Seismic Frequency Enhancement Processing Based on Multi-Layer Residual Network and Its Application to Identification of Thin Reservoirs

doi:10.7657/XJPG20240114

Abstract

Abstract:

The seismic frequency enhancement processing method based on multi-layer residual network combines high-frequency well logging information with seismic data through an intelligent network. This method effectively improves vertical resolution while maintaining lateral continuity，facilitating the identification of thin reservoir beds. In the AMH area，the seismic data processed by conventional techniques enable only the identification of carbonates thicker than 30 m，but not of thinner beds. The seismic frequency enhancement processing method based on multi-layer residual network was proposed for application in this area. First，a training was performed using the multi-layer residual network，a deep learning network，with the near-wellbore seismic amplitudes as training data and the relative wave impedance data from well logging as training labels. Thus，a predictive model for relative wave impedance curve was obtained. By using seismic data as input，the deep network training model was solved to obtain a relative wave impedance data cube，and then a data cube of reflection coefficient corresponding to the frequency-enhanced seismic data cube was obtained. After analyzing the geological conditions of the target area，appropriate wide-frequency wavelet was extracted after calibration，and then convolved with the reflection coefficient cube，so that a frequency-enhanced seismic data cube was obtained. Reservoir inversion was performed using the frequency-enhanced seismic data cube. The inversion results are of high resolution vertically，well matching the main target beds，and can be identifiable and traceable laterally. Ultimately，the identification of thin beds in the AMH area was realized through the application of high-resolution seismic inversion results. The seismic frequency enhancement processing based on multi-layer residual network together with the corresponding high-resolution model inversion can identify beds thicker than 10 m in the AMH area. This method effectively addresses the problem of infeasible thin bed identification using low-resolution seismic data，and improves the accuracy in predicting thin beds. It is referential for identifying similar thin beds.

Key words: carbonate, seismic data, frequency enhancement processing, thin bed, multi-layer residual network, relative wave impedance, high-resolution inversion, deep learning

CLC Number:

TE122

ZHANG Wenqi, LI Chunlei. Seismic Frequency Enhancement Processing Based on Multi-Layer Residual Network and Its Application to Identification of Thin Reservoirs[J]. Xinjiang Petroleum Geology, 2024, 45(1): 102-108.

Figures/Tables 7

Fig. 1.

Fig. 2.

Fig. 3.

Fig. 4.

Fig. 5.

Fig. 6.

Fig. 7.

