Crosswell Electromagnetic Inversion Constrained by the Fluid-Flow Simulator

Abstract

AbstractCrosswell electromagnetic technology has been developed for reconstructing the formation resistivity distribution of the subsurface. By looking at the resistivity changes over time, one can monitor the fluid front movement and estimate saturation changes in the reservoir. Traditionally, the resistivity distribution is directly inverted from the crosswell electromagnetic data through an inversion process. However, this inversion process can be non-unique and has limited spatial resolution.In this paper, we introduce several inversion approaches for improving the interpretation of the crosswell electromagnetic data. The inversion is constrained by a multiphase fluid flow simulator that models the fluid flow in the reservoir and constructs the spatial distribution of the water saturation and salt concentration, which are in turn transformed into the formation resistivity using a resistivity-saturation formula. In these approaches, instead of directly inverting the resistivity distribution, we invert the formation petrophysical parameters such as permeability and porosity.Three inversion approaches were explored. The first is the model-based approach that inverts a layer-by-layer distribution of permeabilities and porosities. The second is the pixel-based approach in which we assume that the formation permeabilities are known and hence we invert only a cell-by-cell distribution of porosities. The third is a hybrid approach in which we invert a cell-by-cell distribution of porosities together with a layer-by-layer distribution of permeabilities. By using these approaches, we obtain a reconstructed resistivity distribution that is consistent with the fluid flow physics, and the spatial resolution is significantly improved. Additional outcomes of the inversion include the saturation spatial distribution, which provides an explicit indication of fluid movement within the reservoir. The results of synthetic case studies show significant improvements compared with those acquired by the traditional approach.