基于长短期记忆网络的CO_(2)气层识别方法  

CO_(2)Gas Layer Recognition Method Based on Long Short-Term Memory Network

作　　者：何丽娜 吴文圣[1] 王显南 张伟[3] 张传举[3] 宋孝雨 HE Lina;WU Wensheng;WANG Xiannan;ZHANG Wei;ZHANG Chuanju;SONG Xiaoyu(State Key Laboratory of Petroleum Resources and Prospecting,China University of Petroleum(Beijing),Beijing 102249,China;Shenzhen Branch,CNOOC(China),LTD.,Shenzhen 518054,China;China Oilfield Services Limited,Shenzhen 518054,China)

机构地区：[1]中国石油大学(北京)油气资源与探测国家重点实验室,北京102249 [2]中海石油(中国)有限公司深圳分公司,深圳518054 [3]中海油田服务股份有限公司,深圳518054

出　　处：《测井技术》2024年第1期1-7,共7页Well Logging Technology

摘　　要：CO_(2)监测是油气开采中的关键环节,传统的CO_(2)监测方法面临很多挑战,在人工智能逐渐兴起的当下,深度学习技术被广泛应用于地球物理测井。珠江口盆地恩平凹陷深层CO_(2)气藏发育,传统测井方法无法准确评价储层流体。构建了基于长短期记忆网络(LSTM)的CO_(2)气层识别模型,采用m×2正则化交叉验证优选CO_(2)敏感测井参数,并对模型进行训练。利用该模型对珠江口盆地恩平凹陷L2井CO_(2)气层进行识别,并与支持向量机和K近邻算法识别结果进行对比。结果表明,3种深度学习算法对CO_(2)气层的识别效果良好,其中LSTM算法对CO_(2)气层的识别效果最好,准确度达93.4%,为深层CO_(2)气层识别工作提供了新思路。CO_(2)monitoring is a crucial part of oil and gas extraction,and traditional methods for monitoring CO_(2)face many challenges.With the gradual rise of artificial intelligence,deep learning technology is widely used in geophysical logging.Due to the development of deep CO_(2)gas layer in Enping sag,the Pearl River Mouth basin,traditional logging methods cannot accurately evaluate reservoir fluids.CO_(2)gas layer recognition model based on Long Short-Term Memory Network(LSTM)is constructed,and CO_(2)sensitive logging parameters are optimized through m×2 regularized cross validation to train the model.This model is used to identify the CO_(2)gas layer of well L2 in Enping sag,the Pearl River Mouth basin,and is compared with the recognition results of support vector machine and K nearest neighbor algorithm.The results show that the three deep learning algorithms have good recognition effects on CO_(2)gas layer,among which the LSTM algorithm has the best identification effect on CO_(2)gas layer,with an accuracy of 93.4%,providing new ideas for deep CO_(2)gas layer recognition work.

关 键 词：CO_(2)气层识别 长短期记忆网络(LSTM) 深度学习 珠江口盆地 

分 类 号：P631.84[天文地球—地质矿产勘探]