New Coreflood Interpretation Method for Relative Permeabilities Based on Direct Processing of In-Situ Saturation Data

Abstract

Abstract Generation of good quality relative permeability data depends on carefully designed core flooding programmes, combined with appropriate interpretation of the laboratory data. Where capillary pressure is important simulation based methods are needed. However, experience shows that the conventional fixed capillary pressure outlet boundary condition may fail to capture the development of the in-situ saturation profiles satisfactorily. This paper describes a new interpretation approach that fully utilises in-situ saturation data and reduces the sensitivity to boundary condition assumptions. The method extends previously published techniques for interpreting gravity drainage floods by including viscous and capillary forces. Relative permeabilities are calculated directly from in-situ saturation and pressure drop data, without the need to use simulation methods, by applying the Darcy flow equations for the mobile phases at each position and time in the core. Darcy's equations are recast to give separate pressure and fractional flow equations. The fractional flow equation is solved locally, independently of any boundary condition assumptions, and the pressure equation (written in a global integral form) is solved for the mobility. The algorithm is applicable to a wide range of water-oil displacements or gasoil displacements with an immobile initial water saturation, including multi-rate and constant pressure drop unsteady-state floods, steady-state and gravity drainage floods. The algorithm is applied to synthetic saturation and pressure drop data sets generated using a conventional simulator for a selection of water/oil and gas/oil core floods. The underlying relative permeabilties are correctly constructed, including a significant fraction of the oil relative permeability tail. A range of smoothing methods have been developed to allow the basic algorithm to be successfully applied to datasets in which noise has been added to the basic saturation and pressure profiles generated by simulation.