摘要
中国页岩气资源类型多样,形成地质条件与成藏模式复杂,如何针对不同类型与级别的页岩气资源展开具有针对性的地质评价,是勘探开发工作的重点与难点。为解决此问题,在页岩气地质评价过程中,通过智能化软件和测量仪器的研发,提高了定量数据的获取精度;建立概率密度曲线,识别了不同类型页岩气相关地质参数的分布特征,进一步开展了资源量计算方法的优选;以机器学习中多种方法如知识图谱、随机森林、支持向量机、聚类分析、概率赋值等背后的数学原理为支撑,从数据本身的自然规律出发,对参数的异同性和内在关联度开展研究。结果表明:不同类型页岩气资源的地质参数分布特征差异较大,结合勘探开发程度共同决定资源量计算方法的选取;对数据开展全面定量地解释与分析,极大程度挖掘了数字背后的地质信息,并帮助调整地质认识的客观性与真实性,同时能够发现评价结果产生的误差与偏差。该研究内容与成果建立并完善了从参数获取到综合评价的技术方法与流程,极大程度上提高了评价结果的可信度,拓展了人工智能在该领域应用的广度与深度,为最终能够达到页岩气地质评价智能化的目标奠定基石。
Based on various types of Chinese shale gas resource with complex formation geological conditions and accumulation patterns,targeted geological evaluation for different types and classes of shale gas resources is the focus and challenge of current exploration and development.To solve this problem,in the process of shale gas geological evaluation,acquisition accuracy of quantitative data is improved through development of intelligent software and measurement instrument.Probability density curves are established to identify distribution of related geological parameters of different types of shale gas resource,and furtherly to perform optimization of resource calculation methods.Starting from natural laws of data,and supported by mathematical principles of a variety of machine learning methods including knowledge graphs,random forests,support vector machines,cluster analysis and probability assignment,the similarities and differences even internal correlations of parameters are studied.The results show that there is much variation in geological parameters distribution of different types of shale gas resource,together with exploration and development degree,may influence selection of resource calculation methods.Comprehensive quantitative interpretation and analysis of data greatly explore geological information behind data,contribute to adjust objectivity and authenticity of geological understanding,and can find errors and deviations in evaluation results.The research content and achievements modify and build a technique process from parameter acquisition to comprehensive evaluation,which greatly improves credibility of evaluation results,expands application breadth and depth of artificial intelligence in this field,and lay a foundation for finally achieving ultimate goal of shale gas intelligent geological evaluation.
作者
郎岳
张金川
王焕第
武向真
李谦超
王东升
李兴起
陈世敬
LANG Yue;ZHANG Jinchuan;WANG Huandi;WU Xiangzhen;LI Qianchao;WANG Dongsheng;LI Xingqi;CHEN Shijing(School of Energy,China University of Geosciences,Beijing 100083,China;Key Laboratory of Shale Gas Resource Strategic Evaluation,Ministry of Natural Resources,Beijing 100083,China;Key Laboratory of Unconventional Natural Gas Energy Geological Evaluation and Development Engineering,Beijing 100083,China;Petroleum Industry Press Co.Ltd,Beijing 100011,China)
出处
《大庆石油地质与开发》
CAS
CSCD
北大核心
2022年第1期166-174,共9页
Petroleum Geology & Oilfield Development in Daqing
基金
国家自然科学基金项目“页岩含气性关键参数测试及智能评价系统”(41927801)
国家科技重大专项“页岩气分类分级资源评价方法研究”(2016ZX05034-002-001)。
关键词
页岩气
地质评价
智能评价
数据处理
数据获取
shale gas
geological evaluation
intelligent evaluation
data processing
data acquisition
作者简介
第一作者:郎岳,女,1991年生,在读博士,从事页岩气藏形成及资源评价研究。E-mail:gouqilier@126.com;通信作者:张金川,男,1964年生,教授,博士生导师,从事非常规天然气地质、页岩形成机理及油气资源评价等方面的教学与研究。E-mail:zhangjc@cugb.edu.cn。