References 19

[1]	李庆忠. 地震勘探分辨率与信噪比谱的关系:答云美厚“地震分辨力新认识”一文[J]. 石油地球物理勘探, 2008, 43(2):244-245.
	LI Qingzhong. Relationship between resolution of seismic exploration and spectrum of S/N ratio[J]. Oil Geophysical Prospecting, 2008, 43(2):244-245.
[2]	单祥, 窦洋, 晏奇, 等. 玛南斜坡区风城组致密油藏储集层特征及控制因素[J]. 新疆石油地质, 2022, 43(6):704-713.
	SHAN Xiang, DOU Yang, YAN Qi, et al. Characteristics and controlling factors of tight oil reservoirs of Fengcheng formation in southern slope of Mahu sag[J]. Xinjiang Petroleum Geology, 2022, 43(6):704-713.
[3]	OPPENHEIM A V, KOPEC G E, TRIBOLET J M. Signal analysis by homomorphic prediction[R]. America:IEEE Transactions on Acoustics,Speech,and Signal Processing, 1976.
[4]	段南. 叠前地震波形指示反演在薄互储层预测中的应用[J]. 地球物理学进展, 2019, 34(2):523-528.
	DUAN Nan. Application of prestack seismic waveform indicates inversion in thin reservoir prediction[J]. Progress in Geophysics, 2019, 34(2):523-528.
[5]	韩淼, 白忠凯, 周新桂, 等. 新疆塔里木盆地温宿凸起区新近系薄储集层储层预测方法研究与应用[J]. 地球学报, 2018, 39(6):691-697.
	HAN Miao, BAI Zhongkai, ZHOU Xingui, et al. Research and application of reservoir prediction method to Neogene thin sand-shale interbed in Wensu uplift,Tarim basin,Xinjiang[J]. Acta Geoscientica Sinica, 2018, 39(6):691-697.
[6]	胡玮, 齐鹏, 杨江峰, 等. 波形指示反演在超深层致密砂岩薄储层中的应用[J]. 地球物理学进展, 2018, 33(2):620-624.
	HU Wei, QI Peng, YANG Jiangfeng, et al. Application of seismic motion inversion in identification of tight thin super deep reservoirs[J]. Progress in Geophysics, 2018, 33(2):620-624.
[7]	罗泽, 谢明英, 涂志勇, 等. 一套针对高泥质疏松砂岩薄储层的识别技术:以珠江口盆地X油田为例[J]. 岩性油气藏, 2019, 31(6):95-101.
	LUO Ze, XIE Mingying, TU Zhiyong, et al. A set of recognition techniques for thin reservoirs with unconsolidated high-argillaceous sandstone:A case study from X oilfield in Pearl River Mouth basin[J]. Lithologic Reservoirs, 2019, 31(6):95-101.
[8]	李奎周, 郑绪瑭, 赵海波, 等. 基于波形分解的重构地震数据体技术在致密油薄储层预测中的应用[J]. 大庆石油地质与开发, 2022, 41(4):131-137.
	LI Kuizhou, ZHENG Xutang, ZHAO Haibo, et al. Application of reconstructed seismic data volume technology based on waveform decomposition in prediction of thin tight oil reservoir[J]. Petroleum Geology & Oilfield Development in Daqing, 2022, 41(4):131-137.
[9]	孙鹏霄. 海上薄储集层砂岩油藏层间干扰规律及调整对策[J]. 特种油气藏, 2022, 29(5):119-125. doi: 10.3969/j.issn.1006-6535.2022.05.017
	SUN Pengxiao. Interlayer interference pattern and adjustment countermeasures of offshore thin interbedded sandstone reservoirs[J]. Special Oil & Gas Reservoirs, 2022, 29(5):119-125.
[10]	廉桂辉, 朱亚婷, 王晓光, 等. 叠前反演技术在玛湖油田储层预测中的应用[J]. 特种油气藏, 2022, 29(1):80-84. doi: 10.3969/j.issn.1006-6535.2022.01.012
	LIAN Guihui, ZHU Yating, WANG Xiaoguang, et al. Application of pre-stack inversion technology to reservoir prediction of Mahu oilfield[J]. Special Oil & Gas Reservoirs, 2022, 29(1):80-84.
[11]	王林生, 艾建华, 伍顺伟, 等. 基于扩展弹性阻抗反演的致密砂砾岩储层定量预测技术:以玛湖凹陷达13井区为例[J]. 油气地质与采收率, 2022, 29(3):36-44.
	WANG Linsheng, AI Jianhua, WU Shunwei, et al. Quantitative prediction technology for tight glutenite reservoirs based on EEI inversion:A case of Well Da13 area in Mahu sag[J]. Petroleum Geology and Recovery Efficiency, 2022, 29(3):36-44.
[12]	METWALLI F I, SHENDI E H, FAGELNOUR M S. Seismic facies analysis of thin sandstone reservoirs,north western desert,Egypt[J]. Journal of Petroleum Exploration and Production Technologies, 2019, 9:793-808.
[13]	董恩清, 刘贵忠, 张宗平. 自适应Kalman滤波反褶积的快速实现方法[J]. 地球物理学报, 2001, 44(2):255-262.
	DONG Enqing, LIU Guizhong, ZHANG Zongping. Fast implementation technique of adaptive Kalman filtering deconvolution via dyadic wavelet transform[J]. Chinese Journal of Geophysics, 2001, 44(2):255-262.
[14]	彭作磊, 陈铭. 多元地震属性分析提高薄储层预测精度[J]. 断块油气田, 2020, 27(3):318-322.
	PENG Zuolei, CHEN Ming. Multivariate seismic attribute analysis improve prediction accuracy of thin reservoirs[J]. Fault-Block Oil & Gas Field, 2020, 27(3):318-322.
[15]	赵波, 俞寿朋, 聂勋碧, 等. 谱模拟反褶积方法及其应用[J]. 石油地球物理勘探, 1996, 31(1):101-116.
	ZHAO Bo, YU Shoupeng, NIE Xunbi, et al. Spectral-modeled deconvolution and its application[J]. Oil Geophysical Prospecting, 1996, 31(1):101-116.
[16]	曹思远, 袁殿. 高分辨率地震资料处理技术综述[J]. 新疆石油地质, 2016, 37(1):112-119.
	CAO Siyuan, YUAN Dian. A review of high-resolution seismic data processing approaches[J]. Xinjiang Petroleum Geology, 2016, 37(1):112-119.
[17]	黄忠来, 张建中, 邹志辉. 二阶同步挤压S变换及其在地震谱分解中的应用[J]. 地球物理学报, 2017, 60(7):2833-2844. doi: 10.6038/cjg20170728
	HUANG Zhonglai, ZHANG Jianzhong, ZOU Zhihui. A second-order synchrosqueezing S-transform and its application in seismic spectral decomposition[J]. Chinese Journal of Geophysics, 2017, 60(7):2833-2844.
[18]	张璐, 汪毓铎. 地震数据分频谱蓝化算子计算方法及应用[J]. 北京信息科技大学学报(自然科学版), 2020, 35(1):63-68.
	ZHANG Lu, WANG Yuduo. Calculation and application of spectral blue operator of the frequency division seismic data[J]. Journal of Beijing Information Science & Technology University, 2020, 35(1):63-68.
[19]	陈雨茂, 赵虎, 杨宏伟, 等. 基于小样本数据深度学习的砂体厚度预测方法及应用[J]. 新疆石油地质, 2023, 44(2):231-237.
	CHEN Yumao, ZHAO Hu, YANG Hongwei, et al. A sand body thickness prediction method based on deep learning from small sample data and its application[J]. Xinjiang Petroleum Geology, 2023, 44(2):231-237.