基于 MAE 神经网络的测井曲线地层自动识别方法

白薷; 王世玉; 张璐; 张亮; 杜炜; 耿代; 姚振杰

doi:10.12055/gaskk.issn.1673-3177.2024.04.007

天然气勘探与开发 >

2024 , Vol. 47 >Issue 4: 63 - 71

DOI: https://doi.org/10.12055/gaskk.issn.1673-3177.2024.04.007

资源勘探

基于 MAE 神经网络的测井曲线地层自动识别方法

白薷 ,
王世玉 ,
张璐 ,
张亮 ,
杜炜 ,
耿代 ,
姚振杰

展开

1.陕西延长石油（集团）有限责任公司研究院陕西西安 710065;
2.延长油田股份有限公司杏子川采油厂陕西延安 717400

白薷,女,1984年生,博士,高级工程师;主要从事智慧油田及人工智能工作。地址：（710065）陕西省西安市唐延路61号。E-mail：546130242@qq.com

修回日期: 2023-12-21

网络出版日期: 2024-09-30

基金资助

国家重点研发计划项目（编号：2022YFE0206700）、陕西延长石油（集团）有限责任公司科技项目（编号：ycsy2022jcts-B-45, ycsy2022jcts-B-47）

收起

An automatic identifying method for strata via logging curves based on MAE neural network

BAI Ru ,
WANG Shiyu ,
ZHANG Lu ,
ZHANG Liang ,
DU Wei ,
GENG Dai ,
YAO Zhenjie

Expand

1. Research Institute, Shaanxi Yanchang Petroleum (Group) Co. Ltd., Xi’an, Shaanxi 710065, China;
2. Xingzichuan Oil Production Plant, Yanchang Oilfield Co. Ltd., Yan’an, Shaanxi 717400, China

Revised date: 2023-12-21

Online published: 2024-09-30

Fold

摘要

针对目前计算机自动分层识别准确率低,人工分层效率不高的问题,提出了一种基于掩码自编码器（MAE）神经网络算法的地层分层识别新方法。通过测井曲线优选,以自然电位、自然伽马、声波时差、电阻率、井位坐标和补心海拔作为输入特征变量进行模型训练及预测,再利用损失函数、准确率、精确率、召回率、F1 值进行模型性能评估。研究结果表明：①基于 MAE 神经网络训练得到地层分层识别模型,预测准确率能够达到 95.54% ;②MAE神经网络模型与卷积神经网络和关注分层边界的卷积神经网络模型进行地层分层实验对比,MAE神经网络模型的性能和预测精度均较高,准确率分别提高了8.30% 和 6.32%,且无地层紊乱情况出现,具有明显分层优势; ③MAE神经网络模型应用于未知井的地层分层中,自动分层预测准确率为98.07%。表明该方法具有较高的地层识别效果,为油田地层识别提供了一种基于自监督神经网络算法的理论支持和有益探索。

关键词： 掩码自编码器; 自监督; 神经网络; 测井曲线; 地层识别

本文引用格式

白薷 , 王世玉 , 张璐 , 张亮 , 杜炜 , 耿代 , 姚振杰 . 基于 MAE 神经网络的测井曲线地层自动识别方法[J]. 天然气勘探与开发, 2024 , 47(4) : 63 -71 . DOI: 10.12055/gaskk.issn.1673-3177.2024.04.007

Abstract

This paper presents a new method to automatically identify strata based on masked autoencoder (MAE) neural network in an effort to address two problems of low accuracy in computerized automatic identification and poor efficiency in artificial identification. As characteristic variables, the spontaneous potential, natural gamma, acoustic slowness, resistivity, well coordinates, and kelly bushing are input into the trained simulation which is predicted by optimizing logging curves. Furthermore, all simulation performance is evaluated using loss function, accuracy and precision, as well as recall and F1 values. It is found that, this trained simulation makes the prediction accuracy up to 95.54%. Some experimental contrasts demonstrate that, with the increasing accuracy by 8.3% and 6.32%, respectively, the simulation is better in its performance and prediction accuracy than those of both convolutional neural network (CNN) and boundary-guided CNN. Without any stratigraphic disturbance, this new method enjoys an obvious cutting edge in stratigraphic identification. In addition, after its application to unknown wells, the accuracy achieved 98.07%, showing a better effect. In general, this method may provide theoretical support and helpful discussion for stratigraphic identification based on self-supervised neural network algorithms.

