Changes

The shale smear factor (SSF) is dependent on the shale bed thickness and the fault throw but not on the smear distance (Lindsay et al., 1993) ([[:file:M91Ch13FG90.JPG|Figure 12]]). Smaller values of the SSF correspond to a more continuous development of smear on the fault plane. A large fault is likely to seal where the SSF is equal to or less than 4.<ref name=Farseth_2006 />

The shale smear factor (SSF) is dependent on the shale bed thickness and the fault throw but not on the smear distance (Lindsay et al., 1993) ([[:file:M91Ch13FG90.JPG|Figure 12]]). Smaller values of the SSF correspond to a more continuous development of smear on the fault plane. A large fault is likely to seal where the SSF is equal to or less than 4.<ref name=Farseth_2006 />

The shale gouge ratio works on the assumption that the sealing capacity is related directly to the percentage of shale beds or clay material within the slipped interval (Yielding et al., 1997). The shale gouge ratio is the proportion of the sealing lithology in the rock interval that has slipped past a given point on the fault ([[:file:M91Ch13FG90.JPG|Figure 12]]). To calculate the shale gouge ratio, the proportion of shale and clay in a window equivalent to the throw is measured.

The shale gouge ratio works on the assumption that the sealing capacity is related directly to the percentage of shale beds or clay material within the slipped interval.<ref name=Yieldingetal_1997 /> The shale gouge ratio is the proportion of the sealing lithology in the rock interval that has slipped past a given point on the fault ([[:file:M91Ch13FG90.JPG|Figure 12]]). To calculate the shale gouge ratio, the proportion of shale and clay in a window equivalent to the throw is measured.

The prediction of fault seal is based on the assumption that if there is enough shale in the section undergoing faulting, then sealing is likely. There is often a continuous shale gouge or shale smear along fault planes where there is sufficient mudstone material available to be incorporated (Lindsey et al., 1993; Foxford et al., 1998). Nevertheless, a number of field studies show that fault zones can have a significant degree of complexity and variation in deformation style along their lengths (Childs et al., 1997; James et al., 1997). For example, Foxford et al. (1998) examined good exposures of the Moab fault in Utah. They found that the structure and content of the fault zone was so variable that it was impossible to predict the nature of the fault zone over even a 10-m (33-ft) distance. Doughty (2003) found that the clay smear along the Calabacillas fault in New Mexico showed numerous gaps particularly where minor faults within the fault zone complex cut out the shale smear associated with the major slip plane. The implication of these field studies is that fault seal can be predicted but is subject to chance factors affecting the reliability of the prediction. Because of this, any fault seal prediction should be calibrated against actual evidence that fault compartmentalization is present. Yielding et al. (1999) made a fault seal analysis for the Gullfaks field in the Norwegian North Sea. Areas of higher shale gouge ratios (>20%) were more likely to seal on the basis of pressure history and chemical tracer movement between wells.

The prediction of fault seal is based on the assumption that if there is enough shale in the section undergoing faulting, then sealing is likely. There is often a continuous shale gouge or shale smear along fault planes where there is sufficient mudstone material available to be incorporated.<ref name=Lindseyetal_1993 /> <ref name=Foxfordetal_1998 /> Nevertheless, a number of field studies show that fault zones can have a significant degree of complexity and variation in deformation style along their lengths.<ref name=Childesetal_1997 /> <ref name=Jamesetal_1997 /> For example, Foxford et al.<ref name=Foxfordetal_1998 /> examined good exposures of the Moab fault in Utah. They found that the structure and content of the fault zone was so variable that it was impossible to predict the nature of the fault zone over even a 10-m (33-ft) distance. Doughty<ref name=Doughty_ 2003 /> found that the clay smear along the Calabacillas fault in New Mexico showed numerous gaps particularly where minor faults within the fault zone complex cut out the shale smear associated with the major slip plane. The implication of these field studies is that fault seal can be predicted but is subject to chance factors affecting the reliability of the prediction. Because of this, any fault seal prediction should be calibrated against actual evidence that fault compartmentalization is present. Yielding et al.<ref name=Yieldingetal_1999 /> made a fault seal analysis for the Gullfaks field in the Norwegian North Sea. Areas of higher shale gouge ratios (>20%) were more likely to seal on the basis of pressure history and chemical tracer movement between wells.

