1. Rainbow B time-lapse results
Monitoring fluid injection using time-lapse analysis: a Rainbow Lake case study
By Hannah Tran Ng,
Larry Bentley, Ed Krebes
November 19, 2004
Significance: time-lapse is not usually done for carbonate pools. Usually done for clastic reservoirs in the North sea.
Data used: seismic data and CMG simulation data and geological data: now combine all three to make an interpretation
I did some modeling but that didn’t work. Gassmann equation doesn’t work for this but there is no problem because the pore type affects the results
Hannah T Ng

2. Gas and solvent were injected into the Rainbow B pool, a carbonate reservoir, to help extract the remaining oil
OBJECTIVE
To determine the locations of the injected gas and solvent using time-lapse analysis

3. What is time-lapse analysis?
4D seismic - 4th dimension is calendar time
Refers to repeating a seismic survey after a period of time in an effort to image changes that could have occurred in a reservoir
Time-lapse analysis is useful:
Improve production by finding bypassed oil
reservoir changes in between wells can be detected

4. Significance of Study
Not much time-lapse work done on carbonates because the fluid changes are difficult to detect
BUT, time-lapse analysis of Rainbow B shows that the fluid changes are bigger than expected due to the pore geometry

5. Main result
Time-lapse analysis appears to detect the presence of gas and solvent in some, but not all locations

6. Outline Background Time-lapse results:
Time-delay map
Amplitude change map
Compare time-lapse results to geology and engineering data

7. BACKGROUND

8. Rainbow B pool: 5.6 km * 2.1 km Thickness of ~200 m Depth ~ 1800 m
Rainbow B time-lapse results
Alberta
Doesn’t matter about injection sites b/c things always get injected and produced so don’t really know where the oil and gas is located. Maybe take out well and because nothing changed between those times.
Rainbow B pool: km * 2.1 km Thickness of ~200 m Depth ~ 1800 m
Hannah T Ng

9. Seal rock: evaporitic Muskeg member
Reservoir rock: Keg River formation (mostly dolomitic) is producing oil
Seal rock: evaporitic Muskeg member
(Laflamme, 1993)

10. Pore geometry Pore geometry affects the velocity changes
Fluid substitution in low pore aspect ratio (cracks) rock causes greater velocity change than high pore aspect ratio (round) rock. (Kuster & Toksoz, 1974)
Rainbow B reef is mostly vuggy and has a low pore aspect ratio.
The Gassmann equation underpredicts the velocity changes

11. Rainbow B time-lapse results
Core from well 7-10: reef mostly dolomitized
High porosity
Vuggy
Intergranular
Hannah T Ng

12. Rainbow Pool Timeline
1965 – pool was discovered and oil produced by natural drives: primary production
1968 – pool waterflooded: secondary production
1984 – miscible gas and solvent injection: tertiary production
1987 – 3D seismic data acquired in area
2002 – 3D seismic data acquired again

13. Fluid contacts from 1987 to 2002 Red: gas & solvent Green: oil
Blue: water

14. Gas Plus Solvent Thickness (m)
1987
2002
Difference

15. TIME-LAPSE RESULTS

16. Seismic changes expected from the injection of gas and solvent

17. Rainbow B time-lapse results
1987 seismic data (Base)
Crossline 130
Raw data, 1987
Hannah T Ng

18. Rainbow B time-lapse results
2002 seismic data (Monitor)
higher frequency content
2002 data
Hannah T Ng

19. Rainbow B time-lapse results
Difference between the 1987 and the 2002 survey
Differece between shifted and 1987
Hannah T Ng

20. Time-delay map (ms)

21. Difference RMS amplitude map from 1987 to 2002

22. COMPARE SEISMIC TIME-DELAY MAPS WITH OTHER MAPS

23. Isochron map (Keg River to Cold Lake)
Time-delay map (ms)

24. Fluid thickness difference (m)
Time-delay map (ms)

25. Gassmann calculated time-delay map (s)
Time-delay map (ms)

26. Porosity type Red: large vugs
Green: vuggy with reservoir quality porosity
Blue: intergranular
Porosity type map
Time-delay map (ms)

27. Conclusions
Time-lapse analysis detected the injected fluids in some, but not all locations.
Time-delay results are most useful
Vuggy areas show more response than intergranular areas
Amplitude change and impedance change results were not useful
Gassmann equation underpredicts the velocity change

28. Acknowledgements Husky Energy for the data U of C CREWES group
Larry Mewhort has helped me tremendously in supplying data and suggestions
Ken Hedlin for his suggestions
Andre Laflamme – geologist who showed me the cores
Larry Carr and the Husky B Pool Asset Team Engineers and Geologists – for CMG simulation results
U of C CREWES group
John Zhang, Ying Zou
Hampson-Russell
Keith Hirsche
Francis Ma