Geoscience Reference
In-Depth Information
wavefield produces zero-phase reflection data, with a more accurate control of the phase
than can be achieved for surface seismic data.
Fortunately, it is quite straightforward to achieve the separation between up- and
downwaves. As shown schematically in
fig. 3.4(b)
,
there is a difference in change of
arrival time with depth between the up- and downgoing waves. When the geophone is
located at the depth of a particular reflector, then the direct travel-time is the same as
the reflected time to the event. As the geophone is moved up away from the reflector,
the direct travel-time decreases and the reflection time increases. For the simple case of
a vertical well with source at the wellhead, it is obvious that the decrease in travel-time
for the direct arrival will be the same as the increase in travel-time for the reflection.
In practice, it is fairly straightforward to measure the travel-times of the direct arrivals,
though it is sometimes hard to identify the exact time at which the trace begins to deflect
owing to the presence of noise. If traces are statically shifted by subtracting these times,
then the direct arrivals will line up horizontally across a trace display; if the traces are
shifted by adding the first arrival times (doubling the slope of the first arrival travel-time
curve) then the upward travelling events will be horizontal (for horizontal reflectors, or
nearly so if they are dipping). Filtering the first type of display to enhance laterally con-
tinuous events will result in an estimate of the downgoing wavefield, which can then be
subtracted from the data to leave only the upgoing wavefield. Applying the second type
of trace shift to these upwaves will then give us a display on which seismic reflectors
are near-horizontal and can be enhanced by median filtering, which emphasises near-
horizontal lineups in the dataset. A filter operator can also be applied to convert the
wavelet (as measured in the downwaves) to zero-phase. It is then possible to form a
corridor stack
trace by stacking together the parts of the upwave dataset immediately
following the direct arrival. This trace should then be zero-phase and free of multiples,
and thus ideally suited for comparison with well synthetic and surface seismic. Since
the VSP averages seismic response over a distance of a few tens of metres around the
borehole, the problems of formation invasion and very small-scale lithological changes
are not present. It is therefore usually helpful; the main problem is the rather low signal
to noise ratio.
Figure 3.5
shows an example display of deconvolved upwaves, after the
traces have been shifted to make the reflected events line up horizontally. The corre-
filter has brought out some consistent events that are scarcely visible in
fig. 3.5
; however,
reliability of these events would need careful thought in a practical application.
Another advantage of the VSP is the ability to give good results in deviated wells,
where synthetic seismograms are often unreliable, perhaps because anisotropy makes
the sonic log readings (which measure velocity along the borehole) differ from the
vertical seismic velocity in the formation; thus the calculated impedance contrasts are
not those seen by a nearly vertically travelling ray. A useful VSP technique is the walk-
above geometry, where the surface source is placed vertically above the geophone at
a series of levels in the deviated hole. In this way, an image is produced of the zone