Gibson (1994) provided a case history for fault seal analysis from the Columbus Basin, offshore Trinidad. Oil and gas fields occur in upper Miocene to Pleistocene deltaic sandstones of the Columbus Basin, located offshore to the southeast of the island of Trinidad. Numerous small faults dissect these reservoirs, and fault seal appears to be a critical feature controlling the size of these petroleum pools. Allan diagrams show that juxtaposition sealing is insufficient to explain the fault control on fluid contacts.

Gibson<ref name=Gibson_1994 /> provided a case history for fault seal analysis from the Columbus Basin, offshore Trinidad. Oil and gas fields occur in upper Miocene to Pleistocene deltaic sandstones of the Columbus Basin, located offshore to the southeast of the island of Trinidad. Numerous small faults dissect these reservoirs, and fault seal appears to be a critical feature controlling the size of these petroleum pools. Allan diagrams show that juxtaposition sealing is insufficient to explain the fault control on fluid contacts.

[[file:M91Ch13FG91.JPG|thumb|300px|{{figure number|13}}Schematic illustration showing the character of fault zones in siliciclastic strata based on outcrop and core observations from onshore and offshore Trinidad (from Gibson, 1994).]]

[[file:M91Ch13FG91.JPG|thumb|300px|{{figure number|13}}Schematic illustration showing the character of fault zones in siliciclastic strata based on outcrop and core observations from onshore and offshore Trinidad (from Gibson<ref name=Gibson_1994 />).]]

The sediments that form the reservoirs offshore are also exposed onshore along the east coast of Trinidad. Outcrops onshore and cores offshore provide control on the nature of the fault rock. In these outcrops, shale smears are found where shale beds have been displaced along the fault. The shale smears range in thickness from millimeter- to centimeter-thick shale partings to complex zones up to several meters thick ([[:file:M91Ch13FG91.JPG|Figure 13]]).

The sediments that form the reservoirs offshore are also exposed onshore along the east coast of Trinidad. Outcrops onshore and cores offshore provide control on the nature of the fault rock. In these outcrops, shale smears are found where shale beds have been displaced along the fault. The shale smears range in thickness from millimeter- to centimeter-thick shale partings to complex zones up to several meters thick ([[:file:M91Ch13FG91.JPG|Figure 13]]).

Offshore, hydrocarbon columns up to 200 m (656 ft) thick are found within compartments interpreted as being sealed by clay smears along faults. The general observation is that the blanket of clay smear along faults only appears to be continuous and effective where the shale content of the displaced section exceeds 25%. The shale smear factor was estimated for faults from two of the fields in the basin. SSF values of between 1 and 4 were found for faults with throws more than 150 m (492 ft) that sealed the longest hydrocarbon columns. It was concluded that faults in this area could be modeled as sealing along their length provided the SSF did not exceed a value of 4.

Offshore, hydrocarbon columns up to 200 m (656 ft) thick are found within compartments interpreted as being sealed by clay smears along faults. The general observation is that the blanket of clay smear along faults only appears to be continuous and effective where the shale content of the displaced section exceeds 25%. The shale smear factor was estimated for faults from two of the fields in the basin. SSF values of between 1 and 4 were found for faults with throws more than 150 m (492 ft) that sealed the longest hydrocarbon columns. It was concluded that faults in this area could be modeled as sealing along their length provided the SSF did not exceed a value of 4.

[[file:M91Ch13FG92.JPG|thumb|300px|{{figure number|14}}Comparison between (a) depth-converted seismic interpretation from the Gullfaks field, Norwegian North Sea, and (b) a plaster model deformed by plane strain extension. The plaster model shows that many small-scale faults are expected to exist in the Gullfaks structure but are below seismic resolution (from Fossen and Hesthammer, 1998). Reprinted with permission from the Geological Society.]]

