Coal body structure identification method based on Bayesian-optimized CNN-BiLSTM-Attention

Authors

BIAN Huiyuan, College of Geology and Environment, Xi’an University of Science and Technology, Xi’an 710054, ChinaFollow
JI Jiajun, College of Geology and Environment, Xi’an University of Science and Technology, Xi’an 710054, China
DUAN Chaowei, China Petroleum Logging Co. Ltd., Xi’an 710077, China
ZHOU Jun, College of Geology and Environment, Xi’an University of Science and Technology, Xi’an 710054, China
LI Kun, College of Geology and Environment, Xi’an University of Science and Technology, Xi’an 710054, China
MA Yuhan, College of Geology and Environment, Xi’an University of Science and Technology, Xi’an 710054, China

Abstract

Background Coal-bearing basins contain primary and tectonically deformed coals due to multistage tectonic deformations. However, the gas-bearing properties of coal seams differ significantly due to varying pore and fracture densities, permeability, and mechanical properties. This makes coal structure assessment critical to coalbed methane (CBM) exploration and production. Objective and Method To enhance the accuracy and intelligence of coal structure identification, this study constructed a CNN-BiLSTM-Attention hybrid model that integrated a Bayesian optimization strategy. This model allowed for efficient fusion and representation of multi-scale log data by combining the local feature extraction capability of the convolutional neural network (CNN), the temporal sequence modeling strength of the bidirectional long short-term memory (BiLSTM), and the feature focusing ability of the Attention mechanism. Moreover, this model showed elevated stability and high training efficiency thanks to automatic parameter tuning through Bayesian optimization. Focusing on coal seams in the Shanxi and Benxi formations within the Ordos Basin, this study constructed a dataset of primary, primary-cataclastic, and cataclastic coals based on conventional log data, subjected to normalization, outlier removal, and interpolation of missing values, as well as data from cores. Then, the hybrid model was trained and assessed using the cross-entropy loss function. Results and Conclusions The CNN-BiLSTM-Attention hybrid model yielded an accuracy of 95.12%, outperforming isolated BiLSTM and CNN models. This hybrid model yielded precision and recall rates above 93% for various coal structures. Furthermore, it yielded a uniform error distribution, as indicated by the confusion matrices. This model was applied to well X2, demonstrating high consistency and discriminative ability for transition zones between varying coal structures. This significantly reduces misclassification between primary-cataclastic and cataclastic coals. Additionally, the hybrid model exhibits strong robust performance in processing noise in log data. This study offers a reliable and effective approach for fine-scale CBM assessment.

Keywords

coal structure, deep learning, CNN-BiLSTM-Attention, Bayesian optimization, log data

DOI

10.12363/issn.1001-1986.25.05.0387

Recommended Citation

BIAN Huiyuan, JI Jiajun, DUAN Chaowei, et al. (2025) "Coal body structure identification method based on Bayesian-optimized CNN-BiLSTM-Attention," Coal Geology & Exploration: Vol. 53: Iss. 11, Article 5.
DOI: 10.12363/issn.1001-1986.25.05.0387
Available at: https://cge.researchcommons.org/journal/vol53/iss11/5

Reference

[1] 魏迎春，孟涛，张劲，等. 不同煤体结构煤储层与煤层气井产出煤粉特征的关系：以鄂尔多斯盆地东缘柳林区块为例[J]. 石油学报，2023，44(6)：1000−1014.

WEI Yingchun，MENG Tao，ZHANG Jin，et al. Relationship between coal reservoirs with different coal structures and the characteristics of coal fines produced in CBM wells：A case study of Liulin Block at the eastern margin of Ordos Basin[J]. Acta Petrolei Sinica，2023，44(6)：1000−1014.

[2] SHU Yong，SANG Shuxun，ZHOU Xiaozhi，et al. Coal body structure detection based on logging and seismic data and its impacts on coalbed methane development：A case study in the Dahebian Block，western Guizhou，Southern China[J]. Natural Resources Research，2023，32(2)：691−716.

