Print this page
Click to view presentation in PDF format.
An Integration of Fault Rock Properties through Time with Basin Modeling*
Marek Kacewicz1, Russell K. Davies2, Michael Welch3, and Rob J. Knipe3
Search and Discovery Article #40349 (2008)
Posted November 11, 2008
*Adapted from extended abstract prepared for poster presentation at AAPG Annual Convention, San Antonio, Texas, April 20-23, 2008
1 Chevron ETC, Sugar Land, TX ([email protected])
2Rock Deformation Research USA. Inc., McKinney, TX
3Rock Deformation Research Ltd., Leeds, United Kingdom
The flow pathways in stratigraphically and structurally complex areas require knowledge of the architecture of these geological systems, but also their flow properties such as the permeability and capillary threshold pressure. Typical basin models in such areas are built based on a series of structural restorations that provide a basic geometric description of the evolving system. However, structural restorations do not address dynamically changing fault and host rock properties. We address the changes of the properties throughout the burial history of the basin from models described in this paper that provide an important predictive capability as to flow pathways and pressure communication within the system. If combined with structural restorations and classical basin modeling, they may become an integrated part of basin modeling workflow in structurally complex areas.
In this article, we discuss a model for the fault rock properties or fault gouge and the change in the properties over time in siliciclastic sediments of sands and shales. Presented examples demonstrate how the improved fault and host rock properties lead to better charge and pressure predictions.
|
|
Modeling Fault Rock Properties The flow across and along faults is strongly controlled by the lithology contrasts, the bed thicknesses, and mechanical strengths. The mechanical properties in turn are a function of the mean effective stress related to burial, the mineralogy, particularly clay content, the fluid pressures, and diagenetic changes with burial. Fault properties change with time although models for fault rock properties are typically based on controls on the properties observed in the present day. This often leads to wrong charge predictions and is not sufficient if effective stress modeling is performed in near-fault areas for reservoir quality prediction, which requires a good handle on effective stress history. A general work flow to estimate fault properties for inclusion into basin models is shown in Figure 1.Diagenesis affects carbonate platforms.In the model for the prediction of fault rock properties, the undeformed host rock properties are determined first from models of their compaction and diagenesis. The fault style and their flow properties are subsequently determined from the modeled properties of the host rock. Models for compaction in a mixed clay and sand sediment are based on the porosity distribution between the quartz and clay, as discussed by Marion et al. (1992) and further developed by Revil et al. (2002), where the macroporosity decreases with an increase in the clay content when grain supported and increases when the clay content is large enough to support the grains. The mechanical compaction and diagenesis of the rock depends strongly on the clay content. Figure 2 shows the modeled compaction for end-members 100% and 0% clay with experimental data fitted to the trends. The dashed line is the compaction curve for an intermediate clay content of 54% based on the model of the end member curves. In low clay content quartzarenites, an important diagenetic effect on the porosity reduction in addition to compaction is the quartz cementation. Walderhaug (1996) modeled the kinetics of the quartz precipitation process and quantified the increase in the quartz precipitation, which depends on the grain size and clay content. At higher clay contents, the quartz cementation is retarded. The increase in the quartz precipitation further reduces the porosity relative to that in the host shown from the compaction curves in Figure 2. Thus the porosity reduction is modeled as a function of the compaction and diagenesis. The compaction trends are a function of the effective stress, which may reduce the effect of the overburden on the porosity loss for higher pore fluid pressures. Thus a rock buried deeply with high fluid pressures may have a porosity equivalent to a shallower depth of burial. The porosity changes are modeled on the stress history and not the depth.Initial Rock Properties and Consolidation State The deformation and fault properties are determined by the state of the undeformed rocks at the time of faulting. Permeability and porosity, for example, are strongly dependent on the clay content and porosity of the rocks and the mean effective stress. Undeformed host rocks that lie along a normal compaction trend or normal consolidation, as shown by the compaction curves in Figure 2 and Figure 3, will deform with a more ductile distributed deformation and are less likely to develop discrete faults. Faulting, however, occurs in rocks that are overconsolidated. Overconsolidation may occur in rocks that follow a normal consolidation path to some depth and are then uplifted. The rocks do not regain their original porosity at a similar mean effective stress or depth and are more brittle and overconsolidated. The greater the uplift or overconsolidation, the more brittle the rock. Strongly