基于深度神经网络的断层高分辨率识别方法

丰超; 潘建国; 李闯; 姚清洲; 刘军

doi:10.3799/dqkx.2022.276

基于深度神经网络的断层高分辨率识别方法

doi: 10.3799/dqkx.2022.276

中国石油勘探开发研究院西北分院, 甘肃兰州 730020

基金项目:

中国石油天然气股份有限公司十四五上游领域前瞻性基础性项目《海相碳酸盐岩成藏理论与勘探技术研究》 2021DJ05

详细信息

作者简介:
丰超（1990-），男，硕士，主要从事储层预测、断层检测方向研究. ORCID：0000-0003-3517-5647. E-mail：feng_chao@petrochina.com.cn

中图分类号: P631
计量
- 文章访问数: 76
- HTML全文浏览量: 52
- PDF下载量: 34
- 被引次数: 0
出版历程
- 收稿日期: 2022-10-22
- 刊出日期: 2023-08-25

Fault High-Resolution Recognition Method Based on Deep Neural Network

Research Institute of Petroleum Exploration & Development-Northwest, PetroChina, Lanzhou 730020, China

摘要

摘要: 传统地震属性断层识别技术多基于数据不连续性识别断裂，干扰因素多，越来越难以满足深层精细勘探的需求. 为了提高断层识别精度，提出一种断层高分辨率智能识别方法，在深度学习方法从地震数据预测断层属性的基础之上，建立高分辨率与低分辨率断层标签库，训练深度神经网络，获得高分辨率检测模型.通过模型与实际数据证实，方法解决了深度学习中卷积神经网络存在上采样造成高频损失，使断层分辨率有所下降的问题，提高了分辨能力，模拟数据均方根误差下降40.02%.方法不仅相对传统算法更加准确地检测了断层特征，而且比一般的深度学习断层识别分辨率高.
- 卷积神经网络 /
- Unet网络 /
- 断层识别 /
- 深度学习
Abstract: Fine fault identification is of great significance to improve the efficiency of exploration and development. Traditional seismic attribute fault recognition technologies identify fractures based on data discontinuity, and there are many interfering factors, making it more and more difficult to meet the needs of fine exploration in deep area. In order to improve the fault identification accuracy, this paper proposes a high-resolution intelligent identification method of fault. Based on the deep learning method to predict fault attributes from seismic data, a high-resolution and low-resolution fault label library is established, and a deep neural network is trained. It is confirmed by the model and actual data that the method solves the problem of high-frequency loss caused by up sampling in the convolutional neural network in deep learning, which reduces the resolution of faults, and improves the resolution ability. The root mean square error of the simulated data decreased by 40.02%.Compared with traditional algorithms, the method not only detects fault features more accurately, but also has a higher resolution than common deep learning fault recognition.
- convolution neural network /
- Unet network /
- fault detection /
- deep learning

HTML全文

图 1 断层与地震数据标签

a. 地层速度模型；b. 合成地震数据；c. 断层标签

Fig. 1. Fault and seismic data labels

下载: 全尺寸图片幻灯片

图 2 基于Unet的深度卷积神经网络

Fig. 2. Unet-based deep convolutional neural network

下载: 全尺寸图片幻灯片

图 3 断层高分辨率识别流程

Fig. 3. Process of high-resolution identification on fault

下载: 全尺寸图片幻灯片

图 4 高低分辨率断层模型标签

a. 合成地震数据；b. 低分辨率断层数据；c. 高分辨率断层数据

Fig. 4. Labels of high-resolution and low-resolution fault models

下载: 全尺寸图片幻灯片

图 5 断层高分辨率识别深度卷积神经网络

Fig. 5. Deep convolutional neural network for high-resolution fault detection

下载: 全尺寸图片幻灯片

图 6 地震数据与断层非线性关系训练过程中训练集与验证集损失监测图

Fig. 6. Loss monitoring diagram of training set and validation set during the training process of the nonlinear relationship between seismic data and fault

下载: 全尺寸图片幻灯片

图 7 低分辨率断层与高分辨率非线性关系训练过程中训练集与验证集损失监测图

Fig. 7. Loss monitoring diagram of training set and validation set during the training process of the nonlinear relationship between low-resolution fault and high-resolution fault

下载: 全尺寸图片幻灯片

图 8 深度学习断层识别检测

a. 合成数据；b. 断裂标签；c. 预测结果

Fig. 8. Deep learning tomography detection

下载: 全尺寸图片幻灯片

图 9 高分辨率断层预测

a. 低分辨率断层模型；b. 高分辨率断层预测结果

Fig. 9. High-resolution fault prediction

下载: 全尺寸图片幻灯片

图 10 原始地震剖面

Fig. 10. The original seismic section

下载: 全尺寸图片幻灯片

图 11 第三代相干算法断层属性与地震数据叠合剖面

Fig. 11. The superimposed section of fault attributes and seismic data of the third-generation coherent algorithm

