Mechanistic Model Enhances Estimates of Solvent-Assisted Oil Recovery in Unconventional Reservoirs
Abstract
In this project we developed a simulator to forecast solvent-assisted enhanced oil recovery using miscible solvents in unconventional reservoirs. We modeled oil recovery using a huff-n-puff process in ultra-low permeability oil reservoirs by including diffusion-dominated flow. We analyzed the swelling of dead and live oils, estimated the thermodynamic-limited recovery given solvent-in-oil solubility, assessed the impact of effective diffusion coefficient, flow length, pressure, and temperature on recovery, and explored the uncertainty and impact of fracture surface area.
Current conventional reservoir simulators do not consider all the complex rock-fluid interactions in unconventional reservoirs. Alternative published modeling techniques are computationally expensive and often require geological data that may not be available. Since flow is diffusion dominated, our simulator, by design, does not require geological data. Further, it was coded using only VBA and Microsoft Excel, making it viable for routine use in industry.
Prior to running the simulator, we estimate the ultimate achievable (thermodynamics-limited) recovery for a specified oil-solvent, which is rooted in complex interactions related to molecular size, individual critical points, and mutual solvent/oil solubility. Depending on the type of oil and solvent, increasing solvent concentration might result in a decrease in saturation pressure at the reservoir temperature, allowing for solvent to be added indefinitely and reach full miscibility at the specified reservoir pressure. And, a saturation point might not exist with increasing solvent concentration, as solvent might reach full miscibility. For these scenarios maximum concentration must be restricted such that realistic injected pore volumes and gas oil ratios are achieved.
Increase in fracture surface area is equivalent to an increase in effective diffusion coefficient, allowing for recoveries to be estimated at any time. We can achieve the same result by decreasing fracture spacing, which can be modeled by decreasing stimulated flow length.
Subject
Molecular DiffusionRecovery
Injection
Unconventional Reservoirs
CO2
Solvent
Numerical Simulation
Citation
Messias, Rodrigo Moreira (2023). Mechanistic Model Enhances Estimates of Solvent-Assisted Oil Recovery in Unconventional Reservoirs. Master's thesis, Texas A&M University. Available electronically from https : / /hdl .handle .net /1969 .1 /198878.