[First Hit]
‹header›
‹date/time›
‹footer›
Page ‹#›
Body Text
Second Level
Third Level
Fourth Level
Fifth Level
•This is a fracture characterization analysis using the unconventional method of multicomponent surface seismic data.
•This technology is still very new but continuing to progress.
-The goal is to incorporate our seismic observations with reservoir models.
•The area of interest is the Pinedale Field anticline in Wyoming.
•We gratefully acknowledge JEBCO Seismic LP, owner of the 3C data over the Pinedale anticline, for permission to use their data in this study. •Special thanks are due Jonathan Fried and John Markert at WesternGeco for their help in processing the converted-wave data. •Thanks also to Dean DuBois with Encana Corporation and Barbara Luneau with Schlumberger Consulting Services for providing valuable insights into the nature of Pinedale reservoir, and Randy Koepsell for helping with the FMI log interpretation. •Location information for the Pinedale field survey and well locations provided by Jeff Thompson with WesternGeco is much appreciated.
•Fractures in the earth tend to be aligned and vertical because of the stress regime. •ADVANCE: If these fractures are smaller than the seismic wavelength we don’t see individual fractures but we get an average response. •ADVANCE: This averaging leads to a directionally dependent response, i.e. it is azimuthally anisotropic, e.g. a HTI medium. •ADVANCE: By measuring suitable seismic attributes of this Previous HitanisotropyNext Hit we can infer information about these fracture systems.
•It is well known that fracture properties are Fractal by Nature. So in terms of scale, we have …
-Cores and Image logs provide the small scale features of the reservoir
-And seismic provides some large scale features.
-Both of these are direct observations ... We can put our finger on these features.
-Each is important and gives part of the overall fracture picture.
-Clearly production is not controlled only by these end members.
-If it were, our reservoir models and fluid simulations would be perfect.
•So how do we fill this gap?
-Modeling has traditionally been used to fill this gap.
-Such as geomechanical modeling to reconstruct paleo-strain fields.
-[ADVANCE] But I propose that Previous HitanisotropyNext Hit is the logical candidate
•for this Sub-seismic resolution.
-Fractures are smaller than the seismic wavelength
•So we don’t see the individual fractures but we get an average response.
•This averaging leads to a directionally dependent response, i.e. it is azimuthally anisotropic.
-Can measure Previous HitanisotropyNext Hit at borehole with VSPs and with P-wave surface seismic data.
•But I want to focus on PS-waves.
•So in this talk I’ll …
-Briefly cover some highlights of the processing and analysis of the PS-wave seismic data,
-to extract the azimuthal Previous HitanisotropyNext Hit.
-Show this application to the Pinedale survey.
-Finish with a discussion of the importance of calibrating these observations,
•and  conclusions.
•I have to include this just as a quick reminder of Converted wave Geometry
•One very important property is the azimuth or the propagation direction from source to receive.
-Need to sample many azimuths for azimuthal Previous HitanisotropyNext Hit measurements.
•In addition to the P-waves <ADVANCE>
•We record P to S converted wave using 3C Geophones. <ADVANCE>
•It’s the source to detector azimuth that controls the direction of the created shear wave
•However the upgoing S-waves split and then travel as a pair of S-waves,
•polarized parallel and perpendicular to fractures.
•
•The complexity of S-wave splitting increases with the distance of travel.
-The separate fast and slow waves produced in the first anisotropic layer encountered
- each split within the next layer,
-giving a total of four waves from the single conversion.
-All are combined together when detected by the horizontal geophones.
•The goal of multicomponent processing is to unravel these waves layer stripping.
-Before Previous HitanisotropyNext Hit can be measured in our targets at depth,
-The overburden Previous HitanisotropyNext Hit must be determined and removed.
•
•Lets look at one way this can be done.
•Early estimation of azimuthal Previous HitanisotropyNext Hit is taken into account in the processing.
•The data are processed in azimuth limited volumes.
-Basically this is an extension of offset dependent processing.
-we divided the data into the 8 azimuth sectors shown here
-Processing the radial and transverse components
-Which results in 16 unique data volumes.
•Four azimuths are oriented in the observed principal fast and slow directions.
•But some azimuths are not.
-So lets look at how the processing and analysis is done for these cases.
•The method used is an Alford rotation and consists of four data trace components.
•In this cartoon, we’re looking at P-wave source direction P1’ for the prestack geometry.
-Oblique to the fast and slow earth coordinate frame.
-ADVANCE: Three different offsets: sources and receivers.
