지구물리와물리탐사 (Geophysics and Geophysical Exploration)

  • 제26권4호
  • /
  • Pages.199-210
  • /
  • 2023
  • /
  • 1229-1064(pISSN)
  • /
  • 2384-051X(eISSN)
한국지구물리·물리탐사학회 (Korean Society of Earth and Exploration Geophysicists)
권호 표지
DOI QR코드

DOI QR Code

심층학습을 이용한 전이대 두께 예측

Thickness Estimation of Transition Layer using Deep Learning

  • 장성형 (한국지질자원연구원 해저지질에너지연구본부) ;
  • 이동훈 (한국지질자원연구원 해저지질에너지연구본부) ;
  • 김병엽 (한국지질자원연구원 해저지질에너지연구본부)
  • 투고 : 2023.09.21
  • 심사 : 2023.11.06
  • 발행 : 2023.11.30
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

CO2 주입 후 저류층은 암석물리 특성이 변하므로 이 연구에서는 저류층을 물성이 선형으로 변하는 전이대 지층모델로 구성한다. 울프 반사계수 함수는 전이대 상하지층의 속도비, 주파수, 전이대 두께 함수로 구성되어 있어 저류층 두께나 해저면 전이대 두께를 추정하는데 활용할 수 있다. 이 연구에서는 심층학습을 이용하여 전이대 두께를 예측 방법을 제안한다. 심층학습을 적용하기 위해 사암 저류층, 셰일 덮개암으로 구성한 인공 전이대 지층모델에 두께에 따른 울프 반사계수 모델링을 수행하고 시간-스펙트럼 영상자료를 확보하였다. 두께별 시간-주파수 스펙트럼 영상과 중합단면도 트레이스에서 구한 시간-주파수 스펙트럼 비교로부터 구한 두께 추정결과는 항상 정확하게 전이대의 두께를 제시하지는 못하였다. 그러나 다양한 환경에서 학습자료를 확보하고 정확도를 높이면 현장자료적용이 가능할 것으로 본다.

The physical properties of rocks in reservoirs change after CO2 injection, we modeled a reservoir with a transition zone within which the physical properties change linearly. The function of the Wolf reflection coefficient consists of the velocity ratio of the upper and lower layers, the frequency, and the thickness of the transition zone. This function can be used to estimate the thickness of a reservoir or seafloor transition zone. In this study, we propose a method for predicting the thickness of the transition zone using deep learning. To apply deep learning, we modeled the thickness-dependent Wolf reflection coefficient on an artificial transition zone formation model consisting of sandstone reservoir and shale cap rock and generated time-frequency spectral images using the continuous wavelet transform. Although thickness estimation performed by comparing spectral images according to different thicknesses and a spectral image from a trace of the seismic stack did not always provide accurate thicknesses, it can be applied to field data by obtaining training data in various environments and thus improving its accuracy.

키워드

과제정보

이 논문은 2023년도 정부(산업통상자원부)의 재원으로 한국에너지기술평가원의 지원을 받아 수행한 연구입니다(20226A10100030).

