FinalPresentation_200630888 (1)
- 1. A CRITICAL ASSESSMENT OF A NEW POST-STACK BATCH-Q-ESTIMATION ALGORITHM Daniel Woods MSc Exploration Geophysics Module: SOEE5110M Student ID: 200-630-888
- 2. WHAT IS 𝑄? Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary 1 𝑄eff = 1 𝑄int + 1 𝑄app Quantifies the progressive loss of amplitude with time. Fraction of energy loss per radian is proportional to 𝑄−1 - so the lower the value of 𝑄, the more attenuative the medium. Estimating 𝑄 from surface seismic measures the Effective Quality factor - the inseparable combination of Intrinsic and Apparent attenuation.
- 3. 𝑄APP VS 𝑄INT Intrinsic Prominent mechanism is Wave Induced Fluid Flow. Absorption of elastic energy due to frictional forces. Müller, T. M., Gurevich, B., & Lebedev, M. (2010). Seismic wave attenuation and dispersion resulting from wave-induced flow in porous rocks—A review. Geophysics, 75(5), 75A147-75A164. Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary Apparent Frequency dependant effects such as Scattering. Can be determined from well log information.
- 4. 𝑄 EFFECT Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary Loss of amplitude Lower dominant frequency Poorer resolution Preferential loss of higher frequencies
- 5. AIMS 1. Test and assess the feasibility of a freely released post-stack 𝑄 estimation tool, which is named ‘𝑄-est’, on both synthetic and real data. 2. Assess the application of an inverse−𝑄 filter to real data using the estimated 𝑄 fields in order to compensate for attenuation. Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary
- 6. 𝑄-EST METHOD Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary 1. Start on one trace, calculate power spectra of two windows separated by time 𝛿𝑡 (sample rate) and each of window length 𝑛. 2. 3. 4. 𝑛 Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary 𝑛 𝑙𝑛 𝐴2(𝑓 𝐴1(𝑓 = −𝜋 𝛿𝑡 𝑄 𝑓 + ln(𝑅 ∙ 𝐺 𝐴2(𝑓 𝐴1(𝑓 Where R and G are the reflection coefficient and geometrical spreading factor,
- 7. 𝑄-EST METHOD Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary 1. 2. Calculate the natural log of the spectral ratios. 3. 4. 𝑙𝑛 𝐴2(𝑓 𝐴1(𝑓 = −𝜋 𝛿𝑡 𝑄 𝑓 + ln(𝑅 ∙ 𝐺 𝑦 = 𝑚𝑥 + 𝑐 Where R and G are the reflection coefficient and geometrical spreading factor,
- 8. 𝑄-EST METHOD Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary 1. 2. 3. Compute the slope between a specified bandwidth to solve for 𝑄. Output this value at the top of the second window. 4. 𝑦 = 𝑚𝑥 + 𝑐 𝑙𝑛 𝐴2(𝑓 𝐴1(𝑓 = −𝜋 𝛿𝑡 𝑄 𝑓 + ln(𝑅 ∙ 𝐺 Bandwidth Slope Where R and G are the reflection coefficient and geometrical spreading factor,
- 9. 𝑄-EST METHOD Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary 1. 2. 3. 4. Repeat for each time sample and trace. 𝑦 = 𝑚𝑥 + 𝑐 𝑙𝑛 𝐴2(𝑓 𝐴1(𝑓 = −𝜋 𝛿𝑡 𝑄 𝑓 + ln(𝑅 ∙ 𝐺 Bandwidth Slope Where R and G are the reflection coefficient and geometrical spreading factor,
- 10. DATA Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary Modern long offset 2D data. Data were acquired with 8,000 m offsets, shot as regional cross border lines and recorded 9 seconds of data. Batch 1: 18,649 km (Red lines), covers full extent of survey. Processing sequence is consistent and images are of high-quality broadband across the North Sea.