overconsolidated rocks, including sands and shales, will deform by open fractures and are more likely to leak, especially if this deformation is focused along the fault. Faults that form in a more ductile regime and have a lower overconsolidation are less likely to develop open fractures and may provide a cross-fault seal. The cross fault seal capacity is a function of the clay contents in the faults and the effective stress conditions at failure. Overconsolidation may also occur due to cementation. This has been referred to as pseudo-overconsolidation. A sandstone, for example, that is buried to temperatures great enough for quartz cementation will have a porosity lower than the normal compaction trend without the cements (Figure 3). The rocks will, therefore, deform with open dilatants fractures that will have a tendency to enhance flow along their paths. Thus the overconsolidation is an important control on the faulting style and flow parameters. The degree of overconsolidation is a measure of the brittleness. The brittleness can be modeled from the unconfined compressive strength of the rocks, the porosity prediction for a given depth relative to the normal consolidation, or the effective stress history from modeling. A common measure of the magnitude of the overconsolidation is the overconsolidation ratio or ratio between the maximum applied stress to the measured mean effective stress. In the red and black curves in Figure 3, the overconsolidation ratio is similar for both rock types although their porosity is different. The greater the difference between the maximum stress and the measured stress, the greater the brittleness of the rock. The rock that is cemented will have a measured maximum effective stress equivalent to a rock with the porosity at the measured depth. In this study, we consider the overconsolidation effect relative to the ratio of the maximum stress, p*, and the measured mean effective stress, p. This p/p* is a convention described by Fisher et al. (2007) for deformation in clean sandstones. For p/p* greater than 0.5 we expect a more ductile behavior and below 0.5, a more brittle deformation. The lower the ratio, the greater is the brittleness. The fault properties controlling flow across faults cutting a section of sands and shales can be estimated from a database of hundreds of measurements that are upscaled to the faults of interest by assuming a simple clastic stratigraphic classification that includes clean sandstones with low clay contents, impure sands with moderate clay contents, and shales with high clay contents. Scanning electron microscope images of these different fault styles are show in Figure 4.For a relatively clean quartzarenite with low clay contents less than 15%, the deformation is controlled by the porosity and grain size. For porous sandstones discrete bands of shear or deformation bands develop by the rearrangement of the quartz grains. At higher mean effective stress these grains crush and reduce the permeability of the rock. For lower porosity quartzarenites, the rock develops open fractures that link to form a through-going fault along a discrete surface. These faults with open pathways form in low-porosity, cemented sandstones (pseudo-overconsolidated). At higher average clay contents between 15 and 45%, the clays mix with the quartz grains occluding the pore spaces and reducing the permeability, forming phyllosilicate framework fault rocks. The higher clay content shales and mudrocks have clay contents exceeding 45% and tend to smear along the fault plane. The change between the mixing of clays and smearing is dependent on the change from a grain- to matrix-supported rock as modeled in the compaction. Figure 5, for example, shows a model for porosity reduction with moderate clay content in the host and fault over time. The fault porosity is reduced relative to that of the host shallow in the section, but deeper the compaction and cementation of the host is similar to the fault. Deeper in the section the impact of the fault for cross-fault flow will be minimal, but with the higher cementation, the section is overconsolidated and brittle open fractures will control the flow. The porosity may be related to the permeability and capillary entry pressures to provide a model for the flow input to basin models. Similar plots provide the estimates of flow through modeled stratigraphic section with a range of average clay contents. The workflow described provides a predictive capability to apply fault rock properties through time based on the known mechanical behavior of rocks for different clay contents and effective stress conditions. The cross-fault flow properties are determined from the clay content and properties that determine the rock brittleness, such as porosity. Lower porosity rocks that are overconsolidated separate regions of brittle and ductile behavior where the more brittle rocks are more likely to develop open fractures that will enhance fluid flow. The approach described above has been applied to model hydrocarbon charge and pressure in the Gulf of Mexico shelf area. The area is characterized by a high sedimentation rate resulting in compaction disequilibrium and in fault and host rocks being exposed to rapidly changing temperature and effective stress conditions.In the model, for a relatively clean sand (clay content <15%) buried and compacted, faulting is shown to occur with a porosity in the faults lower than in the sand reservoir at a similar depth. Quartz