S-Cube delivers its breakthrough waveform inversion through the XWITM toolbox which now integrates the AWI and RWI objective functions. This latest result on the SEG14 blind test dataset demonstrates why this combination is theoretically and practically potentially the most accurate and efficient algorithm in existence to determine the final velocity model deep below the diving wave limit from a highly inaccurate starting model.
Original Starting Model
Original starting velocity model
Simplified Starting Model
Change relative to original start
Simplified starting velocity model
Full Waveform Inversion
Vanilla Technique
Single-Pass FWI
Cumulative update after 1 pass of FWI
Velocity model after 1 pass of FWI
Triple-Pass FWI
Cumulative update after 1st pass with smoothed FWI
Velocity model after 1st pass with smoothed FWI
Cumulative update after 2nd pass with lightly smoothed FWI
Velocity model after 2nd pass with lightly smoothed FWI
Cumulative update after 3rd pass FWI
Velocity model after 3rd pass of FWI
Adaptive Waveform Inversion
Overcomes Cycle Skipping
Single-Pass AWI
Cumulative update after 1 pass of AWI
Velocity model after 1 pass of AWI
Triple-Pass AWI
Cumulative update after 1st pass with smoothed AWI
Velocity model after 1st pass with smoothed AWI
Cumulative update after 2nd pass with lightly smoothed AWI
Velocity model after 2nd pass with lightly smoothed AWI
Cumulative update after 3rd pass with FWI
Velocity model after 3rd pass with FWI
Reflection Waveform Inversion
Extends Update Deeper
Harnessing reflection moveout
During RWI, near offset inversion is used to image reflectors generating temporary virtual sources in the model. Offsets are then opened up to update background velocity values with a scattered wavefield objective function. This misfit minimisation isolates and inverts the moveout misfit between the observed and predicted traces of the reflected events.
Virtual source data
Initial background data
Updated background data
AWI-RWI integrated
Cumulative update after 1st pass integrated AWI-RWI
Velocity model after 1st pass integrated AWI-RWI
Cumulative update after 2nd pass AWI-RWI
Velocity model after 2nd pass AWI-RWI
Cumulative update after 3rd pass with final FWI
Velocity model after 3rd pass with final FWI
Fitting deep reflected events
Deep reflected events in the predicted data shifting into positions through the inversion.
Interleaved predicted (start) and field data
Interleaved predicted (iteration 160) and field data
Observed field data
Run Full Waveform Inversion on the Cloud
Use XWITM on AWS to discover an unprecedented increase in the resolution of your velocity model.