基于深度学习的页岩气压裂缝网预测方法  

Shale pressure fracture network prediction method based on deep learning

作　　者：李兵 吴建军 孟令鹏 赵海峰[2] LI Bing;WU Jianjun;MENG Lingpeng;ZHAO Haifeng(Engineering Technology Research Institute of PetroChina Coalbed Methane Co.,Ltd.,Xi’an 710082,China;College of Petroleum Engineering,China University of Petroleum(Beijing),Beijing 102249,China)

机构地区：[1]中石油煤层气公司工程技术研究院,西安710082 [2]中国石油大学(北京)石油工程学院,北京102249

出　　处：《中国科技论文》2024年第9期987-1001,共15页China Sciencepaper

基　　金：中国石油天然气股份有限公司前瞻性、基础性技术攻关项目(2021DJ2004)。

摘　　要：压裂评估是水力压裂的一项重要内容,而压裂后缝网参数的获取是压裂评估的关键。针对目前常用裂缝诊断方法(微地震监测技术、测斜仪裂缝监测技术和分布式光纤监测技术)存在成本高、操作难及易受影响等问题,基于在非线性回归方面具有显著优势的深度学习算法,从施工压裂曲线入手,综合考虑获取目标参数的各种理论模型和经验方程,对压裂曲线进行量化,选择相关参数作为输入,以微地震监测的缝网参数(缝网长度、宽度及高度)作为目标参数,通过深度学习算法(BP神经网络)构建其与目标参数之间的关系。结果表明:该方法具有较高的预测精度,平均相对误差在12%以内。此外,结果还受输入参数的种类和数量、网络结构及超参数的影响。Fracturing evaluation is an important part of hydraulic fracturing,in which the acquisition of the fracture network param⁃eters after fracturing is the key section.In view of the problems such as high cost,operating difficulty and vulnerability of the com⁃monly used fracture diagnosis methods(microseismic monitoring technology,clinometer crack monitoring technology and distributed optical fiber monitoring technology),the target parameters herein were obtained based on the deep learning algorithm with significant advantages in nonlinear regression.The process was started with the construction fracturing curve,and comprehensively considered various theoretical models and empirical equations.For the quantization of fracturing curve,relevant parameters were selected as input,and the fracture network parameters(length,width and height of fracture network)of microseismic monitoring were taken as target parameters,and the relationship between them and target parameters was constructed by deep learning algorithm(BP neural network).The results show that the method has high prediction accuracy,and the average relative error is less than 12%.In addition,it is also found that the results are affected by the type and number of input parameters,network structure,and hyperparameters.

关 键 词：页岩压裂曲线量化 BP神经网络 缝网参数反演 压裂评估 

分 类 号：TD712[矿业工程—矿井通风与安全]