Seismic Modeling, Migration and Velocity Inversion Review and - - PowerPoint PPT Presentation

Seismic Modeling, Migration and Velocity Inversion

Review and Summary Bee Bednar

Panorama Technologies, Inc. 14811 St Marys Lane, Suite 150 Houston TX 77079

May 18, 2014

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 1 / 27

Outline

1

Modeling

2

Migration

3

Review of Algorithmic Differences

4

Migration Velocity Analysis

5

Migration Summary Strategies Conclusions

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 2 / 27

Modeling

Outline

1

Modeling

2

Migration

3

Review of Algorithmic Differences

4

Migration Velocity Analysis

5

Migration Summary Strategies Conclusions

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 3 / 27

Modeling

Why and Where

Why model

For tying geology to measurements For understanding wave propagation For assessment and verification

Three Earth Models

The most practical is probably anisotropic The others simplify the problem

What we Model

Particle motion and velocity Propagation directions Sources

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 4 / 27

Modeling

Modeling Methods

Three Earth Types

Isotropic (Acoustic), Isotropic Elastic, Anisotropic

Up to nine volumes (Orthorhombic) TTI currently most prevalent

Methods

Raytrace based integral methods Direct wave equation methods

Point sources and receivers

Source - Receiver reciprocity

Huygens Principle

Exploding Reflectors

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 5 / 27

Modeling

Modeling Issues

Model Construction

From Logs From migrated data

Acquisition

Arrays vs point sources and receivers Aperture Areal arrays finely sampled

Computation

Acoustic and TTI realistic

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 6 / 27

Migration

Outline

1

Modeling

2

Migration

3

Review of Algorithmic Differences

4

Migration Velocity Analysis

5

Migration Summary Strategies Conclusions

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 7 / 27

Migration

A Hierarchy of Migration Options

Reverse Time Gaussian Beam, Multiple Arrival Kirchhoff, Multiple Arrival Beam Beam Single Arrival Kirchhoff Single Arrival PSPI Extended Split-Step Phase Screen

Computational intensity tends to decrease from top to bottom Everything below the horizontal line is a one-way method Velocity sensitivity tends to increase from top to bottom Basis for both poststack and prestack algorithms

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 8 / 27

Migration

A Hierarchy of Migration Options

Reverse Time Gaussian Beam, Multiple Arrival Kirchhoff, Multiple Arrival Beam Beam Single Arrival Kirchhoff Single Arrival PSPI Extended Split-Step Phase Screen

Computational intensity tends to decrease from top to bottom Everything below the horizontal line is a one-way method Velocity sensitivity tends to increase from top to bottom Basis for both poststack and prestack algorithms

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 8 / 27

Migration

A Hierarchy of Migration Options

Reverse Time Gaussian Beam, Multiple Arrival Kirchhoff, Multiple Arrival Beam Beam Single Arrival Kirchhoff Single Arrival PSPI Extended Split-Step Phase Screen

Computational intensity tends to decrease from top to bottom Everything below the horizontal line is a one-way method Velocity sensitivity tends to increase from top to bottom Basis for both poststack and prestack algorithms

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 8 / 27

Migration

A Hierarchy of Migration Options

Reverse Time Gaussian Beam, Multiple Arrival Kirchhoff, Multiple Arrival Beam Beam Single Arrival Kirchhoff Single Arrival PSPI Extended Split-Step Phase Screen

Computational intensity tends to decrease from top to bottom Everything below the horizontal line is a one-way method Velocity sensitivity tends to increase from top to bottom Basis for both poststack and prestack algorithms

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 8 / 27

Migration

A Hierarchy of Migration Options

Reverse Time Gaussian Beam, Multiple Arrival Kirchhoff, Multiple Arrival Beam Beam Single Arrival Kirchhoff Single Arrival PSPI Extended Split-Step Phase Screen

Computational intensity tends to decrease from top to bottom Everything below the horizontal line is a one-way method Velocity sensitivity tends to increase from top to bottom Basis for both poststack and prestack algorithms

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 8 / 27

Migration

The Major Difference Between Time and Depth

The major difference between time and depth.

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 9 / 27

Review of Algorithmic Differences

Outline

1

Modeling

2

Migration

3

Review of Algorithmic Differences

4

Migration Velocity Analysis

5

Migration Summary Strategies Conclusions

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 10 / 27

Review of Algorithmic Differences

Prestack migration differences

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 11 / 27

Review of Algorithmic Differences

A Complex Salt Model

The KM velocity model and three surface shot locations.

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 12 / 27

Review of Algorithmic Differences

Impulse Responses from a Complex Salt Model

A comparison of impulse responses from the KM model of the previous slide.