[3] 师庆民，耿旭虎，王双明，等. 基于煤体真密度和自然伽马响应规律的富油煤判识[J]. 煤田地质与勘探，2024，52(7)：85−96.

SHI Qingmin，GENG Xuhu，WANG Shuangming，et al. Identification of tar–rich coal based on the true density and natural gamma ray response of coals[J]. Coal Geology & Exploration，2024，52(7)：85−96.

[4] CHEN Qiang，YAO Huifang，CHANG Suoliang，et al. Coalbody structure classification method based on dual–lateral and RXO crossplot analysis[J]. Journal of Coal Science and Engineering (China)，2013，19(4)：522−529.

[5] 师素珍，郭家成，白洁玢，等. 利用消除调谐效应的AVO属性进行煤体结构识别[J]. 地球物理学报，2025，68(4)：1542−1554.

SHI Suzhen，GUO Jiacheng，BAI Jiebin，et al. Eliminate the tuning effect of AVO attribute for coal structure identification[J]. Chinese Journal of Geophysics，2025，68(4)：1542−1554.

[6] 付玉通，张伟，李永臣，等. 鄂东南地区深部煤层气煤体结构测井评价研究[J]. 中国煤炭，2017，43(9)：31−34.

FU Yutong，ZHANG Wei，LI Yongchen，et al. Logging evaluation research on the structure of deep coal body with CBM in the southeast of Hubei Province[J]. China Coal，2017，43(9)：31−34.

[7] DUAN Hongyue，MA Yulong，WANG Jiyao，et al. Coal structure characteristics of the 2# coal seam in the Jiaozuo mining area and its geological dependence[J]. ACS Omega，2023，8(42)：39242−39249.

[8] 李存磊，杨兆彪，孙晗森，等. 多煤层区煤体结构测井解释模型构建[J]. 煤炭学报，2020，45(2)：721−730.

LI Cunlei，YANG Zhaobiao，SUN Hansen，et al. Construction of a logging interpretation model for coal structure from multi–coal seams area[J]. Journal of China Coal Society，2020，45(2)：721−730.

[9] 巩泽文. 古叙矿区煤体及其组合的测井曲线识别技术[J]. 科学技术与工程，2019，19(20)：77−84.

GONG Zewen. Logging curve identification technology of coal body and combination in Guxu mining area[J]. Science Technology and Engineering，2019，19(20)：77−84.

[10] 张建国，韩晟，张聪，等. 基于聚煤环境分区的煤体结构测井判别及应用：以沁水盆地南部马必东地区为例[J]. 煤田地质与勘探，2021，49(4)：114−122.

ZHANG Jianguo，HAN Sheng，ZHANG Cong，et al. Coal body structure identification by logging based on coal accumulation environment zoning and its application in Mabidong Block，Qinshui Basin[J]. Coal Geology & Exploration，2021，49(4)：114−122.

[11] WANG Dengke，LI Lu，ZHANG Hongtu，et al. Intelligent identification of coal fractures using an improved U–shaped network[J]. Advances in Geo–Energy Research，2025，15(2)：129−142.

[12] 彭苏萍，魏文希，杜文凤，等. 基于AVO反演与叠前同步反演预测构造煤分布[J]. 中国矿业大学学报，2018，47(5)：929−934.

PENG Suping，WEI Wenxi，DU Wenfeng，et al. AVO inversion associated with prestack simultaneous inversion in determining the distribution of tectonic coal[J]. Journal of China University of Mining & Technology，2018，47(5)：929−934.

[13] 陈晶，黄文辉，陆小霞，等. 沁水盆地柿庄地区构造煤定量分析及其物性特征[J]. 煤炭学报，2017，42(3)：732−737.

CHEN Jing，HUANG Wenhui，LU Xiaoxia，et al. Quantitative analysis on tectonically deformed coal and its physical properties in Shizhuang area of Qinshui Basin[J]. Journal of China Coal Society，2017，42(3)：732−737.

[14] 师庆民，赵奔，王双明，等. 三塘湖盆地侏罗系富油煤特征及沉积环境控制[J]. 石油学报，2024，45(5)：787−803.

