theory_practice_migration_1_8.pptx

SEISMIC IMAGING 2021-22 edition Source: https://www.oedigital.com/news/462656-seabed- acquisition-seeing-healthier-spend nicola.biena...

SEISMIC IMAGING 2021-22 edition Source: https://www.oedigital.com/news/462656-seabed- acquisition-seeing-healthier-spend [email protected] April 2021 1 Seismic Imaging - Course objective  Basic but comprehensive introduction to Seismic Imaging  the technology to transform reflection seismic data into subsurface images  We’ll describe both principles & applications, with a balance between  the description of how Seismic Imaging algorithms work, what are their assumptions and limitations,  and what are the results that can be expected according to the geological complexity of the investigated areas.  The course is (hopefully) self-contained.  Course material: these slides & references therein  Textbook: E. Robein, Seismic Imaging: A Review of the Techniques, their Principles, Merits and Limitations, EAGE  Upon completion you will be able to evaluate Seismic Imaging outcomes, and decide the most effective Seismic Imaging workflow for any new project according to project goals and available budget 2 Seismic Imaging - Exams  Oral examination (as long as we are at home)  Written examination (2 problems + 4 multiple choice questions)  As an alternative project:  code development required (Matlab/Octave)  Team max 3 students  Project presentation (via MS-Teams, 30 minutes, max 21 slides) + project report 3 Seismic Imaging - Schedule 27/04 1D imaging, NMO-stack 28/04 Zero-offset (post-stack) migration 29/04 Pre-stack time migration 30/04 Time imaging: Exercises 04/05 Introduction to depth imaging, Kirchhoff migration 1/2 Theory 05/05 (& Kirchhoff migration 2/2, beam migration, one way migration 06/05 practice) Reverse Time Migration of 07/05 Migratio Depth vs Time Imaging, 2D imaging 11/05 n Anisotropy & Amplitude preservation 12/05 Depth imaging: Exercises 13/05 Velocity analysis in 1D media 14/05 Inverse problems 18/05 Cross-well tomography 19/05 Velocity Migration Velocity Analysis 1/2 20/05 Analysis Migration Velocity Analysis 2/2 21/05 Uncertainty in velocity analysis 25/05 Full Waveform Inversion 4 26/05 Velocity analysis: exercises Theory (& practice) of Migration 1/8 [email protected] May 2020  Motivations better having a look before drilling… 7 Motivations 8  Motivations SEISMIC ACQUISITION SEISMIC SUBSURFACE DATA IMAGE SEISMIC IMAGING Can you see any fault here? 9  The beginning of the story? https://en.wikipedia.org/wiki/Lazzaro_Spallanzani 10  The beginning of reflection seismic https://digital.libraries.ou.edu/sooner/articles/p27-30_1952v24n10.pdf http://blogoklahoma.us/place.aspx?id=745 https://library.seg.org/doi/pdf/10.1190/1.2112392 11 Waves 12  Waves www.compadre.org/Informal/images/features 13 Seismic waves: 1D P (pressure) wave 15 Seismic waves: 1D S (shear) wave 16  The complexity of wave propagation source receiver x time Data recorded on the well z Data recorded on the surface Does recorded data resemble the subsurface? time 17 Figuring out reality from shadows 18 Medical Imaging, the hard approach: the image IS the data http://www.thefullwiki.org/Basic_Physics_of_Nuclear_Medicine/X-Ray_CT_in_Nuclear_Medicine 19 Medical Imaging, the soft approach: step 1, from the object to the data The data do not resemble the image 20 Medical Imaging, the soft approach: step 2, from the data to the image An algorithm is needed to transform the data into the image 21 Remote sensing, the broader frame 22 Ground Penetrating Radar Transmitting Transmitting & & receiving receiving antenna antenna 23 Medical Ecography Transmitting Transmitting & & receiving receiving antenna antenna 24 3D Computerized Axial Tomography X-ray source Detectors 25 Guided waves in complex structures 26 http://www.monitorail.eu/home/index.jsp Principles of seismic imaging Recorded data Image J. Bee Bednar, A brief history of seismic migration, Geophysics, May 2005, v. 70, p. 3MJ-20MJ 27  The simplest imaging experiment: echo sounding distance = time speed/2 speed ≈ 340m/sec 28 What if the speed is unknown? speed? distance? offset 2 measurements for 2 unknowns 29  Learning check 1. What is the essence of seismic imaging? 