多尺度时频空三域特征联合下的储层岩性识别方法

doi:10.11720/wtyht.2023.1020

Abstract
Figure/Table
References
Related Citation (15)

Download: PDF (6331 KB) HTML (1 KB)
Export: BibTeX | EndNote (RIS)

Abstract

Conventional methods for reservoir lithology identification suffer low precision and efficiency since reservoir lithologies have various types and complex compositions and alternate frequently.This study proposed a reservoir lithology identification method based on multi-scale time-frequency-space feature combination.Based on the original logging characteristics,this method introduced the multi-scale frequency-domain components from the complementary ensemble empirical mode decomposition (CEEMD) to improve the longitudinal resolution of log curves.Moreover,a multi-scale convolutional neural network-bidirectional gated recurrent unit-attention mechanism (CNN-BiGRU-AT) model was constructed to extract the spatio-temporal features of log data containing multi-scale frequency-domain components.In this way,the joint learning of time-frequency-space features of log data was realized.Finally,the model output was optimized using the attention mechanism to reduce the propagation of error information.To verify the reliability of this method,an experimental analysis was conducted using the data from five wells that have relatively complete data.As revealed by the analysis results,the identification accuracy of training and verification sets containing multi-scale frequency-domain components was increased by 9.50% and 8.66%,respectively in the comparative experiments of different data combinations.The method proposed in this study yielded sample identification accuracy of 94.11%.Compared with support vector machine (SVM),backpropagation (BP) neural network,convolutional neural network (CNN),bidirectional gated recurrent unit (BiGRU),and CNN-BiGRU fusion models, the identification accuracy of this method increased by 16.21%,14.54%,11.69%,5.05%,and 3.38%,respectively.

Key words： CEEMD attention mechanism multi-scale neural network lithology identification

Received: 25 January 2022 Published: 24 February 2023

ZTFLH:

Corresponding Authors: WEN Chang E-mail: wangzongren1999@163.com;wenchang2016paper@163.com

	Service

	E-mail this article
	Add to my bookshelf
	Add to citation manager
	E-mail Alert
	RSS
	Articles by authors
	Zong-Ren WANG
	Chang WEN
	Kai XIE
	Guan-Qun SHENG
	Jian-Biao HE

Cite this article:

Zong-Ren WANG,Chang WEN,Kai XIE, et al. Reservoir lithology identification method based on multi-scale time-frequency-space feature combination[J]. Geophysical and Geochemical Exploration, 2023, 47(1): 81-90.

URL:

https://www.wutanyuhuatan.com/EN/10.11720/wtyht.2023.1020 OR https://www.wutanyuhuatan.com/EN/Y2023/V47/I1/81

步骤	描述
Step 1:	将高斯白噪声 $n i t$ 加入待分解 $s t$ 中得到新信号 $x t$ , $i$ 为加入白噪声的次数, $i = 1,2, …, N$ : $x t = s t ± ε n i t$
Step 2:	对加入噪声的新信号进行EMD分解,并得出一阶本征模态分量 $C 1 t$ : $E x t = C 1 i t + r i$
Step 3:	对N个模态分量进行总体平均得到CEEMD分解的第一个本征模态分量: $C 1 t ˉ = 1 N ∑ i = 1 N C 1 i t$
Step 4:	去除第一个模态分量后的残差: $r 1 t = s t - C 1 t ˉ$
Step 5:	在 $r 1 t$ 中加入正负的高斯白噪声,再次进行EMD分解,得到一阶模态分量 $D 1$ ,由此可得到第2个本征模态分量 $C 2$ : $C 2 t ˉ = 1 N ∑ i = 1 N D 1 i t$
Step 6:	去除第二个本征模态分量后的残差: $r 2 t = r 1 t - C 2 t ˉ$
Step 7:	重复以上步骤,直到 $r n t$ 不能再次分解,算法结束。此时本征模态分量数量为K,原始信号 $s t$ 被分解为: $s t = ∑ k = 1 K C k t ˉ + r k t$

