Bandwidth Extension applied to 3D seismic data on Heather and Broom Fields, UK North Sea

Transcription

1 Bandwidth Extension applied to 3D seismic data on Heather and Broom Fields, UK North Sea Tim Trimble 1., Clare White 2., Heather Poore EnQuest Plc 2. Geotrace Technologies Ltd DEVEX Maximising Our Diverse Resources 156 th May 213

2 Heather Field Background: Seismic Dataset Heather and Broom fields are situated in UK Block 2/5 and 2/4a at the western margin of the East Shetlands Basin Heather Field 1% EnQuest, Broom Field partners EnQuest, Wintershall Norge ASA, Ithaca Energy (UK) Limited 316km 2 pre stack depth migration Shot by PGS 26 stretched back to time. East-West shooting 51m solid streamers at 6m depth Source depth 5m. Processed 27 also PGS much effort on multiple suppression below Base Cretaceous 2

3 Heather-Broom: Ongoing Development Issues Thin Brent Sandstones 125ft-375ft, typically ~1 cycle on conventional seismic Shallow marine to deltaic clastic system individual formations exhibit highly variable reservoir qualities Heavily faulted reservoir Faults with minor throws can operate as baffles/barriers to flow Complex Diagenetic History Kaolinite, Quartz, Illite and Calcite all occur Variable Acoustic Impedances Above and within the Brent Variable pick at Top (and Base) Brent H H61Z H H58Z 4 2/53Z H11Z H H38 2/ H18 H63 3 H5 8 H1Z H1 H25 Heather Platform H1 H /55 H48 H55 2 H32 H36 H36Z 2/5-26Z 3 H43 H H13 2/5 H51 H22 H7 H9 7 H4 2 H52 H23Z H35 H53 H6 H44 H15 H H62Y H17 1 H14 H37 H2 1 H H6Z 2/57 7 H59Z /5-24 (BR1) H4 7 H H57 2/ /5-8B 2/5-7 H24 H16 2/5-25 (BR2) 2/5-2 (BR5) -99 H49 H2 2/5-23 (BR4) 5 H57Z /5-6 2/5-5 H82/5-4 H47 H3-97 2/5-21 (BW 3) H3 H5Y H H H H H H31Z /5-3 2/5-26 2/ /59Y 1 H54 2/52A 4 H42 H31 16 H39Z /54Z Seismic resolution is critical for further development H / /5-22Z (BW 6) Broom STOOIP ~152mmbbls Recovered ~33mmbbls Heather STOOIP ~5mmbbls Recovered ~12mmbbls m 1:

4 A simplified rendition of Bandwidth Extension theory 1. BE utilises the Continuous Wavelet Transform (CWT) to perform a time series analysis of the input seismic trace that decomposes the trace into its respective amplitude and phase components in both frequency and time. Smith et al, 28 (Leading Edge) 4

5 A simplified rendition of Bandwidth Extension theory 1. BE utilises the Continuous Wavelet Transform (CWT) to perform a time series analysis of the input seismic trace that decomposes the trace into its respective amplitude and phase components in both frequency and time. One Octave Smith et al, 28 (Leading Edge) 2. The frequencies present in the bandwidth of the input seismic trace (the fundamental frequencies) are used to predict harmonics (and possibly subharmonics) beyond the chosen pivot frequency. 5

6 A simplified rendition of Bandwidth Extension theory 1. BE utilises the Continuous Wavelet Transform (CWT) to perform a time series analysis of the input seismic trace that decomposes the trace into its respective amplitude and phase components in both frequency and time. One Octave First Harmonic Smith et al, 28 (Leading Edge) 2. The frequencies present in the bandwidth of the input seismic trace (the fundamental frequencies) are used to predict harmonics (and possibly subharmonics) beyond the chosen pivot frequency. 3. A convolution-like process is employed to convolve the predicted harmonic (and sub-harmonic) information onto the initial seismic trace. If reflectivity at low amplitudes is present in the input data that corresponds to the harmonic (and sub-harmonic) predictions, the extended frequencies will remain in the result. However, if extended harmonics or sub-harmonics frequencies do not correspond to reflectivity in the input data, these extended harmonics will drop out of the result. 6

7 A simplified rendition of Bandwidth Extension theory 1. BE utilises the Continuous Wavelet Transform (CWT) to perform a time series analysis of the input seismic trace that decomposes the trace into its respective amplitude and phase components in both frequency and time. Smith et al, 28 (Leading Edge) 2. The frequencies present in the bandwidth of the input seismic trace (the fundamental frequencies) are used to predict harmonics (and possibly subharmonics) beyond the chosen pivot frequency. 3. A convolution-like process is employed to convolve the predicted harmonic (and sub-harmonic) information onto the initial seismic trace. If reflectivity at low amplitudes is present in the input data that corresponds to the harmonic (and sub-harmonic) predictions, the extended frequencies will remain in the result. However, if extended harmonics or sub-harmonics frequencies do not correspond to reflectivity in the input data, these extended harmonics will drop out of the result. 4. The resulting broader bandwidth amplitude and phase spectra are used to reconstruct the modified seismic trace. The resulting modified seismic traces has broader bandwidth and thus increased temporal resolution. Result is less noisy than traditional bandwidth enhancement techniques. 7

