基于叠后地震几何属性融合的煤层大尺度裂缝预测
Large-scale Fracture Prediction of Coal Seam Based on Post-stack Seismic Geometric Attributes Fusion
李智 1吴海波 2刘钦节 3白泽1
作者信息
- 1. 安徽理工大学地球与环境学院,安徽淮南 232001
- 2. 安徽理工大学地球与环境学院,安徽淮南 232001;合肥综合性国家科学中心能源研究院(安徽省能源实验室),安徽合肥 230031
- 3. 合肥综合性国家科学中心能源研究院(安徽省能源实验室),安徽合肥 230031
- 折叠
摘要
煤层大尺度裂缝是导致煤矿瓦斯突出和矿井突水的主要因素之一,准确预测煤层中的大尺度裂缝对于煤层气开发和矿井水害治理至关重要.本文以叠后地震数据为基础,提取能够反映目标层位裂缝几何特征和分布信息的叠后地震几何属性;为避免单一地震几何属性在预测煤层大尺度裂缝时出现的多解性或不确定性问题,通过PCA(Principal Component Analysis)-BP(Back Propagation)神经网络(主成分分析与反向传播神经网络相结合)模型对相干、曲率、倾角、方差等地震几何属性进行融合,预测得到目标煤层大尺度裂缝分布特征.通过对比预测结果与实际揭露情况,90％的断层位置与大尺度裂缝发育区相一致,融合结果与构造解释结果吻合较好.表明基于地震几何属性融合的预测方法具有一定的效果,能在一定程度上刻画煤层大尺度裂缝的特征.
Abstract
The large-scale fracture in coal seam is one of the main factors leading to gas out-burst and water inrush in coal mines.The accurate prediction of large-scale fracture in coal seam is very important for coal seam gas development and mine water damage control.Based on the post-stack seismic data,this paper extracts the post-stack seismic geometric attributes that can reflect the geometric characteristics and distribution information of fractures in the target horizon.In order to avoid multiple solutions or uncertainties of single seismic geometry attributein predicting large-scale fractures of coal seams,PCA(principal component analy-sis)-BP(back propagation)neural network(combining principal component analysis and back propagation neural network)model are used to fuse seismic geometric properties such as coherence,curvature,inclination and varianceto predictthe distribution characteristics of large-scale fractures in target coal seam.By comparing the predicted results with the actual disclosure,90％of the fault locations are consistent with the large-scale fracture develop-ment area,and the fusion results are in good agreement with the structural interpretation re-sults.It shows that the prediction method based on the fusion of seismic geometric attributes has a desirable effect in depicting the characteristics of large-scale fractures in coal seams.
关键词
煤层/大尺度裂缝/叠后地震几何属性/PCA-BP神经网络Key words
coal seam/large-scale fracture/post-stack seismic geometric attribute/PCA-BP neural network引用本文复制引用
基金项目
安徽省高校协同创新项目(GXXT-2021-016)
合肥综合性国家科学中心能源研究院重大培育项目(21KZS215)
安徽省自然科学基金项目(2208085QD114)
安徽省高等学校自然科学研究项目(2022AH050831)
出版年
2024
NETL
NSTL
维普
万方数据