![Page 1: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/1.jpg)
Deep learning Q inversion from reflection seismic data with strongattenuation using an encoder-decoder convolutional neural
network: an example from South China Sea
![Page 2: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/2.jpg)
Outline
Introduction
Method and theory
Field data application
Conclusion
![Page 3: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/3.jpg)
Outline
Introduction
Method and theory
Field data application
Conclusion
![Page 4: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/4.jpg)
➢ Amplitude decay➢ Poor illumination➢ Unreliable AVO
Problems of attenuation
(Zhou ,2011)
A seismic image with strong Q effect
![Page 5: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/5.jpg)
Quality factor that quantifies seismic attenuation § Small Q means large attenuation§ Strong attenuation: Q ~ 10-50§ Mild attenuation: Q ~ 70-300§ Nearly no attenuation: Q >1000
The Q effect
Attenuation classification
![Page 6: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/6.jpg)
Effect of attenuation on amplitudes
![Page 7: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/7.jpg)
Effect of attenuation on phase
![Page 8: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/8.jpg)
Without Q compensation With Q compensation
Effect of attenuation on imaging
Migration without Q compensation – Damps amplitudes – Lowers resolutions – Disperses phases
Courtesy of CNOOC
![Page 9: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/9.jpg)
Effect of attenuation on reservoir characterization
![Page 10: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/10.jpg)
Effect of attenuation on reservoir characterization
![Page 11: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/11.jpg)
1. Filtering method
Nonstationary Deconvolution
(Dasgupta and Clark,1998;Margrave et al.,2003,2011;van der Baan, 2012)
Poststack inverse Q filtering
(Bickel and Natarajan,1985;Hargreaves and Calvert,1991;Wang,2002)
Prestack inverse Q filtering (Wang,2006; Cavalca et al.,2011)
Q inversion and compensation
(Causse et al.,1999;Reine et al., 2012; Chen et al.,2013;Wang and Chen,2014; Li and Liu ,2015; Chai et al.,2016 )
Limitation : Simple Q model used, can not handle heterogeneous Q model well.
11/120
Approach to compensate Q effect
![Page 12: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/12.jpg)
12
2. Q compensation through Pre-stack migration
Ray-based(Ribodetti et al.,1998),
One way wave equation
(Dai and West,1994; Mittet et al.,1995; Yu et al.,2002; Mittet.,2007; Zhang et al,.2013; Shen et al,.2014)
Two way wave equation
( Causse and Usin,2000; Deng and McMechan,2007,2008; Zhang et al.,2010; Yan and Liu,2013; Zhu et al,.2014)
Challenge : Needs a fine heterogeneous Q model in depth domain
Approach to compensate Q effect
![Page 13: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/13.jpg)
13/120
Back-project the amplitude variations along raypaths
−=
raypath vQ
l
2expationAmp_attenu
zx
y source
receiver
xyzQ
ijkI
xyzl
( )
−=
0
0lnexp
2expAA
vQ
li
vQ
ll
Q-PSDM: accumulated Q effect along raypath
(Zhou,2011)
PSDM
Q-PSDM
![Page 14: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/14.jpg)
Outline
Introduction
Method and Theory
Field data application
Conclusion
![Page 15: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/15.jpg)
Traditional Q estimation approach-- Spectral ratio method
![Page 16: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/16.jpg)
Traditional Q estimation approach-- Centroid frequency shift method
(Quan and Harris., 1997; Li et.al.,2015 )
![Page 17: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/17.jpg)
Recent Q estimation approach-- Image domain WE migration Q analysis
(Shen et al., 2018)
![Page 18: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/18.jpg)
Recent Q estimation approach-- image domain WE migration Q analysis
(Shen et al., 2018)
Ground truth Inversion result
➢ Large scale industry problem➢ Sensitive to noise
![Page 19: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/19.jpg)
Good at solving the problems of classification, clustering, regression and dimensionality reduction of high-dimensional data
Yuan et al,2018 , Araya-Polo et al, 2018,Lewis and Vigh,2017, Wu et al, 2016
ML and DL in Geophysics
First break picking VA and FWI Classification of phasesFault, horizon and salt dome identification
…
![Page 20: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/20.jpg)
Work Flow
Migration to output seismic image
Dividing datasets to training, testing, Validation set.
Constructing the structure of neural network, choosing number of network layers, input neurons, the activation function, loss function and optimization method.
Labelling the data by hand picking
Training the network using the training set with labels, adjust the network structure based on the performance of the cost function.
Verify network parameters, complete network training, using test data to check generalization effect.
Input the whole dataset, finish automatic Q inversion and imaging with the Q field.
Data and Network preparation Network training and data validation
Network training is most time-consuming
![Page 21: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/21.jpg)
21/120
CNN architecture for Q inversion
![Page 22: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/22.jpg)
Training evaluation
.
The depth and width of hidden layers decide the learning ability of a NN
Too simple NN causes underfitting.
Over complicated NN causes overfitting.
Through testing, we choose the number of layers at 4
Compare the training error and the validation error with training time
![Page 23: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/23.jpg)
Outline
Introduction
Method and Theory
Field data example
Conclusion
![Page 24: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/24.jpg)
The 3D seismic data
200 km
200 k
m
Survey
location
![Page 25: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/25.jpg)
Inline location
Atten
uatio
n In
tensity (%)
100
0
The Q inversion result
![Page 26: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/26.jpg)
The Q-PSDM method to verify
2
0ln( )( , , ) ( ) exp exp ( ) exp ( )
2 2
gs s gss g
s gg s g
AI x y d F j j d
Q QA Q Q
= − + − + +
Weights Phase correction Amplitude compensation
Imaging result
3D effect Traveltime
Compensating Q effect
![Page 27: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/27.jpg)
The migration gather w/o Q compensation
![Page 28: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/28.jpg)
The imaging result w/o Q compensation
![Page 29: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/29.jpg)
Result comparison : Spectrum
—— PSDM—— Q-PSDM
About 15 Hz main frequency lifting
About 15 Hz
![Page 30: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/30.jpg)
Outline
Introduction
Method and Theory
Field data application
Conclusion
![Page 31: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/31.jpg)
Conclusions
• The DL method can help to capture the Q anomaly automatically after network
training.
• The proposed Q model building workflow is less affective by the noise and suitable for large-scale industrial problems.
• Automatic labeling is the topic that needs further study.
![Page 32: Deep learning Q inversion from reflection seismic data](https://reader031.vdocument.in/reader031/viewer/2022012511/61898c8bbb65b0625a23fa82/html5/thumbnails/32.jpg)