煤田地质与勘探2024，Vol.52Issue(8)：145-164,20.DOI:10.12363/issn.1001-1986.23.12.0813
人工智能测井:基础、原理、技术及应用
Artificial intelligence logging:Fundamental,principle,technique,and application
摘要
Abstract
[Background]Intelligent hydrocarbon exploration and exploitation have become a trend and hot research topic in the oil and gas industry.Artificial intelligence logging(AIL)exhibits considerable potential to address chal-lenges in the explore-exploit of unconventional hydrocarbon resources,as well as resources in complex environments in the deep earth and deep ocean.However,the driving mode,fundamental,implementation principle,structure,and applic-ation scenarios of AIL remain understudied.[Objective and Methods]To build a comprehensive ecology of the AIL system and thoroughly explore and reveal the potential and value of AIL,this study employed methods like literature analysis,theoretical research,technical analysis,and verification using cases.First,this study delved into the critical factors influencing the integrated development of logging technology and AI from multiple dimensions,defining AIL ac-cordingly.Subsequently,it systematically explored the general theoretical framework,hardware arithmetic requirements,and data and physical models of AI.From the perspective of knowledge discovery,this study detailed the function im-plementation mechanisms of logging technology,instrumentation,petrophysics,and interpretation in the AIL system.Furthermore,it conducted an in-depth analysis of several critical technologies including log-related big data techniques,intelligent and fast algorithms,log knowledge graph,digital twins,intelligent instrumentation,and the Internet of things(IoT)of logs.Accordingly,this study posited that physical models and intelligent algorithms emerge as the core force driving the development of AIL.Based on the principles and characteristics of AI algorithms,this study systematically organized critical AIL technologies in terms of logging technology,instrumentation,acquisition operations,and inter-pretation,constructing the dendrogram and solving process of log knowledge graph.[Results and Conclusions]The empirical research reveals that AIL enjoys advantages in terms of the lithologic identification of tight sandstones and logging simulations,accuracies of up to 93.8%,respectively,significantly exceeding those of conventional methods.Re-garding log-based assessment,AIL can simultaneously identify reservoirs and fluids,sufficiently proving the consider-able development potential and application advantages of AIL.Based on the critical links of AIL,this study envisions the fifth development stage of logging technology,i.e.,artificial intelligence logging.The results of this study provide a solid theoretical foundation and practical guidance for the deep integration and extensive application of AI in the field of logging,holding great significance for the promotion and development of AIL technology.关键词
人工智能测井/测井大数据/机器学习/地层参数反演/复杂岩性识别Key words
artificial intelligence logging(AIL)/log-related big data/machine learning(ML)/stratigraphic parameter inversion/complex lithologic identification分类
天文与地球科学引用本文复制引用
程希,任战利..人工智能测井:基础、原理、技术及应用[J].煤田地质与勘探,2024,52(8):145-164,20.基金项目
油气藏地质及开发国家重点实验室(西南石油大学)开放基金项目(PLN2022-14) （西南石油大学）
国家自然科学基金项目(42272152) （42272152）