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 13 / 27

Review of Algorithmic Differences

Impulse Responses from the SEG AA′ model

A simple comparison of Kirchhoff, one-way and two-way impulse responses

• ver the SEG/EAGE AA’ data set.

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 14 / 27

Migration Velocity Analysis

Outline

1

Modeling

2

Migration

3

Review of Algorithmic Differences

4

Migration Velocity Analysis

5

Migration Summary Strategies Conclusions

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 15 / 27

Migration Velocity Analysis

Migration Velocity Analysis

Migration sends sources and receivers offsets to zero

Any gather we build should take this into account Offset-shift and time-shift gathers do The others don’t

Gathers

Common offset, Common Angle, Common Image Point, Offset-shift, Time-shift

Tomography

Not Dix

Short-spread vs Long-spread velocity analysis

Anisotropy Probably requires log information

Horizon vs no horizon based Full waveform inversion (FWI)

Somewhat in its infancy

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 16 / 27

Migration Summary

Outline

1

Modeling

2

Migration

3

Review of Algorithmic Differences

4

Migration Velocity Analysis

5

Migration Summary Strategies Conclusions

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 17 / 27

Migration Summary

Styles

Two basic algorithm styles

Raytrace (Integral) based

Beam, Beam Stack, Kirchhoff, Gaussian Beam So called high frequency approximation

Partial Differential Equation Based

Finite Difference, Finite Element, Dual Domain (FKX)

Multiple Combinations

Raytrace with PDE in shot-profile Delayed shot beam using PDE’s

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 18 / 27

Migration Summary

Anisotropy

Kirchhoff, Beam, and Gaussian Beam

Function only of raytracer

One-way

Limited by approximation

Two-way

All versions of anisotropy

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 19 / 27

Migration Summary

Velocity Sensitivity

Kirchhoff

Single Arrival very sensitive Multiple Arrival should not be so sensitive

Beam

Smearing tends to reduce sensitivity somewhat

One-way

Approximations make it sensitive to large changes

Not as severe as Kirchhoff Unusual subsurface positioning

Gaussian Beam

Almost none

Two-way

None

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 20 / 27

Migration Summary

True Amplitude Accuracy

Kirchhoff

Single arrival limits accuracy

Beam

Single arrival limits accuracy

One-way

Approximations

Not as severe as Kirchhoff Loss mostly due to one-way approach That darn square root

Gaussian Beam

Almost as good as you can get

Two-way

As good as you can get

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 21 / 27

Migration Summary Strategies

Which one should I use?

Dependent on exploration stage

Earth model estimation

Velocity analysis Anisotropic analysis

Geological declination

Stratigraphic detail Structural style

Interpretation

Initial trap delineation True depth Reservoir size Well placement

Reservoir Characterization

Each algorithm has its use

Be prepared to use them all if necessary

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 22 / 27

Migration Summary Strategies

Algorithmic Issues

Beam

Potential loss of stratigraphy Cannot be beaten

For initial Earth model estimation Hypothesis testing with various parameters For mapping large structural traps

Suspect amplitudes

Kirchhoff

Velocity estimation workhorse for many years Better than beam at revealing stratigraphy Excellent for target line imaging Suspect amplitudes Initial reservoir characterization

Particularly in time

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 23 / 27

Migration Summary Strategies

Algorithmic Issues

One-way — PSPI, Extended Split-step, Phase Screen

Can be faster than full two-way Not good for velocity analysis

Exception is common azimuth implementations

Very good at stratigraphic imaging Very good amplitudes Loss of dip limits structural accuracy

Gaussian Beam

Good for velocity analysis

But slow

Excellent stratigraphic imaging Excellent amplitudes Excellent structural accuracy Reservoir characterization

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 24 / 27

Migration Summary Strategies

Algorithmic Issues

Two-way — RTM, Pseudo Spectral

By far the most accurate mathematically Excellent for Earth model refinement

Only method accurate enough for full waveform inversion Turnaround speed is issue GPU versions and new computer systems will change this Becoming very cheap and available to even small contractors

Excellent stratigraphic imaging Excellent amplitudes Excellent structural accuracy

Faults imaged from both sides Absolutely no dip limits

Final image

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 25 / 27

Migration Summary Conclusions

Conclusions

Algorithm choice

Function of basin, play, economics, objective Oil found with all of them In your exploration career you will see them all — and hear a lot of BS from your favorite contractor

RTM most accurate Beam least accurate Rest fall in between

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 26 / 27

Migration Summary Conclusions

Questions?

Bee Bednar (Panorama Technologies) Seismic Modeling, Migration and Velocity Inversion May 18, 2014 27 / 27