Abstract

DME (di-methyl-ether) is a water-soluble solvent with clear potential for enhanced oil recovery (EOR) through its use in DME enhanced water-flooding (DEW). DEW is a novel technology where DME is dissolved in the brine that is injected in the reservoir. Upon contact with oil, DME preferentially partitions into the oil (hydrocarbon) phase. This results in swelling and viscosity reduction of the oil, hence enhancing oil production.

Recovery enhancement for DME enhanced water-flooding (DEW) strongly depends on the phase behavior of the system. Therefore, it is imperative to accurately describe the phase behavior in the dynamic DEW models. This paper presents the PVT experiments and workflow used to model the phase behavior in DEW processes.

The PVT modelling workflow consists of the following steps.

  • Gather basic experimental data for the DME-brine and DME-crude phase behavior required to model phase behavior during the DEW displacement process.

  • Characterize, tune and lump the hydrocarbon fluids using the SRK (Soave-Redlich-Kwong) EoS (equation-of-state).

  • Model the interaction between hydrocarbons, brine and DME using the CPA (Cubic-Plus-Association) EoS.

  • Generate composition dependent partitioning coefficients (K-values) and phase properties based on mixing rules.

Experimental data required as input for the workflow includes routine PVT data, DME-brine solubility data and DME partitioning data between reservoir fluids. This paper provides an overview of how experimental data is obtained and utilized to represent the phase behavior in the modeling workflow. The essential elements of the modeling workflow fall into two segments:

  • representation of the phase behavior with a proper fit-for-purpose model,

  • utilization of the phase behavior representation for dynamic reservoir modeling.

The phase behavior during the DEW displacement process is represented using the CPA EoS. CPA combines SRK with an association term to account for the interaction between DME and brine. As it is relatively computationally costly to run DEW simulations with the CPA EoS, the proposed modelling workflow is based on compositionally dependent partitioning coefficients.

In simulation studies the phase behavior is represented using a K-value model. The K-value model is validated for light and heavy oils in various reservoir models by comparison with conventional water-flooding simulation based on a black-oil PVT description. Results show that the K-value model is accurate for reasonably constant phase compositions, numerically stable and computationally efficient. Furthermore, it is shown that differences in history match obtained with the black-oil model and the K-value model are marginal, as the compositional path is far away from the critical locus. The results give confidence that the K-value model can be used to assess DEW performance in full field models.

You can access this article if you purchase or spend a download.