참고문헌

  1. Addison, P. S., 2018, Introduction to redundancy rules: the continuous wavelet transform comes of age, Phillosophical Transasctions of the Royal Society A, 376(2126), 1-5, doi: http://doi.org/10.1098/rsta.2017.0258
  2. Allen, J., 1977, Short Time Spectral Analysis, Synthesis, and Modification by Discrete Fourier Transform, IEEE Transactions on Acoustics, Speech, and Signal Processing, 25(3), 235-238, doi: https://doi.org/10.1109/TASSP.1977.1162950
  3. Chakraborty, A., and Okaya, D., 1995, Frequency-time decomposition of seismic data using wavelet-based methods, Geophysics, 60(6), 1906-1916, doi: https://doi.org/10.1190/1.1443922
  4. Cho, S. I., and Pyun, S. J., 2023, Comparison of CNN and GAN-based Deep Learning Models for Ground Roll Suppression, Geophysics and Geophysical Exploration, 26(2), 37-51, doi: https://doi.org/10.7582/GGE.2023.26.2.037 (In Korean with English abstract)
  5. Clay, C. S., and Medwin, H., 1977, Acoustical oceanography: Principal & applications, John Wiley & Sons Inc. doi: 10.1016/S0022-460X(78)80104-7
  6. Deng, L., and Yu, D., 2014, Deep Learning: Methods and Applications, Foundations and Trends in Signal Processing, 7(3-4), 197-387, doi: http://doi.org/10.1561/2000000039
  7. Duchi, J., Hazna, E., and Singer, Y., 2011, Adaptive Subgradient Methods for Online Learning and Stochastic Optimization, Journal of Machine Learning Research, 12(7), 2121-2159, doi: https://dx.doi.org/10.5555/1953048.2021068
  8. Dutta, N. C., and Ode, H., 1983, Seismic reflections from a gas-water contact, Geophysics, 48(2), 148-162, doi: https://dx.doi.org/10.1190/1.1441454
  9. Fang, W., Fu, L., Zhang, M., and Li, Z., 2021, Seismic data interpolation based on U-Net with texture loss, Geophysics, 86(1), V41-V54, doi: https://dx.doi.org/10.1190/geo2019-0615.1
  10. Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., WardeFarley, D., Ozair, S., Courville, A., and Bengio, Y., 2020, Generative adversarial networks, Commun. ACM, 63(11), 139-144, doi: https://doi.org/10.1145/3422622
  11. Goupillaud, P., Grossmann, A., and Morlet, A., 1984, Cycleoctave and related transforms in seismic signal analysis, Geoexploration, 23(1), 85-102, doi: https://dx.doi.org/10.1016/0016-7142(84)90025-5
  12. Harsuki, R., and Alkhalifah, T., 2022, StorSeismic: A new paradigm in deep learning for seismic processing, IEEE Transactions on Geoscience and Remote sensing, 60, 1-15, doi: https://doi.org/10.1109/TGRS.2022.3216660
  13. He, K., Zhang, X., Ren, S., and Sun, J., 2016, Deep residual learning for image recongnition, IEEE Conference on Computer Vision and Pattern Recognition, 770-778, doi: https://doi.org/10.1109/CVPR.2016.90
  14. He, Q., and Wang, Y., 2021, Reparameterized full-waveform inversion using deep neural networks, Geophysics, 86(1), V1-V13, doi: https://doi.org/10.1190/geo2019-0382.1
  15. Jia, Y., and Ma, J., 2017, What can machine learning do for seismic data processing? An interpolation application, Geophysics, 82(3), V163-V177, doi: https://dx.doi.org/10.1190/geo2016-0300.1
  16. Jo, J., and Ha, W., 2023, Deep-Learning Seismic Inversion using Laplace-domain wavefields, Geophysics and Geophysical Exploration, 26(2), 84-93, doi: https://doi.org/10.7582/GGE.2023.26.2.084 (In Korean with English abstract)
  17. Kaur, H., Pham, N., and Formel, S., 2020, Seismic data interpolation using deep learnging with generative adversarial networks, Geophysical Prospecting, 69(11), 307-326, doi: https://doi.org/10.1111/1365-2478.13055
  18. Kim, S., and Jun, H., 2022, The Use of Unsupervised Machine Learning for the Attenuation of Seismic Noise, Geophysics and Geophysical Exploration, 25(2), 71-84, doi: https://doi.org/10.7582/GGE.2022.25.2.071
  19. Krizhevsky, A., Sutskever, I., and Hinton, G. E., 2017, ImageNet classification with deep convolutional neural networks, Communications of the ACM, 60(6), 84-90, doi: https://dx.doi.org/10.1145/3065386
