next up previous [pdf]

Next: Sparse data problem Up: Numerical Examples Previous: Numerical Examples

Undershoot problem

This example demonstrates the undershoot problem , where obstructions caused by latter development facilities prevent complete recording of monitor data sets. Here, we consider an obstruction with changes in its location and size (Figure 2) for the three monitor data sets. This is a common scenario in seismic monitoring applications where the construction, addition and alteration of production facilities create obstructions. In each monitor survey, neither shot nor receivers were located within the undershoot area (Figure 2).

Migrated images of the target area for all four data sets are shown in Figure 4(a), and the corresponding Hessian diagonals (so-called illumination) in Figures 4(b). Figure 4(c) shows the cumulative time-lapse image of the target area obtained from the full data sets, and Figure 4(d) shows the illumination ratio between monitor and baseline data sets. The migrated, normalized, separately and jointly inverted time-lapse images are shown in Figure 5. Note that time-lapse images obtained from joint inversion [Figure 5(d)] contain fewer artifacts relative to those from migration, normalization and separate invertion [Figures 5(a) to 5(c)].

migs-1w-h illums-hole-1w-h migs-4d-1w illums-hole-r-1w-h
migs-1w-h,illums-hole-1w-h,migs-4d-1w,illums-hole-r-1w-h
Figure 4.
(a) Migrated (i) baseline and (ii)-(iv) monitor images. (b) Hessian diagonal corresponding to images in (a). (c) Cumulative time-lapse images from full baseline and monitor shot-receiver coverage [Figure 2(a)]. (d) Illumination-ratio for each monitor survey [Figures b(ii)-(iv)] relative to the baseline [Figure b(i)].[CR]
[pdf] [pdf] [pdf] [pdf] [png] [png] [png] [png]

migs-4d-1w-h invd-4d-1w-h invs-4d-1w-h inv2-4d-1w-h
migs-4d-1w-h,invd-4d-1w-h,invs-4d-1w-h,inv2-4d-1w-h
Figure 5.
Cumulative time-lapse images at four production stages (with increasing production from top to bottom) obtained from (a) migration, (b) Hessian-diagonal illumination correction, (c) separate inversion, and (d) RJMI. Compare these results to those from the full data sets [Figure 4(c)].[CR]
[pdf] [pdf] [pdf] [pdf] [png] [png] [png] [png]


next up previous [pdf]

Next: Sparse data problem Up: Numerical Examples Previous: Numerical Examples

2009-05-05