Use of Pore-Network Models to Simulate Laboratory Corefloods in a Heterogeneous Carbonate Sample
- Baomin Xu (Chevron Petroleum Technology Co.) | Jairam Kamath (Chevron Petroleum Technology Co.) | Y.C. Yortsos (U. of Southern California) | S.H. Lee (Chevron Petroleum Technology Co.)
- Document ID
- Society of Petroleum Engineers
- SPE Journal
- Publication Date
- September 1999
- Document Type
- Journal Paper
- 178 - 186
- 1999. Society of Petroleum Engineers
- 5.8.7 Carbonate Reservoir, 4.3.4 Scale, 5.4.1 Waterflooding, 5.3.2 Multiphase Flow, 5.5.2 Core Analysis, 4.1.5 Processing Equipment, 5.4.9 Miscible Methods, 5.5 Reservoir Simulation, 5.1 Reservoir Characterisation, 1.6.9 Coring, Fishing, 5.3.4 Reduction of Residual Oil Saturation, 5.1.5 Geologic Modeling, 4.1.2 Separation and Treating, 5.2.1 Phase Behavior and PVT Measurements, 5.3.1 Flow in Porous Media
- 2 in the last 30 days
- 461 since 2007
- Show more detail
- View rights & permissions
|SPE Member Price:||USD 10.00|
|SPE Non-Member Price:||USD 30.00|
Many carbonate rocks have pore features that are large at the core plug scale. Conventional laboratory assessment of recovery behavior in these carbonates can be unreliable. Pore-network modeling offers an approach to improve our analysis of flow and displacement in such heterogeneous rocks. This paper reports on a systematic approach to link pore model inputs and laboratory data to a calibration methodology, and the subsequent use of a pore-network model to match laboratory corefloods in a heterogeneous dolomite. It is found that spatial correlation effects, determined from core samples using thin section and computed tomography (CT) data, are important for a satisfactory match.
Carbonate rocks often display heterogeneities significant at the core plug scale. Conventional analysis is likely to be inadequate, and it is not surprising that the literature contains few references to multiphase flow studies on such rocks.
Archer and Wong1 found that conventional relative permeability calculation methods were inadequate for their experiments on a vugular, full diameter carbonate sample. The water relative permeability showed anomalous humps and plateaus. Ehrlich2 measured drainage relative permeabilities for vugular, low permeability samples from the Westerose reservoir. He used the steady-state method and full diameter cores. The relative permeability curves obtained also showed significant plateau regions. Narayanan and Deans3 studied miscible displacement in a full diameter, vugular carbonate sample from Gulf-Connor Fenn-Big Valley, Canada. They concluded that centimeter-scale heterogeneities control the flow distribution. In a related study on carbonate systems, the review by Espie et al.4 reported a wide range of process efficiency. For example, they found that the residual oil saturation Sor to water flood may vary between 28 and 80% OOIP and that to tertiary miscible flood between 0 and 50% OOIP.
Pore-network (PN) models provide the ability to relate macroscopic behavior of porous media to the physics of flow and displacement at the pore and pore-network scale. They are useful for understanding multiphase flow mechanisms5 and the effects of pore scale heterogeneity.6,7 A general PN model should contain capillary, viscous, and gravitational forces. Therefore, it should be possible, in principle, to accurately simulate flow and displacements and to account for heterogeneities at various scales in the core sample. This is particularly true for processes, such as drainage, which is one of the simplest displacement processes. However, except for a few cases,8 pore-network models have not been linked directly to data on real rocks. Technological advances in computing and flow visualization are now allowing us to connect laboratory tests and network models more closely. For example, we can now run 500,000 node models9 to recreate fine-scale CT flow images. In this paper, we report on an integrated effort to utilize real laboratory data to calibrate pore-network models and, subsequently, to model displacement processes.
Experimental Data on Carbonate Sample.
In an earlier paper,10 we have discussed our experimental data on a dolomite rock from the Beaverhill Lake formation, Canada (k=300 md, ?=14%, L=4.9 cm, d=3.8 cm). The highly heterogeneous nature of this core is clearly apparent in Figs. 1 to 3. Figs. 1 and 2 show the CT-derived porosity distributions along cross-sections parallel and perpendicular to flow, respectively. It is evident that the porosity is heterogeneously distributed and that it is spatially correlated. The porosity distribution is wide, with a nontrivial tail of large-pore features. The clustering of larger porosity is also evident in the thin section picture of Fig. 3, where pore features of various sizes can be identified. Unit mobility, equal density miscible floods as well as waterfloods were conducted in this laboratory sample as detailed in our previous paper6,10 and its references. CT images revealed that the flow was nonuniform and channeled along connected high permeability paths. It is apparent that the large heterogeneity at the core scale must be carefully accounted for to successfully model displacements in such a core.
Snapshots of cross-sectional images of water saturations obtained at the beginning of a waterflood (at Swi and at the end of different rate water floods, are shown in Fig. 4. These images correspond closely to the porosity image of Fig. 1. In our earlier paper,6 we found that the average water-oil contact angle in this core was 114°. As a result, the waterflood is a drainage process. At the flow rate of 0.5 feet/day, the water mostly invades the high porosity, high permeability channels. A high oil saturation of 67% PV remains in large unswept areas. As the flow rate is increased, completely new channels are invaded and the residual oil saturation decreases. For example, at the high flow rate of 120 feet/day the residual saturation has decreased to 37% PV. These features reveal that under these wettability conditions, the waterflood is mostly a drainage process, with new channels being invaded as the flow rate increases. The data were subsequently analyzed using conventional methods. However, meaningful information could not be extracted from such data.10 This led us to the use of pore-network models.
Calibrating Pore-Network Models.
Since the pioneering work of Fatt11 to simulate flow in porous media by a network of tubes, many researchers have advanced pore-network model approaches to study various aspects of multiphase flow (e.g., see reviews by Blunt12 and Celia et al.13). However, the pore structure used in these models was typically not calibrated with experiments. Other investigators have developed methods for obtaining pore and throat size distributions from mercury injection data,14-16 from a combination of thin sections and capillary pressure data,17,18 and from pore casts.19 However, such results were not used to link these data back to displacement processes.
|File Size||221 KB||Number of Pages||8|