  20. LeCun, Y., Boser, B., Denker, J. S., Henderson, D., Howard, R. E., Hubbard, W., and Jackel, L. D., 1989, Backpropagation applied to handwritten zip code recognition, Neural Computation, 1(4), 541-551, doi: https://doi.org/10.1162/neco.1989.1.4.541
  21. LeCun, Y., Bengio, Y., and Hinton, G. E., 2015, Deep learning, Nature, 521, 436-444, doi: https://doi.org/10.1038/nature14539
  22. Lewis, W., and Vigh, D., 2017, Deep learning prior models from seismic images for full-waveform inversion. In: SEG Technical Program Expanded Abstracts 2017, Society of Exploration Geophysicist, 1512-1517, doi: https://doi.org/10.1190/segam2017-17627643.1
  23. Li, H., Li, X., Dong, H., Han, F., and Wang, C., 2022, Full-waveform inversion with adversarial losses via deep learning, Journal of Applied Geophysics, 205, 1-11, doi: https://doi.org/10.1016/j.jappgeo.2022.104763
  24. Li, Y., and Ma, Z., 2021, Deep learning-based noise reduction for seismic data, Journal of Physics: Conference Series, 1861 012011 IWAACE 2021, doi: https://doi.org/10.1088/1742-6596/1861/1/012011
  25. Liner, C. L., and Bodmann, B. G., 2010, The Wolf ramp: Reflection characteristics of a transition layer, Geophysics, 75(5), A31-A35, doi: https://doi.org/10.1190/1.3476312
  26. Liner, C., 2012, Elements of Seismic Dispersion: A Somewhat Practical Guide to Frequency-Dependent Phenomena, Society of Exploration Geophysicists, 109-124, doi: https://doi.org/10.1190/1.9781560802952.ch6
  27. Liu, J., and Marfurt, K. J., 2006, Thin bed thickness prediction using peak instantaneous frequency, SEG/New Orleans annual meeting, 968-972, doi: https://doi.org/10.1190/1.2370418
  28. Ma, Y., and Luo, Y., 2018, Automatic first-arrival picking with Reinforcement Learning, SEG Global Meeting Abstracts, 493-497, doi: https://doi.org/10.1190/IGC2018-121
  29. Marfurt, K. J., and Kirlin, R. L., 2001, Narrow-band spectral analysis and thin-bed tuning, Geophysics, 66(4), 1274-1283, doi: https://doi.org/10.1190/1.1487075
  30. Mosser, L., Dubrule, O., and Blunt, M. J., 2020, Stochastic seismic waveform inversion using generative adversarial networks as a geological prior, Mathematical Geosciences, 52(1), 53-79, doi: https://doi.org/10.1007/s11004-019-09832-6
  31. Naeini, E. Z., and Prindle, K., 2018, Machine learning and learning from machines, The Leading Edge, 37(12), 886-893. doi: https://doi.org/10.1190/tle37120886.1
  32. Nithyashree, V., 2021, https://github.com/Nithyashree-2022/VGG-19-for-Rock-Paper-and-Scissors-classification (July 18, 2023 Accessed)
  33. Oliveira, D. A., Ferreira, R. S., Silva, R., and Brazil, E. V., 2018, Interpolating seismic data with conditional generative adversarial networks. IEEE Geoscience and Remote Sensing Letters, 15(12), 1952-1956, doi: https://doi.org/10.1109/LGRS.2018.2866199
  34. Ovcharenko, O., Kazei, V., Kalita, M., Peter, D., and Alkhalifah, T., 2019, Deep learning for low-frequency extrapolation from multioffset seismic data, Geophysics, 84(6), R989-R1001, doi: https://doi.org/10.1190/geo2018-0884.1
  35. Ovcharenko, O., and Hou, S., 2020, Deep learning for seismic data reconstruction: Opportunities and challenges, in Proc. 1st EAGE Digitalization Conference Exhibition, no. 1, 1-5, doi: https://doi.org/10.3997/2214-4609.202032054
  36. Ozawa, M., 2023, Automated picking of seismic first arrivals using a single-to multidomain self-trained network, Geophysics, 89(1), WA25-WA38, doi: https://doi.org/10.1190/geo2022-0666.1
  37. Park, J., Choi, J., Seol, S. J., Byun, J., and Kim, Y., 2021, A method for adequate selection of training data sets to reconstruct seismic data using a convolutional U-Net, Geophysics, 86(5), V375-V388, doi: https://doi.org/10.1190/geo2019-0708.1
  38. Partyka, G. A., Gridley, J., and Lopez, J., 1999, Interpretational applications of spectral decomposition in reservoir characterization, The Leading Edge, 18(3), 353-360, doi: https://doi.org/10.1190/1.1438295
  39. Plotnitskii, P., Alkhalifah, T., Ovcharenko, O., and Kazei, V., 2019, Seismic model low wavenumber extrapolation by a deep convolutional neural network, ASEG Extended Abstracts, 2nd Australasian Exploration Geoscience, 2019(1-5), doi: https://doi.org/10.1080/22020586.2019.12073206
