A Quantitative Study to Assess the Value of Pressure Data Acquired with In-Situ Permanent Sensors in Complex 3D Reservoir Models Subject to Two-Phase Fluid Flow

Abstract

Abstract Recent advances in smart well completion technologies have enabled the dynamic acquisition of in-situ pressure measurements with sensors in direct hydraulic contact with rock formations. In addition to the immediate impact of in-situ sensors as tools for real-time, reactive reservoir management and control, usage of in-situ pressure sensors has long-term benefits. Devising an optimal macro-management strategy for hydrocarbon reservoirs requires more than a tool for instantaneous monitoring and control. Precise knowledge of the spatial distribution of petrophysical properties is essential for accurate reservoir delineation, management, and production forecasting. From the formation evaluation viewpoint, large volumes of flow-related data constitute an attractive prospect for robust and accurate characterization of reservoirs. In addition to static information in the form of geostatistical, seismic, and geologic data, usage of dynamic measurements remains imperative to construct accurate reservoir models amenable to production forecast. In this paper, we address the quantitative estimation of three-dimensional (3D) spatial distributions of permeability and porosity from pressure measurements acquired with in-situ permanent sensors. A pilot waterflood operation conducted with a conventional five-spot pattern is chosen as example for our numerical experiments. We assume in-situ permanent pressure sensors to be an integral part of the production-well completion and to remain completely isolated from the hydraulics of the wellbore. Therefore, these sensors perform uncorrupted measurements of in-situ formation fluid pressures. Quantitative estimation of spatial distributions of permeability and porosity is approached with a novel subspace approach and a modified Gauss-Newton inversion algorithm. This inversion strategy incorporates an adjoint formulation for the efficient computation of model sensitivities. In our inversions, the physics of two-phase fluid-flow in the 3D spatial domain is rigorously incorporated into the assessment of distributions of petrophysical properties from in-situ permanent-sensor pressure data. Comparisons are shown of the enhancement in spatial resolution and reduction of uncertainty when using in-situ permanent sensors with respect to estimations performed via standard history matching techniques.