Introduction to Reflection Seismics: Chapter 5

  • Published on
    26-May-2015

  • View
    3.059

  • Download
    5

Embed Size (px)

Transcript

<ul><li> 1. Overview ta3520Introduction to seismics Fourier Analysis Basic principles of the Seismic Method Interpretation of Raw Seismic Records Seismic Instrumentation Processing of Seismic Reflection Data Vertical Seismic ProfilesPractical: Processing practical (with MATLAB)</li></ul> <p> 2. Signal and NoiseSignal: desiredNoise: not desiredSo for reflection seismology:- Primary reflections are signal- Everything else is noise! 3. Signal and Noise (2) Direct wave: noise Reflection: (desired) signalRefraction: noise 4. Signal and Noise (3) Direct wave: noise Multiply reflected : noise Reflection: signal Refraction: noise 5. Signal and Noise for P-wave surveyDesired signal: primary reflected P-wavesNoise: direct wave through first layer direct air wave direct surface wave S-wave Multiply reflected wave Refraction / Head wave 6. Signal and Noise for P-wave survey Signal Primary P-wave Reflected Energy= NoiseAll but Primary Reflection EnergyGoal of Processing:Remove effects of All-but-Primary-Reflection Energy 7. Processing of Signal(Primary-reflected energy)Goal of processing:Focus energy to where it comes from 8. Understanding signal and noise:wave theoryBasic physics underlying signal is captured by wave equationRay theory: approximation of wave equation (high-frequency)Resonances: modes expansion of wave equationS-waves, P-waves: elastic form of wave equation 9. Seismic Processing Basic Reflection and Transmission Sorting of seismic data Normal Move-Out and Velocity Analysis Stacking (Zero-offset) migration Time-depth conversion 10. Basic Reflection and Transmission (pdf-file with eqs) 11. Processing flowshot dataCMP sorting velocity model NMO correctionstack velocity model(zero-offset) migrationimage 12. ProcessingInput: Multi-offset shot recordsResults of processing: 1. Structural map of impedance contrasts 2. Velocity model 13. Sorting:Common Shot gatherSeismic recording in the field: Common Shot data(Each shot is recorded sequentially)Nomenclature: - common-shot gather - common-shot panel 14. Common Shot gather 15. Sorting:Common Receiver gatherGather all shots belonging to one receiver position in the fieldAnalysis/Processing: shot variations(e.g., different charge depths)(Also in common-shot gathers: receiver variations, e.g.,geophones placed at different heights) 16. Common Receiver gather 17. Sorting 18. Sorting:Common Mid-Point gather 19. Sorting:Common Mid-Point gatherMid-points defined as mid-points between source and receiverin horizontal planeSince reflections are quasi-hyperbolic: Seismograms not so sensitive to laterally varying structures Good for velocity analysis in depth Stacking successful (noise suppression)In practice, not really a point but an interval: BIN 20. CMP gather over structure 21. Sorting: Common Offset gatherPurpose: Very irregular structures (in which stacking does not work) Application of Dip Move-Out (correction for dip of reflector) Checking on migration: small and large offsets should give the same picture : otherwise velocities are wrongIn practice, not really a point but an interval: BIN 22. Zero-offset gather over structure 23. SortingCommon-Mid-Point (CMP) gathers: xs + xr = constantCommon-Offset gathers (COG): xs xr = constantMultiplicity = Fold:Nrec2 xs / xr Nrec = Number of receivers xs = Spacing between subsequent shots xr = Spacing between subsequent receivers 24. Sorting 25. Processing flowshot dataCMP sorting velocity model NMO correctionstack velocity model(zero-offset) migrationimage 26. Reflection 1 boundary 27. Normal Move-Out: 1 reflectorR(4d2 + x2)1/2T = =cc x = source-receiver distance R = total distance travelled by ray d = thickness of layer c = wave speedWe do not know distance, but we know time: x2 T = T0 ( 1 +)1/2c2 T02 where T0 is zero-offset (x=0) traveltime: T0 = 2d/c 28. x2T = T0 ( 1 +)1/2 c2 T02Extra time shift compared to T0 called: NMO- Normal Move-Outx2 TNMO = T T0 = T0 ( 1 + 2 2 )1/2 T0 c T0 29. Normal Move-Out (NMO) 30. Normal Move-Out (NMO) 31. x2 TNMO = T T0 = T0 ( 1 + 2 2 )1/2 T0c T0 Larger TNMO for larger offset Smaller TNMO for larger T0 (deeper layers have smaller move-out) Smaller TNMO for larger wave speed c (deeper layers usually larger velocities so smaller move-out) 32. Normal Move-Out (NMO)Input CMP-gather NMO-corrected CMP gather (with right velocity) 33. NMO: effect velocityNMO with right NMO with too small NMO with too largevelocity correction: too high correction: too small velocity velocity 34. NMO: 1-layer approximationx2 TNMO = T T0 = T0 ( 1 + 2 2 )1/2 T0 c T0 Use Taylor expansion of square root:x2 x2 TNMO = T T0 T0 ( 1 + ) T0 = 2c 02T 22 c2 T0 35. NMO: 1-layer approximation 36. NMO: 2-layer Position depends on velocities of layer 1 and 2 (via Snells Law) 37. NMO: 2-layer (pdf-eqs) 38. NMO: multi-layer approximation 39. Velocity model: RMS model velocityInterval velocity(velocity of layer)time Root-mean-square velocity (weighted-average velocity of layers above) 40. Velocity AnalysisApproaches: T2 x2 analysis Alignment of reflectors: visually or mathematical expression of coherenceWith T2 X2 analysis we depend on picking travel-times,and thus signal-to-noise ratio 41. Velocity Analysis:Original CMP gatherStarting CMP gather 42. T 2-x2 analysis 43. Velocity Analysis: aligning reflectorsStarting CMP gather 44. Constant-velocity NMOVelocity too low: Velocity too high:correction too much correction too little 45. Velocity panelsCMP gatherwith velocity CMP gatherCMP gather 1300 m/s with velocity with velocity 1700 m/s2100 m/s 46. Velocity as function of time 47. Velocity panels:real data example 48. Coherence measures for velocity analysis Stacked amplitude (normalized or not) Cross-correlation (normalized or not) Semblance:related to cross-correlation1 ( m A (xm , t , c) )2S (t , c) =Mm A2 (xm , t , c)A = amplitudeSum m over traces m in CMP 49. Semblance for CMP gather 50. Velocity panels:real data example 51. Velocity panels:real data example 52. Velocity panels:real data example 53. Factors affecting velocity estimation (Yilmaz, 1988) Spread length Stacking fold / Signal-to-Noise ratio (fold = multiplicity in CMP) Choice of coherence measure Departures from hyperbolic move-out NMO stretch 54. Effect spread length on velocity estimation 55. Effect spread length on velocity estimation 56. Effect spread length on velocity estimation 57. Effect spread length on velocity estimation 58. Velocity model: RMS model velocityInterval velocity(velocity of layer)time Root-mean-square velocity (weighted-average velocity of layers above) 59. Velocity model: RMS modelTime (m-seconds) CMP location 60. Processing flowshot dataCMP sorting velocity model NMO correctionstack velocity model(zero-offset) migrationimage 61. Applying NMOAmount x2/(c2 T02) never exactly on a sample:INTERPOLATION 62. NMO stretch (via picture)Since T0 is smaller: shift is larger TTSince T0 is larger:shift is smallerHyperbolic shift 63. NMO stretch (mathematically)Due to differential working on T as function of T0:x2 TNMO =T02 c2 T0 T0x2 = 2 c2 T02This is called NMO-stretch 64. NMO stretch 65. NMO stretchAmplitude spectra 66. Mute: too much NMO-stretchWe do not want too much distortion: setting it zero.This called muting 67. NMO-stretch onfield data 68. Processing flowshot dataCMP sorting velocity model NMO correctionstack velocity model(zero-offset) migrationimage 69. StackingAdd traces from NMO-corrected, CMP gather into ONE traceNumber of traces = stack foldEvents that are not hyperbolic, do not add up nicely anddestructively interfereGoal of stacking : to increase signal-to-noise ratio 70. Primaries and multipleprimary multiple primary</p>