摘要
煤体结构直接影响煤储层孔裂隙发育,其准确判识对煤层压裂及煤层气开采具有重要指导价值。以鄂尔多斯盆地榆林地区本溪组8号煤为例,其煤体结构复杂,引入机器学习方法可解决煤层气储层测井中的非线性问题。采用区内已完成预处理的取心井数据,采用BP神经网络、随机森林以及XGBoost算法进行训练和全区煤体结构反演,并结合区内煤层顶底板及煤厚,剖析构造控制下的煤体结构发育特征。结果表明:①随机森林以及XGBoost算法相较于BP神经网络,对目标煤层煤体结构的反演结果更接近于岩心观测的真实情况,准确度更高;②榆林地区8号煤从NW向SE,煤体破碎程度逐渐加剧;③区内由中部至东南部发育构造带,在构造带影响下煤厚减小且原生结构煤逐渐转变为糜棱结构煤。本研究可为研究区实际煤层气生产中的煤体结构识别以及构造带分析提供参考。
[Objective]Coal structure directly affects the pore and fracture system of coal reservoirs.Therefore,theaccurate identification of coal structure is crucial for guiding coal seam fracturing and coal bed methane extraction.Taking No.8 coal of the Benxi Formation in the Yulin area of the Ordos Basin as an example,the complex coalstructure necessitates the introduction of machine learning methods to address the nonlinear challenges in loggingdata interpretation.[Methods]In this study,Back Propagation(BP)neural network,Random Forest,and XGBoostalgorithms are used to train on pre-processed core well data from the study area to invert coal structure across thisregion.By integrating the top and bottom plates of the coal seams and the coal thickness,we explore thedevelopment of coal structure under tectonic control.[Results]The results indicate that:(1)Compared to the BPneural network,Random Forest and XGBoost algorithms provide more accurate inversion results,aligning moreclosely with real core observations.(2)The degree of coal structure fragmentation in No.8 coal in the Yulin areaincreases progressively from northwest to southeast.(3)Tectonic zones,developed from the central to southeasternpart of the study area,cause a decrease in coal thickness,with the coal structure transitioning from primary coal tomylonitic coal under tectonic influences.[Conclusion]The three machine learning algorithms employed in thisstudy successfully inverted the complex coal structure,with Random Forest and XGBoost achieving higherinversion accuracy.Additionally,the relationship between coal structural variations and the development of tectoniczones was analyzed,providing valuable insights for identifying coal structures and evaluating tectonic zones incoalbed methane production.
作者
李安
蔡益栋
王子豪
刘大锰
LI An;CAI Yidong;WANG Zihao;LIU Dameng(School of Energy Resources,China University of Geosciences(Beijing),Beijing 100083,China;Beijing Key Laboratory of Unconventional Natural Gas Geological Evaluation and Development,Beijing 100083,China)
出处
《地质科技通报》
北大核心
2025年第4期2-15,共14页
Bulletin of Geological Science and Technology
基金
国家自然科学基金项目(42130806,42372195)
中央高校基本科研业务费深时数字地球前沿科学中心“深时数字地球”中央高校科技领军人才团队项目(2652023001)。
