Data processing: measuring – calibrating – or testing – Measurement system in a specific environment – Earth science
Reexamination Certificate
2000-05-05
2002-12-10
Lefkowitz, Edward (Department: 2862)
Data processing: measuring, calibrating, or testing
Measurement system in a specific environment
Earth science
C702S016000
Reexamination Certificate
active
06493634
ABSTRACT:
FIELD OF THE INVENTION
This invention relates generally to the field of seismic data processing, and specifically to determining reflection geometry in seismic gather data. This information can be used to determine stacking velocity parameters.
BACKGROUND OF THE INVENTION
In seismic data processing, different pieces of data (time samples from seismic traces) presumed to originate from the same location on a subsurface reflector are combined, or stacked, to enhance the amplitude of desired reflected information relative to noise. This process of stacking requires a parameter, called the stacking velocity, which describes the change in traveltime of the reflected energy as a function of the distance between the seismic source and receiver. This stacking velocity somehow must be selected (or “picked”) throughout the subterranean spatial region represented in the seismic survey. Picking stacking velocities for seismic data is often the most time consuming portion of seismic data processing.
Existing methods for picking stacking velocities consist of first producing velocity semblance, velocity sweep, or velocity events displays at each velocity analysis location in the survey. A brief description of each of these types of display follows:
Velocity Semblance Display
To produce a velocity semblance display, one first measures the trace-to-trace coherence (or coherency), within a common midpoint (CMP) gather of seismic traces, as a function of both moveout velocity and time. This coherency data may be plotted, on a computer monitor for example, as an image with time (or depth) as the vertical axis, velocity as the horizontal axis, and with the pixel color spectrum used as a measure of the coherency value. Peaks in this coherency, as a function of time and velocity, are taken to be potential points on the velocity versus time (or depth) function at the x-y position of that common midpoint. The coherency peaks are assumed to represent subterranean reflectors or horizons, and the time (or depth) value for each peak represents the zero-offset travel time (or depth) of the horizon at that x-y position.
The coherency can be computed using a number of methods. Common methods include simply moveout correcting the traces and computing the trace-to-trace semblance.
Velocity Sweep Display
A velocity sweep is a stack, of several adjacent CMP gathers, for a suite of different stacking velocities. The stacks, of several adjacent CMP gathers, at one velocity in the sweep are called a velocity sweep panel. The velocity sweep displays are again interpreted, in a manner similar to that described above, to produce a velocity versus time function. However, the velocity interpreter can usually pick better velocities from a velocity sweep display than from a semblance display, because:
The interpretation is based not only on discerning trace-to-trace coherency within a CMP gather, but also on the spatial (i.e., lateral) trace-to-trace coherency between the CMP locations. This spatial trace-to-trace coherency, rather than the coherency within a CMP gather, is much more important to the geologic interpreter when performing geologic interpretation of seismic data. Thus, optimizing spatial trace-to-trace coherency should be the primary goal of stacking velocity analysis.
The human eye can often observe optimal spatial trace-to-trace coherency in areas where poor signal-to-noise ratios make discerning optimal trace-to-trace coherency within a single CMP gather impossible.
Velocity sweep displays provide some geologic information that the velocity interpreter can use to separate signal from noise. For example, multiple reflections (“multiples”) often have different dip than the primary geologic reflections one wishes to optimize.
Velocity Events Display
In order to improve the efficiency of the velocity interpretation process, velocity semblance or velocity sweep data are often pre-analyzed using an automatic picking program. See, for example J. H. Bodine, J. N. Gallagher, and J. H. Wright, “Geophysical Exploration Using Velocity Spectra Regional Coherency Peaks”, U.S. Pat. No. 4,984,220 (1991). This automatic picking program searches the velocity semblance for peaks, and outputs a file containing only the time, velocity and coherency value of the peaks. These (time, velocity, coherency) data are called velocity “events” because each of them may represent reflected energy from a specific subsurface reflection or series of reflections. These data are plotted as a scatter plot with time as the vertical axis, velocity as the horizontal axis, and some aspect of the scatter symbol (e.g., color or size) used to indicate coherency. The automatically picked events themselves do not produce a satisfactory velocity function, because they are usually quite noisy. The events must be interpreted to produce a velocity function for the velocity analysis location.
Interpretation of Velocity Displays
After producing the velocity semblance, sweeps or events, these data are interpreted at each location in the survey individually. The only extra information, regarding adjacent velocity analysis locations, that is usually displayed are velocity functions that may previously have been picked at those adjacent locations. This conventional method for picking stacking velocities has several problems:
The simple fact that velocities must be analyzed at each location individually makes this a time consuming process.
Since the velocity interpreter cannot view the velocity analysis displays for a large region, it is difficult to pick velocities that do not have some unreasonable lateral velocity variations. Therefore, the interpreted velocity model must often be edited several times to remove these unreasonable lateral variations.
Velocity semblance and events displays show only coherency within a CMP gather, not spatial coherency. On the other hand, spatial coherency is the attribute that has primary importance to the geologic interpreter, and is more robust in the presence of low signal-to-noise ratios.
Velocity sweep displays have poor velocity precision, because they display panels of several traces for each velocity. This implies that only a few velocities (on the order of 25) can be displayed before the display becomes unwieldy.
Since conventional velocity analysis techniques use 2-D displays, with time being one dimension of the display, it is difficult to pick more than one “attribute” (i.e., any measurement based on seismic data, such as velocity, reflector dip or a non-hyperbolic moveout parameter) of the seismic data. The attribute that is usually picked is hyperbolic moveout velocity. Conventional methods have difficulty picking attributes beyond this hyperbolic moveout velocity, such as non-hyperbolic moveout parameters that may result from anisotropy or lateral velocity variation.
Bodine et al., in their patent referenced above, attempt to overcome the problem of unreasonably large lateral velocity variations. Their patent treats the velocity data, from one seismic line, as a cube with vertical axis being time, one lateral axis being velocity, and the other lateral axis being location along the seismic line. By extracting various slices from this cube, and interpreting those slices, they can produce smoother velocity models than those produced by the conventional method discussed above. However, their method has several shortcomings:
Since their method only works with slices of the data, the velocity interpreter is never presented with a single display that shows the velocity information corresponding to a large region. Thus, it is still difficult to produce velocities that vary smoothly in all dimensions.
For 3-D data, each “inline” in the 3-D volume would have to be analyzed individually by their method. Thus, it would be difficult to ensure that the velocity model is smooth in the direction perpendicular to the inline direction (i.e., the “crossline” direction).
The only attribute used by their method to pick velocity is coherency within CMP gathers. As discussed above, other attributes can be more diagnostic. In particul
Bear Lorie K.
Krebs Jerome R.
ExxonMobil Upstream Research Company
Gutierrez Anthony
Lefkowitz Edward
Plummer J. Paul
LandOfFree
Method for determining stacking velocity parameters or other... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for determining stacking velocity parameters or other..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for determining stacking velocity parameters or other... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2965420