  40. Ronneberger, O., Fischer, P., and Brox, T., 2015, U-Net, convolutional net-works for biomedical image segmentation, Proc. Int. Conf. Med. Image Comput. Comput.-Assist. Intervent. Cham, Switzerland: Springer, 234-241, doi: https://doi.org/10.1007/978-3-319-24574-4_28
  41. Roth, G., and Tarantola, A., 1994, Neural networks and inversion of seismic data, J. Geophys. Res.: Solid Earth, 99(B4), 6753-6768, doi: https://doi.org/10.1029/93JB01563
  42. Sak, H., Senior, A., and Beaufays, F., 2014, Long short-term memory based recurrent neural network architectures for large vocabulary speech recognition, Proc. Interspeech, 338-342, doi: https://doi.org/10.21437/Interspeech.2014-80
  43. Sezawa, K., and Kanai, K., 1935, Discontinuity in dispersion curves of Rayleigh-waves, Proceedings of the Imperial Academy, 11, 13-14, https://doi.org/10.2183/pjab1912.11.13
  44. Simonyan, K., and Zisserman, A., 2015, Very deep convolutional networks for large-scale image recognition, 3rd International Conference on Learning Representations (ICLR 2015), 1-14, doi: https://doi.org/10.48550/arXiv.1409.1556
  45. Sun, B., and Alkhalifah, T., 2019, ML-descent: An optimization algorithm for full-waveform inversion using machine learning, Geophysics, 85(6), R477-R492, doi: https://doi.org/10.1190/geo2019-0641.1
  46. Sun, H., Sun, Y., Nammour, R., Rivera, C., Williamson, P., and Demanet, L., 2023, Learning with real data without real labels: a strategy for extrapolated full-waveform inversion with field data, Geophysical Journal International, 235(2), 1761-1777, doi: https://doi.org/10.1093/gji/ggad330
  47. Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V., and Rabinovich, A., 2015, Going deeper with convolutions, 2015 IEEE Conference on Computer Vision and Pattern Recognition, 1-9, doi: https://doi.org/10.1109/CVPR.2015.7298594
  48. Tsai, K., Hu, W., Wu, X., Chen, J., and Han, Z., 2019, Automatic First Arrival Picking via Deep Learning with Human Interactive Learning, IEEE Transactions on Geoscience and Remote Sensing, 58(2), 1380-1391, doi: https://ieeexplore.ieee.org/abstract/document/8880673 https://doi.org/10.1109/TGRS.2019.2946118
  49. Wang, J., Xiao, Z., Liu, C., Zhao, D., and Yao, Z., 2019, Deep Learning for Picking Seismic Arrival Times, J. Geophys. Res. Solid Earth, 124(7), 6612-6624, doi: https://doi.org/10.1029/2019JB017536
  50. Widess, M. B., 1982, Quantifying resolving power of seismic systems, Geophysics, 47(8), 1160-1173, doi: https://doi.org/10.1190/1.1441379
  51. Wolf, A., 1937, The reflection of elastic waves from transition layers of variable velocity, Geophysics, 2(4), 357-363, doi: https://doi.org/10.1190/1.1438104
  52. Yang, F., and Ma, J., 2019, Deep-learning inversion: A next-generation seismic velocity model building method, Geophysics, 84(4), R583-R599, doi: https://doi.org/10.1190/geo2018-0249.1
  53. Yeeh, Z., Park, J., Seol, S. J., Yoon, D., and Byun, J., 2023, Trace-based Interpolation Using Machine Learning for Irregularly Missing Seismic Data, Geophysics and Geophysical Exploration, 26(2), 62-72, doi: https://doi.org/10.7582/GGE.2023.26.2.062 (In Korean with English abstract)
  54. Yu, S., Ma, J., and Wang, W., 2019, Deep learning for denoising, Geophysics, 84(6), V333-V350, doi: https://doi.org/10.1190/geo2018-0668.1
  55. Zeiler, D., and Fergus, R., 2014, Visualizing and understanding convolutional networks, Springer International Publishing, 818-833, https://link.springer.com/chapter/10.1007/978-3-319-10590-1_53
  56. Zhang, M., Liu, Y., and Chen, Y., 2019, Unsupervised seismic random noise attenuation based on deep convolutional neural network, IEEE Access, 7, 179810-179822, doi: https://doi.org/10.1109/ACCESS.2019.2959238
  57. Zhong, T., Cheng, M., Dong, X., and Wu, N., 2021, Seismic random noise attenuation by applying multiscale denoising convolutional neural network, IEEE Trans. Geoscience Remote Sens, 60, 1-13, doi: https://doi.org/10.1109/TGRS.2021.3095922