Improving subsurface imaging in geological complex area: Structure PReserving INTerpolation in 6D (SPRINT6D)

Transcription

1 Improving subsurface imaging in geological complex area: Structure PReserving INTerpolation in 6D (SPRINT6D) Dan Negut* and Mark Ng, Divestco Inc. Copyright 2008, ACGGP This paper was selected for presentation by an ACGGP Technical Committee following review Of information contained in an abstract submitted by the author(s). Abstract SPRINT6D is a refined interpolation technique that improves upon the 5D minimum weighted norm interpolation (MWNI) for seismic processing. It is a frequency domain Fourier inversion method incorporated with a constrained a priori model to guide the inversion process. The constraint is achieved by imposing an angular weight function that is derived from scanning of many different dips of the input data in the f-k frequency-wavenumber domain. The objective is to preserve structural integrity of the data after interpolation. Real data examples from around the world shown in this presentation demonstrate successful structure preservation for interpolation challenges posed by: (1) 3D up-sampling megabin acquisition, (2) Poor 3D offset/azimuth coverage, (3) 3D block gap acquisition, and (4) 2D sparse acquisition. AWMWNI is illustrated in the 2015 paper. Those who are interested are encouraged to refer to them. Examples Case studies from real data examples will be presented during the talk. First example is a 3D dataset from Kurdistan. Imaging problems to be addressed are: randomly missing traces, block gaps in acquisition, lack of uniform illumination, steeply dipping aliased horizons, rough topography. The 6D Interpolation strategy is to use COV domain SPRINT6D. Main targets are preserving structural integrity and AVO effects. Introduction Based on the work of Liu and Sacchi (2004), the 5D interpolation method for 3D prestack seismic data by MWNI has been widely used since its introduction by Trad (2009, 2014). However, in that method there is no global data insight extracted from the frequency-wavenumber domain since an independent 4D MWNI is applied in space for every frequency slice, one slice at a time. In order to correct such deficiency, Ng et al. (2013, 2014, 2015, 2016) suggested integrating a Fourier angular stack concept into MWNI to connect data information across all frequency-wavenumbers, and thereby raising the dimension of the MWNI by one. Applying the extra dimension of angular weight (AW) to 5D MWNI for 3D prestack data results in a 6D AWMWNI. A structure preserving interpolation is desired for regularization and upsampling. Figure 1. A stack inline section without SPRINT6D. Method The methodologies, equations, and the proof of concepts of the angular weighted MWNI have been described in detail in the past few papers of Ng et al. (2013, 2014, 2015, 2016). The 6D Figure 2. Same as fig.1, stack inline with SPRINT6D. Note both the structural preservation and inherent S/N ratio enhancement.

2 2 NEGUT AND NG Figure 3. Inline PSTM without SPRINT6D. Steeply dipping deep reflectors are aliased while dipping shallow reflectors are not continuous. Figure 6. PSTM with SPRINT6D. Timeslice along yellow line in fig. 4. Due to proper illumination, the structural integrity is preserved, while imaging is improved. Second example is from Eastern Alberta, Canada. Four 3D surveys with different acquisiton parameters were seamessly merged. Imaging problems to be addressed are: steep dipping diffraction patterns in order to properly image discontinuities, randomly missing traces, block gaps in acquisition, lack of uniform illumination, shallow (500 to 900 msec) geological targets. The 6D Interpolation strategy is to use COV domain SPRINT6D. Main targets are preserving structural integrity and AVO effects. Figure 4. Inline PSTM with SPRINT6D. Steeply dipping deep reflectors are de-aliased while dipping shallow reflectors are continuous. Note the structural integrity is preserved and faulting patters are sharpened. Figure 7. Fold map before SPRINT6D. Figure 5. PSTM with no SPRINT6D. Timeslice along yellow line in fig. 4.

3 IMPROVING SUBSURFACE IMAGING BY 6D INTERPOLATION 3 Figure 8. Fold map after SPRINT6D. Figure 11. PSTM NO SPRINT6D. Timeslice. Figure 9. Inline PSTM no SPRINT6D. Figure 12. PSTM SPRINT6D. Timeslice. Note the remarkable structural integrity preservation, improved imaging of short wavelength features associated with discontinuities and alleviated footprint patterns. Figure 10. Inline PSTM SPRINT6D. Note the better discontinuities definition, shallow horizon continuity while structural integrity is preserved. Third example is from Llanos Basin, Colombia. Eight orthogonal 3D surveys with different acquisiton parameters (450 to 600 m source line separation, 70 to 120 m source interval, 250 to 300 m receiver line separation, 50 to 60 m receiver interval) were seamessly merged to a 30 by 30 m CDP binning scheme. Imaging problems to be addressed are: subtle diffraction patterns preservation in order to properly image discontinuities edges, very short wavelength geological features, randomly missing traces, block gaps in acquisition, lack of uniform illumination, upsampling to 30 by 30 m CDP bins,medium to deep (1600 to 2200 msec) geological targets. The 6D Interpolation strategy is to use COV domain SPRINT6D. Main targets are preserving structural integrity and AVO effects.

