Local MVA for VSP Data

Progress ReportSanzong Zhang, Xin

Wang and Xian XiaoJan. 7, 2010

Outline Local VSP

Migration Local VSP MVA Challenges Numerical

Examples Future Work Acknowledgements

Local VSP Migration

Conventional VSP Migration

s

x

g

BackwardForward

s

Defocusing in VSP Migration

s

x

g

Errors in the overburden

and salt body velocity model

Defocusing

Limitations in VSP Migration

Overburden or salt velocity model is required, but hard to build.

Errors due to imperfect velocity models.

Local VSP Migration

s

x

R(g’|s)

g’

T(g|s) g

(1) Crosscorrelation imaging condition

Imaging Condition

(2) Deconvolution imaging condition

(a) VSP data: P(g|s)=T(g|s)+R(g|s)

T(g|s)

s

gR(g|s)

x

s

(b) Backward reflection

R(g|s)g

x

R(x|s)= G(x|g)*R(g|s)g

(c) Backward transmission

T(g|s)

s

g

x

T(x|s)= G(x|g)*T(g|s)

g

(d) Crosscorrelation

m(x)= R(x|s)*T(x|s)s

R(g|s)g

x

Steps of Local VSP Migration

Benefits Local VSP migration is oriented to our

target . Only a local velocity model near the well is needed.

Complex overburden and salt body are avoided.

Source statics are automatically accounted for.

Immune to salt-related interbed cross-talk.

Fast and easy to perform.

Local MVA for VSP Data

Local VSP MVA (LVM) LVM combines VSP migration and velocity

model updating

LVM is based on the local VSP migration obtained by using reflected and transmitted waves.

Depth residuals from common image gathers (CIGs) are transferred to traveltime residuals.

Traveltime tomography is used to update the local velocity model near the well.

Challenges in Local VSP MVA

Comparison of Three Migration Methods in Local

VSP MVA.

CIGs using the background velocity model 2000m/s.

Depth Residuals

Numerical Examples

Sigsbee P-wave Velocity Model m/s

0

Dep

th (

km

)

9.2

4500

1500-12.5 12.5Offset (km)

279 shots, interval of 45.7m

150 receivers,Interval of

30m

Pressure component of a common receiver gather for the receiver at the depth of 4.6 km

Local VSP Migration Results4.6

9.2

Dep

th (

km)

-3 3Offset (km)

True modelMigration image

f = fault

f

d

d

d = diffractor

Offset (m)

Marine 2D Offset VSP dataD

ep

th

(m)

48780 1829

0Source @150 m offset

2800 m

3200 m

Salt

82 receivers with 15.3-m interval

@600 m offset @1500 m offset

Velocity Profile

4500

P Wave

Dep

th

(m)

0

0 5000

2800 m

3200 m

Salt

Incorrect velocity model

Velocity (m/s)

Z-Component VSP DataD

epth

(m

)

Traveltime (s)

2652

3887

1.2 3.0

Salt

Direct P

Reflected P

150 m offset

3.3

3.9

0 100

Dep

th (

km)

Offset (m) 0 100Offset (m)

Without deconvolution

With deconvolution

600 m offset

3.3

4.4

0 600

Dep

th (

km)

Offset (m) 0 600Offset (m)

Without deconvolution

With deconvolution

1500 m offset

3.3

4.4

0 600

Dep

th (

km)

Offset (m) 0 600Offset (m)

Without deconvolution

With deconvolution

Future work Method to recognize minor

depthresiduals

Interactive velocity updating based

on time residual tomography

Conversion between depth residual

and time residual

Acknowledgements

Thanks to the 2009 sponsors of UTAM Consortium for their support. Thanks to Jerry for providing me excellent working conditions at KAUST. Thanks to Xian Xiao for providing me his data and code.