Joint inversion of simultaneous source time-lapse seismic data sets

discussion and conclusions

We have presented a joint least-squares wave-equation inversion method (J-PLSI) for imaging simultaneous source (or blended) time-lapse seismic data sets. Because it is difficult to repeat both the shot-receiver positions and relative shot-timing for different surveys, there will be significant non-repeatability in simultaneous source data sets. J-PLSI directly inverts the simultaneous source data sets without need for prior separation, combining the cost savings advantages of simultaneous source acquisition and phase-encoded migration.

marm-encoded-dat-1,marm-encoded-dat-2
Figure 2. Encoded (a) baseline and (b) monitor data sets.[CR]

marm-encoded-wav-1,marm-encoded-wav-2
Figure 3. Downgoing source wavefields at 0.52 s for (a) the baseline and (b) monitor data sets. Note the difference in shot-timing.[CR]

marm-nocoded-mig-2,marm-encoded-mig-2,marm-encoded-inv-2
Figure 4. (a) Single source migrated monitor image. Simultaneous source (b) migrated and (c) inverted monitor images.[CR]

Direct migration of simultaneous source data generates strong cross-term artifacts (Figure 4(b)) relative to conventional single source data (Figure 4(a)). It is unlikely that conventional cross-equalization methods will adequately attenuate these cross-term artifacts while preserving the production-related time-lapse amplitude changes. J-PLSI attenuates the artifacts giving images with better resolution and more balanced amplitudes than migration (Figure 4). This translates to relatively high-quality, high-resolution time-lapse images (Figure 5(c)).

Because simultaneous source acquisition reduces the overall data acquisition cost and the acquired data can be efficiently processed, we recommend shorter survey intervals. By acquiring time-lapse seismic data sets in this manner and processing them using the J-PLSI method, we can get closer to the goal of continuous seismic reservoir monitoring. The J-PLSI formulation also provides a suitable framework for simultaneous inversion of multiple seismic and other reservoir monitoring data (e.g. production).

J-PLSI provides gives good-quality inverted time-lapse images at a fraction of cost of least-squares migration of single source data sets. Therefore, it can be applied to conventional single source time-lapse data sets, specifically encoded for computational cost savings. Where such an approach is taken, similar encoding schemes can be used for all data sets and the suitable encoding scheme (Romero et al., 2000) can be chosen to minimize the cross-term artifacts.

Joint inversion of simultaneous source time-lapse seismic data sets