4 4 NEGUT AND NG Figure 13. Fold map before SPRINT6D. Note the difference in acquisition parameters. Figure 16. Inline PSTM. No SPRINT3D Figure 14. Inline Stack. No SPRINT3D. Figure 15. Inline Stack. SPRINT3D. Note greatly improved horizon continuity, sharp discontinuity edges and illumination coverage. Figure 17. Inline Stack. SPRINT3D. Note greatly improved horizon continuity, sharp discontinuity edges, illumination coverage and amplitude normalization. Fourth example is from Western Canadian Basin. Seventeen orthogonal and megabin 3D surveys with different acquisiton parameters (120 to 400 m source line separation, 50 to 120 m source interval, 120 to 240 m receiver line separation, 50 to 100 m receiver interval, natural fold ranging from 16 to 196 coverage) were seamessly merged to a 25 by 25 m CDP binning scheme. Imaging problems to be addressed are: subtle diffraction patterns preservation in order to properly image discontinuities edges, very short wavelength geological features, randomly missing traces, block gaps in acquisition, lack of uniform illumination, shallow to medium (400 to 1000 msec) geological targets. The 6D Interpolation strategy is to use COA domain SPRINT6D. Main targets are preserving structural integrity and AVO effects while merging so many different surveys.

5 IMPROVING SUBSURFACE IMAGING BY 6D INTERPOLATION 5 Figure 18. Fold map before SPRINT6D. Note the difference in acquisition parameters, as well as between orthogonal and megabin surveys. Figure 21. Inline Stack. SPRINT3D. Note the vast overall improvement Figure 22. Timeslice along the yellow line in Fig. 19. Stack.No SPRINT3D. Figure 19. Fold map after SPRINT6D. Note the difference in acquisition parameters, as well as between orthogonal and megabin surveys is solved, leading to an uniform illumination. Figure 23. Timeslice along the yellow line in Fig. 19. Stack.SPRINT3D. Note the remarkable structural integrity preservation, improved imaging and footprint attenuation. Figure 20. Inline Stack. No SPRINT3D.

6 6 NEGUT AND NG Figure 24. Inline Stack.No SPRINT3D. Orthogonal and Megabin survey merge zone. Figure 27. Horizon slice along the yellow line in Fig. 23. Stack. SPRINT3D. Note the remarkable, seamless merge between the othogonal, and megabin surveys, structure integrity preservation, vastly improved imaging quality and interpretability. Conclusions Figure 25. Inline Stack. SPRINT3D. Note the vast overall improvement Figure 26. Horizon slice along the yellow line in Fig. 23. Stack. No SPRINT3D. After examining a handful of case studies presented in this paper, structure preserving interpolation in 6D is demonstrated, and uplifts in interpretation can be seen, leading the way to better, lower risk imaging. References Liu, B. and M. D. Sacchi, 2004, Minimum weighted norm interpolation of seismic record: Geophysics, 69, no. 6, Ng, M. and D. Negut, 2013, Data interpolation by angular weighted MWNI: Annual GeoConvention meeting in Calgary, Expanded Abstracts. Ng, M. and D. Negut, 2014, Minimum weighted norm interpolation with an angular weighted deconvolved prior: 84 th Annual International Meeting, SEG, Expanded Abstracts, Ng, M., X. Wang and D. Negut, 2015, 6D interpolation by incorporating angular weight constraints into 5D MWNI: 85 th Annual International Meeting, SEG, Expanded Abstracts, Ng, M. and D. Negut, 2016, Structure-preserving 6-D interpolation: AAPG EXPLORER, Geophysical Corner, July issue, 26. Trad, D., 2009, Five-dimension interpolation: Recovering from acquisition constraints: Geophysics, 74, no. 6, V123-V132. Trad, D., 2014, Five-dimension interpolation: New directions and challenges: CSEG Recorder, 39, no. 3,