摘要
煤层气水平井是提高煤层气单井产量的有效技术措施,为确保煤层气水平井的有效钻遇率,需要进行随钻地层判识,进而对钻井施工过程实施地质导向。现有的煤层气水平井随钻地层判识技术大都以井下随钻数据进行分析,判识准确率有待提高。针对目前中国煤层气水平井钻井过程中普遍采用的仪器组合方式,在满足煤层气低成本开发的客观要求的前提下,选取了自然伽马、声波时差、补偿密度、补偿中子4个参数进行分析,采用支持向量机算法对上述测井参数进行归一化处理,根据随钻地层判识要求选择了符合条件的核函数,对相关参数进行了优化,通过对输出结果的分类实现了煤层气水平井随钻地层判识,判识准确度满足现场应用的要求。
Existing methods of logging-while-drilling stratigraphie identification for horizontal wells for eoalbed methane are mainly based on the logging-while-downhole drilling data. However, the accuracy of these methods needs to be improved. This study employs horizontal drilling jig configurations widely adopted for eoalbed methane exploration in China, and takes the cost-effectiveness into account. Four parameters are selected and analyzed, namely natural gamma, inlerval transit time, compensated density, and compensated neutron. The support vector machine algorithm is used to normalize data collected for these parameters, and to optimize the parameters with the appropriate kernel functions. By classifying the final outputs, this study realizes logging-while-drilling stratigraphie identification for horizontal wells for eoalbed methane. Such an identification method is accurate enough for field applications.
作者
张禾
张弓
ZHANG He;ZHANG Gong(School of Mechatronic and Electrical Engineering,Southwest Petroleum University,Chengdu,Sichuan 610500,China)
出处
《西南石油大学学报（自然科学版）》
CAS
CSCD
北大核心
2018年第6期131-138,共8页
Journal of Southwest Petroleum University(Science & Technology Edition)
基金
四川省应用基础研究基金(2016JY0049)
关键词
煤层气
水平井
支持向量机
测井数据
随钻地层判识
coalbed methane
horizontal well
support vector machine
well-logging data
logging-while-drilling stratigraphic identification