8 Bandwidth Analysis Deterministic Wavelets Bandwidth Extension: Key Steps Well Data Well-toseismic Tie Phase Information BE Well Tie Final BE Result Post-Stack Normal Bandwidth Input Zero Phase Input Data BE Result Horizons Interpretation? Windows for BE parameterisation 8

9 Bandwidth Extension: Key Steps Phase Analysis Well Data Well-toseismic Tie Phase Information Post-Stack Normal Bandwidth Input Zero Phase Input Data VSP Corridor Stack Normal Bandwidth Stack Average Average phase of the input volume was assessed to be Seismic data must be rotated in the opposite direction (23 ) in order to zero phase the data to a polarity convention in which an acoustic impedance is a peak, as required by the algorithm. (Example: Well 2/57 Tie)

10 Bandwidth Extension: Key Steps Frequency Analysis Post-Stack Normal Bandwidth Input Horizons Interpretation? Windows for BE parameterisation Spatially the frequency content appeared relatively constant across the survey (multiple inlines and xlines tested), and temporally the frequency content supported the use of a three window parameterisation 1

11 Results 11

12 12 Normal Bandwidth Stack, Inline 355 Window 3 (Primary Target Interval) 42 Hz

13 Bandwidth Extension Stack, Inline 355 Window 3 (Primary Target Interval) 42 Hz Hz

14 PSDM data with Noise Cancellation North 5m South

15 PSDM data with Noise Cancellation and Spectral Whitening North 5m South

16 PSDM data after Blueing North 5m South

17 PSDM data (Noise Cancelled) with 1 Octave Bandwidth Extension North 5m South

18 PSDM data (Noise Cancelled) with filtered 2 Octave Bandwidth Extension North 5m South

19 PSDM data (Noise Cancelled) with unfiltered 2 Octave Bandwidth Extension North 5m South

20 Combined Amplitude Spectra (Xline 771, secs)

21 TWT (s) 2/57 Synthetic-Seismic Tie: Input PSDM-Bandwidth Extended Comparison 2/57 Synthetic Trace- 25Hz Ricker 2/57 Synthetic Trace- 4Hz Ricker NW Sonic SE NW SE Top Brent 5m INPUT PSDM DATA BANDWIDTH EXTENDED DATA 3ms-3ms Window

22 TWT (s) Well 2/5-3: Synthetic ties before and after 1 Octave Bandwidth Extension Input PSDM Data 1 Octave Extended Data

23 23 Well 2/5-3 Seismic Tie: Input PSDM-Bandwidth Extended Comparisons Extracted Wavelet Extracted Wavelet 45Hz Ricker Wavelet INPUT PSDM DATA BANDWIDTH EXTENDED DATA (1 Octave) BANDWIDTH EXTENDED DATA (2 Octave- Filtered)

24 Summary Synthetic to Stack Correlation and Phase Analysis: 1 Octave Bandwidth Extension stack 24

25 25 Well 2/57: Phase behaviour of deterministic wavelets before and after Bandwidth Extension Well 2/ ms Normal Bandwidth 1 Octave Bandwidth Extension 1 Octave Bandwidth Extension with Stretch-Squeeze

26 26 Improved fault plane definition on Bandwidth Extended data INPUT PSDM DATA BANDWIDTH EXTENDED DATA (1 Octave) Top Brent Top Brent 1km

27 27 Similarity Extractions Comparison at Top Brent INPUT PSDM DATA BANDWIDTH EXTENDED DATA (1 Octave)

28 Broom Field: Comparison of Top Brent fault mapping with Bandwidth Extended data INPUT PSDM DATA BANDWIDTH EXTENDED DATA (1 Octave)

29 29 Conclusions Bandwidth Extension has significantly improved the resolution of the Heather-Broom seismic dataset Synthetic ties confirm little loss of Signal/Noise Improvement over previous frequency enhancement techniques Most improvement in fault definition Brent formation properties and diagenesis remain a challenge But more work to be done

30 3 Acknowledgements Thanks to EnQuest and Geotrace Technologies for permission to publish this paper Also to Broom partners - Wintershall (UK North Sea) Limited and Ithaca Energy (UK) Limited All colleagues at EnQuest and Geotrace for input and help throughout Contact: Clare White, cwhite@geotrace.com Awarded for Bandwidth Extension

31 Geotrace Technologies, Ltd May, 213