下载: 全尺寸图片幻灯片

图 12 深度学习算法断层属性与地震数据叠合剖面

Fig. 12. The superimposed section of the fault attributes and seismic data of the deep learning algorithm

下载: 全尺寸图片幻灯片

图 13 高分辨率深度学习断层属性与地震数据叠合剖面

Fig. 13. Superimposed section of high-resolution deep learning fault attributes and seismic data

下载: 全尺寸图片幻灯片

图 14 Kerry3D地震资料三维切片图

Fig. 14. 3D slice map of Kerry3D seismic data

下载: 全尺寸图片幻灯片

图 15 Kerry3D数据第三代相干三维切片图

Fig. 15. 3D slice map of third generation coherence of Kerry3D data

下载: 全尺寸图片幻灯片

图 16 Kerry3D数据蚂蚁追踪三维切片图

Fig. 16. 3D slice map of data ant tracking of Kerry 3D data

下载: 全尺寸图片幻灯片

图 17 Kerry3D数据深度学习断层识别三维切片图

Fig. 17. 3D slice map of fault detection on deep learning of Kerry3D data

下载: 全尺寸图片幻灯片

图 18 Kerry3D数据深度学习断层高分辨率识别三维切片图

Fig. 18. 3D slice map of fault high-resolution detection on deeplearning of Kerry3D data

下载: 全尺寸图片幻灯片

参考文献(31)

