransform laboratory measured curves into those consistentrnwith numerical simulation approach is presented. It is alsornshown that the use of rock relative permeability curves,rnpreviously adapted to the particular geometry of the systemrnunder study, drastically reduces the number of grid blocksrnrequired and overcomes numerical dispersion.
展开▼
