Coreflood Chromatography in Limestones

Abstract

Most of the remaining oil reserves are stored in carbonate reservoirs. Improving recovery from these reserves is paramount to uninterrupted energy supply and any future energy transition, should it happen. One of the aspects of recovery improvement is crude oil-brine-carbonate rock interface chemistry. The study of these interactions must be supported by: 1) a correct reactive transport model that describes the advective-dispersive forces in porous media; and 2) convincing experimental evidence consonant with the reactive transport model.

As a part of such a larger study, this work addresses some modeling and experimental aspects of dispersive tracer transport through porous media. At both field and laboratory scales, tracer data allows estimation of brine-accessible pore volume and hydrodynamic dispersion. In addition to tracer transport during laboratory-scale coreflood experiments, one typically observes adsorption, dissolution/precipitation, ion-exchange, or all. Without proper tracer data interpretation, these phenomena cannot be distinguished. Therefore, tracer experiments are vital for a meaningful interpretation of any coreflood experiment.

The modeling part of the work first discusses choice of boundary conditions that capture the peculiarities of laboratory-scale experiments. We find that Robin or third-kind boundary condition in both inlet and outlet is appropriate for description of the experimental results. Next, the modeling part presents a validated numerical approach to simulate dispersive tracer transport through the experimental system with high hydrodynamic dispersion.

Tracer coreflood experiments are performed on $1.5''$ by $3''$ Indiana limestone core plugs with high hydrodynamic dispersion due to extreme rock heterogeneity. Chloride ion is used as the tracer in these experiments. Effluent from the cores is collected each 0.1 pore volume, and the concentration history data are recorded for at least 3 pore volumes.

Finally, in the last part of this work, we present how the recorded concentration history data were scaled to the dimensionless model output and assess fit quality. The scaling parameters yield porosity and Péclet number that allow estimation of fluid accessible pore volume and hydrodynamic dispersion. The porosity of the core samples estimated from tracer data is close to those measured using helium pycnometer and/or brine imbibition. Very high hydrodynamic dispersion coefficients obtained from the tracer points correlate well with the scarcely published data on Indiana limestone.