Algorithm process of CEEMD

Flow chart of multi-scale CNN-BiGRU-AT lithology identification algoritm

CNN multi-scale space feature extraction

BiGRU network structure diagram

Cell structure of GRU

Attention mechanism optimization diagram

Partial logging lithology data

Logging lithology box diagram of Well Su 27

Experimental hyperparameter configuration

Relationship between accuracy rate and iteration times

Relation diagram of loss value and iteration times

GR curve decomposition diagram of by different methods

Experimental results of different combinations of data

Lithology identification bar charts for groups 1 and 3

Lithology identification results of different models in Well Su 27

Histogram of lithology identification results of different models in Well Su 27

Confusion matrix of this method and CNN-BiGRU model for Well Su 27

[1]	付光明, 严加永, 张昆, 等. 岩性识别技术现状与进展[J]. 地球物理学进展, 2017, 32(1):26-40.
[1]	Fu G M, Yan J Y, Zhang K, et al. Current status and progress of lithology identification technology[J]. Progress in Geophysics, 2017, 32(1):26-40.
[2]	Adeniran A A, Adebayo A R, Salami H O, et al. A competitive ensemble model for permeability prediction in heterogeneous oil and gas reservoirs[J]. Applied Computing and Geosciences, 2019, 1:100004.
[3]	袁照威, 段正军, 张春雨, 等. 基于马尔科夫概率模型的碳酸盐岩储集层测井岩性解释[J]. 新疆石油地质, 2017, 38(1):96-102.
[3]	Yuan Z W, Duan Z J, Zhang G Y, et al. Interpretation of logging lithology in Carbonate reservoirs based on Markov Chain probability model[J]. Xinjiang Petroleum Geology, 2017, 38(1):96-102.
[4]	成大伟, 袁选俊, 周川闽, 等. 测井岩性识别方法及应用——以鄂尔多斯盆地中西部长7油层组为例[J]. 中国石油勘探, 2016, 21(5):117-126.
[4]	Cheng D W, Yuan X J, Zhou C M, et al. Logging-lithology identifi cation methods and their application:A case study on Chang 7 Member in central-western Ordos Basin,NW China[J]. China Petroleum Exploration, 2016, 21(5):117-126.
[5]	王泽华, 朱筱敏, 孙中春, 等. 测井资料用于盆地中火成岩岩性识别及岩相划分:以准噶尔盆地为例[J]. 地学前缘, 2015, 22(3):254-268.
[5]	Wang Z H, Zhu X M, Sun Z C, et al. Igneous lithology identification and lithofacies classification in the basin using logging data:Taking Junggar Basin as an example[J]. Earth Science Frontiers, 2015, 22(3):254-268.
[6]	Bai X L, Zhang S N, Huang Q Y, et al. Origin of dolomite in the Middle Ordovician peritidal platform carbonates in the northern Ordos Basin,western China[J]. Petroleum Science, 2016, 13(3):434-449.
[7]	Bressan T S, Souza M, Girelli T J, et al. Evaluation of machine learning methods for lithology classification using geophysical data[J]. Computers & Geosciences, 2020, 139:104475.
[8]	Corina A N, Hovda S. Automatic lithology prediction from well logging using kernel density estimation[J]. Journal of Petroleum Science and Engineering, 2018, 170:664-674.
[9]	安鹏, 曹丹平. 基于深度学习的测井岩性识别方法研究与应用[J]. 地球物理学进展, 2018, 33(3):1029-1034.
[9]	An P, Cao D P. Research and application of logging lithology identification based on deep learning[J]. Progress in Geophysics, 2018, 33(3):1029-1034.
[10]	蔡泽园, 鲁宝亮, 熊盛青, 等. 基于自适应核密度的贝叶斯概率模型岩性识别方法研究[J]. 物探与化探, 2020, 44(4):919-927.
[10]	Cai Z Y, Lu B L, Xiong S Q, et al. Lithology identification based on Bayesian probability using adaptive kernel density[J]. Geophysical and Geochemical Exploration, 2020, 44(4):919-927.
[11]	谷宇峰, 张道勇, 鲍志东, 等. GBDT识别致密砂岩储层岩性[J]. 地球物理学进展, 2021, 36(5):1956-1965.
[11]	Gu Y F, Zhang D Y, Bao Z D, et al. Lithology prediction of tight sandstone reservoirs using GBDT[J]. Progress in Geophysics, 2021, 169(5):1956-1965.
[12]	苏赋, 马磊, 罗仁泽, 等. 基于改进多分类孪生支持向量机的测井岩性识别方法研究与应用[J]. 地球物理学进展, 2020, 35(1):174-180.
[12]	Su F, Ma L, Luo R Z, et al. Research and application of logging lithology identification based on improve multi-class twin support vector machine[J]. Progress in Geophysics, 2020, 35(1):174-180.
[13]	杨柳青, 陈伟, 查蓓. 利用卷积神经网络对储层孔隙度的预测研究与应用[J]. 地球物理学进展, 2019, 34(4):1548-1555.
[13]	Yang L Q, Chen W, Cha P. Prediction and application of reservoir porosity by convolutional neural network[J]. Progress in Geophysics, 2019, 34(4):1548-1555.
[14]	武中原, 张欣, 张春雷, 等. 基于LSTM循环神经网络的岩性识别方法[J]. 岩性油气藏, 2021, 33(3):120-128.
[14]	Wu Z Y, Zang X, Zhang C L, et al. Lithology identification based on LSTM recurrent neural network[J]. Lithologic Reservoirs, 2021, 33(3):120-128.
[15]	周恒, 张春雷, 张欣, 等. 基于胶囊网络的碳酸盐岩储层岩性识别方法[J]. 天然气地球科学, 2021, 32(5):685-694.
[15]	Zhou H, Zhang C L, Zhang X, et al. Lithology identification method of carbonate reservoir based on capsule network[J]. Natural Gas Geoscience, 2021, 32(5):685-694.
[16]	王逸宸, 柳林涛, 许厚泽. 基于卷积神经网络识别重力异常体[J]. 物探与化探, 2020, 44(2):394-400.
[16]	Wang Y C, Liu L T, Xu H Z. The identification of gravity anomaly body based on the convolutional neural network[J]. Geophysical and Geochemical Exploration, 2020, 44(2):394-400.
[17]	梁立锋, 刘秀娟, 张宏兵, 等. 超参数对GRU-CNN混合深度学习弹性阻抗反演影响研究[J]. 物探与化探, 2021, 45(1):133-139.
[17]	Liang L F, Liu X J, Zhang H B, et al. A study of the effect of hyperparameters GRU-CNN hybrid deep learning EI inversion[J]. Geophysical and Geochemical Exploration, 2021, 45(1):133-139.