-Propagation direction is in the P1’ direction.
-Two receivers: S’1 and S’2.
-One reflection point.
•Here’s the P1’ data stacked to zero offset.
•Even though it is zero offset there is still the P1’ azimuth associated with it.
•There are two receivers S’1 and S’2
-These are the two traces with the single reflector: PS’11 and PS’12.
-Since P’1 is oblique to principal directions there is energy on both.
•If the medium were isotropic,
-There would be energy only on the PS’11 component.
•Perform the same thing for the P’2 direction.
•ADVANCE: Here is the prestack data of three offsets.
-And important to note: the same receiver configuration for S’1 and S’2.
•After stack now we have our four component data for Alford rotation.
•Again the traces for the PS’21 and PS’22.
•Principal (inline) components and crossline components.
-Crossline are the same but not the inlines.
•These data are often referred to as 2C by 2C data because,
-There are two source components
-And two receiver components – the same orientation.
•Now they can treated exactly the same as S-wave source data
-And rotated with Alford's 4C algorithm,
-Which is a simultaneous 2C source and 2C receiver rotation.
•This diagram illustrates the data after 2C x 2C Alford rotation,
-over angle theta, to the principal directions
-(indicated here as S1 and S2 or fast and slow).
•The resulting traces show the characteristic
-minimization of energy on the off-diagonal components,
-and principal components maximized.
•If we apply this rotation to all orthogonal pairs within our survey
-they can then be summed to create a single set of four components
-oriented in the principal directions.
•
•So in this talk I’ll …
-Briefly cover some highlights of the processing and analysis of the PS-wave seismic data,
-to extract the azimuthal Previous HitanisotropyNext Hit.
-Show this application to the Pinedale survey.
-Finish with a discussion of the importance of calibrating these observations,
•and  conclusions.
•
•The area of interest is shown here in Sublette county
-In the Green River Basin.
•ADVANCE: So the Jebco multicomponent survey is just east of the Jonah field
-at the tip of the Pinedale anticline,
-In the Antelope area.
•Full-azimuth geometry for this survey was a diagonal shot pattern.
-But oriented in the NW-SE direction.
-Indicated by the red symbols.
•This provided good wide azimuth illumination
-Particularly in the NW-SE and NE-SW directions.
-Not good fold in the NS and EW directions.
•As a result, the final azimuthal Previous HitanisotropyNext Hit analysis for this survey
-was limited to the NW-SE and NE-SW directions.
•This is a time structure map of an event near the upper Ft. Union fm – above the resvoir to give you a sense of the regional structure within the survey
•Similar processing was performed
-And the data were azimuth limited in 8 directions
-Oriented with the survey.
•The well is the Antelope Tail: 15-23.
•ADVANCE: Lets look at the inline and crossline sections near the well.
•Here’s the P-wave data and shows good S/N.
•ADVANCE: On the left is the E-W inline section
-Where I have interpreted the structure for for several formations.
-A low angle disturbance/fault below the Pinedale anticline
-And the faults over the crest.
•The horizons are a rough estimate from the Antelope Tail 15-23 sonic log
-To near Base of Fort Union
-And for the Lance formation.
•The anticline is less well defined and
•Possible disturbances off the crest.
•ADVANCE: On the right side is the N-S cross line section.
-That shows the high angle bounding fault
-At the southern end of the Pinedale Anticline.
•Now lets look at the PS-wave data for comparison.
•Here’s the E-W PS-wave inline.
•Graphically compressed to a Vp/Vs value of 2.
-1.5 seconds PS time corresponds to 1.0 seconds PP time.
-3.0 to 2.0 seconds.
-There appears to be a very close match and similar features agree nicely.
•ADVANCE: Here’s the same interpretation,
-Just translated rigidly over to PS-wave data.
-Anticline – disturbance – sags below
-S/N is poorer – shadow zones,
-But clearly the same structure.
•The N-S crossline comparison shows similar characteristics.
-Vertical fault
-And other subtle features.
•So here are the three layers where layer stripping was performed
-Shown for the fast S-wave propagation direction, N145E.
-The lines are oriented the same as the previous examples.
•The events picked are close to the horizons for the Fort Union and Lance,
-But are guided by reflectors with good S/N
-At the base of each layer,
-And they are continuous (without overlap).
•First lets look at a comparison of the fast and slow PS-wave data.
•We’ll do this at the vertical line shown here.
-So we’re still looking at the fast PS-wave data.
•ADVANCE: On right now is the slow PS-wave superimposed
-These are PS-waves that have propagated in the NE-SW direction.
