Progress Report on Separation of Multisource Data

Progress Report on Separation of Progress Report on Separation of Multisource Data Multisource Data

Xin WangXin Wang

February 5, 2009February 5, 2009

OutlineOutline

• MotivationMotivation::

• Numerical Results Numerical Results

• MethodologyMethodology

• ConclusionsConclusions

Separating multisource data in order to get single shot gathersSeparating multisource data in order to get single shot gathers

Applying a combination of several local adaptive filtersApplying a combination of several local adaptive filters

Synthetic tests of 2, 5 and 10 sources on 2D SEG/EAGE modelSynthetic tests of 2, 5 and 10 sources on 2D SEG/EAGE model

01

MotivationMotivationGoal:Goal:Shoot more times in a certain time to Shoot more times in a certain time to

improve the efficiency of seismic improve the efficiency of seismic

survey, especially in wide azimuth survey, especially in wide azimuth

angles surveys. angles surveys.

02

Constraint:Constraint:Avoid time overlaps in seismic records Avoid time overlaps in seismic records

from different sources.from different sources.

CSG without time overlap0

81 129Trace #ChallengeChallenge::

Can we separate multisource Can we separate multisource

data with time overlaps to get data with time overlaps to get

single shot gathers?single shot gathers?

Multisource CSG with time overlap

0Single CSG 1 Single CSG 2

81 129Trace # 1 129Trace # 1 129Trace #

0

8

0

8

OutlineOutline





Separation multisource data to get single shot gathersSeparation multisource data to get single shot gathers



03

Random time shooting with known time delay (Chevron, 2008).

MethodologyMethodology

0481 Trace #

Multisource CSG 1

129

0

81 Trace #

Multisource CSG 2

129

0

Resort data from CSGs to CRGs.

0

1 Trace #

Multisource CSG

129

0

81 Trace #

Multisource CRG

1298

Apply a combination of slant stacking, median filtering, threshold mute filtering and adaptive subtraction matching filtering.

Trace #1 9

0

1Trace #1 9

0

1

Original dataApplying slant stacking

and median filtering

Trace #1 9

0

1Trace #1 9

0

1

Original dataApplying threshold

mute filterOriginal data

Slant stacking and median filtering

Local adaptive subtraction matching

filtering 0

1

Trace #1 10

1 Trace #

Filtered CRG

45

Filtered CSG

451 Trace #

0

8

0

8Resort filtered data from CRGs to CSGs.

WorkflowWorkflow

Multisource data

Resort CSG to CRG with known time shift

Apply slant stacking and median filter

Apply threshold mute filter

Apply local adaptive subtraction matching

filter

Apply slant stacking and median filter

Near offset: Far offset:

Apply local adaptive subtraction matching

filter

Resort CRG to CSG

05

Resort CSG to CRG with known time shift

OutlineOutline








06

Numerical ResultsNumerical Results

Model Size: 5.9 (km) * 1.4 (km)

Receiver number: 129

Range of random shooting time : 0~4s

SEG/EAGE model

5.9X (km)01.4

0

2 sources number of shots: 120, shot interval: 27.4 m, multisource interval: 2600 m

5 sources number of shots : 45, shot interval: 27.4 m, multisource interval: 1152 m

10 sources number of shots : 65, shot interval: 9.14 m, multisource interval: 585 m

07

Multisource CSG Multisource CRG

1 Trace # 129

0

81 Trace # 120

0

8

1 Trace # 129

0

81 Trace # 120

0

8

Filtered CRGFiltered CSG

08

Numerical ResultsNumerical Results2 simultaneous sources2 simultaneous sources


Multisource CRG Filtered CRG

1 Trace # 120

0

81 Trace # 120

0

8

1 Trace # 120

0

81 Trace # 120

0

8

Filtered CRGMultisource CRG

Near offset traces

Far offset traces

09


Multisource CRG Filtered CRG

1 Trace # 45

0

81 Trace # 45

0

8

1 Trace # 45

0

81 Trace # 45

0

8

Filtered CRGMultisource CRG

Near offset traces

Far offset traces

10


1 Trace # 129

0

8

Multisource CSG

1 Trace # 129

0

8

Filtered CSG

1 Trace # 129

0

8

Actual CSG

11

Numerical ResultsNumerical Results5 simultaneous sources5 simultaneous sourcesKirchhoff Migration Image Kirchhoff Migration Image

KM image of simultaneous data

5,9X (km)01.4

0

KM image of filtered data

5.9X (km)0

0

KM image of actual data

5.9X (km)0

0

12

1.4

Numerical ResultsNumerical Results5 simultaneous sources5 simultaneous sourcesKirchhoff migration image only with far offset dataKirchhoff migration image only with far offset data

KM of far offset multisource data

5.9X (km)01.4

0

KM of far offsetfiltered data

5.9X (km)01.4

0

KM of far offset actual data

5.9X (km)01.4

0

13

5.9X (km)01.4

0

40 Receivers , 1.82 km45 shots, 1.23 km 45 shots, 1.23 km40 Receivers , 1.82 km


Near offset traces:Near offset traces:

1 Trace # 65

0

8

Multisource CRG

1 Trace # 658

Filtered CRG

1 Trace # 65

0

8

Actual CRG 0

14


Far offset traces:Far offset traces:

1 Trace # 65

0

8

Multisource CRG

1 Trace # 658

Applying threshold mute filter

1 Trace # 65

0

8

Applying 10 iterations of median filter and adaptive

subtraction filter0

Applying slant stacking and median filter

Applying adaptive subtraction filter

15

OutlineOutline








16

ConclusionsConclusions

17

• Separation of multisource data can greatly improve the efficiency of seismic surveys, especially with wide azimuth angle.

• By applying several filters, we can separate 2 and 5 multisource data with acceptable results. • With increasing number of multisource, the effectiveness of separation degrades.

• The effectiveness of separation method degrades with increasing offsets. More effort is needed for the far offset data.

• Future Work: improving filters, more multisource, and realistic field data.

Acknowledgement

We would like to thank the UTAM 2008 sponsors for their support.

Thank You

Documents

Progress Report on Separation of Multisource Data