基于地震属性的淮南潘集外围太原组煤系气储层含气性预测  

A prognosis for the gas content in gas reservoirs of coalmeasures of the Taiyuan Formation in the outskirts of the Panji mining area of Huainan based on seismic attributes

作　　者：丁海[1,2] 臧子婧 吴海波 DING Hai;ZANG Zi-jing;WU Hai-bo(Exploration Research Institute,Bureau of Coal Geology of Anhui Province,Hefei,Anhui 230088,China;Anhui Unconventional Natural Gas Engineering and Technology Research Center,Hefei,Anhui 230088,China;School of Earth and Environment,Anhui University of Science and Technology,Huainan,Anhui 232001,China)

机构地区：[1]安徽省煤田地质局勘查研究院,安徽合肥230088 [2]安徽省非常规天然气工程技术研究中心,安徽合肥230088 [3]安徽理工大学地球与环境学院,安徽淮南232001 [4]安徽能科工程科技有限公司,安徽合肥230088

出　　处：《安徽地质》2022年第2期119-123,共5页Geology of Anhui

基　　金：安徽省重点研究与开发计划项目“煤系天然气优质储层精细识别技术研究与示范”(编号:1804a0802203)、“两淮矿区废弃矿井综合调查及资源化利用潜力评价”(编号:2021-g-2-14)资助。

摘　　要：煤系气是一种重要的非常规天然气资源,具有巨大的开发潜力。但已有的勘探经验表明该类型储层的地层组成与结构均较为复杂,难以精确判别其含气性特征。因此,为了有效预测煤系地层含气性,本文提出一种基于地震属性的太原组煤系气储层含气性预测方法。针对含气性精准预测这一目标,先利用聚类分析优选出对煤系气含气量变化反应最敏感的地震属性;再利用井位置处的含气量样本数据,结合优选地震属性训练BP神经网络预测模型;最后,基于训练好的预测模型,以研究区的优选地震属性为输入,从横向和纵向两个维度预测研究区的煤系储层含气量。将井位置的预测与实测结果对比表明,本文采用的预测技术及流程精度高,可有效用于煤系气储层含气量的多维度大范围预测。Gas in coal measures,as an important unconventional natural gas resource,has great potential for development.However,according to exploration experiences,the reservoir of this kind of gas is complicated in the stratigraphic composition and structure,and it is hard to precisely tell its gas-bearing characteristics.Therefore,in order to effectively predict the gas-bearing capacity of coal measure strata,this paper gives an approach to the gas reservoir of the Taiyuan Formation based on seismic attributes,which,selected by cluster analysis,are most sensitive to gas content change in coal measures.Then we train the BP neural network prediction model by using gas content sample data from the well location and the optimized seismic attributes.Finally,based on the well-trained prediction model and taking the preferred seismic attributes of the study area as input,we predict the gas content in the gas reservoir of coal measures in the study area from both transverse and longitudinal dimensions.The comparison between the prediction and measured result for the well location shows that the prediction method and technical process adopted in this paper are high in precision and can be effectively used for multi-dimensional and large-scale prediction of gas content in gas reservoir of coal measures.

关 键 词：煤系气 储层 BP神经网络 聚类分析 含气量 

分 类 号：P624[天文地球—地质矿产勘探]