-The S/N is a bit poorer but,
-We can see that the same events come in later.
-TOGGLE
•ADVANCE: On the left shows the same relationship for the inline section.
-TOGGLE
•Here is the azimuthal Previous HitanisotropyNext Hit analysis for this shallow layer.
-The isochron is shown again in the lower left,
-And in the analysis at the right, north is straight up.
-The orientation is shown by the small vectors: N45W.
-The percent azimuthal Previous HitanisotropyNext Hit is shown in color,
-Where the range is 0 to 7.5%.
•So on the anticline we see areas where the Previous HitanisotropyNext Hit is higher
-And south of the fault it is generally lower.
-There is a hint of this feature south of the E-W fault.
•There is a bit of a geometry foot print in the receiver line direction.
-But clearly there appears to be more azimuthal Previous HitanisotropyNext Hit over the anticline.
-NOTE: N20W orientation of fast S-wave at the well location.
•So here is a simple interpretation.
•The anticline stresses associated with folding
-Results in fractures aligned NW-SE
-And this produces zones of increased azimuthal Previous HitanisotropyNext Hit.
•These might be areas of more intense fracturing.
30 s
•Here’s the same E-W PS-wave inline we saw earlier.
•Arrows point to the zone in the with FMI images.  Let’s look at these next.
•
•Here’s one example of the FMI log at 7900 feet.
-These are paper logs unoriented so the quality is not very good.
-There are clearly two fractures present here.
-One has a strike of N10W and the other N30W.
•Other example have orientations N20W.
•
•Here’s another example of the FMI log at 8500 feet.
-These both show a strike of N20W.
-Experienced log analysts will probably interpret these as
-induced fractures and not natural fractures.
-But induced fracture orientation will be controlled by the principal horizontal stress direction.
•So we see that these FMI results
-agree with the fast S-wave direction of the PS-waves.
-Suggest that PS-waves are sensitive to horizontal stresses.
•
•So lets look at the results for the target Lance formation.
•Here is the azimuthal Previous HitanisotropyNext Hit analysis containing the Lance formation layer where the majority of the production occurs in this area.
-The isochron is shown again in the lower left,
-And in the analysis at the right, north is straight up.
-The orientation is shown by the small vectors: predominantly N45W
-But varies north of the E-W fault.
-The percent azimuthal Previous HitanisotropyNext Hit is shown in color,
-Where the range is 0 to 10.5%.
•Over the anticline, areas of high Previous HitanisotropyNext Hit,
-some have shifted to the east compared to the overburden in the lower Fort Union.
-And south of the fault it is still generally lower.
-There is still a hint of this feature south of the E-W fault.
•There is still a bit of a geometry foot print in the receiver line direction.
-Overall a lower S/N: halo around the edges.
-But clearly there is a similar pattern as the overburden,
-And there appears to be more azimuthal Previous HitanisotropyNext Hit over the anticline.
•Within the survey area at Pinedale there is a well (Antelope #15-23) on the apex of the anticline.
•It is important to use well log information, such as the FMI and dipole sonic logs to corroborate these Previous HitanisotropyNext Hit results. For example, the FMI logs show mostly bedding planes near horizontal dip, but at many levels the dip exceeds 30˚ and 40˚, which could be related to low-angle fracturing associated with the thrust fault tectonics at Pinedale. It is interesting that at these levels, the dip direction is qualitatively in agreement with the principal PS-wave directions.
•However, for optimal calibration with the PS-wave data, the dipole sonic log should be acquired with both orthogonal shear sources to obtain quantitative estimates of the azimuthal Previous HitanisotropyNext Hit orientations and magnitude of Previous HitanisotropyNext Hit.
•In addition, near-offset S-wave source vertical seismic profiles could be acquired. These can provide valuable S-wave splitting properties at the same seismic wavelengths as the surface data.
•Early estimation of the principal S-wave orientation is critical to optimize processing for fracture characterization.
•Rotating to radial and transverse and processing common-azimuth volumes allows all the data to be combined using 2Cx2C Alford rotation. This increased fold helped improve the signal quality.
•2Cx2C rotation and layer-stripping analyses leads to quantifying the fracture properties at target horizons. •Azimuthal Previous HitanisotropyTop analyses at the southern tip of the Pinedale anticline in the Antelope area indicate potential sweet spots of more intense fracturing. These occur along the flanks of the anticline and appear to be controlled by faulting.
•Strike of near vertical fractures agrees with the fast S-wave direction.