Numerical simulation of miscible displacement experiments in fractured porous media was carried out to analyze the data. It was found that both the location and the magnitude of the fracture permeability heterogeneity were the main factors in determining the miscible displacement efficiency. The stripping of solute from matrix to fracture, which is caused by the permeability variation in the fracture, significantly increases the recovery efficiency. By taking the fracture permeability heterogeneity into account, the experiments were successfully simulated.
The initial (also the maximum) gravity drainage rate, qo of a homogeneous rock matrix for an immiscible process is
In Eq. 1, is the threshold capillary pressure, and L is the height. Other symbols are defined in the Nomenclature. Assumptions made in the derivation of Eq. 1 include infinite gas mobility. In fractured porous media comprised of matrix blocks and fracture network, the initial (also the maximum) gravity drainage rate does not exceed that of the unfractured porous media, provided the fracture storage is negligible. After the initial period, due to the contrast in capillary pressures of the matrix and the fracture, the gravity drainage rate of the fractured media is less than that of the unfractured media. For zero capillary pressure, Eq. 1 reduces to. The critical rate for frontal instability (viscous fingering) in a homogeneous medium is given by. Therefore, the maximum drainage rate is less thanprovided the displacing fluid has a negligible viscosity. In the experiments presented in Ref. 5, even when the rate of injection or production exceeded the critical rate, an efficient displacement was achieved. The main reason for such high recovery efficiency is a pronounced crossflow mechanism between the fracture medium and the matrix blocks.
The primary objective of this work is to analyze the laboratory measurements reported in Ref. 5 and to investigate the mechanism which results in a very pronounced crossflow of fluids between the matrix and the fractures.
Ref. 5 provides data on miscible displacement tests for different fluids at various injection rates (different V/V the critical velocity, and for a unit cross sectional area) in the two-, three- and twelve-slab assembly. The data reveal that the recovery performance in these tests are more efficient than in the corresponding experiments of Thompson and Mungan, performed at the same V/Vc. The effluent concentration is small after the primary breakthrough, then gradually increases to about 1 PV of injection. This characteristic is completely different from the data of Thompson and Mungan, which show a sharp primary breakthrough, plateau of solvent cut for a long period of time, then a sharp secondary breakthrough.