- 11. TESTING SYNTHETIC DATA Key a) Original synthetic (no 𝑄 applied) b) Assumed 𝑄 field c) Synthetic after forward modelling for 𝑄 d) Recovered 𝑄 field from 𝑄-est tool e) 𝑄 compensated image Synthetic seismograms were generated from wells that tied to the dataset. Attempting to recover an assumed 𝑄 field. Recovered field shows good correlation to assumed with percentage error of 16%. Amplitudes were sufficiently recovered in 𝑄 compensation and had a correlation coefficient of 0.98 to original synthetic. Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary Key a) Original synthetic (no 𝑄 applied) b) Assumed 𝑄 field c) Synthetic after forward modelling for 𝑄 d) Recovered 𝑄 field from 𝑄-est tool e) 𝑄 compensated image
- 12. ESTIMATED 𝑄 FIELDS Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary
- 13. LATERAL CONSISTENCY Zoomed in image shows high temporal resolution Lateral consistenancy between formation tops is good
- 14. ERRONEOUS HIGH ANOMALIES Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary Anomalous high 𝑄 values are associated with anomalous high dominant frequencies or upwards shifts in frequency content. Can be caused by unresolvable thin beds. Results in slope of natural log of spectral ratios to be very shallow (High 𝑄) or positive (Negative 𝑄 – output null value). Resolved by narrowing bandwidth of which slope is fit or clipping values down the High dominant frequency
- 15. SHALLOW GAS Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary Low dominant frequency Anomalous low 𝑄 values are associated with anomalous downwards shifts in frequency content. Causes slope of natural log of spectral ratios to be relatively high (low 𝑄). Indicates highly attenuating bodies such as gas reservoirs. NMO stretch broadens the wavelet and lowers the dominant frequency causing inaccurate low 𝑄 values. This
- 16. 𝑄 VERSUS OFFSET Propogating wave spends longer in attenuating medium at farther offsets. Stacking procedure can cause bias in 𝑄 estimates and ‘smeared’ results. Magnitude of smearing can be observed in near, mid and far stack estimated 𝑄 fields. Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary Decreasing 𝑄 estimates
- 17. 𝑄-COMP – RAW FIELD Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary
- 18. 𝑄-COMP - SMOOTHED Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary
- 19. 𝑄-COMP – NO COMPENSATION Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary
- 20. 𝑄-COMP – ESTIMATED FIELD Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary
- 21. 𝑄-COMP – SINGULAR VALUE Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary
- 22. 𝑄-COMP – ESTIMATED FIELD ZOOM Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary
- 23. 𝑄-COMP – SINGULAR VALUE ZOOM Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary
- 24. 𝑄-COMP – SPECTRA Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary Preferential boosting of higher frequencies Better resolution Lower ambient noise Flatter spectrum No compensation Estimated Q field Singular value
- 25. 𝑄-COMP – TRACE Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary No Compensation Estimated field compensation Single value of 110 compensation Better resolution Increasing amplitudes
- 26. SUMMARY Concept of Q AIMS Q Est Data Testing Estimated Q Fields Q-comp Summary The 𝑄-est tool is simple to implement but requires care in that scaling, demultiple and NMO have been applied properly. Parameterisation is required of the window length and signal bandwidth over which the slope of the natural log of the spectral ratios is computed. Anomalous 𝑄 values can be observed to correlate with the seismic data structure as well as hydrocarbon reservoirs. Using the clipped and smoothed 𝑄 fields in 𝑄 compensation successfully recovers amplitudes and provides better resolution than using a singular 𝑄 value of 110 in 𝑄 compensation.
- 27. ACKNOWLEDGEMENTS
- 29. Q-COMP – NO COMPENSATION
- 30. Q-COMP – ESTIMATED FIELD
- 31. Q-COMP – SINGULAR VALUE
- 32. Q-COMP - SPECTRA
Editor's Notes
- #2: Not actually new algorithm, created 20 years by some chaps working for Amoco, but it’s been freely released and now aiming to incorporate this into the PRIMA software.
- #3: Using surface seismic to quantify attenuation has a drawback in a thinly layered Earth in that it cannot distinguish between 𝑄 int , the intrinsic quality factor describing anelastic absorption, and 𝑄 app , the apparent quality factor describing frequency-dependant attenuation effects such as scattering.
- #4: There are many mechanisms that influence attenuation such as mechanical compression, grain boundary friction and bubble compression. However, the most dominant attenuation mechanism in porous media is that of wave-induced fluid flow. The most prominent contributer to Qeff is often unknown. Therefore this can be a problem if using Qe as a seismic attribute ie pore fluid characterisation
- #5: So areas under highly attenuating bodies such as gas reservoirs, will have lower frequencies and amplitudes. If we can quantify the magnitude of attenuation – what does it tell us about the subsurface (relate to previous slide ie cause) and how can we compensate or it
- #6: By doing this, TGS hope to incorporate the 𝑄 eff estimation tool into their proprietary processing software and utilise it in their processing flows.
- #7: Simple to use with no special processing required and few parameters required for input. Estimates effective seismic Q Employs an algorithm based on a method of spectral ratios. Ratios are computed one sample at a time using a continuously time-variant DFT (Discrete Fourier Transform). Simple to use with no special processing required and few parameters required for input. Important that data have not already been Q compensated.
- #8: Simple to use with no special processing required and few parameters required for input. Estimates effective seismic Q Employs an algorithm based on a method of spectral ratios. Ratios are computed one sample at a time using a continuously time-variant DFT (Discrete Fourier Transform). Simple to use with no special processing required and few parameters required for input. Important that data have not already been Q compensated.
- #9: Simple to use with no special processing required and few parameters required for input. Estimates effective seismic Q Employs an algorithm based on a method of spectral ratios. Ratios are computed one sample at a time using a continuously time-variant DFT (Discrete Fourier Transform). Simple to use with no special processing required and few parameters required for input. Important that data have not already been Q compensated.
- #10: Simple to use with no special processing required and few parameters required for input. Estimates effective seismic Q Employs an algorithm based on a method of spectral ratios. Ratios are computed one sample at a time using a continuously time-variant DFT (Discrete Fourier Transform). Simple to use with no special processing required and few parameters required for input. Important that data have not already been Q compensated.
- #16: Still good to show testing even if it didn’t work. Can always have 20 major slides then loads more prepared at the end of your presentation ready to go into more details if questions are asked