Prediction of Terrestrial Gas Hydrate Accumulation Based on Support Vector Regression
-
摘要: 天然气水合物，俗称“可燃冰”，自2008年在青海木里冻土区首次钻获天然气水合物以来，人们已在此地开展了大量的勘查工作，但该区地质情况复杂，天然气水合物成藏规律不清，单一的地球物理方法难以充分利用信息，因此难以有效地圈定出天然气水合物异常区。本文选取木里地区为主要研究区域，综合区内勘查已获得的地球物理、地球化学和地质资料，分析和提取对水合物成藏有利的特征，给出相应的预测变量转化规则。采用支持向量回归方法进行成藏预测研究，并对结果进行评估。结果显示，钻遇水合物的钻井与预测得到的高有利度区吻合，未遇水合物的钻井基本落于低有利度区，算法有效实用，能够提供一定的指导意义。Abstract: Natural gas hydrate, commonly known as "combustible ice", had been extensively surveyed. Since the discovery of natural gas hydrate in Muli area of Qinghai province for the first time. However, the geological structure in this area was complex and gas hydrate accumulation rule is unclear. It's difficult for a single geophysics method to make full use of information to effectively find gas hydrate anomalies. In this paper, the features which were favorable for gas hydrate accumulation were extracted from geology, geophysics and geochemistry data In Muli area, and the corresponding transformation rules were proposed. Support vector regression was used to carry out the study of gas hydrate prediction, and the results were evaluated. The results showed that the drilling with hydrates and the drilling without hydrates were consistent with the prediction. The algorithm was effective and practical, and it could offer the guide for future exploration.
-
Keywords:
- support vector regression /
- gas hydrate /
- hydrate prediction /
- permafrost
-
-
期刊类型引用(2)
1. 范婕，许欣怡，周诗岽，周年勇. 基于PSO-SVM的天然气水合物生成条件预测. 天然气化工—C1化学与化工. 2022(05): 171-176 . 
百度学术
2. 叶智慧，宁禹强，张敏，李晓蓉. 基于机器学习分类算法的地层水合物识别方法研究. 海洋技术学报. 2021(05): 51-61 . 百度学术
其他类型引用(4)
计量
- 文章访问数: 287
- HTML全文浏览量: 6
- PDF下载量: 11
- 被引次数: 6
下载:
