漠河冻土区天然气水合物科学钻探MK 2孔地层岩性的测井识别

岩性识别是天然气水合物储层测井评价的基础，准确的岩性识别结果可以为天然气水合物的勘探提供可靠的依据，在寻找天然气水合物和评估天然气水合物储量方面发挥着巨大的作用。针对漠河冻土区天然气水合物科学钻探 2孔（MK 2孔）的钻探情况，利用已钻井段地层的岩心资料和常规测井资料，分别采用交会图法和支持向量机法对研究区的地层开展了岩性识别研究。结果表明：研究区内有砂岩、泥岩、石灰岩、糜棱岩和泥质板岩５种岩石，其不同岩性的测井响应差异能够定性识别岩性；自然电位与电阻率测井参数的交会，能够有效的、定量识别研究区地层的岩性；支持向量机法所建立的岩性识别模型，对研究区地层的岩性识别率可达96.67%。所建立的测井识别方法较好地解决了该区地层岩性识别问题，也为类似地区的天然气水合物地层测井评价提供了一种重要的手段。

Well-log lithology identification in well MK-2 for scientific drilling and exploration of gas hydrate in Mohe permafrost, China

Lithology identification is the basis of logging evaluation of gas hydrate reservoirs and its accurate information will provide a reliable foundation for gas hydrate exploration, thus it plays a big role in searching for gas hydrate and estimating gas hydrate reserves. In view of the well MK 2 for scientific drilling and exploration of gas hydrate in Mohe permafrost, we performed lithology identification by using the cross plotting method and the Support Vector Machine (SVM) method based on the core data and conventional logging data. Five lithologic categories were identified in the study area, including sandstone, mudstone, limestone, mylonite and argillaceous slate. These lithologies can be qualitatively identified according to their logging response differences and they can also be quantitatively identified through the cross plotting of SP vs. resistivity logging. The lithology identification model built with the SVM method can recognize the lithologies in the study area with an accuracy ratio of up to 96.67%. These methods successfully solve the problem of lithology identification in the study area and can be also used in logging evaluation of gas hydrate reservoirs in such areas with similar geologic conditions.