Seismic Data Processing, Ahmed Osama
Download as PPT, PDF14 likes2,854 views
This document discusses seismic data processing workflows. It begins with an introduction and agenda. The general workflow includes reformatting, trace editing, geometry handling, amplitude recovery, noise attenuation through techniques like frequency and FK filtering, deconvolution, multiple removal, migration, velocity analysis, NMO correction, muting, stacking, and post-stack filtering and amplitude scaling to produce a final image for geological interpretation. The document emphasizes that the proper workflow selection depends on processing environment, targets, costs, and client preferences. It concludes with time for questions.
Seismic Data Processing, Ahmed Osama
- 1. Seismic Data Processing Schlumberger – WesternGeco Ahmed Osama
- 2. • Personal Introduction. • Objective of Seismic Processing. • General Workflow of Seismic Processing. • Selection of Proper Processing Workflow. • Q & A 2 Agenda
- 3. • Personal Introduction. • Objective of Seismic Processing. • General Workflow of Seismic Processing. • Selection of Proper Processing Workflow. • Q & A 3 Agenda
- 4. 4 Ahmed Osama Ahmed • 3rd Level Student at Faculty of Science-ASU, Geology Dept. • President of AAPG-ASUSC 2016/2017. • Online Freelance (Arabic <-> English) Translator.
- 5. • Personal Introduction. • Objective of Seismic Processing. • General Workflow of Seismic Processing. • Selection of Proper Processing Workflow. • Q & A 5 Agenda
- 6. 6 Seismic JobSeismic Job Acquisition Processing Interpretation Field Raw DataField Raw Data GeologyGeology
- 7. • Personal Introduction. • Objective of Seismic Processing. • General Workflow of Seismic Processing. • Selection of Proper Processing Workflow. • Q & A 7 Agenda
- 8. 8 Simple Processing WorkflowSimple Processing Workflow 1.1. Reformat.Reformat. 2.2. Trace Edit.Trace Edit. 3.3. Geometry.Geometry. 4.4. Amplitude Recovery.Amplitude Recovery. 5.5. Noise Attenuation.Noise Attenuation. 6.6. Deconvolution.Deconvolution. 7.7. CMP Gather.CMP Gather. 8.8. Demultible.Demultible. 9.9. Migration.Migration. 10.10.NMO Correction.NMO Correction. 11.11.Mute.Mute. 12.12.Stack.Stack. 13.13.Filtering.Filtering. 14.14.Amplitude Scaling.Amplitude Scaling. 15.15.Final Image.Final Image. Shot DomainShot Domain CMP DomainCMP Domain Offset DomainOffset Domain Stack DomainStack Domain Receiver DomainReceiver Domain
- 12. 12
- 13. 13
- 14. 14
- 18. 18 AfterBefore
- 20. 20 Types of NoiseTypes of Noise Random Noise:Random Noise: •WindWind •TrafficTraffic •Falling DebrisFalling Debris •Instrument NoiseInstrument Noise Coherent Noise:Coherent Noise: •Ground RollGround Roll •MultiplesMultiples •Direct ArrivalsDirect Arrivals •Rig NoiseRig Noise
- 21. 21
- 22. 22 Aim of Noise Attenuation ProcessAim of Noise Attenuation Process S/N RatioS/N Ratio
- 23. 23 Noise Attenuation Techniques :Noise Attenuation Techniques : • Frequency FiltersFrequency Filters • FK Filter.FK Filter. • FXCNS Filter.FXCNS Filter. • AAAAAA
- 24. 24 Frequency FiltersFrequency Filters High Cut FilterHigh Cut Filter Low Cut FilterLow Cut Filter Band Bass FilterBand Bass Filter Passband F1 F2 Passband F3 Notch FilterNotch Filter TT FF TT
- 25. 25 FK FilterFK Filter FourierFourier TransformTransform FourierFourier TransformTransform 2D Filter2D Filter T-XT-X F-KF-K
- 26. 26
- 28. 28 FXCNS FilterFXCNS Filter FourierFourier TransformTransform FourierFourier TransformTransform 1D Filter1D Filter T-XT-X F-XF-X
- 29. 29 InputInput
- 30. 30 OutputOutput
- 31. 31 Difference
- 32. 32 AAAAAA • Time Windows. • Average Amplitude. • Removal or Replacement.
- 33. 33
- 34. 34
- 36. 36 • Algorithm process to reverse the effect of Convolution. • Aims to the following: 1. Shaping the wavelet. 2. Improving the resolution of the data. 3. Attenuate some multiples. DeconvolutionDeconvolution
- 37. 37 BeforeBefore
- 38. 38 AfterAfter
- 40. 40 DemultipleDemultiple Events which have been reflected more than once before being recorded by receivers. Source Receiver
- 41. 41 T-X DomainT-X Domain Tau-P DomainTau-P Domain Radon TransformRadon Transform
- 42. 42 BeforeBefore
- 43. 43 AfterAfter
- 45. 45 MigrationMigration The aim of this major process is to remove the effect of the beds’ curvature and to relocate the energy to its true position.