SHI Qingmin，ZHAO Ben，WANG Shuangming，et al. Characteristics and sedimentary environment control of Jurassic tar–rich coal in Santanghu Basin[J]. Acta Petrolei Sinica，2024，45(5)：787−803.

[15] 姚军朋，司马立强，张玉贵. 构造煤地球物理测井定量判识研究[J]. 煤炭学报，2011，36(增刊1)：94−98.

YAO Junpeng，SIMA Liqiang，ZHANG Yugui. Quantitative identification of deformed coals by geophysical logging[J]. Journal of China Coal Society，2011，36(Sup.1)：94−98.

[16] 张超，黄华州，徐德林，等. 基于测井曲线的煤体结构判识[J]. 煤炭科学技术，2017，45(9)：47−51.

ZHANG Chao，HUANG Huazhou，XU Delin，et al. Coal structure identified based on logging curve[J]. Coal Science and Technology，2017，45(9)：47−51.

[17] 田瀚，李宁，王双明，等. 富油煤焦油产率测井评价方法研究[J]. 煤田地质与勘探，2024，52(7)：97−107.

TIAN Han，LI Ning，WANG Shuangming，et al. A log–based method for evaluating the tar yield of tar–rich coal[J]. Coal Geology & Exploration，2024，52(7)：97−107.

[18] 李全中，赵凌云，胡海洋. 基于测井响应的煤体结构识别及开发效果评价[J]. 煤矿安全，2021，52(1)：13−19.

LI Quanzhong，ZHAO Lingyun，HU Haiyang. Coal body structure identification and development effect evaluation based on logging response[J]. Safety in Coal Mines，2021，52(1)：13−19.

[19] CHANG Xiangchun，HAN Runye，ZHANG Junjian，et al. Prediction of coal body structure of deep coal reservoirs using logging curves：Principal component analysis and evaluation of factors influencing coal body structure distribution[J]. Natural Resources Research，2025，34(2)：1023−1044.

[20] 丁阳阳，赵军龙，李兆明，等. 基于XGBoost算法的煤体结构测井识别技术研究[J]. 地球物理学进展，2022，37(3)：998−1006.

DING Yangyang，ZHAO Junlong，LI Zhaoming，et al. Research on logging recognition technology of coal structure based on XGBoost algorithm[J]. Progress in Geophysics，2022，37(3)：998−1006.

[21] 肖航，张占松，郭建宏，等. 基于随机森林结合地球物理测井资料的煤体结构识别方法及应用[J]. 科学技术与工程，2021，21(24)：10174−10180.

XIAO Hang，ZHANG Zhansong，GUO Jianhong，et al. Coal structure identification method based on random forest combined with geophysical logging data and its application[J]. Science Technology and Engineering，2021，21(24)：10174−10180.

[22] TANG Xiaoyan. Quantitative evaluation of CBM reservoir fracturing quality using logging data[J]. Journal of Geophysics and Engineering，2017，14(2)：226−237.

[23] TONG Zhongzheng，MENG Yanjun，ZHANG Jinchuan，et al. Coal structure identification based on geophysical logging data：Insights from Wavelet Transform (WT) and Particle Swarm Optimization Support Vector Machine (PSO–SVM) algorithms[J]. International Journal of Coal Geology，2024，282：104435.

[24] 匡亮，赵万强，喻渝. BP神经网络法预测隧道瓦斯突出的模型与实例[J]. 铁道工程学报，2018，35(2)：56−61.

KUANG Liang，ZHAO Wanqiang，YU Yu. Research on the prediction model and case of coal and gas outburst in tunnel by using BP neural network[J]. Journal of Railway Engineering Society，2018，35(2)：56−61.

[25] 汪子祺，吴朝容，黄开兴，等. 基于卷积神经网络的页岩TOC三维定量预测方法[J]. 石油地球物理勘探，2025，60(2)：273−282.

WANG Ziqi，WU Chaorong，HUANG Kaixing，et al. 3D quantitative prediction method for shale TOC based on convolutional neural network[J]. Oil Geophysical Prospecting，2025，60(2)：273−282.

