基于统计学与专家系统的测井相自动识别  

Logging facies automatic identification based on statistics and expert system

作　　者：李存磊[1] 任伟伟[1] 闫伟[2] 刘婷[2] 霍晓敏[2] 

机构地区：[1]北京师范大学资源学院,北京100875 [2]中国海洋大学海洋地球科学学院,青岛266100

出　　处：《地球物理学进展》2011年第4期1400-1408,共9页Progress in Geophysics

基　　金："十一五"国家科技支撑项目(2006BAC18B05);国际重大资源态势评价与预测技术开发项目联合资助

摘　　要：通过测井相来研究沉积微相的平面展布是油藏描述研究工作的重要手段,尤其是针对油藏开发后期密集井网区的沉积微相精细研究尤为重要.在测井相自动识别的研究中,目前主要采用神经网络和Bayes准则判别分析等方法,但这些方法过于依赖数学模型而忽略了地质专家的判相经验,常导致识别结果准确率较低.本文从研究区沉积环境以及岩性特点出发,通过分析60余口井的单井相层序与测井曲线特征的关系,编程提取了测井曲线的7个特征参数.根据取心段不同微相测井曲线特征参数的数理统计规律,采用"专家系统"的"产生式规则表示法"建立沉积微相判别规则,并编程实现测井相的自动识别.判别结果是建立在统计学和专家经验的基础上,判别准确率可达90%.The study of sendimentary microfacies distribution and comparison with logging facies is important to reservoir description, especially for fine reservoir description with huge amounts of well-logs. At present, the main methods used in automatic identification of logging facies are neural network and Bayes criterion, but these methods often lead to low accuracy of recognition results due to excessively relying on mathematical models and ignoring the geological expertsr experience. In this paper, on the basis of sedimentary and lithological characters, by analyzing the relational matching between the single well facies and logs, 7 characteristic parameters reflecting sedimentary facies characteristics are extracted. Drill core data of 60 wells gives a lot of information about sedimentary microfacies and promotes to set up the relation between sedimentary microfacies and logs in the same cored intervals. A table revealing the distribution characters of characteristic parameters matching the particular microfacies is set up by selective extracting and statisticsing the characteristic parameters of these logs. It can be used for the inference engine programes of expert system. The discrimination rules of sedimentary facies based on production rules of expert system and that to achieve the automatic identification of logging facies are established by programme. The system can export the microfacies codes in the different well sections when importing the logging data and horizon data. It can get a high accuracy about 90% by contrasting the expert＇s identification results. It was necessary to state that this system is used to distinguish microfaeies aimed at certain sedimentary systems. When using the system, defining the sedimentary system firstly by experts is important.

关 键 词：测井相 自动识别 三角洲 统计学 专家系统 

分 类 号：P631[天文地球—地质矿产勘探]