Page 260 - Fundamentals of Gas Shale Reservoirs

240 PASSIVE SEISMIC METHODS FOR UNCONVENTIONAL RESOURCE DEVELOPMENT

scattered throughout it. The microseismic data therefore It is important to remember that SRV is simply a volume
indicates two very different stress and natural fracture defined by one or more measures of seismic activity—SRV
regimes. Because of these stress and natural fracture differ­ is not a direct measure of the volume of the reservoir that is
ences, an infill well drilled parallel to and west of the exist­ producing. Parts of the reservoir that exhibit microseismic
ing well would frac very differently than an infill well drilled activity may not contribute to production and the natural
to the east. Both wells would likely frac differently than the fracture system in the reservoir may contribute to production
well in the diagram which was drilled down a fault zone. even if it did not produce seismic activity. For these reasons,
Also, the natural fracture permeability should be very different Doe et al. (2013) introduced the term Tributary Drainage
in both regions. Volume (TDV). TDV is the volume of the reservoir that con­
The likely geological explanation is that the region west of tributes production to the wellbore. Like SRV, TDV can be
the wells is a complex fault stepover, while the region east of the computed for individual stages or for the entire stimulated
well is a less damaged block in a normal faulting stress regime. length of the wellbore. The TDV is estimated by integrating
Note that this study is from an actively propagating fold‐thrust microseismic data with all available geomechanical and geo­
belt. This shows the variability in space and time of geological logical data to model the productive volume of the reservoir.
stress fields. Irregularities in faults and jostling of large fault Note that TDV is not constant over time because the volume
blocks can produce complex and variable stress states very dif­ of produced reservoir increases with production (Lacazette
ferent from those indicated by the overall tectonic regime and et al., 2014). Moos et al. (2011) take a conceptually similar
regional geology. Also note that the active fractures were approach to estimate frac effectiveness and productive
vertical and that the treatment pressure was well below Sv, volume albeit with alternative modeling procedures (and
which also suggests that the region around this well is not in a without using the term TDV).
thrust‐faulting stress state. However, because reactivation of
preexisting vertical fractures appears to be the primary mecha­
nism of stimulation, treatment pressure equal to slightly above 10.6.4 Using Passive Seismic Results for Frac,
Sv is not required to propagate hydraulic fractures. Well‐Test, and Reservoir Simulation
The full value of passive seismic data is obtained from
learning as much as possible about the effects of hydraulic
10.6.3 Fracture Width, Height, SRV, and Tributary fracture treatments and reservoir properties. Simple mea­
Drainage Volume
sures of fracture width, height, and volume; identification of
The width and height of a cloud of MEQs is often used as an major features (e.g., faults) that take frac fluid; and stress
approximation of the extent of a hydraulic fracture. The SRV state are immediately useful, but do not directly reveal reser­
is the volume of the part of the reservoir that produced strong voir properties. The last step in the chain of extracting value
microseismic signal during a hydraulic fracture treatment. from passive seismic data is achieved by using the data to
Various methods are used to compute the SRV. The simplest model the hydraulic fracture process, well performance, and
methods are simply counting the number of events per stage reservoir properties and performance. Passive seismic data
or measuring the dimensions of a cloud of MEQs. A more provides important geological and engineering constraints
sophisticated and commonly used approach is to compute a on such models that make them more realistic, and hence
density cloud by defining a 3D counting grid in the region of more accurate. Better models lead to better fracture treatments,
interest, then centering a counting volume (usually a sphere) better production, better estimated ultimate recovery (EURs),
on each MEQ hypocenter and incrementing the grid nodes and better production forecasts and reserve estimates—all of
that are contained in the counting volume. The grid may be which directly impact the bottom line.
incremented simply by the number of MEQs, or by a scheme Passive seismic data can be used in several ways to enhance
that weights the count by the magnitude of each MEQ. A frac, well performance, and reservoir simulations. Modeling
density value is then defined as the outer boundary of effec­ efforts are generally focused on producing discrete fracture
tive stimulation so that the volume within this isosurface is network (DFN) models of the hydraulic fractures, natural
taken as the SRV. Other approaches to determining SRV fractures stimulated by hydraulic fracturing, and the natural
include fitting a shrink wrap (minimum estimate) or convex fracture network in the reservoir. DFN simulations informed
hull (maximum estimate) to a microseismic cloud after by passive seismic results can provide better production fore­
removing events thought to be outliers. A non‐MEQ‐derived casts than simple decline‐curve analysis (e.g., Williams‐
measure of SRV is determined from cumulative seismic Stroud et al., 2013). The most basic approach relies on a
activity volumes by directly mapping the cumulative seismic correspondence between clouds of MEQs observed during
activity at each voxel within a study volume, then defining a the fracture treatment and simulated clouds of MEQs from
cutoff based on relative activity and requiring a direct DFN hydraulic fracture simulations (e.g., see Dershowitz
connectivity to the wellbore (Section 10.5.4; Fig. 10.20; et al., [2010] for geomechanical details; Rogers et al., 2010).
Lacazette et al., 2014). Additional detail is provided by using focal mechanism