[26] JIA Cunde，KONG Xiangdong，WANG Minghui，et al. Application of attention mechanism–based LSTM neural network in stratigraphy identification[J]. Results in Engineering，2025，27：105267.

[27] GAO Yining，TIAN Miao，GRANA D，et al. Attention mechanism–assisted recurrent neural network for well log lithology classification[J]. Geophysical Prospecting，2025，73(2)：628−649.

[28] 闫霞，熊先钺，徐凤银，等. 深部煤岩储层地质力学影响机制及控制因素：以鄂尔多斯盆地大吉区块为例[J]. 煤炭学报，2025，50(5)：2550−2566.

YAN Xia，XIONG Xianyue，XU Fengyin，et al. Deep coal geomechanical influence mechanism and its control factors of Daji Block in Ordos Basin[J]. Journal of China Coal Society，2025，50(5)：2550−2566.

[29] 赵喆，徐旺林，赵振宇，等. 鄂尔多斯盆地石炭系本溪组煤岩气地质特征与勘探突破[J]. 石油勘探与开发，2024，51(2)：234−247.

ZHAO Zhe，XU Wanglin，ZHAO Zhenyu，et al. Geological characteristics and exploration breakthroughs of coal rock gas in Carboniferous Benxi Formation，Ordos Basin，NW China[J]. Petroleum Exploration and Development，2024，51(2)：234−247.

[30] 陈跃，汤达祯，许浩，等. 应用测井资料识别煤体结构及分层[J]. 煤田地质与勘探，2014，42(1)：19−23.

CHEN Yue，TANG Dazhen，XU Hao，et al. Application of logging data in recognition of coal structure and stratification[J]. Coal Geology & Exploration，2014，42(1)：19−23.

[31] 傅雪海，姜波，秦勇，等. 用测井曲线划分煤体结构和预测煤储层渗透率[J]. 测井技术，2003，27(2)：140−143.

FU Xuehai，JIANG Bo，QIN Yong，et al. Classification of coalbody structure and prediction of coal reservoir permeability with log curves[J]. Well Logging Technology，2003，27(2)：140−143.

[32] 梅放，康永尚，李喆，等. 中高煤阶煤层煤体结构识别测井法适用性评价研究[J]. 煤炭科学技术，2019，47(7)：95−107.

MEI Fang，KANG Yongshang，LI Zhe，et al. Applicability evaluation of logging method for recognition of coal body structure in medium and high rank coal seams[J]. Coal Science and Technology，2019，47(7)：95−107.

[33] ZHAO Lipeng，CAO Yunxing，LYU Shuaifeng，et al. Coal structure characteristics in the northern Qinshui Basin and their discrimination method based on the particle size of drilling cuttings[J]. ACS Omega，2022，7(26)：22956−22968.

[34] 陶传奇，王延斌，倪小明，等. 基于测井参数的煤体结构预测模型及空间展布规律[J]. 煤炭科学技术，2017，45(2)：173−177.

TAO Chuanqi，WANG Yanbin，NI Xiaoming，et al. Prediction model of coal–body structure and spatial distribution law based on logging parameters[J]. Coal Science and Technology，2017，45(2)：173−177.

[35] 王新领,祝新益,张宏兵,等. 基于随机树嵌入的随钻测井岩性识别方法[J]. 吉林大学学报 (地球科学版),2024,54(2):701−708

WANG Xinling,ZHU Xinyi,ZHANG Hongbing,et al. Lithology identification method for logging while drilling basedon random tree embedding[J]. Journal of Jilin University (Earth Science Edition),2024,54(2):701−708

[36] 李安,蔡益栋,王子豪,等. 基于机器学习测井反演的煤体结构评价:以鄂尔多斯盆地榆林地区本溪组8号煤为例[J]. 地质科技通报,2025,44(4):2−15

LI An,CAI Yidong,WANG Zihao,et al. Evaluation of coal structure based on machine learning logging inversion:A case from No. 8 coal of Benxi Formationin Yulin area of Ordos Basin[J]. Bulletin of Geological Science and Technology,2025,44(4):2−15