[[file:M91Ch13FG92.JPG|thumb|300px|{{figure number|14}}Comparison between (a) depth-converted seismic interpretation from the Gullfaks field, Norwegian North Sea, and (b) a plaster model deformed by plane strain extension. The plaster model shows that many small-scale faults are expected to exist in the Gullfaks structure but are below seismic resolution (from Fossen and Hesthammer<ref name=Fossenandhesthammer_1998 />). Reprinted with permission from the Geological Society of London.]]

==Subseismic faults==

==Subseismic faults==

Only the faults that the geophysicist can pick from seismic data will be mapped, that is, those faults with vertical displacements down to the limit of seismic resolution. As mentioned in chapter 6, this can be about 20–40 m for reservoirs at moderate depths. However, a significant number of subseismic faults will probably be present with vertical displacements less than this ([[:file:M91Ch13FG92.JPG|Figure 14]], [[:file:M91Ch13FG93.JPG|Figure 15]]). Thus, the true degree of the structural complexity of a reservoir will be underrepresented.

Only the faults that the geophysicist can pick from seismic data will be mapped, that is, those faults with vertical displacements down to the limit of seismic resolution. As mentioned in [[Data: sources]], this can be about 20–40 m for reservoirs at moderate depths. However, a significant number of subseismic faults will probably be present with vertical displacements less than this ([[:file:M91Ch13FG92.JPG|Figure 14]], [[:file:M91Ch13FG93.JPG|Figure 15]]). Thus, the true degree of the structural complexity of a reservoir will be underrepresented.

[[file:M91Ch13FG93.JPG|thumb|300px|{{figure number|15}}Fault maps of the East Pennine coalfield, United Kingdom. In map (a), only faults with throws of 20 m (64 ft) or more are shown. These are equivalent to faults that are detectable by seismic surveys at reservoir depths. In map (b), every mapped fault is shown, with fault throws of between 10 cm (4 in.) and 180 m (590 ft) (from Watterson et al., 1996). Reprinted with permission from the Journal of Structural Geology.]]

[[file:M91Ch13FG93.JPG|thumb|300px|{{figure number|15}}Fault maps of the East Pennine coalfield, United Kingdom. In map (a), only faults with throws of 20 m (64 ft) or more are shown. These are equivalent to faults that are detectable by seismic surveys at reservoir depths. In map (b), every mapped fault is shown, with fault throws of between 10 cm (4 in.) and 180 m (590 ft) (from Watterson et al.<ref name=Wattersonetal_1996 />). Reprinted with permission from the Journal of Structural Geology.]]

It is possible to input subseismic faults into a reservoir model using stochastic methods (Munthe et al., 1993; Hollund et al., 2002). In summary, this is a computerized procedure for randomly inserting shapes representing geological features into a 3-D model while still honoring predefined rules and statistics controlling the global distribution of the data.

It is possible to input subseismic faults into a reservoir model using stochastic methods.<ref name=Muntheetal_1993 /> <ref name=Hollundetal_2002 /> In summary, this is a computerized procedure for randomly inserting shapes representing geological features into a 3-D model while still honoring predefined rules and statistics controlling the global distribution of the data.

The first part of the method involves making an estimate of the number of subseismic faults by extrapolating from statistics on the length versus frequency of known seismic faults into the subseismic region. Fractal analysis has been used on the assumption that fault-size populations approximate to fractal distributions (Gauthier and Lake, 1993). Statistics are also compiled on fault orientations, length to throw ratios, and fault densities per square kilometer. A further step is to determine those areas of the field where subseismic faults are more likely to be present than elsewhere. One method is to predict the paleostrain regime of the reservoir at the time of faulting (Maerten et al., 2006). On this basis, a model will be made, which will include both the seismic and subseismic faults. Fault seal analysis can be applied to the subseismic faults in the model to determine whether they are sealing or not.

