摘要
针对水淹层测井解释符合率难以满足开发需求的现状,通过水淹层测井解释技术文献检索分析认为:解释的实质就是利用数学手段建立地质特性与测井响应间的一种变换关系;随着油田开发的深入,油层物理性质与测井响应关系向不确定性、模糊性发展,从而使测井解释中数学思想也向不确定的数理统计、模糊数学方向发展;数理统计与模糊数学方法适用于不同的模式结构,水淹层解释中必须区别对待;统计变量向复杂分布的非参数方法转变.所以,应开拓适应“线状覆盖结构”延拓的新思维,充分发挥几何分类方法的优势,提高分类效率;准确分析研究对象的“模式结构”,有针对性地选择数学方法;加强非参数理论在测井解释中应用的基础分析,提高认识问题、解决问题的能力.
Aiming at the current situation of the development requirements that are difficult to be met by the coinci- dence rate of the watered-out layer interpretation, the analyses of tile literature retrieval for the well logging inter- pretation indicate that the essence of the interpretation is to establish the transformation relationship between the ge- ological characteristics and logging responses by meas of the mathematical methods; with the deepening of the oil- field development, the relationship between the reservoir physical properties and the responses is changed to uncer- tainty and fuzziness1 thus the mathematical thoughts in the logging interpretation are inverted to uncertain mathematical statistics and fuzzy mathematics; these two mathematical methods are suitable to different mode structures, so they must be treated individually in the interpretation of the watered-out reservoirs; the statistical variables are alerted to the complexly-distributed nonparametrie methods. So the new thought adapted to the continuation of "linear overlay structure" should be initiated, and moreover giving full play to the advantages of geometric classification to improve the efficiency of the classification~ accurately analyzing the "model structure" of the research objects and choosing the responding mathematical methods; strengthening the basic analyses on the nonparametric theory applied in the well logging interpretation so as to enhance the abilities to acknowledge and solve the problems.
出处
《大庆石油地质与开发》
CAS
CSCD
北大核心
2014年第6期134-139,共6页
Petroleum Geology & Oilfield Development in Daqing
基金
国家科技重大专项“大型油气田及煤层气开发”(2011ZX05014)
关键词
水淹层
测井解释
数学思想
数理统计
模糊数学
神经网络
watered-out layer
logging interpretation
mathematical thought
mathematical statistics~ fuzzy mathematics
neural network