cementation occurs in both the reservoir and the fault, enhancing the porosity reduction. Experimental results on deformed sandstones show that the brittleness of the sandstones increases for porosity less than 15% (Wong et al., 1999). The depth at which the fault and host rocks reach brittle conditions, as shown in Figure 6A, is different because of the lower porosity of the faults at a shallower depth. Thus the faults reach brittle behavior with expected open fractures at 2100 meters, but similar conditions do not exist for the reservoir until 3950 meters. The p/p* is here calculated only for the reservoir and shows an increase in brittle behavior with depth (decrease in p/p*). The brittleness indicated at shallower depths is associated with the development of cataclastic deformation bands and not open fractures. These results show faults subjected to differential stress conditions are likely to provide flow and pressure communication within the system from 2100 meters and deeper. Below 3950 meters the fault and host rock reach similar brittleness conditions and the mechanical behavior of the fault and host rocks should be similar. Thus between 2100 and 3950 meters the flow will be more focused along fault pathways and is likely to be more distributed at depths greater than 3950 meters. The brittleness as a function of the porosity is extended to all grain-supported rocks or sediments, as shown in Figure 6B for a sandstone with 25% clay with similar Gulf of Mexico burial histories used for the quartzarenite. Here the brittleness is greater in the faults at a shallower depth as in the more clay-poor quartzarenite described above. The results for a matrix-supported rock are controlled by the expected unconfined compressive strength of the overconsolidated section. The results in Figure 6C show that at depths greater than 3500 meters the shale is normally consolidated, but with shallower burial the rocks are overconsolidated and brittle with a greater brittleness more shallow in the section. Thus in this case, the brittle-ductile transition occurs at 3500 meters, which is a critical depth for the basin modeling. Determination of the expected cross fault flow in the rocks with a range of clay contents is a function of the mechanical deformation as described above. These processes have a greater control on the flow in the more shallow “ductile” regime, where fewer open fractures are expected, but they will also control cross-fault flow in the deeper section. Figure 7 shows the predicted permeability of the reservoir and fault with depth and the corresponding capillary threshold pressures for 25% clay that were used in the basin model. The calculated fault and host rock properties were incorporated into basin models using Petromod 2D (IES) software with TectonicLink. A series of structural restorations depicting structural evolution of the area were imported and incorporated into a 2D basin model. Faults in the model were treated as zones represented by a series of fault rock facies. Treating faults as zones rather than fault lines gave us more flexibility as for iterative modifications of fault rock properties. An analysis of fault and host rocks, performed using techniques described in the previous sections, allowed us to select realistic initial rock properties and helped us to identify depth ranges where host and fault rocks become brittle and prone to open fractures and enhanced flow. A series of iterative modifications of fault facies led to a quick convergence to well calibration data. Figure 8 presents an extract from a much longer transect going through a Gulf of Mexico shelf area. Calculated effective stress reflects fault and host rock evolution and honors well data. These results may be used for analysis of reservoir quality and pressure away from well locations. This article describes an integrated basin modeling workflow incorporating an analysis of diagenetic and mechanical history of fault properties. The workflow proved to be very useful in areas characterized by a high level of stratigraphic and structural complexity. In our Gulf of Mexico experiments, the approach led to much more realistic charge and pressure predictions. An important constraint is the stress and pore pressure history, which can change the mechanical behavior of the sediments and rocks and strongly influence the fluid flow. Figure 9, for example, shows how the brittleness in the rock may be impacted by the unloading due to erosion changing the depth and mean effective stress, but that the rock behavior is dependent on the pressure history and not strictly on the depth. An increase in the pore pressure will alter the rock behavior as well.Fisher, Q.J., S.D. Harris, M. Casey, and R.J. Knipe, 2007, Influence of grain size and geothermal gradient on the ductile-to-brittle transition in arenaceous sedimentary rocks: implications for fault structure and fluid flow: Geological Society, London, Special Publications 289, p. 105 - 121. Marion, D., A. Nur, H. Yin, and D-H. Han, 1992, Compressional velocity and porosity in sand-clay mixtures, Geophysics, v. 57, p. 554-563. Revil, A., D. Grauls, and O. Brévart, 2002, Mechanical compaction of sand/clay mixtures: Journal Geophysical Research, v. 107(B11), p. 2293. Walderhaug, O., 1996, Kinetic modeling of quartz cementation and porosity loss in deeply buried sandstone reservoirs: AAPG Bulletin, v. 80, p. 731-745. We wish to thank Chevron Corporation for the permission to publish this work.
|