The first part of the method involves making an estimate of the number of subseismic faults by extrapolating from statistics on the length versus frequency of known seismic faults into the subseismic region. Fractal analysis has been used on the assumption that fault-size populations approximate to fractal distributions.<ref name=Gauthierandlake_1993 /> Statistics are also compiled on fault orientations, length to throw ratios, and fault densities per square kilometer. A further step is to determine those areas of the field where subseismic faults are more likely to be present than elsewhere. One method is to predict the paleostrain regime of the reservoir at the time of faulting.<ref name=Maertenetal_2006 /> On this basis, a model will be made, which will include both the seismic and subseismic faults. Fault seal analysis can be applied to the subseismic faults in the model to determine whether they are sealing or not.

General experience with inserting subseismic faults into simulation models is that they will influence the flow behavior (Damsleth et al., 1998; England and Townsend, 1998; Ottesen et al., 2005). The critical feature seems to be whether the faults are sealing or not. Sealing faults can create an open framework of short baffles, which helps to improve sweep. The baffles increase the tortuosity of the flood front and delay water breakthrough. A large number of sealing subseismic faults in a reservoir will, on the other hand, create numerous dead ends, which will reduce the sweep efficiency of a waterflood. Nonsealing subseismic faults form cross-fault juxtapositions, which can improve vertical connectivity and enhance sweep.

General experience with inserting subseismic faults into simulation models is that they will influence the flow behavior.<ref name=Damslethetal_1998 /> <ref name=Englandandtownsend_1998 /> <ref name=Ottesenetal_2005 /> The critical feature seems to be whether the faults are sealing or not. Sealing faults can create an open framework of short baffles, which helps to improve sweep. The baffles increase the tortuosity of the flood front and delay water breakthrough. A large number of sealing subseismic faults in a reservoir will, on the other hand, create numerous dead ends, which will reduce the sweep efficiency of a waterflood. Nonsealing subseismic faults form cross-fault juxtapositions, which can improve vertical connectivity and enhance sweep.

[[file:M91Ch13FG94.JPG|thumb|300px|{{figure number|16}}Reservoir intervals thicken markedly across growth faults. They are common in areas with thick delta sequences and mobile substrates such as shale or salt. This example is from Upper Triassic deltaic sediments exposed in the coastal cliffs of Svalbard (from Edwards, 1976).]]

[[file:M91Ch13FG94.JPG|thumb|300px|{{figure number|16}}Reservoir intervals thicken markedly across growth faults. They are common in areas with thick delta sequences and mobile substrates such as shale or salt. This example is from Upper Triassic deltaic sediments exposed in the coastal cliffs of Svalbard (from Edwards<ref name=Edwards_1976 />).]]

==Growth faults==

==Growth faults==

Growth faults are faults that were active at the same time as the sediments were being deposited ([[:file:M91Ch13FG94.JPG|Figure 16]]). Many show a listric geometry with the fault soling out into shale horizons. They are common in areas with thick delta sequences. Growth faults can be recognized because sediments thicken into the hanging wall of a growth fault and the throw of the fault increases with depth. All the individual reservoir units may thicken up across a mapped growth fault. Alternatively, growth can be taken up by additional layers filling the accommodation space in the hanging wall (Hodgetts et al., 2001).

Growth faults are faults that were active at the same time as the sediments were being deposited ([[:file:M91Ch13FG94.JPG|Figure 16]]). Many show a listric geometry with the fault soling out into shale horizons. They are common in areas with thick delta sequences. Growth faults can be recognized because sediments thicken into the hanging wall of a growth fault and the throw of the fault increases with depth. All the individual reservoir units may thicken up across a mapped growth fault. Alternatively, growth can be taken up by additional layers filling the accommodation space in the hanging wall.<ref name=Hodgettsetal_2001 />

==Faults as flow conduits==

==Faults as flow conduits==

It is known that faults can conduct flow along the fault plane. Brittle rocks such as carbonates are more likely to contain conductive faults by comparison to shallow buried siliciclastic sediments, for example.

It is known that faults can conduct flow along the fault plane. Brittle rocks such as carbonates are more likely to contain conductive faults by comparison to shallow buried siliciclastic sediments, for example.