- 46. 46 BeforeBefore
- 47. 47 AfterAfter
- 49. 49 Velocity AnalysisVelocity Analysis We make the Velocity Analysis process to obtain a velocity model which will be used later in NMO Correction.
- 50. 50
- 51. 51
- 52. 52 NMONMO CorrectionCorrection Normal Move Out Correction is used to remove the effect of time variation with offset.
- 53. 53
- 54. 54 MuteMute After the NMO Correction the data will be stretched so some noisy parts of it will make no sense and here comes the role of the Muting process to remove these parts.
- 55. 55
- 56. 56 StackStack
- 57. 57 StackStack In this process we sum all CMP gathers altogether to increase S/N ratio and improve the data quality.
- 58. 58
- 60. 60 FilteringFiltering Amplitude ScalingAmplitude Scaling Balancing the amplitudesBalancing the amplitudes of the stacked sectionof the stacked section horizontally or vertically tohorizontally or vertically to facilitate the interpretationfacilitate the interpretation process laterprocess later High Cut FilterHigh Cut Filter Low Cut FilterLow Cut Filter Band Bass FilterBand Bass Filter Passband F1F2 Passband F3 Notch FilterNotch Filter
- 62. 62 Selection of Proper Processing Workflow • Environment.Environment. • Target.Target. • Cost.Cost. • Client Preference.Client Preference.
- 63. 63 Q & AQ & A
Editor's Notes
- #3: 1- Introducing Myself 2- Seismic Stages & Why Processing? 3- Simple Workflow with Domains 4- How we select the best proper sequence of Processing ? 5- Any Questions !!
- #5: Introduction
- #7: Seismic is performed on 3 main stages .. Acquisition is Recording the data in the field. Processing is enhancing the clearance of the field data to make a better image of the subsurface to facilitate the interpreter job. Interpretation is conversion of seismic into geological model.
- #9: We start with reformatting the field data .. And end with final stacking for CMPs and see geology.
- #11: We begin with reformatting the field data from seg-d or seg-y into the internal format .. As for WG we use Omega software which deals with DIO format .. So we change the field format into DIO to start dealing with it by Omega software.
- #12: In the field during recording the data .. We might have some instrumental failures which leads into bad recording forming dead traces or maybe bad shots We edit traces by reversing the polarity of them (1st method) Or by the removal of the whole trace using AAA (2nd method) Or by deletion of the whole shot in case we couldn’t remove the dead trace by aaa (3rd method) Taking notes of the numbering of dead traces within shots to check them after we apply the noise attenuation methods, if they are removed so be it, but if not, we’ll remove the whole shot.
- #16: We receive 2 parts of data from the field … SPS/UKOOA files & Seismic traces .. SPS/UKOOA files contain coordinates and elevations of shot points, lines and receiver points, lines…. We merge these files with the seismic traces so we update each trace with his location information in the field ..
- #17: We’re sorting the seismic data (traces) into different domains .. Coz in each domain we apply some process which can’t be applied in any other domain..
- #18: When the seismic source transmits waves (energy) … with increasing the depth there’s a loss in the submitted energy due to several factors such as spherical divergence, absorption and decay .. So we need to balance the energy within the whole shot gather .. Which leads to the amplitude recovery process but with keeping the original amplitudes untouched we just make a compensation for the amplitudes in the deeper part of the shot..
- #21: Random Noise: energy that doesn’t exhibit correlation from trace to another trace (has no specific shape), can’t be predicted Coherent Noise: source generated, can be predicted though the shot traces.
- #23: To increase S/N ratio.. Attenuation of noises with preserving the data as possible as we can..
- #25: According to the frequency content, we design a filter to remove all frequencies of the noise and keep only the frequency of the required data
- #26: 1- We transform the data from time-offset domain to frequency-wave no. domain by Fourier Transform: The Fourier Transform is simply a mathematical process that allows us to take a function of time (a seismic trace) and express it as a function of frequency (a spectrum). 2- According to the dip of the data and the dip of the noise, we design a filter to remove noise with preserving the data untouched ((( We determine the dip of the data and we design the filter to remove anything else )))
- #29: This filter removes the linear noise with certain range of dips and velocities
- #33: Anomalous Amplitude Attenuation .. We make time windows out of the traces .. We determine the amplitude range for each window and anything higher the average amplitude is either removed or replaced by the average amplitude..
- #37: When the source send the energy into the subsurface, the submitted wave has the form of Wavelet and when it hits the geology (RC) it reflects and be recorded by receivers in the form of traces which result from interaction between wavelet and RC. So we’re here to reverse this process .. We aim to remove the wavelet from the trace to get a better resolution of the RC (Geology)
- #42: We transform the data from TX domain to Tau-P domain using the radon transform .. Then we separate the real data from the multiples and remove the multiples with preserving the data.