Deriving Relative Permeability from Capillary Pressure Using Gaussian and Rational Equations

摘要

While dynamic data are necessary for a robust reservoir characterization, measuring these type of data in a laboratory is time consuming and very expensive. On the other hand, if dynamic data, especially relative permeability and capillary pressure, are available for discrete grids, they might lead to a more promising simulation model. In the following study, capillary pressure is predicted by artificial neural networks for distinct flow units. Then, two methods are introduced for estimating relative permeability: the first one is based on using Gaussian and rational equations for deriving relative permeability from capillary pressure data and the second one is by utilizing ANN.