2. What is the ingredient needed to convert time into distance? 3. What is the attribute of recorded data that contains most of the information about distance and velocity? 30  Learning check 1. The essence of seismic imaging consists in converting echo traveltimes into distances 2. To convert time into distance WE NEED VELOCITY! 3. Traveltimes provide information about BOTH distance and velocity. 31 Imaging of flat reflectors Flat reflector and constant velocity 𝑽𝟏 z [m] 𝑽𝟐 x [m] Question In this example is it V1 > V2? V2 > V1? Why? 33 The law of reflections: Snell’s law qI=qR qI qR V1 𝒔𝒊𝒏 𝜽 𝑰 𝒔𝒊𝒏 𝜽 𝑻 V2 𝑽𝟏 = 𝑽𝟐 qT 𝑽𝟐 𝒔𝒊𝒏 𝜽 𝑻 = 𝐬𝐢𝐧 𝜽 𝑰 ≤ 𝟏 𝑽𝟏 𝑽𝟏 𝐬𝐢𝐧 𝜽 𝑰 = (𝑪) 𝑽𝟐 Flat reflector and constant velocity If you know that the subsurface is flat & velocity is constant then: 1. reflection points lie under the midpoint between source and receiver 2. Ray paths consist of straight lines h = = t NMO (h)  Midpoint 2 2 1  h 1  h  Z 2    2 Z   V  2 V  2 Z V 2 2 h2 2  h 2  Z     t0  2 V  2 V h=offset=sou-rec distance 35  Flat reflector and constant velocity If you know that the subsurface is flat & velocity is constant then: 1. to get an image: sort data into CMPs, apply NMO then stack Common MidPoint (CMP) gather 𝑡0 2 h time √ 𝑉 2 ∆ 𝑇 = 𝑡 + 2 −𝑡 0 0 time time NMO Stack Correction H (sou-rec distance) H (sou-rec distance) 𝑠 ( 𝑡 ) =∑ 𝑑 𝑁𝑀𝑂 (𝑡 , h) h h2 ( √ 2 𝑑 𝑁𝑀𝑂 ( 𝑡 0 , h ) =𝑑 𝑡= 𝑡 + 2 , h 0 𝑉 ) 36 Multiple layers & constant velocities V=2000 m/s V=2250 m/s z [m] V=3500 m/s V=4000 m/s x [m] 38 Multiple layers & constant velocities 2 h2 t NMO (h)  t  2 1 𝑡1 vrms (t1 ) 2 h2 𝑡2 t NMO (h)  t  2 2 vrms (t 2 ) 𝑡3 time [msec] 2h2 t NMO (h)  t  23 vrms (t3 ) Offset [m] 39  Stack NMO corrected CMP Gather STACK S time time [msec] [msec] offset xmidpoint 2 𝑑 𝑁𝑀𝑂 ( 𝑡 ,h )=𝑑 (√ 2 𝑡 + h 𝑉 2𝑅𝑀𝑆(𝑡 ) ,h ) 42  Matlab/Octave code function dNmo=nmo( data, V, offset, t) % dNmo=nmo( data, V, offset, t) % % size(data)=[length(t), length(offset)] for h=1:length(offset) tNmo=sqrt(t.^2+(offset(h)/V)^2); dNmo(:,h)=interp1(t ,data(:,h),tNmo); endfor endfunction >> help interp1 -- YI = interp1 (X, Y, XI) One-dimensional interpolation. Interpolate input data to determine the value of YI at the points XI. https://www.gnu.org/software/octave/ 43  Stack: SNR improvement NMO corrected CMP Gather STACK S time time [msec] [msec] offset Midpoint x position 44 Common offset section after NMO offset 45  STACK Notice the SNR improvement Question: In this example from onshore or offshore data? Why? http://www.ahay.org/RSF/book/geo384s/hw3/paper_html/node3.html 46  Depth stretching after NMO-stack offset offset Depth stretching 𝒛 𝒛 𝒗 𝒕𝟎 𝒕 𝟎=𝟐 𝒗 NMO 𝒕 𝟎= 𝟐 𝒗 Stack 𝒛= 𝟐 correction Additional delay due to offset time time time Depth stretching maps the sample at time t0 NMO removes into the sample at additional delay depth depth due to offset 47 For flat earth NMO-stack is just fine http://homepage.tudelft.nl/t4n4v/9_Theses_students/Schoot.pdf 48  NMO-stack: Imaging for a flat earth NMO-stack is the imaging algorithm under two assumptions: – reflectors are flat – layer (interval) velocities are constant When assumptions are met, use NMO-stack: it’s just fine One CMP gather contributes to the image for that specific MP only NMO-stack is extremely fast NMO-stack is used also when layers are almost (how much?) flat NMO-stack resembles geology: used for quick processing QC 49 If you wanna play a little bit Mobil Avo Viking Graben Line 12 http://s3.amazonaws.com/open.source.geoscience/open_data/ Mobil_Avo_Viking_Graben_Line_12/mobil_avo.html Improving resolution of NMO stack using shaping regularization https://repositories.lib.utexas.edu/bitstream/handle/2152/39535/ REGIMBAL-THESIS-2016.pdf?sequence=1&isAllowed=y 50 Two possible projects 1. Develop the Octave/Matlab code to apply RMS velocity analysis NMO Stack to the Viking dataset 2. Develop the Octave/Matlab code to apply Post-stack time migration Depth conversion to the Viking dataset 51