Key words： MAE; Self-supervised; Neural network; Logging curve; Stratigraphic identification

参考文献

[1] 覃瑞东, 潘和平, 郭博, 等. 基于Hilbert-Huang变换的测井曲线自动分层方法[J]. 地质科技情报, 2017, 36(2): 258-264.
QIN Ruidong, PAN Heping, GUO Bo, et al.Automatic stratification of well logging curves with Hilbert-Huang transform[J]. Geological Science and Technology Information, 2017, 36(2): 258-264.
[2] 张强, 李家金, 王毛毛, 等. 基于改进主成分分析法的测井曲线岩性分层技术[J]. 吉林大学学报(地球科学版), 2022, 52(4): 1369-1376.
ZHANG Qiang, LI Jiajin, WANG Maomao, et al.Logging curve rock layering technology based on improved principal component analysis[J]. Journal of Jilin University (Earth Science Edition), 2022, 52(4): 1369-1376.
[3] 肖波, 韩学辉, 周开金, 等. 测井曲线自动分层方法回顾与展望[J]. 地球物理学进展, 2010, 25(5): 1802-1810.
XIAO Bo, HAN Xuehui, ZHOU Kaijin, et al.A review and outlook of automatic zonation methods of well log[J]. Progress in Geophysics, 2010, 25(5): 1802-1810.
[4] 井元帅. 致密砂岩含气储层预测方法优化及应用——以苏53区块为例[J]. 天然气勘探与开发, 2019, 42(3): 78-85.
JING Yuanshuai.Methods to predict tight sandstone gas-bearing reservoirs and their application to Su 53 block[J]. Natural Gas Exploration and Development, 2019, 42(3): 78-85.
[5] 尚福华, 李金成, 原野, 等. 基于改进BP神经网络的地层划分方法[J]. 计算机技术与发展, 2020, 30(9): 148-153.
SHANG Fuhua, LI Jincheng, YUAN Ye, et al.Stratigraphic division method based on improved BP neural network[J]. Computer Technology and Development, 2020, 30(9): 148-153.
[6] 徐朝晖, 刘钰铭, 周新茂, 等. 基于卷积神经网络算法的自动地层对比实验[J]. 石油科学通报, 2019, 4(1): 1-10.
XU Zhaohui, LIU Yuming, ZHOU Xinmao, et al.An experiment in automatic stratigraphic correlation using convolutional neural networks[J]. Petroleum Science Bulletin, 2019, 4(1): 1-10.
[7] 韩科龙. 测井自动分层方法研究及其在岩性识别中的应用[D]. 北京: 中国地质大学(北京), 2011.
HAN Kelong.Research on the automatic zonation methods of well logs and its application on the lithology identification[D]. Beijing: China University of Geosciences (Beijing), 2011.
[8] 刘家瑾, 陆国纯. 测井曲线的有序最优化极差分层[J]. 测井技术, 1987, 11(3): 59-67.
LIU Jiajin, LU Guochun.Ordered optimization range stratification of logging curves[J]. Well Logging Technology, 1987, 11(3): 59-67.
[9] 易觉非. 利用活度分层法实现测井自动地质分层[J]. 石油天然气学报, 2007, 29(1): 78-80.
YI Juefei.Automatic geologic zonation using activity layering method[J]. Journal of Oil and Gas Technology, 2007, 29(1): 78-80.
[10] 薛波, 杨青, 张超虹. 基于形态学滤波与小波变换的测井曲线自动分层方法[J]. 地球物理学进展, 2020, 35(1): 203-210.
XUE Bo, YANG Qing, ZHANG Chaohong.Automatic stratification method of logging curve based on morphological filtering and wavelet transform[J]. Progress in Geophysics, 2020, 35(1): 203-210.
[11] 初颖, 吕堂红. 基于极值法和聚类分析法的测井曲线自动分层模型——以山东省胜利油井为例[J]. 长春理工大学学报(自然科学版), 2017, 40(6): 105-110.
CHU Ying, LV Tanghong.Logging-curve automatic layered model based on the extreme value method and clustering analysis—Taking Shandong province, Shengli oilfield as an example[J]. Journal of Changchun University of Science and Technology (Natural Science Edition), 2017, 40(6): 105-110.
[12] 万应明, 高峻, 董建平, 等. 多测井曲线合成应用方法初探[J]. 石油物探, 2005, 44(1): 71-75.
