一种新型冷热地震数据分离的方法研究  

Research on A New Method for Separating Cold and Hot Seismic Data

作　　者：冯永照 张亮 陈冠毅 FENG Yongzhao;ZHANG Liang;CHEN Guanyi(China Oilf ield Services Co.,Ltd.Geophysical BU,Zhanjiang 524057,China;Zhanjiang Branch of CNOOC Ltd.,Zhanjiang 524057,China)

机构地区：[1]中海油田服务股份有限公司物探事业部,广东湛江524057 [2]中海石油(中国)有限公司湛江分公司,广东湛江524057

出　　处：《化工管理》2024年第26期69-71,共3页Chemical Management

摘　　要：随着勘探新技术的发展,机器学习应用越来越多,地震数据体量越来越大,占用储存资源不断增加。作为数据库底层基本技术,冷热数据分离技术将数据分为热数据和冷数据。基于缓存替换策略,结合LRU算法和LFU算法,根据地震数据处理规律,设计了一种针对机器学习的地震资料处理中地震数据的冷热分离策略。在地震数据上应用冷热数据分离技术能有效降低存储硬件成本,并可减轻HPC集群的储存负担,提升数据处理效率。With the development of new exploration technology,the application of machine learning is increasing,the volume of seismic data is increasing,and the storage resources are increasing.As the basic technology of database,cold and hot data separation technology divides data into hot data and cold data.Based on the cache replacement strategy,combined with LRU algorithm and LFU algorithm,and according to the seismic data processing law,a cold and hot separation strategy for seismic data processing in machine learning is designed.The application of cold and hot data separation technology on seismic data can effectively reduce the cost of storage hardware,reduce the storage burden of HPC cluster,and improve the eff iciency of data processing.

关 键 词：冷热数据分离 地震数据处理 缓存替换策略 

分 类 号：TP333[自动化与计算机技术—计算机系统结构]