[1]	Bahorich, M., Farmer, S., 1995. 3-D Seismic Discontinuity for Faults and Stratigraphic Features: The Coherence Cube. The Leading Edge, 14(10): 1053-1058. https://doi.org/10.1190/1.1437077
[2]	Bahorich, M., Lopez, J., Haskell, N. L., et al., 1995. Stratigraphic and Structural Interpretation with 3D Coherence. SEGTechnical Program Expanded Abstracts. SEG, 97-100. https://doi.org/10.1990/1.1887435
[3]	Bakker, P., van Vliet, L. J., Verbeek, P. W., 1999. Edge Preserving Orientation Adaptive Filtering. IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Fort Collins, CO: IEEE, 535-540. https://doi.org/10.1109/CVRP.1999.786989
[4]	Bergbauer, S., Mukerji, T., Hennings, P., 2003. Improving Curvature Analyses of Deformed Horizons Using Scale–Dependent Filtering Techniques. AAPG Bulletin, 87(8): 1255-1272. https://doi.org/10.1306/0319032001101
[5]	Chehrazi, A., Rahimpour-Bonab, H., Rezaee, M. R., 2013. Seismic Data Conditioning and Neural Network-Based Attribute Selection for Enhanced Fault Detection. Petroleum Geoscience, 19(2): 169-183. https://doi.org/10.1144/petgeo2011-001
[6]	Gong, Y. J., Zhang, K. H., Zeng, Z. P., 2021. Origin of Overpressure, Vertical Transfer and Hydrocarbon Accumulation ofJurassic in Fukang Sag, Junggar Basin. Earth Science, 46(10): 3588-3600(in Chinese with English abstract).
[7]	Hinton, G. E., Salakhutdinov, R. R., 2006. Reducing the Dimensionality of Data with Neural Networks. Science, 313(5786): 504-507. https://doi.org/10.1126/science. 1127647 doi: 10.1126/science.1127647
[8]	Huang, L., Dong, X. S., Clee, T. E., 2017. A Scalable Deep Learning Platform for Identifying Geologic Features from Seismic Attributes. The Leading Edge, 36(3): 249-256. https://doi.org/10.1190/tle36030249.1
[9]	LeCun, Y., Bengio, Y., Hinton, G., 2015. Deep Learning. Nature, 521(7553): 436-444. https://doi.org/10.1038/nature14539
[10]	Lecun, Y., Bottou, L., Bengio, Y., et al., 1998. Gradient-Based Learning Applied to Document Recognition. Proceedings of the IEEE, 86(11): 2278-2324. https://doi.org/10.1109/5.726791
[11]	Lisle, R. J., 1994. Detection of Zones of Abnormal Strains in Structures using Gaussian Curvature Analysis. AAPG bulletin, 78(12): 1811-1819. https://doi.org/10.1306/A25FF305-171B-11D7-8645000102C1865D
[12]	Marfurt, K. J., Kirlin, R. L., Farmer, S. L., et al., 1998. 3-D Seismic Attributes Using a Semblance‐Based Coherency Algorithm. Geophysics, 63(4): 1150-1165. https://doi.org/10.1190/1.1444415
[13]	Marfurt, K. J., Sudhaker, V., Gersztenkorn, A., et al., 1999. Coherency Calculations in the Presence of Structural Dip. Geophisics, 65(1): 304-320. https://doi.org/10.1190/1.1444508
[14]	Roberts, A., 2001. Curvature Attributes and Their Application to 3D Interpreted horizons. First Break, 19(2): 85-100. https://doi.org/10.1046/j.02635046200100142x
[15]	Ronneberger, O., Fischer, P., Brox, T., 2015. U-Net: Convolutional Networks for Biomedical Image Segmentation. International Conference on Medical Image Computing and Computer-Assisted Intervention, 234-241. https://doi.org/10.1007/978331924574428
[16]	Wang, H. X., Fu, X. F., Fu, G., et al., 2014. Vertical Segmentation Growth of Fault and Oil Source Fault Determination in Fuyang Oil Layer of Sanzhao Depression. Earth Science, 39(11): 1639-1646(in Chinese with English abstract).
[17]	Wu, X. M., Liang, L. M., Shi, Y. Z., 2019. FaultSeg3D: Using Synthetic Data Sets to Train an End-to-End Convolutional Neural Network for 3D Seismic Fault Segmentation. Geophysics, 84(3): 35-45. https://doi.org/10.1190/geo2018-0646.1
[18]	Yan, W. B., Wang, G. C., Li, L., et al., 2015. Deformation Analyses and Their Geological Implications of Carboniferous-Permian Tectonic Transformation Period in Northwest Margin of Junggar Basin. Earth Science, 40(3): 504-520(in Chinese with English abstract).
[19]	Yang, B. L., Ye, J. R., Wang, Z. S., et al., 2014. Hydrocarbon Accumulation Models and Main Controlling Factors in Liaodong Bay Depression. Earth Science, 39(10): 1507-1520(in Chinese with English abstract).
[20]	Yang, W. Y., Yang, J. R., Chen, S. Q., 2021. Fault Detection of Seismic Data Based on U-Net Deep Learning Network. Oil Geophysical Prospecting, 56(4): 688-697(in Chinese with English abstract).
[21]	Yun, L., Zhang, J., Xu, W., et al., 2021. Geometry, Kinematics and Regional Tectonic Significance of the Huahai Fault in the Western Hexi Corridor. Earth Science, 46(1): 259-271(in Chinese with English abstract).
[22]	Zhou, B. W., Chen, H. H., Yun, L., et al., 2022. The Relationship between Fault Displacement and Damage Zone Width of the Paleozoic Strike‐Slip Faults in Shunbei Area, Tarim Basin. Earth Science, 47(2): 437-451 (in Chinese with English abstract).
[23]	Zhou, W. W., Wang, W. F., An, B., et al., 2014. Identification of Potential Fault Zones and Its Geological Significance in Bohai Bay Basin. Earth Science, 39(11): 1627-1638(in Chinese with English abstract).
[24]	宫亚军, 张奎华, 曾治平, 等, 2021. 准噶尔盆地阜康凹陷侏罗系超压成因、垂向传导及油气成藏. 地球科学, 46(10): 3588-3600. doi: 10.3799/dqkx.2020.366
[25]	王海学, 付晓飞, 付广, 等, 2014. 三肇凹陷断层垂向分段生长与扶杨油层油源断层的厘定. 地球科学, 39(11): 1639-1646. doi: 10.3799/dqkx.2014.146
[26]	晏文博, 王国灿, 李理, 等, 2015. 准噶尔西北缘石炭-二叠纪构造转换期变形分析及其地质意义. 地球科学, 40(3): 504-520. doi: 10.3799/dqkx.2015.040
[27]	杨宝林, 叶加仁, 王子嵩, 等, 2014. 辽东湾断陷油气成藏模式及主控因素. 地球科学, 39(10): 1507-1520. doi: 10.3799/dqkx.2014.133
[28]	杨午阳, 杨佳润, 陈双全, 等, 2021. 基于U-Net深度学习网络的地震数据断层检测. 石油地球物理勘探, 56(4): 688-697. https://www.cnki.com.cn/Article/CJFDTOTAL-SYDQ202104003.htm
[29]	云龙, 张进, 徐伟, 等, 2021. 河西走廊西段花海断裂几何学、运动学及区域构造意义. 地球科学, 46(1): 259-271. doi: 10.3799/dqkx.2019.227
[30]	周铂文, 陈红汉, 云露, 等, 2022. 塔里木盆地顺北地区下古生界走滑断裂带断距分段差异与断层宽度关系. 地球科学, 47(2): 437-451. doi: 10.3799/dqkx.2021.073
[31]	周维维, 王伟锋, 安邦等, 2014. 渤海湾盆地隐性断裂带识别及其地质意义. 地球科学, 39(11): 1627-1638. doi: 10.3799/dqkx.2014.145