WANG Yingming, GAO Jun, DONG Jianping, et al.The primary discussion to the application of synthetic multiple-logging[J]. Geophysical Prospecting for Petroleum, 2005, 44(1): 71-75.
[13] 金燕. 人工神经网络在测井地质领域中的应用[J]. 天然气勘探与开发, 1999, 22(1): 1-13.
JING Yan.Application of artificial neural network in logging geology[J]. Natural Gas Exploration and Development, 1999, 22(1): 1-13.
[14] 贺斌, 马立涛, 蔡瑞豪, 等. 基于非电阻率的低阻致密砂岩气层含水饱和度预测[J]. 非常规油气, 2019, 6(6): 29-40.
HE Bin, MA Litao, CAI Ruihao, et al.Water saturation prediction method in low resistivity tight sandstone reservoirs based on non-resistivity[J]. Unconventional Oil & Gas, 2019, 6(6): 29-40.
[15] 唐振国, 迟博, 吕金龙, 等. 基于多属性神经网络地震相的扶余油层砂体精细刻画及应用[J]. 非常规油气, 2020, 7(2): 41-48.
TANG Zhenguo, CHI Bo, LV Jinlong, et al.Fine characterization and application of sand body in Fuyu reservoir based on seismic facies by multi-attribute neural network[J]. Unconventional Oil & Gas, 2020, 7(2): 41-48.
[16] 肖红, 张瑶瑶, 张福禄. 改进的卷积神经网络及在地层识别中的应用[J]. 计算机技术与发展, 2021, 31(9): 167-172.
XIAO Hong, ZHANG Yaoyao, ZHANG Fulu.Improved convolution neural network and its application of stratigraphic identification[J]. Computer Technology and Development, 2021, 31(9): 167-172.
[17] 崔欣锋. 基于深度学习的多井分层与地层对比方法研究[D]. 大庆: 东北石油大学, 2023.
CUI Xinfeng.Research on multi-well stratification and stratigraphic comparison method based on deep learning[D]. Daqing: Northeast Petroleum University, 2023.
[18] HE K.M., CHEN X. L., XIE S. N., et al. Masked autoencoders are scalable vision learners[C]//2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). New Orleans: IEEE, 2022: 15979-15988.
[19] 孙杰光. 基于MAE预训练的深度学习缺陷检测网络研究[J]. 信息与电脑, 2022, 34(24): 161-166.
SUN Jieguang.Deep learning defect detection network base on MAE pre-training[J]. Information & Computer, 2022, 34(24): 161-166.
[20] 周飞燕, 金林鹏, 董军. 卷积神经网络研究综述[J]. 计算机学报, 2017, 40(6): 1229-1251.
ZHOU Feiyan, JIN Linpeng, DONG Jun.Review of convolutional neural network[J]. Chinese Journal of Computers, 2017, 40(6): 1229-1251.
[21] 陈钢花, 梁莎莎, 王军, 等. 卷积神经网络在岩性识别中的应用[J]. 测井技术, 2019, 43(2): 129-134.
CHEN Ganghua, LIANG Shasha, WANG Jun, et al.Application of convolutional neural network in lithology identification[J]. Well Logging Technology, 2019, 43(2): 129-134.
[22] 王华, 张雨顺. 测井资料人工智能处理解释的现状及展望[J]. 测井技术, 2021, 45(4): 345-356.
WANG Hua, ZHANG Yushun.Research status and prospect of artificial intelligence in logging data processing and interpretation[J]. Well Logging Technology, 2021, 45(4): 345-356.
[23] 由婷, 朱天怡, 徐鹏晔, 等. 高分辨岩心曲线构建在测井岩性识别中的应用[J]. 测井技术, 2023, 47(2): 138-145.
YOU Ting, ZHU Tianyi, XU Pengye, et al.Application of high-resolution core curve construction in logging lithology identification[J]. Well Logging Technology, 2023, 47(2): 138-145.
[24] 罗仁泽, 周洋, 康丽侠, 等. 基于DMC-BiLSTM的沉积微相智能识别方法[J]. 石油物探, 2022, 61(2): 253-261.
LUO Renze, ZHOU Yang, KANG Lixia, et al.Intelligent identification of sedimentary microfacies based on DMC-BiLSTM[J]. Geophysical Prospecting for Petroleum, 2022, 61(2): 253-261.

Options

文章导航

模态框（Modal）标题

摘要

本文引用格式

Abstract

参考文献

访问统计