同位素注入剖面测井解释智能化批量处理技术的实现与应用  

Implementation and Applications of Intelligent Batch Processing Technology for Isotope Injection Profile Logging Interpretation

作　　者：孙亮[1] SUN Liang(Logging and Testing Service Company,CNPC Daqing Oilfield,Heilongjiang 163453,China)

机构地区：[1]大庆油田有限责任公司测试技术服务分公司,黑龙江大庆163453

出　　处：《测井技术》2024年第1期67-74,共8页Well Logging Technology

基　　金：中国石油天然气集团有限公司前瞻性基础性技术攻关项目“测井核心装备与软件平台研制”(2021DJ3905)。

摘　　要：同位素注入剖面测井是水驱油田了解各小层吸水状况的主要监测手段,该测井方法在中国油田的年测井工作量在3万井次左右,在注水油藏动态监测中占主导地位。传统的同位素注水剖面测井结果依赖人工解释,工作效率低、工作量大且多解性强,容易受到不同分析员人为因素影响,已经无法满足日益增长的生产需求。运用大数据监测和机器学习等技术,解决了基础数据录入、原始数据解编、深度匹配、工具识别、曲线叠合、综合评价等关键环节技术难题,形成了智能化批量处理技术。通过在CIFLog测井解释平台上的集成运行,单井处理时间从30~60min缩短到30s,人工干预率低于12%,验证了该处理流程及算法的有效性。Isotope injection profile logging is the main monitoring method for understanding the water absorption status of each sub layer in water drive oilfields.The annual logging workload in China’s oilfields is about 30000 wells,which plays a dominant role in dynamic monitoring of water injection reservoirs.The traditional isotope water injection profile relies on manual interpretation,which has low work efficiency,huge workload,and strong multi-solution,and is easily influenced by human factors of different analysts.It is no longer able to meet the growing production demand.By utilizing technologies such as big data monitoring and machine learning,key technical challenges such as basic data input,raw data decomposition,deep matching,tool recognition,curve stacking,and comprehensive evaluation have been solved,forming intelligent batch processing technology.Through integrated operation on the national major scientific and technological integration platform CIFLog,the single well processing time is reduced from 30~60minutes to 30 seconds,and the manual intervention rate is less than 12%,verifying the effectiveness of the processing process and algorithm.

关 键 词：油藏动态监测 同位素注入剖面 智能化 批量处理 机器学习 水驱油田 

分 类 号：P631.84[天文地球—地质矿产勘探]