Specific examples of faults acting as fluid conduits have been described. Production wells located near faults showed rapid water breakthrough in the Fateh field, offshore Dubai. Well tests, production logs, radioactive tracer surveys, and interference tests indicate that aquifer influx is occurring along conductive faults within the reservoir (Trocchio, 1990). A campaign of horizontal drilling in the Prudhoe Bay field in Alaska showed that between 10 and 20% of the faults intersected by the wells were conductive to flow. These caused early water or gas production as a result of fault intersection with the water leg or the gas cap (Pucknell and Broman, 1994). Production, pressure, and production log data indicated that water flowing up faults had resulted in rapid water breakthrough in the crestal area of the Khafji field in the Arabian Gulf (Nishikiori and Hayashida, 1999).

Specific examples of faults acting as fluid conduits have been described. Production wells located near faults showed rapid water breakthrough in the Fateh field, offshore Dubai. Well tests, production logs, radioactive tracer surveys, and interference tests indicate that aquifer influx is occurring along conductive faults within the reservoir.<ref name=Trocchio_1990 /> A campaign of horizontal drilling in the Prudhoe Bay field in Alaska showed that between 10 and 20% of the faults intersected by the wells were conductive to flow. These caused early water or gas production as a result of fault intersection with the water leg or the gas cap.<ref name=Pucknellandbroman_1994 /> Production, pressure, and production log data indicated that water flowing up faults had resulted in rapid water breakthrough in the crestal area of the Khafji field in the Arabian Gulf.<ref name=Nishikioriandhayashida_1999 />

==Stress changes in reservoirs==

==Stress changes in reservoirs==

In stress-sensitive reservoirs, fractures may dilate during injection and close during drawdown. These effects are most pronounced in low-permeability, overpressured, and naturally fractured reservoirs (Lorenz, 1999). Pressure depletion as a result of production will change the stress state of a reservoir (e.g., Hillis, 2001).

In stress-sensitive reservoirs, fractures may dilate during injection and close during drawdown. These effects are most pronounced in low-permeability, overpressured, and naturally fractured reservoirs.<ref name=Lorenz_1999 /> Pressure depletion as a result of production will change the stress state of a reservoir (e.g., Hillis<ref name=Hillis_2001 />).

From a mechanical aspect, sandstone reservoirs are porous structures that form a load-bearing framework supporting the weight of the overburden. Reservoir depletion increases the effective stress on the grain framework; this is the difference between the total stress acting on all sides of the rock and the pore fluid pressure. The effective stress is applied at the grain to grain contacts. This leads to elastic deformation of the rock (recoverable on depletion reversal) and, with increasing stress, inelastic deformation. Inelastic deformation mechanisms include microcrack growth and closure, cement breakage, grain rotation, and sliding as well as deformation in clay, mica, and diagenetically altered feldspar grains (Bernabe et al., 1994; Schutjens et al., 1998, 2004; Wong and Baud, 1999). These mechanisms result in the compaction of the rock and a reduction in the porosity. Because the reservoir remains physically connected to the rock surrounding it, the overburden and underburden will also deform in response to reservoir depletion.

From a mechanical aspect, sandstone reservoirs are porous structures that form a load-bearing framework supporting the weight of the overburden. Reservoir depletion increases the effective stress on the grain framework; this is the difference between the total stress acting on all sides of the rock and the pore fluid pressure. The effective stress is applied at the grain to grain contacts. This leads to elastic deformation of the rock (recoverable on depletion reversal) and, with increasing stress, inelastic deformation. Inelastic deformation mechanisms include microcrack growth and closure, cement breakage, grain rotation, and sliding as well as deformation in clay, mica, and diagenetically altered feldspar grains (Bernabe et al., 1994; Schutjens et al., 1998, 2004; Wong and Baud, 1999). These mechanisms result in the compaction of the rock and a reduction in the porosity. Because the reservoir remains physically connected to the rock surrounding it, the overburden and underburden will also deform in response to reservoir depletion.